Towards Dynamic Master Determination in MEMS-based Micro ...
Post on 30-May-2022
3 Views
Preview:
Transcript
Towards Dynamic Master Determination in MEMS-based
Micro-Scanning LiDAR Systems
Philipp Stelzer1, Andreas Strasser1, Christian Steger1 and Norbert Druml2
Abstract— Automated driving has been expected for decades.The first systems, which at least partially automate the vehicle,have been installed in higher priced vehicles for several years.In the near future, however, many more competencies are tobe transferred to the systems and the vehicle will thus befully automated. Such systems receive their data from varioussensor systems such as Light Detection And Ranging (LiDAR).Therefore, it is essential that this information is transmittedcorrectly and reliably to the environmental perception system.In order to ensure this, redundancy of sensors is a key factorin addition to diversity. For example, multiple, independent-ly controlled MEMS-based LiDAR systems can be operatedsynchronously. This requires the selection of a Master systemwhich can be reliably followed by all Slave systems. In thispublication, an architecture for MEMS-based Micro-ScanningLiDAR systems is proposed to determine the appropriate systemas Master. The architecture has been implemented in an FPGAprototyping platform to demonstrate its feasibility and evaluateits performance.
I. INTRODUCTION
Until now, Advanced Driver-Assistance Systems (ADAS)
have taken over tasks such as lane keeping, but the driver
had to be ready to intervene at any time [1]. For example, in
Article 8 of the Vienna Convention on Road Traffic, some
countries have agreed that the driver must have a continuous
opportunity to gain control of the vehicle [2]. At the moment
there are also some considerations about how, when, how
long, and where a human can intervene in systems of auto-
mated vehicles. Terken et al. [3], for example, have shown in
their article about various use cases where shared control can
occur. Flemisch et al. [4] also made considerations to this
effect in their article. However, these approaches are only
possible up to SAE level 3, since beyond that the driver is
only a passenger without responsibilities. These SAE levels
are well explained in SAE International Standard J3016 [5].
At the moment, this is still a vision for the future, but
in a few years, highly automated vehicles will appear on
the roads. Consequently, there have been several projects
addressing this issue for some time, developing concepts
and systems to enable SAE level 4 and 5 of vehicles. In
the PRYSTINE project, for instance, camera, robust Radio
Detection And Ranging (RADAR), and Light Detection
And Ranging (LiDAR) sensors are deployed to achieve
Fail-operational Urban Surround perceptION (FUSION) [6].
Figure 1 illustrates this concept of PRYSTINE from the
intended FUSION. How such fail-operational architectures
1Institute of Technical Informatics, Graz University of Technology, 8010Graz, Austria {stelzer, strasser, steger}@tugraz.at
2ATV Sense and Control, Infineon Technologies Austria AG, 8020 Graz,Austria {norbert.druml}@infineon.com
for environmental perception systems can be designed is
also described by Kohn et al. [7] in their publication. Fail-
operational behaviour is mandatory for high automated vehi-
cles of SAE level 4 and 5 and is also claimed by Vermesan
et al. [8] and Macher et al. [9]. In the absence of the driver at
SAE levels 4 and 5, the vehicle requires multiple systems re-
sponsible for execution of steering, acceleration/deceleration
and monitoring of driving environment. Also, the fall-back
performance of dynamic driving task is no longer the driver,
but from SAE level 4 the system as well. Such fusion systems
combine data from different sensors and thus provide reliable
information about the driving environment in the fusion of
these data. Koci et al. [10] and De Silva et al. [11] discussed
in their publications the importance of such sensors and
data fusion for automated vehicles. Therefore, affordable
automotive qualified RADAR and LiDAR components that
can be deployed in large vehicle fleets for environmental
detection are important. An automotive qualified MEMS-
based LiDAR has been presented by Yoo et al. [12], which
can be a key for affordable LiDAR sensor systems in highly
automated vehicles. But to achieve the largest possible Field-
of-View (FoV), multiple MEMS-based LiDAR systems must
be synchronised. For this purpose, a Master system has
to be selected that operates a MEMS mirror that has the
appropriate physical properties so that all Slave systems can
reliably follow it. If the LiDAR systems were operated asyn-
chronously, so-called ghost objects could occur, as described
by Baumgart et al. [13]. Therefore, our publication addresses
the determination of an appropriate Master system for the
synchronisation of multiple independently controlled MEMS
Fig. 1. PRYSTINE’s concept view of a Fail-operational Urban SurroundperceptION (FUSION) [6].
mirrors.
With our paper contribution we:
• achieve a determination of an appropriate Master sys-
tem,
• prevent a system crash by determining an ineligible
Master system and
• enhance safety by enabling a safe and reliable synchro-
nisation of MEMS-based LiDAR systems.
The remainder of the paper is structured as follows. The
overview on related work is given in Section II. The method
will be described in detail in Section III and the results
including a short discussion will be provided in Section IV.
A summary of the findings will conclude this paper in
Section V.
II. RELATED WORK
There are already LiDAR systems on the market that can
perceive the driving environment and still have a very large
FoV. Such a system is for example the HDL-64E system from
Velodyne [14], [15]. The major drawbacks of such systems,
however, are that they either perform poorly or are bulky,
expensive, and energy hungry [16]. Additionally, there are
also MEMS-based LiDAR systems available. A distinction is
made between LiDAR systems with 1D MEMS mirrors and
2D MEMS mirrors. Wang et al. [16] describe the advantages
and disadvantages well in their publication. It is also evident
from this publication that 1D scanning MEMS mirrors are
the more mature and typically have wider scanning angles,
larger apertures and higher resonant frequencies. Therefore,
in terms of automotive applications for highly automated
vehicles, a MEMS-based LiDAR system with a 1D MEMS
mirror is arguably the better choice. Thus, Druml et al. [17]
have introduced an automotive qualified, low-cost, long-
range and robust 1D MEMS Micro-Scanning LiDAR system.
However, Druml et al.’s system has either a small and long-
range FoV or a wide and short-range FoV, so a way has to
be found to achieve both.
A. 1D MEMS Micro-Scanning LiDAR
In this Subsection, the MEMS-based LiDAR System ap-
proach by Druml et al. is presented. With this approach, a
vertical laser beam line is typically projected into the scenery.
This laser beam line is then moved through the scenery with
Fig. 2. Functional principle of a 1D Micro-Scanning LiDAR [17].
System
Safety
Controller
(AURIX)
Laser Illumination
MEMS Mirror
MEMS
Driver
ASIC
Actuation
Sensing
Reflected Signal
Photo Diodes
dt
Emitted
Signal Point
Cloud
Data
Trigger and
Laser Power Setting
FPGA / Dedicated
LiDAR Hardware
Accelerators
Receiver
Circuits
Raw Data
Emitter Path
Receiver Path
Trigger and
Gain Setting
Config
and
Status
Fig. 3. System concept of a 1D MEMS-based automotive qualified LiDARsystem by Druml et al. [17].
the help of an oscillating 1D MEMS mirror. This method
is called horizontal scanning and its principle is depicted in
Figure 2. What it depends on which FoV is obtained will
be explained later. First, the architecture of Druml et al.,
shown in Figure 3, will be explained in general. This MEMS-
based LiDAR system basically consists of an Emitter path, a
Receiver path, and a System Safety Controller (AURIX) [18]
that coordinates the Emitter and Receiver paths. As the
mirror frequency is indirectly responsible for the size and
depth of the FoV, we have focused on the Emitter path,
especially the MEMS Driver ASIC, of Druml et al.’s LiDAR
system. The MEMS Driver actuates, senses and controls the
mirror. A MEMS device can be operated either in open-loop
or closed-loop, according to Borovic et al. [19]. Actuation
of the mirror happens by switching a High Voltage (HV) on
and off. When the HV is switched on, the mirror is pulled
towards the zero position. Once the HV is switched off, an
oscillation in direction of the maximum mirror deflection
occurs. Stelzer et al. [20] state that the maximum mirror
Fig. 4. Druml et al.’s MEMS mirror response curve [17].
Master Management System
HV On/OffSensing HV On/OffSensing
f1 f2
HV On/OffSensing
fn
MEMS System
2
MEMS System
nF
allb
ack F
req
ue
ncy
Regis
ter
Fa
llback F
req
ue
ncy
Com
pasrion
Blo
ck
Ma
ster
Dete
rmin
ation B
lock
MEMS System
1
MMSC_1MMSC_2
MMSC_n
FBF_nFBF_2FBF_1
Fig. 5. Block diagram of a MEMS-based LiDAR environment perception system with an additional Master Management block.
deflection depends on the actuation frequency. A low mirror
oscillation frequency results in a narrow but long-range FoV
and a higher mirror oscillation results in a wider but short-
range FoV. Furthermore, the mirror has to be operated in a
certain frequency range to achieve a corresponding maximum
mirror deflection. This is conditioned by the implemented
electrostatic comb drive approach [21] and the associated
mechanical MEMS mirror characteristics. Figure 4 shows
the response curve of the MEMS mirror - the MEMS mirror
corresponds to a nonlinear oscillator. The two threshold
frequencies fjump and ffb can also differ from example to
example due to process variations.
B. Synchronisation of MEMS-based LiDAR Systems
In order to extend the FoV of an environment perception
system and to enhance safety, it is possible to integrate addi-
tional sensors into the system. Figure 6 shows symbolically
how the field of view is extended by two additional MEMS-
based LiDAR systems. However, in the case of MEMS-based
LiDAR systems, these individual sensors must be synchro-
nised to avoid possible interferences in the measurements.
Fig. 6. Icon of multiple independently controlled and synchronised MEMS-based LiDAR systems [22].
How these independently controlled MEMS mirrors can be
synchronised is presented by Stelzer et al. [22] and Strasser et
al. [23] in their publications. Since each mirror has different
fallback frequencies caused by process variations, not every
mirror is equally well suited as a Master mirror. Depending
on the characteristics of the different MEMS mirrors, the
Master system should be selected. The maximum possible
FoV of the entire environment perception system will also
result from this.
From this, the following research questions emerge:
• Is it possible that despite the different characteristics of
MEMS mirrors, the system with the most appropriate
MEMS mirror is always made the Master?
• What happens if the Master crashes for some reason
and becomes unavailable to the environment perception
system?
• Is it necessary to check more often whether the most
appropriate MEMS mirror is the Master mirror or is the
Master determined only once?
III. DYNAMIC MASTER DETERMINATION IN
MEMS-BASED LIDAR SYSTEMS
In this Section, we introduce the novel dynamic Mas-
ter determination architecture and describe the process of
determining the Master system. In Figure 5, the Master
Management System (MMS) is shown in general. It shows
separate MEMS-based LiDAR systems that are synchronised
together to form the LiDAR environment perception system.
In principle, n systems can be synchronised with each other
and thus a larger FoV can be obtained. For the synchroni-
sation, however, a Master is needed, which is determined
and assigned by the MMS. The MMS is implemented on an
external device, for example the System Safety Controller,
and communicates with the individual LiDAR systems. The
various fallback frequencies of the individual MEMS mirrors
are communicated to the MMS via the Fallback Frequency
(FBF) m connections - m stands for 1 to n. These are stored
in a register of the MMS and subsequently compared with
each other. The system with the MEMS mirror that has the
lowest fallback frequency is then designated as the Master.
This is because it ensures that all Slave mirrors can follow
HV On/OffSensing
fm
MMSC_m FBF_m
MEMS System m
Fallback Frequency Detection Block
Frequency
Adaption
Block
FBFDT
……
…
FBFT
……
……
Fallback
Frequency
Determination
Block
Fig. 7. Concept of a MEMS-based LiDAR system with an integratedfallback frequency detection block.
the Master mirror. If a Master were chosen that had a higher
fallback frequency than at least one Slave, there would be a
risk that at least one Slave system would permanently crash.
In the worst case, this can have safety-relevant consequences,
such as an accident due to a lack of environment perception.
Via Master Management System Communication (MMSC)
m the individual MEMS-based LiDAR systems are informed
whether they are the Master or not. Those systems that
are Slaves then switch to Slave mode. How the fallback
frequency of the individual MEMS mirrors is determined
is explained below.
A. Fallback Frequency Determination in MEMS-based Li-
DAR Systems
In this Subsection it is explained how the fallback fre-
quencies of the individual MEMS-based LiDAR systems
are determined. To determine the fallback frequency of the
individual MEMS system, an additional module must be
integrated into the existing system, as shown in Figure 7. In
this module, there is a frequency adaption block, which first
reduces the frequency to reach the jump frequency and then
increases it until the fallback frequency is reached. However,
this process only starts once the Fallback Frequency Determi-
nation Trigger (FBFDT) is set. This trigger can be set either
by the System Safety Controller, the system itself at system
start-up or by another external device on which the MMS
is implemented. When the fallback frequency is reached,
this value is written into a register via Fallback Frequency
Threshold (FBFT) and subsequently transmitted to the MMS.
To determine the jump frequency and fallback frequency, an
analog-to-digital converter (ADC) value of the Driver is used.
By means of this value, significant frequency changes can be
detected. More about this follows in Section IV.
B. Dynamic Master Determination Procedure
Next, the dynamic Master determination procedure is de-
scribed in detail in this Subsection. How the dynamic Master
determination procedure looks like is shown in Figure 8 and
described in the following steps:
1) System Start-Up
In the process of the system start also the trigger
FBFDT is set to get the current fallback frequency
of the MEMS system right at the beginning. During
operation it is not necessary to start another check for
the correct fallback frequency, only at the next system
start.
2) Mirror Frequency Reduction
At the beginning the frequency is reduced step by step
in open-loop mode. This is continued until the jump
Start
System Start-Up
Trigger
Frequency
Adaption
Yes
No
End
Reduce
Mirror
Frequency
Reached Jumping
Frequency not yet
Reached
Jumping
Frequency
Raise
Mirror
FrequencyReached Fallback
Frequency not yet
Reached
Fallback
Frequency
Set
Fallback
Frequency
in Register
Get Fallback
Frequency
from MEMS
System m
Store
Fallback
Frequency
in Register
MEMS System m
Master Management System
Fallback
Frequency
Comparison
Set MEMS
System m
as Slave
Fallback
Frequency
lower than
from another
MEMS System
Set MEMS
System m
as Master
Fallback
Frequency
higher than
from another
MEMS System
Fig. 8. Process flow of the Dynamic Master Determination Procedure forMEMS-based LiDAR environment perception systems.
frequency of the mirror is reached. It is checked by
means of an ADC value of the Driver - explained in
the Section IV.
3) Mirror Frequency Raise
After that, the frequency is increased step by step until
the fallback frequency of the mirror is reached. At
this point, the mirror falls back to the lower resonance
curve. This is also detected by means of the ADC value
of the Driver.
4) Set Fallback Frequency
This value is then written to a register and forwarded
to the MMS via MMSC. At the same time, the mirror
is brought back to the upper resonance curve and put
into closed-loop mode. This will put the mirror into
operation mode and will make it ready to become
Master or to synchronise to the selected Master.
5) Get Fallback Frequency from MEMS Systems
After receiving the fallback frequencies from the
MEMS systems, these are written to a register in the
MMS. The system remains in the waiting state until
all connected MEMS systems have transmitted their
fallback frequencies to the MMS.
6) Comparison of Fallback Frequencies
Subsequently, the entered fallback frequencies are
compared with each other and sorted by means of a
common sorting algorithm. A priority list is created
within the MMS, based on these priorities the Master
system is selected. The system with the mirror that
has the lowest fallback frequency is given the highest
priority. The other priorities are assigned analogously.
7) Set Master System
The priority list is then consulted to decide which
system becomes the Master. This is also communicated
to the MEMS systems via MMSC. As long as the
system with the highest priority is available, this is
now the Master system. However, if this system is not
available for a certain period of time, the system with
0 0.5 1 1.5 2
Mirror Half Periods [1] 104
400
450
500
550
DL
2 a
vera
ge
[1
]
0 0.5 1 1.5 2
Mirror Half Periods [1] 104
4000
4200
4400
4600
4800
5000
Fre
qu
en
cy [
Hz]
Fig. 9. Measurement with the first MEMS mirror sample - fallbackfrequency at 4899 Hz.
0 0.5 1 1.5 2
Mirror Half Periods [1] 104
400
450
500
550
DL2 a
vera
ge [1]
0 0.5 1 1.5 2
Mirror Half Periods [1] 104
4000
4200
4400
4600
4800
5000
Fre
quency
[H
z]
Fig. 10. Measurement with the second MEMS mirror sample - fallbackfrequency at 4850 Hz.
the following higher priority automatically follows as
Master.
With this new dynamic Master determination procedure,
it is now possible to create an enlarged field of view with
multiple MEMS-based LiDAR systems and an appropriate
Master. This significantly increases the safety and reliability
of the entire environment perception system, as there is no
more risk of selecting an inappropriate Master. It is thus
possible to obtain the maximum possible FoV with the given
components without risking a system crash.
IV. RESULTS
In this Section we provide the test results of our novel
dynamic Master determination procedure, which has been
introduced in Section III. Our architecture was implemented
in an FPGA Prototyping Platform by Yoo et al. [12]. This
FPGA Prototyping Platform consists of a MEMS Driver
0 0.5 1 1.5 2
Mirror Half Periods [1] 104
400
450
500
550
DL2 a
vera
ge [1]
0 0.5 1 1.5 2
Mirror Half Periods [1] 104
4000
4200
4400
4600
4800
Fre
quency
[H
z]
Fig. 11. Measurement with the third MEMS mirror sample - fallbackfrequency at 4719 Hz.
0 0.5 1 1.5 2
Mirror Half Periods [1] 104
400
450
500
550
DL
2 a
vera
ge
[1
]
0 0.5 1 1.5 2
Mirror Half Periods [1] 104
4000
4200
4400
4600
4800
5000F
req
ue
ncy
[H
z]
Fig. 12. Measurement with the fourth MEMS mirror sample - fallbackfrequency at 4902 Hz.
Digital Part (FPGA), a MEMS Driver Discrete Analog Part,
and the 1D MEMS mirror. For our test setup, we have used
four different MEMS mirrors. The procedure was performed
for all four MEMS mirror samples. It can be clearly seen in
the measurements at which point in time the jump frequency
or fallback frequency of the respective mirror is reached.
Each plot contains a measurement of the actuation frequency
and a measurement of the Device Layer 2 (DL2) average
value. The frequency plot shows the frequency at which the
Driver switches the HV on and off. The DL2 average value
is an ADC value of the Driver and reflects the currents at
the MEMS mirror combs. The higher the DL2 average value,
the lower the mechanical deflection of the mirror. Between
the jump to the upper resonance curve and the fallback
to the lower resonance curve, it can be clearly seen that
the deflection is much more than on the lower resonance
curve. Using the four different MEMS mirrors, it can also
be clearly seen that there are different fallback frequencies.
The fallback frequencies of samples 1 and 4 are within a few
Hz, as can also be seen in Table I. The fallback frequency
TABLE I
FALLBACK FREQUENCIES OF MEMS MIRROR SAMPLES
Sample Fallback Frequency Determined Master Order
[Hz]
1 4899 3rd2 4850 2nd3 4719 1st4 4902 4th
of mirror number 2 is approximately 50 Hz lower than the
fallback frequency of mirrors 1 and 4. Mirror number 3’s
frequency is almost 200 Hz away. Thus, mirror number 3
would be the Master mirror among these four mirrors. If this
mirror would fail, mirror number 2 would follow as Master.
V. CONCLUSION
In our paper, we have introduced a novel Master deter-
mination procedure for synchronised MEMS-based LiDAR
systems. With this new procedure, we are able to select an
appropriate Master and thus obtain the maximum possible
FoV with the given components. We can also answer the
research question whether it is possible to always select the
most appropriate Master despite different characteristics with
yes. Since the mirror with the lowest fallback frequency is
always the most appropriate one, we always choose the most
appropriate Master. This should also prevent a system crash
of Slaves due to the operation of the mirrors too close to their
fallback frequencies. However, if the Master system crashes
unexpectedly, the Slave systems do not crash as well, but the
next better Slave is changed to be the Master. The FoV is
then more limited, but the perception of the environment is
still possible. We can also clearly answer the last research
question. It is not necessary to check more often than at
system startup which mirror is most appropriate as Master.
At the beginning a check is done and also a priority list
is created. If the Master fails, it is known which system
will replace the old Master. In case of changed fallback
frequencies due to damage or aging, it will be noticed at
the next system startup. Usually, a system start is performed
at least once a day, because no vehicle and its components
are in continuous operation. Finally, our presented procedure
increases the safety and reliability of a synchronised LiDAR
system and related applications. It is another step toward
highly automated respectively autonomous vehicles.
ACKNOWLEDGMENT
The authors would like to thank all national funding
authorities and the ECSEL Joint Undertaking, which funded
the PRYSTINE project under the grant agreement number
783190.
PRYSTINE is funded by the Austrian Federal Ministry
of Transport, Innovation and Technology (BMVIT) under
the program ICT of the Future between May 2018 and
April 2021 (grant number 865310). More information: http-
s://iktderzukunft.at/en/.
REFERENCES
[1] C. Brunglinghaus, “Wie das Recht automatisiertes Fahren hemmt,”ATZ - Automobiltechnische Zeitschrift, vol. 117, no. 4, pp. 8–13, Apr2015. [Online]. Available: https://doi.org/10.1007/s35148-015-0039-0
[2] United Nations Conference on Road Traffic, “19 . Conventionon Road Traffic,” https://treaties.un.org/pages/ViewDetailsIII.aspx?src=TREATY&mtdsg no=XI-B-19&chapter=11&Temp=mtdsg3&clang= en, retrieved: October, 2019. [Online]. Available:https://treaties.un.org/pages/ViewDetailsIII.aspx?src=TREATY&mtdsg no=XI-B-19&chapter=11&Temp=mtdsg3&clang= en
[3] J. Terken and B. Pfleging, “Toward shared control between automatedvehicles and users,” Automotive Innovation, vol. 3, no. 1, pp. 53–61,2020.
[4] F. Flemisch, D. Abbink, M. Itoh, M.-P. Pacaux-Lemoine, andG. Weel, “Shared control is the sharp end of cooperation: Towardsa common framework of joint action, shared control and humanmachine cooperation,” IFAC-PapersOnLine, vol. 49, no. 19, pp. 72 –77, 2016, 13th IFAC Symposium on Analysis, Design, and EvaluationofHuman-Machine Systems HMS 2016. [Online]. Available: http://www.sciencedirect.com/science/article/pii/S2405896316320547
[5] SAE, “SAE International Standard J3016 - Taxonomy and Definitionsfor Terms Related to On-Road Motor Vehicle Automated DrivingSystems,” SAE International, Standard, January 2014.
[6] N. Druml, G. Macher, M. Stolz, E. Armengaud, D. Watzenig, C. Ste-ger, T. Herndl, A. Eckel, A. Ryabokon, A. Hoess, S. Kumar, G. Dim-itrakopoulos, and H. Roedig, “Prystine - programmable systems forintelligence in automobiles,” in 2018 21st Euromicro Conference on
Digital System Design (DSD), Aug 2018, pp. 618–626.
[7] A. Kohn, R. Schneider, A. Vilela, A. Roger, and U. Dannebaum,“Architectural Concepts for Fail-Operational Automotive Systems,” inSAE 2016 World Congress and Exhibition. SAE International, April2016.
[8] O. Vermesan, R. Bahr, R. John, M. Ottella, R. Gjølstad, O. Buckholm,and H.-E. Sand, “Advancing the Design of Fail-Operational Architec-tures, Communication Modules, Electronic Components, and Systemsfor Future Autonomous/Automated Vehicles,” in Intelligent System
Solutions for Auto Mobility and Beyond, C. Zachaus and G. Meyer,Eds. Cham: Springer International Publishing, 2021, pp. 53–71.
[9] G. Macher, N. Druml, O. Veledar, and J. Reckenzaun, “Safety andSecurity Aspects of Fail-Operational Urban Surround perceptION(FUSION),” in Model-Based Safety and Assessment, Y. Papadopoulos,K. Aslansefat, P. Katsaros, and M. Bozzano, Eds. Cham: SpringerInternational Publishing, 2019, pp. 286–300.
[10] J. Koci, N. Jovii, and V. Drndarevi, “Sensors and Sensor Fusionin Autonomous Vehicles,” in 2018 26th Telecommunications Forum
(TELFOR), 2018, pp. 420–425.
[11] V. De Silva, J. Roche, and A. Kondoz, “Robust Fusion ofLiDAR and Wide-Angle Camera Data for Autonomous MobileRobots,” Sensors, vol. 18, no. 8, 2018. [Online]. Available:https://www.mdpi.com/1424-8220/18/8/2730
[12] H. W. Yoo, N. Druml, D. Brunner, C. Schwarzl, T. Thurner, M. Hen-necke, and G. Schitter, “MEMS-based lidar for autonomous driving,”e & i Elektrotechnik und Informationstechnik, vol. 135, no. 6, pp.408–415, Oct 2018.
[13] M. Baumgart, C. Consani, M. Dielacher, and N. Druml, “Opticalsimulation of Time-of-Flight sensor accuracy in rain,” in The Euro-
pean Conference on Lasers and Electro-Optics. Optical Society of
America, 2017, p. CH P 25.[14] Velodyne LiDAR, “HDL-64E,” 2016.[15] R. Halterman and M. Bruch, “Velodyne HDL-64E lidar for unmanned
surface vehicle obstacle detection,” in Unmanned Systems Technology
XII, G. R. Gerhart, D. W. Gage, and C. M. Shoemaker, Eds., vol.7692, International Society for Optics and Photonics. SPIE, 2010,pp. 123 – 130. [Online]. Available: https://doi.org/10.1117/12.850611
[16] D. Wang, C. Watkins, and H. Xie, “MEMS Mirrors for LiDAR: AReview,” Micromachines, vol. 11, no. 5, 2020. [Online]. Available:https://www.mdpi.com/2072-666X/11/5/456
[17] N. Druml, I. Maksymova, T. Thurner, D. Van Lierop, M. Hennecke,and A. Foroutan, “1D MEMS Micro-Scanning LiDAR,” in The Ninth
International Conference on Sensor Device Technologies and Appli-
cations (SENSORDEVICES 2018), 09 2018.[18] Infineon Technologies AG. (2018) AURIX System Safety Controller.
https://iot-automotive.news/aurix-microcontroller-tc3xx-family-infineon-fuels-automated-driving-electromobility.
[19] B. Borovic, A. Q. Liu, D. Popa, H. Cai, and F. L. Lewis, “Open-loopversus closed-loop control of MEMS devices: choices and issues,”Journal of Micromechanics and Microengineering, vol. 15, no. 10, p.1917, 2005.
[20] P. Stelzer, A. Strasser, C. Steger, A. Garcia, and N. Druml, “Fast AngleAdaption of a MEMS-based LiDAR System,” IFAC-PapersOnLine,vol. 52, no. 15, pp. 55 – 60, 2019, 8th IFAC Symposium onMechatronic Systems MECHATRONICS 2019.
[21] P. Deng and W. Ma, “Nonlinearity investigation of the MEMSscanning mirror with electrostatic comb drive,” in The 9th IEEE
International Conference on Nano/Micro Engineered and Molecular
Systems (NEMS), 2014, pp. 212–215.[22] P. Stelzer, A. Strasser, C. Steger, H. Plank, and N. Druml, “To-
wards Synchronisation of Multiple Independent MEMS-based Micro-Scanning LiDAR Systems,” in 2020 IEEE Intelligent Vehicles Sympo-
sium (IV), 2020, pp. 1080–1085.[23] A. Strasser, P. Stelzer, C. Steger, and N. Druml, “Towards Synchronous
Mode of Multiple Independently Controlled MEMS Mirrors,” IFAC-
PapersOnLine, vol. 52, no. 15, pp. 31 – 36, 2019, 8th IFAC Sympo-sium on Mechatronic Systems MECHATRONICS 2019.
top related