Electromagnetic compatibility of unmanned aircraft - DiVA Portal

Electromagnetic compatibility of unmanned

aircraftExamination of legislation and evaluation of two commerical systems

Alex Bergdahl

Engineering Physics and Electrical Engineering, master's level

2022

Luleå University of Technology

Department of Computer Science, Electrical and Space Engineering

Preface

This master thesis was conducted at, and for, Lulea University of Technology(LTU) to aid FieldRobotiX and another anonymous company (henceforth re-ferred to as company A) in assessing the electromagnetic compatibility (EMC)of one of their products. I would like to thank Magnus and Erik from companyA and also Dariusz Kominiak and Martin Blaszczyk from FieldRobotiX. Theywere essential in helping me with any technical questions I’ve had about theirrespective drones and their usage, but also provided vital assistance in settingup, performing and evaluating the tests performed on their respective systems.

As for staff at LTU, my sincerest thanks to Joakim Nilsson and my examinerJonny Johansson for providing their assistance and advice during the course ofthe whole project. This thesis could also not have been completed without thehelp of Andreas Nilsson, who contributed with his expertise on the usage of theEMC lab at LTU and EMC theory.

i

Abstract

Electromagnetic compatibility (EMC), or the field of reducing emissions fromand increasing immunity against electromagnetic interference, is an essentialpart of designing modern electronics. As one would expect, EMC is especiallyimportant for things such as aircraft and aviation equipment where outages ordisturbances could have severe consequences. The problem presented in thisthesis was to consolidate the available legislation regarding EMC for unmannedaircraft and then apply this information onto two commercial systems still underdevelopment, one made by FieldRobotiX and one made by company A (who hasasked to remain anonymous in this thesis). The FieldRobotiX system consistsof a prototype module created by FieldRobotiX themselves which is fitted andinterfaced with a commercial DJI drone and Velodyne LiDAR (Light DetectionAnd Ranging). Company A:s system is an electrical vertical take of and landing(VTOL) drone prototype. Based on the applied rules, pre-compliance measure-ments were then performed to identify problematic areas of their designs inregards to EMC.

The research process for the legislation involved reading through mainly theofficial documents and directives published by the European Commission, theEuropean Parliament and the European Aviation Safety Agency (EASA), look-ing up declarations of conformity made by drone manufacturers and also con-tacting accredited EMC labs for information on how they usually prove compli-ance for drones. The conclusion of this research being that (for EMC purposes)drones need to follow either the EMC directive 2014/30/EU [25], the radioequipment directive 2014/53/EU [26] or the essential requirements of directive2018/1139/EU [27] depending on the intended usage of the drone and its tech-nical specification.

As for application of legislation onto the FieldRobotiX and company A sys-tems, it became clear that as there were no drone-specific EMC standards (i.evoluntary ways to more easily prove conformity) in the EU some simplificationswould need to be made. This took the form of applying parts of both the EN55032 [37] (applicable for multimedia equipment) and EN 301 489-1 [17] (ap-plicable for radio equipment) standards for radiated emissions and immunitytesting respectively. The measurements showed that the FieldRobotiX systemhas issues with its LiDAR (used for 3D mapping) radiating too much interfer-ence and the drone also experiences issues when exposed to continuous radiatedfields, causing the rolling motion of the LiDAR to stutter. As for the company Adrone, it was shown that while all motors provide sizeable interference it seemsthe back row of motors are providing more interference than their front rowcounterparts, which may indicate an issue with the wiring in the back wing(s).Furthermore the 15 V regulator of company A:s drone proved to be a big sourceof interference.

While the application of legislation in the end became more simplified than ini-tially planned, the goal of condensing down the available information has stillbeen achieved. As for the measurements, it should be noted that while most ofthe problematic areas were indeed successfully identified on both systems thereare still measurements that should be done in the future. This includes testing

ii

conducted emissions and immunity against transient electromagnetic phenom-ena such as electrostatic discharge (ESD).

iii

Contents

1 Introduction 11.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Problem description . . . . . . . . . . . . . . . . . . . . . . . . . 21.3 Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.4 Contribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31.5 Outline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 Theory 52.1 Literature review . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.2 EMC testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2.2.1 Anechoic chamber . . . . . . . . . . . . . . . . . . . . . . 52.2.2 Shielding . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.2.3 Detector types . . . . . . . . . . . . . . . . . . . . . . . . 6

3 Method 73.1 Legislation regarding EMC of drones . . . . . . . . . . . . . . . . 7

3.1.1 Finding directives . . . . . . . . . . . . . . . . . . . . . . 73.1.2 Application of standards . . . . . . . . . . . . . . . . . . . 10

3.2 Legislation for the company A drone . . . . . . . . . . . . . . . . 103.2.1 Choosing standards . . . . . . . . . . . . . . . . . . . . . 13

3.3 Legislation for the FieldRobotiX system . . . . . . . . . . . . . . 143.4 Testing environment . . . . . . . . . . . . . . . . . . . . . . . . . 16

3.4.1 Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . 163.4.2 Positioning the company A drone . . . . . . . . . . . . . . 17

3.5 Test plans and experimental setup . . . . . . . . . . . . . . . . . 193.5.1 FieldRobotiX system . . . . . . . . . . . . . . . . . . . . . 193.5.2 Company A drone . . . . . . . . . . . . . . . . . . . . . . 21

4 Results 244.1 FieldRobotiX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

4.1.1 Vertical polarization emissions . . . . . . . . . . . . . . . 254.1.2 Horizontal polarisation emissions . . . . . . . . . . . . . . 294.1.3 Immunity . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

4.2 Company A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324.2.1 Motor emissions . . . . . . . . . . . . . . . . . . . . . . . 334.2.2 15 V regulator emissions . . . . . . . . . . . . . . . . . . . 374.2.3 12 V voltage regulator . . . . . . . . . . . . . . . . . . . . 414.2.4 Immunity . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

5 Discussion 445.1 Legislation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445.2 Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

5.2.1 FieldRobotiX . . . . . . . . . . . . . . . . . . . . . . . . . 455.2.2 Company A . . . . . . . . . . . . . . . . . . . . . . . . . . 46

iv

6 Conclusions 486.1 Legislation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 486.2 FieldRobotiX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 486.3 Company A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49

7 Appendix7.1 Images of measurement hardware . . . . . . . . . . . . . . . . . .7.2 Results from measurements on single motor from company A drone

v

Abbreviations

BVLOS Beyond Visual Line of SightEASA European union Aviation Safety AgencyEMC Electromagnetic CompatibilityEMCD Electromagnetic Compatibility DirectiveEMI Electromagnetic InterferenceESD Electrostatic DischargeEUT Equipment Under TestFAR Fully Anechoic RoomGNSS Global Navigation Satellite SystemIoT Internet of ThingsLiDAR Light Detection And RangingLTU Lulea University of TechnologyRED Radio Equipment DirectiveUAS Unmanned Aircraft SystemVLOS Visual Line of SightVTOL Vertical Take-off and Landing

vi

1 Introduction

1.1 Background

In today’s society, technology moves towards being faster, smaller and more in-telligent. Aspects such as increased usage of internet of things (IoT) devices andwireless products present an increasingly complex electromagnetic environmentfor devices to operate in. Through all of this, it is still expected and demandedthat products should continue to work as intended. This poses unique challengesfor engineers as electromagnetic interference (EMI) is a very real and palpableconcern that greatly affects electronics design. Ideally, the consumer should notneed to concern themselves with whether or not their devices will interfere witheach other, but this is easier said than done. To mitigate and hinder issues stem-ming from EMI, electromagnetic compatibility (EMC) is the field of reducingthe susceptibility to EMI of a piece of equipment and reducing its own emissions.

Although its effects might not be immediately apparent until problems occur,EMC is very much an essential part of modern technology. Article 3, Para-graph 1.4 of the European EMC directive (EMCD) [25] defines EMC as follows:”‘electromagnetic compatibility’ means the ability of equipment to function sat-isfactorily in its electromagnetic environment without introducing intolerableelectromagnetic disturbances to other equipment in that environment;”. Theconsequences of poor EMC can range from slightly annoying to posing serioussafety concerns. A less severe example would be a wireless speaker sounding dis-torted due to its poor electromagnetic immunity and thus high susceptibility toelectromagnetic interference. A more serious example would be when in Swedenan electrical road signs poor EMC caused it to emit undesired electromagneticwaves of a frequency which interfered with flight communications in a nearbyairport during takeoff and landing [p.59, 9].

Needless to say, EMC and immunity to electromagnetic disturbances is espe-cially important for aircraft where disruptions and/or outages can have severeconsequences. In the case of unmanned aircraft (drones), the safety risks mightnot be as immediately intense but with drones predicted to take bigger roles insectors such as agriculture, construction and law enforcement [33], its becomingever more important for drones to be able to function reliably in a multitude ofenvironments. From open fields to urban areas, these environments present dif-ferent amounts of EMI sources with varying intensity. For a drone to be used ina flexible way for multiple applications, considerable effort needs to be put intoensuring the drone is able to function properly regardless of its electromagneticenvironment. Not to mention ensuring that the drone itself does not interferewith the environment around it, which is especially important if used in a moreresidential setting.

While drones and other electrical aircraft are presently not set up to overtakeall facets of the aviation industry, future implementations would certainly helpreduce emissions compared to only using fossil based aircraft. In 2018, CO2

emissions from commercial operations of aircraft totalled to 918 million metrictonnes or 2.4 % of global CO2 emissions from fossil fuel use [28]. Unmanned(and manned) electrical aircraft could help reduce these emissions by provid-

1

ing a more cost-effective and sustainable way of transporting goods and peoplealong shorter flight routes [35]. Although EMC is certainly not the only thingthat needs to be considered for drones and electric aviation to become viable,ensuring proper EMC is an important aspect of making them commerciallyviable and safe for future implementations.

1.2 Problem description

There are many rules and regulations concerning aviation, not to mention thespecific ones regarding unmanned aircraft. This thesis aims to consolidate theinformation available to provide a better overview over the exact rules (i.e di-rectives/standards) regarding EMC that an unmanned electric drone should actin accordance with. The results of this research is to be applied to two differ-ent systems; an electric vertical take-off and landing (VTOL) drone prototypemade by company A and a drone equipped with a 3D mapping module made byFieldRobotiX. Specifically, this means finding directives and standards basedon the intended operation of the drones and their technical specifications. Sowhile the initial research deals with the overarching legislation regarding EMCfor drones, the application then involves using this framework to apply rulesapplicable specifically for the company A and FieldRobotiX systems. Based onthe applied directives/standards, EMC tests are then to be carried out on thedrones to evaluate their EMC performance. However it should be noted thatthese tests should be regarded as pre-compliance measurements more than any-thing else. This is simply because the purpose of the measurements is not tooprove absolute compliance of the drones in regards to EMC, but rather identifyproblematic areas that might reasonably impact future (more official) compli-ance testing. The results of these measurements will also be used to discusspositive as well as negative aspects of the design of the drones along with futureimprovements that could be made. As such the goal of the measurements is tohelp FieldRobotiX and company A improve their designs and make them awareof any shortcomings (or lack thereof) when it comes to the EMC of their drones.

1.3 Scope

The research regarding EMC legislation for drones will be conducted from aEuropean point of view, meaning that all legislation and applicable rules will befor the EU market. This is done to reduce the scope of the research in order tomake it more manageable to write a concise guide on the applicable legislation.As all relevant parties involved in this thesis are based in the EU, it was alsodeemed the most relevant market to cover.

This thesis will also limit itself to only dealing with EMC aspects of the drones.Aspects such as hardware design, component choices and operational efficiencywon’t be considered unless they have (or are suspected to have) a tangible im-pact on the EMC of the drone. This also means that the literature review andfollowing ’guide’ on EMC for drones will only consider the aspects that regardEMC. Certifying drones is an involved process that demands more criteria to bemet than just those regarding EMC. So to keep the scope within realistic bounds,many important aspects of for example drone certification/air-worthiness maybe overlooked.

2

Another limitation is that this work will be focused on dealing with functionality-related EMC. More specifically this means EMC in relation to the performanceand functionality of the equipment rather than health and safety considerations.In this context ’health and safety’ means for example the hazards of exposureto electromagnetic/radio waves on people and animals. Although one couldcertainly argue that the proper function of a drone when exposed to EMI is asafety related topic due to the possible damage caused by for example a crashingdrone. This particular area of EMC is still not discussed at length as the au-thor is not educated enough in the field of aviation safety to provide meaningfulinsight and/or suggestions on design in regards to safety aspects. Instead, thefocus is on analysing functionality of the components in the drones.

1.4 Contribution

Studying EMC and the effects of EMI on equipment is not a foreign conceptto the aviation industry. Papers such as ”EMI/EMC for military aircraft andits challenges” by Shukla, A K and Nirmala, S [36] describe testing methodol-ogy, causes for EMI and countermeasures all in the context of military aircraftapplications. Another project that should be highlighted is the work done byAW-Drones, who’ve created a search portal which ”... is an open repositorythat collects the technical standards, published or under development, for thecommercial use of drones worldwide.” [1]. The work they’ve done could be seenas a much more large-scale version of the work presented in this thesis. Insteadof focusing on EMC for drones they’ve analyzed drone operations as a wholeand proposed ”the most suitable technical standards for all relevant categoriesof drones operations.” [2]. It should however be noted that this work done byAW drones was not done specifically for EMC standards and as such it doeswarrant a further exploration of this particular field of drone usage.

Actually applying EMC regulations onto drones is not a new concept either,multiple accredited EMC labs like TUV SUD and Testups advertise their ser-vices for certifying drones, which involves being compliant with the relevantEMC directives. However the lack of specific EMC standards for drones (in theEU at least) makes this a more complicated process as the rules are not tailormade for drones. While it is possible to find relevant standards that have beenused before (see for example DJI’s declaration of conformity for one of theirdrones [5]), due to differences in hardware and application one cannot alwayssimply copy what has been done before when trying to prove that ones prod-uct is EMC compliant. Instead of covering a single area such as testing, thisthesis aims to analyse the existing legal framework regarding EMC for dronesspecifically and how one can go about choosing relevant standards based on thehardware of the equipment and then evaluating its performance through testing.In other words, the goal is to provide is a full analysis from legislation to testresults.

1.5 Outline

Chapter 2 explains some basic EMC theory that is relevant to understandingthe testing methodology and any suggested improvements of the drone design

3

in regards to EMC.

Chapter 3 presents the overall workflow for determining the tests that shouldbe performed on the company A and FieldRobotiX systems. It takes on thefollowing structure:

1. Explain how to choose the proper directive to follow when showing EMCcompliance for a drone. This information is summarised in the form of a’guide’ which is the result of the initial research into EMC legislation fordrones.

2. Apply the aforementioned guide on the drones to determine what directivethey should adhere to.

3. Based on hardware, intended usage and an EMC risk analysis, determinewhat standards should be used during testing.

4. List testing equipment and perform a practical analysis on what tests arefeasible and relevant to perform.

Chapter 4 showcases the results from the performed EMC tests.

Chapter 5 presents a discussion on the validity of the results as well as thedrones performance in regards to EMC. Here is also where any improvementsare suggested.

Chapter 6 shows the conclusions drawn from the tests but also from the ini-tial research on EMC legislation regarding drones. Along with this, suggestionsfor future works are also presented.

4

2 Theory

2.1 Literature review

The initial stage of the research was focused on condensing the available EMClegislation in the context of drones. In practice this meant reading some pa-pers on EMC for aircraft [42][36] but also articles and advertisements postedby EMC testlabs and aviation agencies/corporations to try and find out whatrules are applicable for unmanned aircraft in general and how they relate toEMC. This also involved reading through the proper European legislation (seefor example [25] and [27]) to see exact wording on how the rules should applyand how different directives interact with each other. The main sources for in-formation regarding drone/aviation legislation came from the European UnionAviation Safety Agency (EASA) and the European Commission/Parliament asthey provide the most official information available in the EU.

In regards to literature, the one book used for this thesis was EMC for ProductDesigners by TimWilliams [41]. Williams has covered many of the essential top-ics when it comes to EMC legislation, theory and testing. The book was mainlyused for the information regarding EMC testing methods and equipment alongwith the presented theory around disturbance coupling. While aviation is men-tioned in the book, it is not explored in detail and drones are not mentioned atall and as such the book was not used as a basis for the legislative research.

2.2 EMC testing

When testing the EMC of a piece of equipment, there are generally four areaswhich need to be tested: Radiated emissions, radiated immunity and conductedemissions and immunity. As the name suggests, radiated emissions are in regardsto ability of the equipment to generate and radiate EM waves to its environment.Radiated immunity on the other hand is about gauging the defense against suchincident electromagnetic waves. Conducted emissions and immunity insteadconcerns interference being transferred through the ports of the equipment dueto its internal interference sources (or those of connected equipment). Thiswould include the conducted interference through the mains port if it is pluggedinto a mains power supply or other source which could be susceptible to suchinterference (not a battery for example). Conducted immunity would includeprotection against things such as electrostatic discharge (ESD) and electricalsurges coming from the mains network which the equipment is connected to.To simplify the matter, one could think of it as radiated interference beingcoupled through for example the surrounding air while conducted interferenceis through for example cabling.

2.2.1 Anechoic chamber

In this thesis the tests on radiated emission and immunity will be carried outin the fully anechoic room (FAR) at Lulea University of Technology (LTU). Ananechoic room is a shielded chamber which not only keeps out outside distur-bances (such as local radio broadcast frequencies) but also makes sure emittedelectromagnetic waves stay within the chamber (from both the equipment un-der test and the measuring antenna). The lab in question uses the method of

5

covering the walls, ceiling and part of the floor (semi-anechoic would entail notcovering the floor and dealing with the reflections) with RF absorbing foampyramids to avoid internal reflections within the chamber. The part of the floorwhich is not covered in foam pyramids instead acts as an absorbing plane.

2.2.2 Shielding

Shielding is a common practice in EMC design when one wants to reduce the ra-diated emissions from a piece of equipment or mitigate incident electromagneticinterference. The most simple example would be placing a metal (or otherwiseconductive) ’box’ around the equipment. A basic example here being that in-cident magnetic fields would induce currents in the shielding which would thencreate a magnetic field counteracting the original one in accordance with Lenzlaw [3]. However, it is usually not feasible to place equipment in a solid metalbox as the equipment needs to interface with the user or other devices in itsvicinity.

Cables can also be screened, as outlined by Tim Williams in [p.387, 41] us-ing a conductive screen on a cable is an effective method to reduce radiatedinterference caused from signals in the cables themselves and hinder commonmode currents from coupling from external sources into the signal source orcoupling common mode currents from the signal source to elsewhere. Choosinga proper screening setup is however not always straightforward, as dependingon the frequency and type of interference that needs to be dealt with the screenmay need to be grounded at both or one end [p.391, 41]. Choosing the wrongsetup may prove ineffective and at worst detrimental to the design. However,designing slits and openings in enclosures to reduce leakage of electromagneticinterference and choosing proper cable screening setups is a science all of itsown and this thesis is not concerned with the finer details.

2.2.3 Detector types

When it comes to measuring radiated emissions, there are three main methods ofdetecting disturbances: Average, Peak and Quasi-peak. The two detectors rele-vant to this thesis are peak and quasi-peak. Peak measurements are as the nameentails used to measure the peaks of the signal at the specific frequency and isgenerally the fastest of the three methods and is useful when performing quickpre-compliance measurements. Generally, peak levels are higher than measure-ments performed with a quasi-peak detector. As defined by Tim Williams in[p.146, 41], quasi-peak measurements dwell longer on each frequency and useweighted charge and discharge times to not be as punishing for disturbanceswhich have a low repetition frequency (i.e the pulses occur with further timebetween them). The reason given for this being that ”low pulse repetition rates(PRFs) is said to be subjectively less annoying on radio reception than at highPRFs”.

6

3 Method

3.1 Legislation regarding EMC of drones

3.1.1 Finding directives

In the EU there exists a council directive 2014/30/EU which outlines the gen-eral laws to follow in regards to EMC [25]. However it is stated in introductoryParagraph 11 of this directive that ”Aircraft or equipment intended to be fit-ted into aircraft should not be covered by this directive, since they are alreadysubject to special Union or international rules governing electromagnetic com-patibility.”.

To get the specifics, regulation 2019/947/EU ”on the rules and procedures ofunmanned aircraft” [24] goes over general rules regarding for example risk as-sessment of an unmanned aircraft systems (UAS) operation. More importantlyit also outlines the different classes for drones, these classifications are open,specific and certified. Exact information on what defines these classificationsare in Article (4) – (6) of 2019/947/EU but in short they are mainly basedon the risk of operation where open is the ’lowest’ classification. As outlinedby the EASA [7], the risk of operation is gauged by both the attributes of thedrone (for example weight) and by its intended usage (for example transportinggoods), note also that they draw no distinction between commercial and per-sonal drones as the classification is based on risk of operation. The descriptionof the classes provided in this thesis should be regarded as a simplification andproper care needs to be taken on a case-by-case basis when determining a dronesclass. Regulation 2019/945/EU on ”unmanned aircraft systems and on third-country operators of unmanned aircraft systems” [23] provides design rules forthe classes defined in 2019/947/EU. For example introductory Paragraph 46 of2019/945/EU states that a drone in the specific class can still need certifica-tion if the risk assessment of the operation deems that the risk ”...cannot beadequately mitigated without the certification of the UAS”. The point herebeing that it is possible for a UAS to be outside the certified class but stillneed certification in certain cases.

The importance of a drone’s class becomes apparent when looking at EU reg-ulation 2018/1139/EU covering ”common rules in the field of civil aviation...”[27]. This regulation covers many topics regarding civil aviation and this in-cludes EMC. Introductory Paragraph 87 of this regulation states that from themoment the UAS is ”subject to certification in accordance with this Regula-tion” (from what has been gathered this includes being classified as ’certified’pursuant to 2019/945/EU and 2019/947/EU) it is also subject to essential re-quirements regarding efficient use of the radio spectrum and immunity to elec-tromagnetic disturbances. These requirements are equivalent to those presentedin the EMCD and in the radio equipment directive 2014/53/EU [26] (RED).More specifically these requirements are (taken from Paragraph 2.5 of Annex IIin 2018/1139/EU)

Those unmanned aircraft, engines, propellers, parts and non-installedequipment shall be designed and produced in such a manner, having re-gard to the state of the art, as to ensure that:

7

(a) the electromagnetic disturbance which they generate does not exceedthe level above which radio and telecommunications equipment orother equipment cannot operate as intended; and

(b) they have a level of immunity to the electromagnetic disturbancewhich allows them to operate without unacceptable degradation oftheir intended use.

Those unmanned aircraft, engines, propellers, parts and non-installedequipment shall be designed and produced in such a manner, having re-gard to the state of the art, as to ensure that they effectively use andsupport the efficient use of radio spectrum in order to avoid harmful in-terference.

These requirements are almost verbatim the requirements presented in the Gen-eral requirements in Annex I of the EMCD and Article 3.2 of the RED. In otherwords, even if the UAS would be exempt from these directives they would stillneed to comply with some requirements hailing from them.

Furthermore, 2018/1139/EU amends both the EMCD and the RED now statingthat aviation equipment and unmanned aircraft are to be excluded from thesedirectives (manned are still always excluded) if they are certified by the Euro-pean union Aviation Safety Agency. According to the introductory Paragraph87 of 2018/1139/EU the motivation behind this is that certification is done inaccordance with the 2018/1139/EU regulation and thus meets the previouslymentioned essential requirements on EMC and usage of the radio spectrum. Asstated in the same paragraph, these exclusions are only applicable if the drone iswithin the scope of the regulation and the UAS/equipment is intended only forairborne use on protected aeronautical frequencies. Otherwise the drone wouldface the rules of both the 2018/1139/EU regulation and the Directives. How-ever as previously mentioned the regulations essential requirements on EMCand radio spectrum usage are very similar to the EMCD and RED, so from anEMC point of view it would be logical to assume that not much would changecompared to only following the EMCD or RED.

Even if a drone in the open or specific category (i.e not certified) falls withinthe scope of 2018/1139/EU it is not automatically exempt from the EMCDor the RED. Specific wording can be found in regulation 2019/945/EU. Alongwith providing design rules for the classifications defined in 2019/947/EU, thisregulation also specifically states conditions for exemption from the EMCD andRED (taken from introductory Paragraph 8 and 9 of 2019/945/EU)

Directive 2014/53/EU should apply to unmanned aircraft that are notsubject to certification and are not intended to be operated only on fre-quencies allocated by the Radio Regulations of the International Telecom-munication Union for protected aeronautical use, if they intentionally emitand/or receive electromagnetic waves for the purpose of radio communi-cation and/or radiodetermination at frequencies below 3 000 GHz.

Directive 2014/30/EU should apply to unmanned aircraft that are notsubject to certification and are not intended to be operated only on fre-quencies allocated by the Radio Regulations of the International Telecom-munication Union for protected aeronautical use, if they do not fall withinthe scope of Directive 2014/53/EU.

8

In other words (from the official Guide for the EMCD [p.14, 29]), if the UASis not subject to certification and does not intend to operate exclusively onprotected aeronautical frequencies, the RED will apply if the aircraft intention-ally emits or receives radio waves (i.e frequencies below 3000 GHz as definedin the RED) for communication or position/velocity determination. Otherwisethe EMCD will apply instead. But what if the UAS is made to operate onprotected frequencies but not subject to certification? As previously mentionedintroductory Paragraph 87 of 2018/1139/EU states that only equipment whichhas been certified by the EASA is exempt from the EMCD/RED. This wouldthen indicate that the UAS would not be exempt from these directives even ifit used only protected frequencies.

So in short, a drones EMC requirements are based on the class of the droneas well as its radio-capabilities and frequency usage. So a proposed first stepwould be to try and gauge which class the drone in question would fall into. Tothis end the EASA provides information regarding classification on their websiteand in their regulations 2019/947/EU and 2019/945/EU. Presented in figure 1below is a simple flowchart to determine which directive(s) should be followedfor a drone. Note that while the flowchart may present for example the EMCDas the most relevant directive, this does not mean the drone is exempt from the2018/1139/EU directive as it contains other regulations for areas besides EMC.However, from a strict EMC point of view the EMCD (in this example) wouldbe the more proper directive to adhere to when wanting to prove conformity.

Figure 1: Flowchart for choice of directive when wanting to prove conformitywith EMC legislation.

9

3.1.2 Application of standards

Before continuing with choosing appropriate standards, a short explanation onharmonised standards. Section 4.3.2 of the Guide for the EMCD [29] defines aharmonised standard as one which is published in the Official Journal of the Eu-ropean Union (OJEU) and thus benefits from presumption of conformity withthe relevant directive that the standard is harmonised under. These standardsinclude testing methodology (i.e how to set up the testing environment and howto test the equipment) and specific limits on radiated/conducted emissions forbeing compliant with the essential requirements of the directive (the EMCDin this case). In other words, if one follows the methodology and meets therequirements presented in a harmonised standard the product will be presumedto meet the associated essential requirement from the directive the standard isharmonised under.

In the list of harmonised standards for the EMCD[30] and the RED [31], thereare no product specific standards for aircraft or aviation equipment. From whathas been gathered, there are also no harmonised standards for drones underdirective 2019/945/EU nor 2018/1139/EU, meaning there are no easy stan-dards to follow for presumption of conformity regardless of whether the dronein question is exempt from the EMCD/RED or not. However, while harmonisedstandards simplify the process of compliance, one is not strictly required to fol-low them. As stated in an article by FORCE TECHNOLOGY [32] ”In mostcases the directives allow the manufacture [sic] to argue compliance based onrelevant standards found outside the lists of HS [Harmonised Standards] for thedirectives or a selection of relevant tests to supplement standards, which applyin part.”. The question is then what standards to apply to drones.

Again, there are no dedicated standards for drones yet so it is up to the tester orEMC lab to judge which kinds of standards to apply based on application andthe technical specification of the drone. As recommended by an EMC test lab,if a drone is to be used for military/defense purposes the US standard DO-160G(or the ED14G standard, which is the EU version) ”Environmental Conditionsand Test Procedures for Airborne Equipment” could be used. This standarddoes not specifically cover drones but goes over many environmental considera-tions regarding aircraft (not just EMC) and presents rigorous requirements fora drone to follow [6]. For civil applications, the test lab suggested EN 301-489-Xseries of standards or the EN 55032 [37] and EN 55035[38] standards. The stan-dard EN 301 489-1 ”ElectroMagnetic Compatibility (EMC) standard for radioequipment and services; Part 1: Common technical requirements; HarmonisedStandard for ElectroMagnetic Compatibility ” is like a ’base’ standard whosemethodology and test setup the latter standards in the series refer to. The latterstandards are applicable based on the type of equipment in question. As for EN55032 and EN 55035, they are applicable for multimedia equipment and presentlimits on emissions and requirements on immunity respectively.

3.2 Legislation for the company A drone

The first drone to discuss is the prototype company A unmanned, electric VTOLdrone as seen in Figure 2. Based on the information presented in the EASA:s

10

”Easy Access Rules for Unmanned Aircraft Systems” [8], this drone is deemedto be within the specific class of operation. This tentative classification is ofcourse subject to change to certified if the risk assessment associated with thedrones operation concludes that certification is needed.

Figure 2: The company A drone fastened inside the EMC chamber.

11

Article 4 of 2019/947/EU [24] states criteria for a drone operating under theopen classification. One of these is that the operation should be one where ”theremote pilot keeps the unmanned aircraft in VLOS [Visual Line of Sight] at alltimes except when flying in follow-me mode or when using an unmanned aircraftobserver as specified in Part A of the Annex;”. The definition of ”remote pilot”is from Article 3, Paragraph 31 in 2018/1139/EU [27]: ” ‘remote pilot’ meansa natural person responsible for safely conducting the flight of an unmannedaircraft by operating its flight controls, either manually or, when the unmannedaircraft flies automatically, by monitoring its course and remaining able to in-tervene and change the course at any time;”. Therefore the previous criteria isnot strictly based around an operator manually controlling the UAS at all times.

This criteria is not fulfilled by the company A drone as it will operate beyondvisual line of sight (BVLOS). Since all criteria of Article 4 in 2019/947/EU needto be fulfilled, this is enough to move the drone from open to specfic. As towhy the drone tentatively isn’t in need of certification, Paragraphs 1 (a) – (c)of Article 6 in 2019/945/EU [23] state the following:

The design, production and maintenance of UAS shall be certified if theUAS meets any of the following conditions:

(a) it has a characteristic dimension of 3 m or more, and is designed tobe operated over assemblies of people;

(b) it is designed for transporting people;

(c) it is designed for the purpose of transporting dangerous goods andrequiring a high level of robustness to mitigate the risks for thirdparties in case of accident;

None of these criteria are met for the company A drone, and as such the opera-tion of the drone is placed within the specific class. Referring to the flowchartin Figure 1, the company A drone should then follow the radio equipment direc-tive 2014/53/EU [26] since it has radio receiving and transmitting capabilities.In regards to EMC, in Paragraph 1 (b) of Article 3 in the RED (essential re-quirements) it is stated that ”[Radio equipment shall be constructed so as toensure:] an adequate level of electromagnetic compatibility as set out in Direc-tive 2014/30/EU.”. Therefore its not unreasonable to assume one needs to lookto the essential requirements of the EMCD to get an idea of what this means.More specifically, this entails looking at the General requirements of the essen-tial requirements, the other part is ”Specific requirements for fixed installations”whereas the previous part also applies for apparatus. The general requirementsof the EMCD are as follows (Paragraph 1 in Annex I of the EMCD)

Equipment shall be so designed and manufactured, having regard to thestate of the art, as to ensure that:

(a) the electromagnetic disturbance generated does not exceed the levelabove which radio and telecommunications equipment or other equip-ment cannot operate as intended;

(b) it has a level of immunity to the electromagnetic disturbance tobe expected in its intended use which allows it to operate withoutunacceptable degradation of its intended use.

Note the similarity to the essential requirements from 2018/1139/EU presentedin Section 3.1.1. As stated in Section 1.4.2.1 of the EMCD guide [29], if a

12

drone is within the scope of the RED it is then exempt from the scope of theEMCD. Meaning that the EMCD needs to be considered when it comes toelectromagnetic compatibility, however one should not refer to the EMCD whenperforming the EU Declaration of Conformity. The proper route to take in thiscase would be to use (ideally) harmonised standards under the RED that arestated to fulfill the requirements of Article 3.1(b) of the RED.

3.2.1 Choosing standards

Before proceeding with choosing standards its beneficial to conduct a risk anal-ysis on the drone from an EMC perspective. This means that the analysis doesnot necessarily cover safety aspects but rather risk areas that might cause poorEMC for the drone and make it violate the essential requirements of the (in thiscase) RED. The company A drone has many planned use cases, including cargotransport, surveillance and scanning of infrastructure. So the drone is plannedto operate in many different types of environments, which in turn places a highimportance on proper EMC as one cannot foresee the amount and intensity ofinterference at each location. These kinds of operations, and really drones op-erating in general, could face interference from things such as cell towers, radarand lightning [39]. While these are indeed serious sources of interference, highintensity radiated fields (HIRF) disturbances are unfortunately not possible tomeasure with the equipment available at the EMC lab at LTU due to theirhigh frequency and also the power needed to generate the desired fields at thosefrequencies during immunity testing. However these potential sources of inter-ference should still be noted.

As one would expect, communication is key when operating a drone and thisis even more important when it comes to BVLOS operations. The company Adrone uses a combination of radio, 4G and satellite communications modulesto handle external communications for the drone. Note that having a systemconsist of EMC compliant parts doesn’t necessarily mean that the system itselfis EMC compliant. As such, in a proper compliance test it is still required totest the parts in their intended configuration as part of the system. Since thedrone is supposed to adhere to the essential requirements of the RED, one wantsto (if possible) use standards that are harmonised under this directive. To thisend, standards out of the EN 301 489-X series of standards for radio equipmenthave been chosen as the series itself was specifically recommended when askingan EMC test lab about applicable standards and the series is specifically statedto meet the requirements of the EMC-part of the REDs essential requirements.These standards have also been used before by companies such as DJI whenshowing compliance for some of their drones [5]. Table 1 shows which standardswere chosen for each part of the drone. The choice here is mainly based aroundfrequency usage, or in other words choosing standards that cover the intendedoperating frequency of the equipment. This of course comes with the restrictionthat the standard should be applicable for the specific equipment to begin with(i.e not choosing a standard for GNSS for a simple radio receiver).

One thing to point out with the standards listed in table 1 is that while theEN 301 489-X series is said to cover the EMC part of the essential requirementsof the RED, it is actually stated in clause 8.1.1 of EN 301 489-1 [17] that radiatedemissions from antenna and enclosure ports are covered in standards meeting

13

Part Applicable standard(s) References

433MHz radio EN 301 489-3, EN 300 220-1 [22] [15]450MHz 4G EN 301 489-50/52, EN 300 328 [12][13][16]GNSS EN 301 489-19, EN 303 413 [18] [19]Other satellite communication EN 301 489-20 [11]

Table 1: Table over relevant standards for parts of the company A drone.

the requirements for usage of the radio spectrum. In other words, the limits forradiated emissions from specific radio equipment are found in standards whichare harmonised to meet the essential requirements for efficient usage of the ra-dio spectrum rather than the requirements on EMC. Though it should still bementioned that EN 301 489-1 still places requirements on the radiated emis-sions from the ancillary equipment connected to the radio equipment, wherethe standard refers to EN 55032 ”Electromagnetic compatibility of multimediaequipment - Emission Requirements” [37] for the limits. For the sake of trans-parency, it should be said that the standards covering efficient use of the radiospectrum have not been thoroughly studied or considered for the company Adrone (or the FieldRobotiX system for that matter) and are only mentionedto show where the specific radiated emissions levels can be found for differentpieces of equipment. The reason for this being that while radiated emissions isa key part of EMC, additionally studying efficient use of complying with theradio spectrum is deemed to be outside the scope of this thesis.

It should also be noted that the standards listed above do not paint a com-plete picture of the actual amount of standards used. As mentioned previously,the EN 301-489-1 standard for radio equipment acts as a sort of ’base’ standardfor the EN 301-489-X series, meaning that the rest of the entries in the seriesgive more specific rules for certain equipment while the ’base’ outlines for ex-ample the general testing methodology. This list of documents needed to applythe standard is usually referred to as ”Normative references”, and in turn theEN 301-489-1 standard also has normative references to (among others) productfamily standard EN 55032 [37] and some generic standards in the CENELECEN 61000-X-X series which outline baseline EMC measurement techniques usedwhen performing compliance testing. The point here is to illustrate that it maynot be as simple as following a single standard when performing EMC compli-ance testing.

3.3 Legislation for the FieldRobotiX system

In addition to the company A drone, testing is also to be done on a systemprovided by FieldRobotiX (shown in Figure 3). While the drone itself wasnot built by FieldRobotiX, they have built a module which when fitted ontothe drone allows autonomous surveillance and 3D-mapping of its surroundings.Although the drone it’s attached to (DJI Matrice 300) is available on the marketand is (at least supposed to be) compliant with the relevant EMC requirements,it is still highly relevant to check to what degree the addition of the module byFieldRobotiX will impact the EMC performance of the drone. As previouslystated, just because a piece of equipment consists of compliant parts does not

14

mean the resulting system is in itself compliant. Therefore it is still of interestto examine whether the addition of the module will make the drone no longerbe compliant with the relevant test levels.

Figure 3: FieldRobotiX system.

In regards to the class of the drone and which directive to follow, the moti-vation is quite similar to the company A drone. In other words, the system isregarded as tentatively being in the specific category of operations for the samereasons of operating in BVLOS but not quite certified as the drone is for exam-ple not designed to transport people. In turn, this leads to the proper directiveagain being the RED and it should be mentioned once again the classificationis subject to change depending on specific usage of the drone. However unlikethe company A drone where the whole system is a prototype, the FieldRobotiXas previously mentioned uses a commercially available DJI drone. With this inmind, it becomes a bit easier to find applicable standards for the system notonly because there is some expected overlap with the standards found applicablefor the company A drone but also because testing has been done before on partsof the system.

As the product in this case is a module which itself is attached to an alreadycommercially available drone, it would be proper to look at the standards againstwhich this drone has previously been tested. Luckily, DJI has an open repos-itory of their declarations of conformity for their products so one can easilyfind what directives and EMC standards have been used [4] (but not how theyhave been applied). Additionally, for the module itself the relevant standardfor the LiDAR seems to again be the EN 55032 [37] and EN 55035 (”Elec-tromagnetic compatibility of multimedia equipment - Immunity requirements”[38]) standards for emissions and immunity measurements on multimedia equip-ment. This is again based on looking up the declaration of conformity for theVelodyne Puck Lite LiDAR (or as close to it as possible) [34]. Note also theoverlap between the standards chosen for the prototype company A drone and

15

the commercially available DJI drone. Table 2 presents the relevant standardsfor the FieldRobotiX system based on looking up the standards applied to thedrone by DJI and the manufacturers applied standards for the LiDAR.

Part Applicable standard(s) Reference

- EN 301 489-1 (Base standard) [17]Radio control EN 301 489-3, EN 300 440 [22] [21]Broadband Transmissions EN 301 489-17, EN 300 328, [20] [16]

EN 301 908-1 [14]GNSS EN 301 489-19, EN 303 413 [18] [19]Multilateration equipment EN 303 213-5 [10]LiDAR EN 55032 and EN 55035 [37] [38]

Table 2: Table over relevant standards for the FieldRobotiX system.

3.4 Testing environment

3.4.1 Equipment

The radiated emissions/immunity tests will be performed in the FAR at LTUsEMC lab. The testing environment consists of the equipment under test (EUT)placed on a table which itself is on a rotating floor tile 3 m from a bilog antenna.The voltage from the measured field is amplified as well as filtered by a separateadd-on connected to the antenna. This is to filter out noise and to be ableto discern smaller sources of interference from the noise floor. The add-on isplaced close to the output of the antenna (i.e beginning of cable) so as to notamplify the noise being picked up when travelling through the cable to themeasurement receiver. The testing environment with a horizontally polarisedmeasuring antenna can be seen in Figure 4 (albeit with the table rotated) alongwith the FieldRobotiX system.

16

Figure 4: Testing environment inside the FAR.

For measuring the radiated fields, a Rohde & Schwarz ESPI test receiver isused along with software on a nearby PC whose functionality includes changingof test settings and displaying the test results. For immunity testing, a test sig-nal is required to create the desired field within the chamber. For this purposean Agilent Technologies E8267D PSG Vector Signal Generator is used. As thesignal generator by itself is not powerful enough to generate the required fieldstrength for each frequency during radiated immunity testing, three Prana poweramplifiers are used. Having a single amplifier cover all the necessary frequencieswould be very costly, so these three power amplifiers each cover a separate fre-quency band (80 MHz - 1GHz, 1GHz - 2 GHz and 2 GHz - 6 GHz respectively)and the amplifiers are automatically switched between (by the computer soft-ware) when sweeping the frequencies during immunity testing. Additionally, aRohde & Schwarz NRVD Dual-Channel power meter is used during immunitytesting to monitor the power going into the antenna and measuring the powerreflected back. These measurements are used to calculate the Voltage StandingWave Ratio (VSWR), which is what’s used to make sure enough power is be-ing supplied to actually generate the desired field strength within the chamber.This is achieved through the computer software communicating with the powermeter and changing the output of the signal generator to compensate for thelosses (images of the equipment can be found in Appendix 7.1). Finally, to ob-serve the EUT inside the FAR, there is a video camera placed within the FARconnected to a screen outside providing real-time visual observation.

3.4.2 Positioning the company A drone

When performing EMC measurements, it is usually only required to test theEUT in its intended orientation or any foreseeable orientation by the manufac-turer. For example, a wall mounted device would be tested in the same positionit would normally be operated in with the display facing the user/antenna. This

17

is not the case when it comes to the company A or FieldRobotiX systems, bytheir nature the drones will be exposed to as well as expose other devices toEMI in all directions. What this means is that it is not entirely sufficient toperform measurements with the drones only in a horizontal position on a table.While relevant for both drones, due to time constraints only the company Adrone received a solution to this problem. The proposed solution was to designand construct a holder to position the drone in a vertical position, this com-bined with the fact that the floor of the EMC chamber has a rotational floortile means nearly all possible angles can be achieved.

Figure 5 shows the holder, where my internal and lab supervisors aided withdesigning it and construction was done by myself and my internal and lab su-pervisors. The design is rather straightforward; The drone simply rests on itstail-end on a piece of styrofoam for support while the body rests on the twovertical pillars where it is also fastened using straps. Thankfully, the drone istall enough for the holder to be floor mounted and still have the relevant elec-tronics reach high enough to face the antenna across the chamber. One thing topoint out is that metal screws were used when constructing the holder, which isnot ideal as the metal can act as antennas (i.e themselves radiate interference)if exposed to electromagnetic fields. However, this contribution should not beexcessive and it was deemed a worthwhile trade-off for easier construction ofthe holder. Seeing as most of the screws are also located close to the floor andnot right in front of the measuring antenna, their impact is further diminished.

Figure 5: The drone holder used to fasten the company A drone vertically.

18

3.5 Test plans and experimental setup

Although there are some differences in the test plans for each drone, some thingsremain the same for both products. For each plan, although necessary, the elec-trostatic discharge (ESD) tests have been purposely omitted due to the verypalpable risk of breaking components inside the equipment. Another aspectthat has not been taken into consideration is conducted emissions testing andvariations/interruptions in supply voltage. This is due to the fact that bothdrones are not operating through an AC mains power source and as such thesekinds of tests are unnecessary. Though, one could argue that the act of hav-ing to charge the batteries could warrant such testing but in this case it couldbe seen more as placing requirements on the charger rather than the drone itself.

It should be noted that while care has been taken into making the test plansprovide useful results, they should by no means be regarded as exhaustive andcomprehensive. Due to the nature of the equipment comprising of multiple com-plex parts which in and of themselves would warrant an EMC examination andthe lack of time and expertise has led to the test plans providing more surfacelevel information than extensive performance information. In other words, eachcomponent such as GPS, radio, Wi-Fi module etc. have their own standards,test limits (see Tables 1 and 2) and constraints, and they should ideally betested separately (but still as part of the system). The test plans have insteadbeen constructed in such a way as to try and cover the system as a whole whilestill providing relevant limits. Furthermore, it should once again be empha-sized that the measurements are not done to prove absolute compliance of theEUTs in regards to EMC. Instead, the purpose of the measurements (or pre-compliance measurements) is to identify problematic areas of the design whichshould be dealt with before moving on to actually having the drones performofficial compliance testing.

3.5.1 FieldRobotiX system

For the emission measurements, the performance of the system is to be gaugedagainst the class B limits defined in Table A.4 of EN 55032 [p.26, 37] as this wasdeemed (from the prior research) to be the most general and broadly applicablestandard for emissions. While the class B limits are for equipment primarilyintended for use in a residential environment and the drone has its main usein industrial environments, its still proper to test the drone against the moredemanding limits of class B to have some margins of error and have it be moreprepared for widespread usage. For comparisons sake, the class A limits for the30-230 MHz band when using the same setup is 52 to 45 dBµV/m and for 230- 1000 MHz it’s 52 dBµV/m.

Freq. range [MHz] Distance Detector/bandwidth Limits [dBµV/m]

30 - 230 3m Quasi-peak, 120 kHz 42 to 35230 - 1000 3m Quasi-peak, 120 kHz 42

Table 3: Class B limits for equipment tested in a FAR up to 1 GHz.

Based on the highest internal frequency of the system, it would actuallybe required to measure radiated emissions up to 6 GHz as detailed in Table 1

19

of EN 55032 [37]. However, the measurements took longer than expected andbecause sources of interference could already be identified below 1 GHz it wasdecided to omit testing above this frequency in favor of spending time perform-ing more tests on different configurations.

In order to find the maximum levels of emissions, multiple tests were performedwith different configurations on the equipment and the antenna inside the cham-ber. More specifically, the methodology is as follows:

1. Vertically polarize the measuring antenna.

2. Perform measurement with the system facing the antenna.

3. Turn parts of the system on and off to identify contributions to the emis-sion spectrum.

4. Perform additional test where the table inside the chamber has been ro-tated in 45 degree increments. See if identified sources of emissions havebeen increased or decreased.

5. Horizontally polarise the measuring antenna and try to identify newlydetected sources of emission (if present).

The experimental setup for the FieldRobotiX system inside the FAR is shownin Figure 6, where the drone is tied to a table using straps and pieces of woodto fasten it to the table. The table itself places the system at a height of 0.8m above the ground plane of the FAR and at a distance of 3 m away fromthe measuring antenna which itself is placed 1.3 m above the ground plane foremission measurements and around 1.8 m for immunity testing.

Figure 6: Placement of the FieldRobotiX system inside the FAR during emissionand immunity measurements.

20

As for the radiated immunity tests, the test levels are based on those pre-sented in the EN 301 489-1 standard [17]. The choice of standard was onceagain based on it being general enough to be applied to the whole system (whilestill being relevant). Of course one could also use the EN 55035 [38] immunitystandard for multimedia equipment (the counterpart to EN 55032 for emissionrequirements), but I wanted to try and use a more diverse set of standards fromdifferent sources so I chose EN 301 489-1. As specifically stated in clause 9.2.2 ofEN 301 489-1 ”the test level shall be 3 V/m (measured unmodulated). The testsignal shall be amplitude modulated to a depth of 80 % by a sinusoidal audiosignal of 1 000 Hz. If the wanted signal is modulated at 1 000 Hz, then an audiosignal of 400 Hz shall be used; ”, additionally the test should sweep throughfrequencies between 80 – 6000 MHz in 1 % intervals. But how is the successof the immunity test verified? There can be some variations in the specificsdepending on the type of equipment, but as a general rule of thumb in this casethe following definition for class A performance criteria from EN 55035 [p.29,38] is used:

The equipment shall continue to operate as intended without operatorintervention. No degradation of performance, loss of function or changeof operating state is allowed below a performance level specified by themanufacturer when the equipment is used as intended. If the minimumperformance level or the permissible performance loss is not specified bythe manufacturer, then either of these may be derived from the productdescription and documentation, and by what the user may reasonablyexpect from the equipment if used as intended.

As no performance requirements were provided prior to testing, the require-ments are derived based on what the customer probably expects when using thesystem. In this case it would probably be that the drone and attached moduleshould continue operating and map its surroundings irregardless of the applieddisturbances from the immunity test. However it should be noted that thesurveillance of the internal systems during these tests was slightly limited dueto lack of time to prepare specific software for such a task. While it is simple touse the camera inside the chamber to see if the drones motors have stopped, itis more difficult to at a glance see if for example stored data has been altered.

3.5.2 Company A drone

In many regards the testplan for the company A drone is the same as the onefor the FieldRobotiX system. The main difference in this case is that a holderfor the drone has been built, allowing it to be propped upright during testing.While this was intended to be used to allow emissions testing in both a horizon-tal and vertical position, to be more efficient and save time the holder was onlyused during radiated immunity testing. This is because in this vertical positionthe applied electromagnetic field would be spread more uniformly across thewhole drone. In regards to test limits, the drone is to again be tested againstthe class B limits on radiated emissions from EN 55032 presented in Table 3(with the same reasoning for stopping at 1 GHz for radiated emissions). Again,the motivation for going with the class B limits as opposed to the class A limitsfor industrial environments is the same as for the FieldRobotiX system. It’sprobably even more important in this case due to the large dimensions andachievable speed of the drone presenting larger safety hazards in the event of a

21

system failure. As for radiated immunity, the same field strength and frequencyrange are used from the EN 55035 [38] standard and the same performancecriterion where the drone should not experience any significant faults or disrup-tions in any of its systems.

The experimental setup is otherwise the same as for the FieldRobotiX sys-tem. That is, the company A drone is placed on a table 0.8 m above the groundplane 3 m away from the measuring antenna which itself is positioned 1.3 mabove the ground plane for emission measurements and approximately 1.8 mabove the ground plane when testing immunity. The company A drone is setby default to have its nose point towards the door of the FAR as it was themost convenient placement for fastening it and the table can be rotated eitherway, so any desired position can still be achieved. As can be seen in Figure7, for emission measurements the drone is fastened by having the body restupon a cradle provided by company A themselves and then fastening the droneusing straps which are tied to a pallet loaded with rocks placed under the ta-ble. Figure 8 instead shows how the drone was fastened to the holder duringthe immunity tests. One additional simplification that should be noted is thatthe drones motors were not turned on during the immunity tests (as opposedto the FieldRobotiX tests). This is because company A was already aware oftheir current method of controlling the motors being susceptible to interference.Combined with the fact that they already have plans to implement a new setupfor controlling the motors inside the drone, there was really no need to test theimmunity of the current implementation. This also is a more safer option asinterference on the motors could cause them to uncontrollably increase theirspeed and possibly break the holder.

22

Figure 7: The placement of the company A drone inside the FAR for radiatedemission measurements.

Figure 8: The placement of the company A drone inside the FAR for immunitytesting.

23

4 Results

The results and measurements setups were done with the help from my internallab supervisors but also from employees of FieldRobotiX and company A re-spectively who additionally assisted with the hardware/software modificationsto enable different testing conditions and also helped monitor the internal statusof their systems. It should also be noted that the results presented are only aselection of the most important and illustrative results from the total numberof measurements made on the two system. Also note that when the topic ofantenna polarisation is brought up, this is mostly for clarity regarding the mea-surement setup as there really wasn’t a motivation behind picking vertical orhorizontal polarisation other than that both should be used.

Although the standard EN 55032 specifically calls for the usage of a quasi-peak detector when gauging the performance below 1 GHz, the measurementson both of the drones were done using the peak detector functionality (with ameasuring bandwidth of 120 kHz) as this is much faster and made it simplerto quickly figure out sources of interference rather than focusing on specificemission levels.

4.1 FieldRobotiX

Presented in Figure 9 is the noise floor for the FAR at LTU during the mea-surements on the FieldRobotiX system. It is measured at a bandwidth of 120kHz on the measurement receiver. The red line drawn in the figure is the classB limits for the EN 55032 [37] standard when measuring radiated emissions ina FAR. Seeing as the system did not use any of the communication interfacesinside the chamber, this noise floor is also the ideal floor where the chamberis as isolated from the outside as possible. The operating mode for the systemduring the measurements presented in this section was an idle state where thepropeller motors were operating at a lower power than usual. This was due tothe fact that the drone was fastened onto the table and increasing the powerwould cause the internal fail-safes of the DJI drone to activate and shut downthe drone as it detects its not lifting although it’s trying to.

24

Figure 9: The Noise floor of the FAR during the FieldRobotiX measurements.

4.1.1 Vertical polarization emissions

Figure 10 presents the emissions from the system when directly facing the an-tenna (i.e table rotated 0◦), measured with a vertically polarised antenna andhaving everything be on. In this case that means the drone (with motors), thecontrol computers of the module, the DC motor allowing the LiDAR to rolland the LiDAR itself are all operating. The red triangles in the figure highlightvalues being within -10 dB of the limits. In contrast, Figure 11 shows the emis-sions when only the drone is turned on and operating, even the motors wereturned off during this test. This graph then represents the contributions to theemissions spectrum from the drone itself.

25

Figure 10: Emissions from the system when using a vertically polarised antenna,no rotation and everything being turned on in the system.

Figure 11: Emissions when only the drone itself was operating.

26

While additionally turning on the central control computer of the moduleadded very little to the spectra, then also turning on the propellers gave quitedifferent emissions as seen in Figure 12. Additionally turning on the LiDARproduces the emissions in Figure 13. In this setup the LiDAR itself is on androtating to capture the surrounding environment in two dimensions, but thesecondary control computer (and associated DC motor) of the module added byFieldRobotiX which allows the LiDAR to roll and capture environments in threedimensions is not powered on (and also the only thing not powered on). As canbe seen in figure 13, the addition of the LiDAR provides noticeable increasesin high-frequency emissions and near 90 MHz, 190 MHz and above 700 MHz.Furthermore, as Figure 13 is the same setup as Figure 10 aside from the DCmotor being turned off, its impact on the spectrum can be directly observed.More specifically the DC motor is shown to contribute emissions around 140MHz and 290 MHz.

The impact of the LiDAR on the emissions spectra can be further illustratedby looking at Figure 14. This shows how rotating the table which the droneis placed upon by 135◦ (positive angle counter-clockwise) reduces the high fre-quency emissions as the LiDAR is now facing away from the measuring antenna(compare to figure 10 with no rotation).

27

Figure 12: Emissions with drone, central control computer and propellers turnedon.

Figure 13: Emissions with everything on except the electronics allowing rollingof LiDAR.

28

Figure 14: Emissions with everything on but with the table rotated 135◦

counter-clockwise.

4.1.2 Horizontal polarisation emissions

Measurements were also made with the measuring antenna inside the chamberbeing horizontally polarised. The resulting spectrum in Figure 15 shows theemissions from the system with 0◦ table rotation and having everything on witha horizontally polarised antenna. Compare this to Figure 10 which shows thesame setup but with a vertically polarised antenna. It can be seen that thehorizontal setup manages to pick up more of the high frequency emissions. Likewhen testing with the vertically polarised antenna, the table was also rotatedwith this setup to try and uncover more patterns in the emission spectra. Themost noticeable change occurred at 315◦ rotation counter-clockwise as can beseen in Figure 16, where there is a considerable increase in the emissions aroundthe 200 MHz mark. From having the previous measurements, it was suspectedthat the LiDAR was the source of these emissions. Turning off the LiDARproduced the spectrum in Figure 17, where it becomes apparent that it indeedwas the LiDAR that was producing the previously mentioned emissions.

29

Figure 15: Emissions with everything and horizontal antenna polarisation.

Figure 16: Emissions with everything on, horizontal antenna polarisation and315◦ rotation.

30

Figure 17: Emissions with everything except LiDAR on, horizontal antennapolarisation and 315◦ rotation.

4.1.3 Immunity

In addition to the planned immunity test, my lab supervisor suggested to con-duct an initial immunity test using a 10-3-1 methodology. In other words, testingwith a 10 V/m field from 80 MHz to 1 GHz then 3 V/m to 2 GHz and thenfinally 1 V/m to 2.7 GHz. Using this methodology (horizontally polarised an-tenna 1.75 meters of the ground plane), the system started experiencing issuesaround 218 MHz where the rolling of the LiDAR started stuttering. Throughthe internal monitoring setup by FieldRobotiX, they observed no noticeable dis-crepancies in the data aside from the measured roll angle of the LiDAR beingrepeatedly set to zero after a while. This is then suspected to be the point intime when the stuttering began.

The initial idea of instead testing with 3 V/m from 80 MHz up to 6 GHz wasthen put into practice. With this field strength the system showed no visibledegradation of performance up to around 4.6 GHz when the tests were aborteddue to the measuring equipment’s power meter measuring too high of a VSWR(too much power ’bouncing back’ from the antenna). This was the case for bothvertical and horizontal polarisation of the antenna in the chamber. Althoughit should be noted that in both of these 3 V/m tests the internal monitoringshowed the same problem with the measured roll angle as when the LiDARstuttered. However the error occurred for a much shorter amount of time andas previously stated no visible degradation was shown.

31

4.2 Company A

Presented in Figure 18 is the noise floor for the FAR at LTU during the measure-ments on the company A drone. It is again measured at a bandwidth of 120 kHzon the measurement receiver. In this case the FAR is not entirely isolated fromthe outside. This is because for one, there was an antenna connector being ledoutside for controlling the drone and monitoring its internal status. Company Aalso wanted to measure immunity and emissions from a single, separate motorfor the drone (see Appendix 7.2), so to control this motor a cable for CAN buscommunication was also fed through a port on the outside wall of the FAR tothe inside. The effect of this interfacing becomes clear when compared to thenoise floor in figure 9. Now there are peaks around approximately 90 MHz and105 MHz which are frequencies commonly used for radio broadcasts in Swedenas detailed by Teracom [40]. As for the spike around 420 MHz, it’s more difficultto say where it originates as unlike the peaks from the radio broadcasts, thisthird peak did not appear (or at least wasn’t as prevalent) in future measure-ments. However, all of these ’new’ emissions come from the fact that the FARis not completely isolated due to the extra methods of communication.

Figure 18: The Noise floor of the FAR during the company A measurements.

32

4.2.1 Motor emissions

With everything turned on in the drone (drone, motors and external communi-cation), no rotation of the table and using a vertical antenna polarisation theemissions were as presented in Figure 19. As can be seen, the emissions fromthe system are clearly above the specified limits.

Figure 19: Radiated emissions from the company A drone when everything wasturned on and with no table rotation.

Figure 20 shows the reference numbers for the motors whose contributionsto the emissions are shown in this section, where motor number 11 (green) is athrusting motor and the other numbers are lifting motors (yellow). The motoremissions were measured when the lifting motors were operating at lifting power(400 W) and the forward thrusting motors at cruising power (800 W).

Figure 20: Placement of company A drone inside the FAR with reference formotor numbers.

33

Keeping the vertical antenna polarisation and 0◦ rotation, Figure 21 shows(in blue) the emissions when everything on the drone was running but withonly motor number 2 turned on. The yellow trace in the figure are the emis-sions with the same setup but with no motors being turned on. It should bepointed out that the huge spike in emissions between 400 MHz – 500 MHz is dueto this being the frequency band which is used when communicating with thebase station outside the chamber. The exact level varies between measurementswith the peak sometimes not appearing at all (see for example figure 19) butstill showing an increase in emissions around the same frequency span. Also,the external communication seems to contribute to the emissions at the end ofthe spectrum as when the communication is on is when the emissions are thelargest. Still, as can be seen in Figure 21 the second motor does not impactthe emissions much. Compare this to Figure 22 where motor 3 is turned oninstead, here the slight increase in emissions around 150 MHz has turned intoa noticeable peak.

Figure 21: Comparison between having the drone with only motor number 2turned on (blue) and having no motors turned on (yellow).

34


However, the highest emissions come from the back row of motors as illus-trated in Figure 23, where only motor number 5 was turned on. Not only has thepeak around 150 MHz become wider, but there are also entirely new increasesbeing detected around 80 MHz and 180 MHz. While the previous figures havebeen for lifting motors, Figure 24 shows the emissions from motor 11 which is amotor used for horizontal movement. As can be seen, the emissions for motor 11are generally lower compared to motor number 5 which is placed right next to it(even lower than motor 3 if comparing peaks at 150 MHz). However the peaksare still present at more or less the same frequencies as in Figure 23, so its worthkeeping in mind that the back row of motors seem to present unique emissionscompared to the front row ones. To further illustrate the contribution from theback row specifically, figure 25 shows the contribution from motor number 7.As can be seen, the emissions are quite similar to those of motor number 5 inFigure 23, which would make sense as it’s placed in the same position as motornumber 7 but on the other wing (refer to Figure 20). The key difference beingthe smaller emissions around 80 MHz and the increased emissions around 180MHz. The increase could be explained by the fact that motor 7 was positionedcloser to the measuring antenna compared to motor number 5, while the thesmaller levels around 80 MHz could indicate a fault in motor number 5 itself orthe wiring in that wing specifically.

35



36


4.2.2 15 V regulator emissions

The company A drone uses three different voltage regulators for stepping downthe voltage from the internal battery powering the drone. Specifically, the droneuses a 5 V, a 12 V and a 15 V regulator for this purpose. Presented in Figure 26are the emissions from the drone when no motors are turned on and the exter-nal communication is disabled (as the motors no longer need to be controlled).The measurements in this section were also done using a horizontal antennapolarisation as opposed to the vertical polarisation used in the previous figuresregarding motor emissions from the company A drone. Compare the emissionsof Figure 26 to those presented in Figure 27 which shows the emissions usingthe same setup but with the 15 V regulator (and thus its associated compo-nents) being disconnected. As can be seen, although the motors themselveshave proven themselves to be significant sources of interference simply discon-necting the 15 V regulator has significantly reduced the emissions across almostthe entire spectrum (low frequencies around 30 MHz remain the same).

To further illustrate this point, Figures 28 and 29 show the same comparisonbut when the table inside the chamber was rotated 135◦ counterclockwise. Ascan be seen there is a substantial decrease in overall emissions.

37

Figure 26: Emissions when the drone was running with no external communi-cations or motors being turned on.

Figure 27: Emissions when the drone was running with no motors or externalcommunications and the 15 V regulator disconnected

38

Figure 28: Emissions when the drone was rotated 135◦ counterclockwise witheverything on except motors and external communication.

Figure 29: Emissions when the drone was rotated 135◦ counterclockwise, nomotors or communication and the 15 V regulator disconnected

39

Figure 30 shows another test, this time with vertical antenna polarisation,0◦ rotation and having only the drone on with no communication or motorsoperating. It also illustrates the difference between only having the 15 V reg-ulator on (yellow) compared to also plugging in its associated components it’sconnected to except the 4G module (blue). The impact of the 4G module isinstead shown in Figure 31. The choice of separating the contributions like thiswas because it was suspected the 4G module might be an important source ofemissions from the components connected to the regulator, however as can beseen in the figure this is not the case. Although the testing setup for the fig-ures was a bit different, Figures 30 and 31 shows that it’s the 15 V regulatoritself which is the biggest contributor to the emissions, and not the componentsconnected to it.

Figure 30: Comparison between emissions when aside from the drone, only the15V regulator is on (yellow) and when all its associated components (except4G) are also on (blue).

40

Figure 31: Emissions when adding the 4G module to the rest of the componentsconnected to the 15V regulator, compared to having everything but the 4Gmodule connected.

4.2.3 12 V voltage regulator

To see the effects of the other regulators on the radiated emissions, Figures32 and 33 shows the emissions when only the 12 V regulator is on and wheneverything connected to it is also on. For these tests, a vertical polarisation wasused for the antenna and no external communication was operating. However,notice that due to the CAN bus interfacing cable still being inside the chamber, apeak is very visible at 90 MHz coming from local radio broadcasts. Furthermore,the regulator by itself does not seem to provide much in the way of radiatedemissions but connecting their respective equipment does increase the emissionssomewhat. Still, this increase is not as dramatic as with the 15 V regulator.

41

Figure 32: Radiated emissions from the drone when only the 12 V regulatorwas on.

Figure 33: Radiated emissions from the drone when the 12 V regulator andeverything connected to it was on.

42

4.2.4 Immunity

The same 10-3-1 test used for the FieldRobotiX system was attempted and thistime no discernible degradation of performance was observed visually or inter-nally within the drone for both antenna polarisations. The internal monitoringwas done through representatives from company A observing packet loss andsensor errors on the inertial measurement unit, accelerometer, gyro and barom-eter of the drone. As the drone had already passed being exposed to 10 V/m upto 1 GHz and then 3 V/m up to 2 GHz, it was only really necessary to attemptthe additional 3 V/m test (the one which ideally goes up to 6 GHz) startingat 2 GHz. This test did not present any issues for either polarisation but onceagain both the 3 V/m tests stopped at around 4.5 GHz due to the power meterdetecting too high of a VSWR.

43

5 Discussion

5.1 Legislation

Although care has been taken into covering all EMC aspects of the drones, itshould be noted that there may be some standards shown that are actuallynot applicable, are misused and even some applicable ones that are missedentirely. This could be the case due to the choices being heavily influenced bywhat has been done before through looking at other declarations of conformityand contacting EMC test labs (whose identities I won’t disclose to keep theiranonymity). While research was done outside the scope of those sources tofind applicable standards, it’s still important to keep in mind that there is acertain bias in the choices of standards towards using what has already beenused before to try and be as realistic as possible in presenting what standardscompany A and FieldRobotiX should be keeping in mind for future compliancemeasurements.

5.2 Measurements

Although stated multiple times, it bears repeating that the approach taken dur-ing the tests was to identify problematic areas of the drone from an EMC pointof view rather than prove absolute compliance with every applicable standardfor the drone. This choice was based mostly on the fact that proper compliancetesting was too complicated and time consuming to finish within the bounds ofthis thesis. This was also the case due to scheduling only allowing a maximumof 2 days of testing with the drones, and therefore simpler test setups wouldallow more testing to be done. So to also provide some tangible benefits for thecompanies testing their drones, a more simplified approach was taken where thedrone was treated as a single piece of equipment rather than multiple pieces ofequipment comprising a single system. As can be seen in the results of sections4.1 and 4.2, this meant really only testing against a single standard for theradiated emissions tests and using a more basic and common setup for testingimmunity. While the used standards (EN 55032 [37] for emissions and EN 55035[38] for immunity) are viable and relevant to the equipment at hand, it would bedisingenuous to say that the performed tests would constitute a comprehensiveEMC assessment of the drone as some parts of the drone would need specifictesting covered in other standards (for example EN 301-489-19 [18] for GNSS).

As both systems proved to be incompliant with the class B limits used fromthe EN 55032, one could ask if maybe using the more stringent class B limitsfor residential use is a bit unfair? While the systems could be said to performbetter if compared against the less demanding class A limits for industrial en-vironments, the purpose of the tests was never to try and prove compliance byany means necessary but instead about identifying potential problems with thedesign of the product. In the end, for company A and FieldRobotiX these testare about making their products the best they can be when it comes to EMC.As such, it would be disingenuous for this thesis to change the framework forthe measurements until the results get ’better’ by making sure the emissions arebelow (or closer to) the limits by any means possible.

44

5.2.1 FieldRobotiX

One important thing to take note of with the measurements is the fact thatthe system was operating in an idle state which did not allow the propellers tooperate at full power. As previously stated this was due to a software safetymechanism within the DJI drone shutting down the motors if it does not de-tect the drone lifting off although its trying to. This means that the measuredemissions, or at least the contributions from the motors, are by all likelihoodsmaller than what they really are during normal operation. With this in mind,it should be noted that although the results show that the system is very closeto meeting the limits, this may actually be slightly misleading.

Another thing to point out is the usage of the peak detector during radiatedemission measurements instead of the requested quasi-peak detector from theEN 55032 standard. This was due to the peak detector offering more rapidtesting and allowing more configurations to be tested during the testing timeframe. The drawback to this approach is that the levels presented in the re-sults are most likely higher than what they would have been if measured using aquasi-peak detector and seeing as the limits are for when using quasi-peak detec-tors, the conclusion of saying the system is incompliant can be a bit misleading.However, while the exact measured levels of radiated emissions may have beendifferent, the conclusions drawn regarding problematic sources of interferencewould most likely not be. Looking at for example Figures 16 and 17, one can seethat the emissions are well above the specified limits and while the levels mayhave been lowered a bit (say 4-5 dBµV/m) with a quasi-peak measurement, thefact that the LiDAR produces this much unwanted emissions is still cause forconcern and something that should be further investigated. To further put theresults into perspective, one would ideally like to have the measured emissionsbe at least 6 dB below the limits during testing to properly ensure complianceat all times. This point regarding the usage of the peak detector is also appli-cable for the company A measurements (as it also used a peak detector), butthe argument is slightly more important here as the emissions are closer to thelimits for the FieldRobotiX system.

As indicated by the results in Section 4.1, the LiDAR is what seems to bethe most obvious source of unwanted interference (that is the most feasible toreduce). The source of this, and possible area of improvement, as suggestedby FieldRobotiX might be the cabling connecting the rest of the module tothe LiDAR. This is a reasonable conclusion to make as one would expect thecommercial LiDAR module by itself to be compliant. In this case, shielding ofthe cables or the associated area where the LiDAR and module are connectedmight be a valid option. Another simple and efficient way to reduce emissionsis to employ the use of twisted pairs of cables. The twisted pair works on theprinciple that the forward and return currents in the pairs would have opposingorientations on their magnetic fields and thus mitigate each other. The twist-ing part is simply to bring the conductors as close to each other as possible toproduce maximum mitigation.

The DC motor allowing rolling of the LiDAR does also present an increasein emissions and should be noted (compare for example levels at 120 MHz and

45

280 MHz between Figures 10 and 13). However the increase is not quite asbad as in the LiDAR case as the increased levels keep themselves below (or atleast almost below) the specified limits. As such, their current implementationis probably sufficient for now and something that could be looked at when theLiDAR has been taken care of. One positive aspect about the design of thesystem is therefore the fact that while the LiDAR attached to the module pro-duces undesired emissions, the module itself and the connected DC motor forrolling the LiDAR doesn’t seem to contribute excessive amounts of interference.Furthermore, as it’s suspected that the control computer being affected is whatcaused the issues during immunity testing, shielding on the control computermight be warranted to try and avoid these issues. As the issues started ap-pearing around 218 MHz (specifically being above 30 MHz), a suggestion forshielding would be to just use a plastic shell covered in a conductive coating asdetailed by Tim Williams in [p.435, 41].

5.2.2 Company A

The results show that the drone is not compliant with the emission limits usedin this thesis. However, the company A drone is still a prototype and as suchthere is more room for tuning the design and addressing issues with the designin regards to EMC to meet the requirements. More importantly, some of theseproblematic areas were successfully identified through the measurements. Whileit may not be shocking to hear that the motors produce sizeable amounts of in-terference, what is more interesting is the fact that the back row of motors thatwere tested seem to provide a disproportionate amount of EMI compared totheir front row counterparts. This may indicate an issue with the cabling in theback wing(s) or the implementation of the thrusting motors as their placementon the outer portion of the wings is the most apparent difference compared tothe front row of motors. However, as the motor emission tests were performedin the configuration shown in Figure 7, it’s possible that the reason for the frontmotors presenting generally lower emissions is because they were blocked by thebody of the drone. This could of course have been confirmed by rotating thedrone or testing the front row motors on the other side of the body. This was notfeasible for this thesis however as this apparent discrepancy was noticed muchtoo late to perform new tests. Still, the apparent change in emissions betweenmotor rows is still something that should be kept in mind as the changes werestill sizeable.

Furthermore, the 15 V voltage regulator proved to be a major source of in-terference and one that should be properly addressed. Exactly what part of theimplementation of the 15 V causes this amount of interference is difficult to say.However, company A has proposed a solution for at least mitigating some ofthe interference. That being applying an input filter to the voltage regulator asdetailed by its manufacturer. One could make the argument that maybe it’s thecomponents connected to the 15 V regulators that’s causing the interference.As shown in Figures 30 and 31 this does not seem to be the case. However, onething to note with those figures is that the emissions above around 600 MHzseem to have greatly decreased although for example Figures 26 and 27 indi-cates that it is the 15 V regulator that has caused the interference above thatfrequency. The reason for this discrepancy is being unknown is of concern as it

46

would be problematic for future measurements if this amount of interference wasa rare occurrence. It should not be due to using different antenna polarisationsas both have exhibited the same variation in results. Purely from guesswork, theonly theory I can propose is that it’s somehow harmonics from the 4G moduleconnected to the 15 V regulator which sometimes creates sporadic emissions inthis frequency span, either that or some other communication was not properlyput out of operation during the tests.

While the emissions from the motors and the 15 V regulator have proven tobe detrimental parts of the design in regards to EMC, there are still positiveswith the design. The 12 V regulator and its associated equipment do not seemto be excessive contributors when it comes to radiated emissions. However theseemissions are still near the limits (see figure 33) and should be kept in mind andmaybe looked at further if the opportunity presents itself. As noted by companyA, an increase in emissions when connecting the associated equipment to a reg-ulator is not surprising as the regulator is then loaded and has to actually powerthe equipment. Although it should be stated that the measurements presentedin this thesis did not measure exactly how much of these added emissions comefrom the regulators and how much is added by the connected equipment itself.The only thing proven is that connecting the equipment does present differentemissions.

47

6 Conclusions

6.1 Legislation

Through the initial research done on the legislation for EMC of drones, someconclusions can be drawn. In regards to strictly EMC, the main directive thatshould be followed is either the EMC directive 2014/30/EU [25], the radio equip-ment directive 2014/53/EU [26] or essential requirements from 2018/1139/EUfor rules in the field of civil aviation [27]. The real question to ask with a droneis whether it is exempt from the EMCD and RED or not, which depends onfactors such as frequency usage, transmission capabilities and intended opera-tion of the drone.

When it comes to standards one can use to conform to the applicable direc-tive, things become more complicated. While work was being done to makethem a reality, at the time of researching for this thesis there were as of yetno finished, fully implemented drone-specific standards that one can use whenwanting to prove conformity in regards to EMC in the EU. Instead, asking EMCtest labs and looking through previous declarations of conformity has shown thatin regards to EMC the EN 301-489-X series and EN 55032/55035 are some ofthe more relevant standards to keep in mind. However, due to the fact that notall drones are built alike it should be noted that other more relevant standardsmay be applicable to certain systems.

While in the end the application of legislation onto the both the drones be-came more simplified than initially intended, the stated goal of this part of thethesis to consolidate the available legal framework has still been achieved.

In regards to the legislative research done in this thesis, further work couldfor example expand the research to cover markets other than the EU. Further-more, as the development of EMC standards for drones is currently ongoing, itwould be proper to revisit the topic when these have been published and areavailable.

6.2 FieldRobotiX

Based on the limits and measurement techniques used in this thesis along withthe results in section 4.1, the FieldRobotiX system consisting of the commercialdrone and LiDAR and their own 3D mapping module is tentatively incompliantwith the class B limits on radiated emissions presented in EN 55032. In thiscase, although the DC motor used for rolling of the LiDAR was a source ofinterference, the most interesting source of interference that was identified wasthe LiDAR attached to the module as it should be compliant as a commerciallyavailable product in the EU. As such this indicates an issue with the imple-mentation rather than the product itself. Exactly what causes this amount ofundesired interference is a topic for further studies and testing.

Furthermore, the system (or more specifically, the module itself) experiencesissues when exposed to radiated electromagnetic fields of strengths 10 V/m and3 V/m. This comes in the form of the LiDAR starting to stutter in its rolling

48

motion when exposed to 10 V/m around 215 MHz and further analysis showingdiscrepancies in sensor values for the DC motor (however these were not com-pletely confirmed to have occurred at the same frequency, but it is the mostreasonable theory). Albeit much less severe, for 3 V/m the same discrepan-cies in data occur (once again exact frequencies were difficult to pinpoint, butdata logs suggests around 150 MHz) and it is theorized by FieldRobotiX thatall theses errors occur due to the control computer attached to the DC mo-tor resetting itself. With these results, the drone cannot be said to meet thepreviously defined performance requirement for immunity testing as defined insection 3.5.1. With all of this said, it should be noted that the system is stillunder development and as such the performance of the system and specificallythe prototype module is subject to change and does not represent the finalizedproduct.

Overall, the measurements have met the goal of the thesis which was to identifyproblematic areas of the design in regards to EMC. Although there are areassuch as conductive emissions and ESD testing which have not been considered,the obtained results are enough to move forward with improving the design ofthe system.

6.3 Company A

While the company A drone did not meet the limits used during testing, themeasurements still provided useful results in the identification of two main prob-lems with the design in regards to EMC. While the tested motors all seem togive noticeable contributions when it comes to radiated emissions, the fact thatthe back row of motors seemingly radiate more emissions than their front rowcounterparts is something that should be looked at closer as this indicates anissue with the cabling/design of the wing rather than issues with the motorsthemselves. Though, one thing to keep in mind is that the configuration (seeFigure 7) might have accidentally been one which blocked some of the emissionscoming from the front row motors as they were partially covered by the bodyof the drone. Still, one motor specifically did give disproportionate amounts ofinterference, which could indicate a fault with that specific unit (refer to motor5 of Figure 20 which wasn’t even blocked by the body of the drone). The secondmain source of interference that should be looked at is the 15 V regulator andits implementation into the system as the regulator proved to be a source ofsignificant interference across almost the entire spectrum up to 1 GHz. Thereis however an unidentified source of radiated emissions which sometimes occurabove around 600 MHz, but seemingly only when the 15 V regulator and itsassociated equipment is plugged in.

The 12 V regulator does seem to present emissions that are close to the lim-its when its associated equipment is connected to it and powered alongside it.However, these levels are not nearly as bad as those for the 15 V regulator andas such it is recommended that this part of the design be left as is for now andproperly dealt with if the opportunity presents itself. As with the FieldRobotiXdrone it should be noted that the performance of this protoype system is subjectto change and does not represent the final product.

49

Topics for further work would involve re-examining the implementation of the15 V regulator and especially the motors on the back wing(s) of the aircraft toidentify the exact source of what might cause the interference. Testing in thefuture should specifically try and target the exact source of the aforementionedunknown high frequency interference that sometimes occur. As in the case ofthe FieldRobotiX system, the measurements performed in this thesis also leavesome further room for testing conducted emissions and immunity to ESD andother electrical transients.

50

References

[1] AW DRONES. DRONE STANDARDS INFORMATION PORTAL. url:https://standards.aw-drones.eu. (accessed: 10.02.2022).

[2] AW DRONES. Harmonising drone standards to support the ongoing EUregulatory process. url: https://www.aw-drones.eu/. (accessed: 07.05.2022).

[3] The Editors of Encyclopaedia Britannica. “Lenz Law”. In: Encyclope-dia Britannica (29 May. 2020). url: https://www.britannica.com/science/Lenzs-law. (accessed: 15.05.2022).

[4] DJI. Fajia Wang. Declaration of Conformity. May 7. 2020. url: https:/ / terra - 1 - g . djicdn . com / 9b266a09f53f4040901d8f8eca40ac3e /

Industrial/M300.pdf. (accessed: 15.04.2022).

[5] DJI. Fajia Wang. Declaration of Conformity. Oct 9. 2020. url: https:/ / terra - 1 - g . djicdn . com / 9b266a09f53f4040901d8f8eca40ac3e /

Mavic/DJI%20MINI%202%20%EF%BC%88Camera%20Drone%20%EF%BC%

89Aircraft.pdf. (accessed: 14.03.2022).

[6] do160.org. RTCA DO-160G - Environmental Conditions and Test Pro-cedures for Airborne Equipment. url: https://do160.org/rtca-do-160g/. (accessed: 14.05.2022).

[7] EASA. Civil drones (unmanned aircraft). url: https : / / www . easa .

europa.eu/domains/civil-drones. (accessed: 15.05.2022).

[8] EASA. Easy Access Rules for Unmanned Aircraft Systems. September2021. url: https://www.easa.europa.eu/downloads/110913/en.(accessed: 22.02.2022).

[9] Elsakerhetsverket. Elektromagnetiska storningar - regeringsuppdrag. De-cember 2020. url: https://www.elsakerhetsverket.se/globalassets/publikationer/rapporter/elektromagnetiska-storningar-regeringsuppdrag-

webb.pdf. (accessed: 23.02.2022).

[10] ETSI. Draft EN 303 213-5-1 Advanced Surface Movement Guidance andControl System (A-SMGCS); Part 5: Harmonised Standard for access toradio spectrum for Multilateration (MLAT) equipment; Sub-part 1: Re-ceivers and Interrogators. V1.0.0. December 2019. url: https://www.etsi.org/deliver/etsi_en/303200_303299/3032130501/01.00.00_

20/en_3032130501v010000a.pdf. (accessed: 02.05.2022).

[11] ETSI. Draft ETSI EN 301 489-20 ElectroMagnetic Compatibility (EMC)standard for radio equipment and services; Part 20: Specific conditionsfor Mobile Earth Stations (MES) used in the Mobile Satellite Services(MSS); Harmonised Standard for ElectroMagnetic Compatibility. V2.1.2.March 2021. url: https://www.etsi.org/deliver/etsi_en/301400_301499/30148920/02.01.02_20/en_30148920v020102a.pdf. (accessed:02.05.2022).

51

https://standards.aw-drones.eu

https://www.aw-drones.eu/

https://www.britannica.com/science/Lenzs-law

https://www.britannica.com/science/Lenzs-law

https://terra-1-g.djicdn.com/9b266a09f53f4040901d8f8eca40ac3e/Industrial/M300.pdf



https://terra-1-g.djicdn.com/9b266a09f53f4040901d8f8eca40ac3e/Mavic/DJI%20MINI%202%20%EF%BC%88Camera%20Drone%20%EF%BC%89Aircraft.pdf




https://do160.org/rtca-do-160g/

https://do160.org/rtca-do-160g/

https://www.easa.europa.eu/domains/civil-drones

https://www.easa.europa.eu/domains/civil-drones

https://www.easa.europa.eu/downloads/110913/en

https://www.elsakerhetsverket.se/globalassets/publikationer/rapporter/elektromagnetiska-storningar-regeringsuppdrag-webb.pdf



https://www.etsi.org/deliver/etsi_en/303200_303299/3032130501/01.00.00_20/en_3032130501v010000a.pdf





[12] ETSI. Draft ETSI EN 301 489-50 ElectroMagnetic Compatibility (EMC)standard for radio equipment and services; Part 50: Specific conditions forCellular Communication Base Station (BS), repeater and ancillary equip-ment; Harmonised Standard covering the essential requirements of arti-cle 3.1(b) of the Directive 2014/53/EU. V2.1.0. June 2016. url: https://www.etsi.org/deliver/etsi_en/301400_301499/30148950/02.01.

00_20/en_30148950v020100a.pdf. (accessed: 02.05.2022).

[13] ETSI. Draft ETSI EN 301 489-52 ElectroMagnetic Compatibility (EMC)standard for radio equipment and services; Part 52: Specific conditions forCellular Communication User Equipment (UE) radio and ancillary equip-ment; Harmonised Standard for ElectroMagnetic Compatibility. V1.1.2.December 2020. url: https : / / www . etsi . org / deliver / etsi _ en /301400_301499/30148952/01.01.02_20/en_30148952v010102a.pdf.(accessed: 02.05.2022).

[14] ETSI. Draft ETSI EN 301 908-1 IMT cellular networks; Harmonised Stan-dard for access to radio spectrum; Part 1: Introduction and common re-quirements Release 15. V15.0.1. June 2021. url: https://www.etsi.org/deliver/etsi_en/301900_301999/30190801/15.00.01_20/en_

30190801v150001a.pdf. (accessed: 02.05.2022).

[15] ETSI. ETSI EN 300 220-1 Short Range Devices (SRD) operating in thefrequency range 25 MHz to 1 000 MHz; Part 1: Technical characteristicsand methods of measurement. V3.1.1. February 2017. url: https://www.etsi.org/deliver/etsi_en/300200_300299/30022001/03.01.01_60/

en_30022001v030101p.pdf. (accessed: 02.05.2022).

[16] ETSI. ETSI EN 300 328 Wideband transmission systems; Data transmis-sion equipment operating in the 2,4 GHz band; Harmonised Standard foraccess to radio spectrum. V2.2.2. July 2019. url: https://www.etsi.org/deliver/etsi_en/300300_300399/300328/02.02.02_60/en_

300328v020202p.pdf. (accessed: 02.05.2022).

[17] ETSI. ETSI EN 301 489-1 ElectroMagnetic Compatibility (EMC) standardfor radio equipment and services; Part 1: Common technical requirements;Harmonised Standard for ElectroMagnetic Compatibility. V2.2.3. Novem-ber 2019. url: https://www.etsi.org/deliver/etsi_en/301400_301499/30148901/02.02.03_60/en_30148901v020203p.pdf. (accessed:14.03.2022).

[18] ETSI. ETSI EN 301 489-19 ElectroMagnetic Compatibility (EMC) stan-dard for radio equipment and services; Part 19: Specific conditions for Re-ceive Only Mobile Earth Stations (ROMES) operating in the 1,5 GHz bandproviding data communications and GNSS receivers operating in the RNSSband (ROGNSS) providing positioning, navigation, and timing data; Har-monised Standard covering the essential requirements of article 3.1(b) ofDirective 2014/53/EU. V2.1.1. April 2019. url: https://www.etsi.org/deliver/etsi_en/301400_301499/30148919/02.01.01_60/en_

30148919v020101p.pdf. (accessed: 02.05.2022).

[19] ETSI. ETSI EN 303 413 Satellite Earth Stations and Systems (SES);Global Navigation Satellite System (GNSS) receivers; Radio equipmentoperating in the 1 164 MHz to 1 300 MHz and 1 559 MHz to 1 610 MHzfrequency bands; Harmonised Standard covering the essential requirements

52









https://www.etsi.org/deliver/etsi_en/300200_300299/30022001/03.01.01_60/en_30022001v030101p.pdf











of article 3.2 of Directive 2014/53/EU. V1.1.1. June 2017. url: https://www.etsi.org/deliver/etsi_en/303400_303499/303413/01.01.

01_60/en_303413v010101p.pdf. (accessed: 02.05.2022).

[20] ETSI. ETSI EN301 489-17 ElectroMagnetic Compatibility (EMC) stan-dard for radio equipment and services; Part 17: Specific conditions forBroadband Data Transmission Systems; Harmonised Standard for Elec-troMagnetic Compatibility. V3.2.4. September 2020. url: https://www.etsi.org/deliver/etsi_en/301400_301499/30148917/03.02.04_60/

en_30148917v030204p.pdf. (accessed: 02.05.2022).

[21] ETSI. Final draft ETSI EN 300 440 Short Range Devices (SRD); Ra-dio equipment to be used in the 1 GHz to 40 GHz frequency range; Har-monised Standard covering the essential requirements of article 3.2 of Di-rective 2014/53/EU. V2.1.1. January 2017. url: https://www.etsi.org/deliver/etsi_en/300400_300499/300440/02.01.01_30/en_

300440v020101v.pdf. (accessed: 02.05.2022).

[22] ETSI. Final draft ETSI EN 301 489-3 ElectroMagnetic Compatibility (EMC)standard for radio equipment and services; Part 3: Specific conditions forShort-Range Devices (SRD) operating on frequencies between 9 kHz and246 GHz; Harmonised Standard covering the essential requirements of ar-ticle 3.1(b) of Directive 2014/53/EU. V2.1.1. March 2017. url: https://www.etsi.org/deliver/etsi_en/301400_301499/30148903/02.01.

01_30/en_30148903v020101v.pdf. (accessed: 02.05.2022).

[23] European Commission, Directorate-General for Mobility and Transport.Commision delegated regulation (EU) 2019/945/EU of 12 March 2019 onunmanned aircraft systems and on third-country operators of unmannedaircraft systems. OJ L 152, 11.6.2019, p. 1–40. url: https : / / eur -

lex.europa.eu/legal- content/EN/TXT/?uri=CELEX:32019R0945.(accessed: 22.02.2022).

[24] European Commission, Directorate-General for Mobility and Transport.Commision implementing regulation (EU) 2019/947 of 24 May 2019 onthe rules and procedures for the operation of unmanned aircraft (Text withEEA relevance.) OJ L 152, 11.6.2019, p. 45–71. url: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32019R0947.(accessed: 22.02.2022).

[25] European Parliament, Council of the European Union. Council directive2014/30/EU of 26 February 2014 on the harmonisation of the laws ofthe Member States relating to electromagnetic compatibility. OJ L 96,29.3.2014, p. 79–106. url: https : / / eur - lex . europa . eu / legal -

content/EN/TXT/?uri=CELEX%3A32014L0030. (accessed: 22.02.2022).

[26] European Parliament, Council of the European Union.Directive 2014/53/EUof the European Parliament and of the Council of 16 April 2014 on theharmonisation of the laws of the Member States relating to the mak-ing available on the market of radio equipment and repealing Directive1999/5/EC Text with EEA relevance. OJ L 153, 22.5.2014, p. 62–106. url:https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:

32014L0053. (accessed: 22.02.2022).

53







https://www.etsi.org/deliver/etsi_en/300400_300499/300440/02.01.01_30/en_300440v020101v.pdf






https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32019R0945

https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32019R0945

https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32019R0947

https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32019R0947

https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32014L0030

https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A32014L0030

https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32014L0053

https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32014L0053

[27] European Parliament, Council of the European Union. European par-liament and Council regulation 2018/1139/EU of 4 July 2018 on com-mon rules in the field of civil aviation and establishing a European UnionAviation Safety Agency, and amending Regulations (EC) No 2111/2005,(EC) No 1008/2008, (EU) No 996/2010, (EU) No 376/2014 and Di-rectives 2014/30/EU and 2014/53/EU of the European Parliament andof the Council, and repealing Regulations (EC) No 552/2004 and (EC)No 216/2008 of the European Parliament and of the Council and Coun-cil Regulation (EEC) No 3922/91. OJ L 212, 22.8.2018, p. 1–122. url:https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%

3A32018R1139. (accessed: 22.02.2022).

[28] Graver, B., Zhang, K. and Rutherford, D. (2019). CO2 emissions fromcommercial aviation, 2018. url: https://theicct.org/publication/co2-emissions-from-commercial-aviation-2018/. (accessed: 04.04.2022).

[29] GROW.DDG1.C.4. Guide for the EMCD (Directive 2014/30/EU). De-cember 2018. url: https : / / ec . europa . eu / docsroom / documents /33601. (accessed: 22.02.2022).

[30] GROW.H.3. Summary of references of harmonised standards published inthe Official Journal – Directive 2014/30/EU of the European Parliamentand of the Council of 26 February 2014 on the harmonisation of the lawsof the Member States relating to electromagnetic compatibility. April 2021.url: https://ec.europa.eu/docsroom/documents/45365. (accessed:22.02.2022).

[31] GROW.H.3. Summary of references of harmonised standards published inthe Official Journal – Directive 2014/53/EU1 of the European Parliamentand of the Council of 16 April 2014 on the harmonisation of the laws ofthe Member States relating to the making available on the market of radioequipment and repealing Directive 1999/5/EC. July 2021. url: https://ec.europa.eu/docsroom/documents/46432. (accessed: 22.02.2022).

[32] Ole Hammer. “What to do when your product does not match a har-monised standard”. In: FORCE TECHNOLOGY (Unknown year). url:https://forcetechnology.com/en/articles/notified-body-products-

standards-risk-assessment. (accessed: 16.02.2022).

[33] Insider Intelligence. “Drone market outlook in 2022: industry growth trends,market stats and forecast”. In: Insider Intelligence (Apr 15. 2022). url:https://www.insiderintelligence.com/insights/drone-industry-

analysis-market-trends-growth-forecasts/. (accessed: 16.05.2022).

[34] Intertek. Jainu Jogani and Jhory Albayalde and Krishna K Vemuri. DATAREPORT. 11 Nov. 2015. url: https://velodynesupport.zendesk.com/hc/en-us/article_attachments/360003936553/VLP-16_EMC_ClassB_

Customer.pdf. (accessed: 20.05.2022).

[35] RISE. Electric aviation in Sweden – how far has development progressed?url: https://www.ri.se/en/our-stories/electric-aviation-in-sweden-how-far-has-development-progressed. (accessed: 04.04.2022).

[36] A K Shukla and S Nirmala. “EMI/EMC for military aircraft and its chal-lenges”. In: 2006 9th International Conference on Electromagnetic Inter-ference and Compatibility (INCEMIC 2006). 2006, pp. 101–107.

54

https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A32018R1139

https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A32018R1139

https://theicct.org/publication/co2-emissions-from-commercial-aviation-2018/

https://theicct.org/publication/co2-emissions-from-commercial-aviation-2018/

https://ec.europa.eu/docsroom/documents/33601





https://forcetechnology.com/en/articles/notified-body-products-standards-risk-assessment

https://forcetechnology.com/en/articles/notified-body-products-standards-risk-assessment

https://www.insiderintelligence.com/insights/drone-industry-analysis-market-trends-growth-forecasts/

https://www.insiderintelligence.com/insights/drone-industry-analysis-market-trends-growth-forecasts/

https://velodynesupport.zendesk.com/hc/en-us/article_attachments/360003936553/VLP-16_EMC_ClassB_Customer.pdf



https://www.ri.se/en/our-stories/electric-aviation-in-sweden-how-far-has-development-progressed

https://www.ri.se/en/our-stories/electric-aviation-in-sweden-how-far-has-development-progressed

[37] Svensk Elstandard. SS - EN 55032, Multimediautrustning – EMC-fordringar– Emission. STD-3336709. August 2015. url: https://www.sis.se/en/produkter / telecommunications - audio - and - video - engineering /

electromagnetic - compatibility - emc / emission / ssen55032/. (ac-cessed: 05.04.2022).

[38] Svensk Elstandard. SS - EN 55035, Multimediautrustning – EMC-fordringar– Immunitet. STD-3337592. September 2017. url: https://www.sis.se/en/produkter/telecommunications-audio-and-video-engineering/

electromagnetic-compatibility-emc/immunity/ss-en-55035/. (ac-cessed: 02.05.2022).

[39] Steve Taranovich. “EMI Shielding for Drones and UAVs”. In: ElectronicDesign (Nov 3. 2021). url: https : / / www . electronicdesign . com /power-management/whitepaper/21179428/electronic-design-emi-

shielding-for-drones-and-uavs. (accessed: 13.03.2022).

[40] Teracom. Frekvenstabeller radio. url: https://www.teracom.se/privat/Support/frekvenstabeller-radio/. (accessed: 05.05.2022).

[41] Tim Williams. EMC for Product Designers. 5:th edition. Newnes, 2016.

[42] Koh Watanabe et al. “Measurements of Electromagnetic Emission nearbya Compact Drone”. In: 2021 Asia-Pacific International Symposium onElectromagnetic Compatibility (APEMC). 2021, pp. 1–4. doi: 10.1109/APEMC49932.2021.9597146.

55

https://www.sis.se/en/produkter/telecommunications-audio-and-video-engineering/electromagnetic-compatibility-emc/emission/ssen55032/



https://www.sis.se/en/produkter/telecommunications-audio-and-video-engineering/electromagnetic-compatibility-emc/immunity/ss-en-55035/



https://www.electronicdesign.com/power-management/whitepaper/21179428/electronic-design-emi-shielding-for-drones-and-uavs



https://www.teracom.se/privat/Support/frekvenstabeller-radio/

https://www.teracom.se/privat/Support/frekvenstabeller-radio/

https://doi.org/10.1109/APEMC49932.2021.9597146

https://doi.org/10.1109/APEMC49932.2021.9597146

7 Appendix

7.1 Images of measurement hardware

Figure 34: Measurement receiver used during radiated emissions measurements.

Figure 35: Signal generator used in immunity testing.

Figure 36: The three power amplifiers used during immunity testing.

Figure 37: Power meter for monitoring during immunity testing.

7.2 Results from measurements on single motor from com-pany A drone

Figures 38 and 39 show the emissions from a single thrusting motor being testedseparately from the drone, aside from the motor being connected to the dronesbattery for power. The figures correspond to having the motor run with a lowRPM and no RPM respectively. To control the motor it is also connected viaCANbus to the outside of the FAR. The motor was tied to the top of the holderused for holding the company A drone vertically and it should be noted that thetests were performed with the drone still inside the chamber (but not poweredon of course), which could have some impact on the results.

Figure 38: Radiated emissions from a single thrusting motor tested separatelyfrom the drone aside from being connected to its battery.

Figure 39: Emissions from the single thrusting motor when it’s not spinning (noRPM).

Electromagnetic compatibility of unmanned aircraft - DiVA Portal

Documents