Y o u r s t r a t e g i c p a r t n e r 1 RESEARCH METHODS: From Problem and Hypothesis to Experiments Dr. AARNE MÄMMELÄ Research Professor (VTT), Docent.

ELECTRONICS

1

Y o u r s t r a t e g i c p a r t n e r

RESEARCH METHODS:From Problem and Hypothesis to Experiments

Dr. AARNE MÄMMELÄResearch Professor (VTT), Docent (HUT)Oulu, Finland 6.11.2001

VTT ELECTRONICSKaitoväylä 1, P.O. Box 1100, FIN-90571 Oulu, Finland

Email: [email protected], http://www. vtt.fi/ele

Tel. 5512111, 5512482 (direct), 040-5762963 (GSM), Fax 5512320

ELECTRONICS

2

Y o u r s t r a t e g i c p a r t n e r

CONTENTS

1 Introduction2 Discovering new knowledge3 Learning process4 Would you like to work in a great group?5 Advisor is your best friend6 Research starts from ideas7 Research methods8 ConclusionsReferences

ELECTRONICS

3

Y o u r s t r a t e g i c p a r t n e r

Phases of research

Review ofliterature

Experiments& discussions

Problem &hypothesis

Paper

Prototype

ELECTRONICS

4

Y o u r s t r a t e g i c p a r t n e r

1 Introduction

•in science/engineering our aim is an organized body of information (= knowledge) collected in the literature

•research means that new knowledge is discovered (an original contribution to the literature)

•elements of research work include ideas, people, funding, culture

•key questions: where to find existing knowledge, how to discover new knowledge, how to publish new knowledge

•we will first discuss how to discover new knowledge

ELECTRONICS

5

Y o u r s t r a t e g i c p a r t n e r

2 Discovering new knowledge

Literature (knowledge)

Researchers

Peer review

Editor

ELECTRONICS

6

Y o u r s t r a t e g i c p a r t n e r

HOW TO DO RESEARCH (1)

Literature (knowledge)

Problem

Peer review

Editor

Manuscript

Experiments Discussion

ELECTRONICS

7

Y o u r s t r a t e g i c p a r t n e r

What do research and development mean?•In research we discover new knowledge

basic research: no specific applications or immediate commercial objectives in mind (scientific papers emphasized)

applied research: ideas into operational form (product ideas and patents emphasized)

•In development we use systematically the existing knowledge

production of useful materials, systems and methods (inc. design and development of prototypes and processes and design documents).

•Discuss. What is a research prototype? Should it be applied research or development?

ELECTRONICS

8

Y o u r s t r a t e g i c p a r t n e r

Comments on research

•high risks, takes long time, usually supported only by the government or very large companies

•abstract and vague aims; must tolerate ambiguity, implies a lot of frustration and special independence and maturity

•ideally research should be free from any standards

•research in science: research into discovering new laws of the nature

•research in engineering: research done for new man-made products, proprietary or standardized, not found in the nature

ELECTRONICS

9

Y o u r s t r a t e g i c p a r t n e r

Why is research important?•New knowledge is discovered•Prestige for yourself and for your employer•Know the state of the art and teach it to your colleagues and customers

•Know the history and see the trends

Example. During the Word War II the smartest people developed radar instead of larger binoculars!

ELECTRONICS

10

Y o u r s t r a t e g i c p a r t n e r

3 Learning process

How do engineering students learn?•The process is quite similar to learning the first language, we learn sounds, words and example sentences (applications) first, grammar (principles) afterwards.

•Professors try to teach principles first and applications later (if ever).

•We must partially reverse the process by starting from simple examples (induction).

•General principles are emphasized later to really master the subject (deduction).

•It is helpful to know at least some simple principles in the beginning.

ELECTRONICS

11

Y o u r s t r a t e g i c p a r t n e r

What a doctoral student should learn?•Excellent oral and written knowledge of native language and English.

•Know the literature of a specific topic (big picture, history, state of the art, future trends).

•Know how to discover new knowledge (research methods).

•Publish some original papers and write a thesis (contribution to the literature).

•Learn to discuss and argue in seminars (public defence).

•Guide master’s students (pedagogical skills).

ELECTRONICS

12

Y o u r s t r a t e g i c p a r t n e r

How to start?

•Find a suitable advisor and a good group.•Look for a good idea, study literature & discuss, do not invent the wheel.

•Define the problem, limit the scope, find the right approach and hypotheses (= possible solutions).

•Analyze the system, make experiments (simulations, prototypes) and discuss the results, use right tools.

•Write a paper or thesis and listen carefully to comments and be prepared to argue and defend your claims (opponents try to find weak points in your reasoning!).

•Remember the value of criticism all the time, argue and discuss.

ELECTRONICS

13

Y o u r s t r a t e g i c p a r t n e r

Systematic work vs. creativity •Use systematic methods to find literature and write your own papers.

•Research itself is a creative (probably chaotic) process and cannot be fully systematized: output cannot be predicted from the various inputs used.

•You must think differently from the way others think; you may loose your creativity if you use only past knowledge.

•“Genius is 99 % of perspiration and 1 % of inspiration” (Edison).

•The aim of science/engineering is the truth, which is actually inaccessible.

ELECTRONICS

14

Y o u r s t r a t e g i c p a r t n e r

4 Would you like to work in a great group?Properties of a good research group•A clear, seductive and attainable vision & strategy.•A strong leader with managerial/engineering skills.•Talented optimistic people co-operating: advisors, idea-generating and marketing people, communicators, need for autonomy & change, comfortable with abstract thinking, ability to conceptualize.

•Characteristic spirit: impersonal judgments, unique communication network, avoid groupthinking.

•Triple hierarchy: management and professional hierarchies not completely separate.

ELECTRONICS

15

Y o u r s t r a t e g i c p a r t n e r

5 Advisor is your best friendLook for a good advisor•Be there for the length of your project•Experience on research in the same area (a doctor)•Pedagogical skills, know the big picture, know literature

•Respected by colleagues, critical, tough methodologist

•Interested in your topic, gives comments, you respect him/her

ELECTRONICS

16

Y o u r s t r a t e g i c p a r t n e r

How to keep your advisor?•Orient your advisor (system model, block diagrams, table of contents)

•Follow instructions (make notes), but also discuss and argue

•Make concise progress reports (organize the material, limit the scope)

•Do not expect ready-made solutions, but ways of thinking

•Advisor needs also credit for his/her work in the form of publications

•Get into the driver’s seat!

ELECTRONICS

17

Y o u r s t r a t e g i c p a r t n e r

6 Research starts from ideas Selection of topic•Know the literature, especially original papers (write brief summaries)

•Do experiments, use your colleagues’ experience

•Discuss with students and colleagues and teach them

•Extend earlier theory, look for gaps and missing links

•Use ideas from other areas, change view point, improve your creativity

•In addition to the right problem you need right timing and right approach.

ELECTRONICS

18

Y o u r s t r a t e g i c p a r t n e r

Big issues guiding our work

History &roadmaps

Fundamentallimits

Systemsengineering

ELECTRONICS

19

Y o u r s t r a t e g i c p a r t n e r

History of engineering is related to the history of mathematics

ELECTRONICS

20

Y o u r s t r a t e g i c p a r t n e r

Example landmark paper

ELECTRONICS

21

Y o u r s t r a t e g i c p a r t n e r

Systems engineering (1)

ELECTRONICS

22

Y o u r s t r a t e g i c p a r t n e r

Systems engineering (2)Examples:•theory of science•knowledge of history, roadmaps (vision), strategy

•value chain, total quality management (TQM)•hierarchy, modularity, reference models•knowledge on fundamental limits, network capacity

•generalized fading channel and noise models with diversity

•unified approach to performance analysis

ELECTRONICS

23

Y o u r s t r a t e g i c p a r t n e r

Systems engineering (3)

Definitions:•Systems engineering: the discipline of managing the development of complex systems; a combination of engineering theories and practice with the theory of architecting and practice.

•Roadmap: An extended look at the future of a chosen field of inquiry composed from the collective knowledge and imagination of the brighest drivers of change in that field.

ELECTRONICS

24

Y o u r s t r a t e g i c p a r t n e r

Fundamental limits: attenuation

ELECTRONICS

25

Y o u r s t r a t e g i c p a r t n e r

Fundamental limits: noise

ELECTRONICS

26

Y o u r s t r a t e g i c p a r t n e r

HOW TO DO RESEARCH (2)

Literature

State-of-the-art

Problem

Roadmap Vision

Hypothesis

Present Future

State-of-the-art

Experiments

ELECTRONICS

27

Y o u r s t r a t e g i c p a r t n e r

Phases of research

Review ofliterature

Experiments& discussions

Problem &hypothesis

Paper

Prototype

ELECTRONICS

28

Y o u r s t r a t e g i c p a r t n e r

Research proposal AbstractProblem and hypothesisReview of the literatureMaterials and methods

plan of operation, experimental procedures analytical etc. tools, anticipated experimental

data

Discussion (can be combined with conclusions)originality (novelty), open questions, limitations

Conclusionsvalidation, significance, applications

Time frame, budget and biographycurriculum vitae (for grant proposal)

AppendicesBibliography

ELECTRONICS

29

Y o u r s t r a t e g i c p a r t n e r

Doctoral thesis

Abstract1. Introduction2. Literature review3. Materials and methods4. Results5. Discussion6. ConclusionsReferencesAppendices

Note. Results and discussion may be combined.

ELECTRONICS

30

Y o u r s t r a t e g i c p a r t n e r

Research proposalHints.•research plan is different from ordinary project plan for development projects

•remember to include some time for literature surveys and documentation (publications must be the basic aim)

•ambiguity and risks must be tolerated, results cannot be predicted accurately, researchers need freedom, projects of 3 years typically optimum

ELECTRONICS

31

Y o u r s t r a t e g i c p a r t n e r

Curriculum vitae•Full names•Date and place of birth•Nationality•Marital status•Address, telephone•Education and training•Present position•Fields of research•Previous professional appointments•Research awards, honours and major grants•Editorial board memberships•Memberships in scientific societies•Other academic and professional merits and activities

ELECTRONICS

32

Y o u r s t r a t e g i c p a r t n e r

7 Research methods •analysis

– creates best scientific papers– simple, mathematically tractable problem, must

be often linear, system model is the key factor

•simulations – complicated systems can be developed rapidly,

but slow to simulate– basic idea: lower level blocks are simplified and

idealized– key problem: realistic models for the

environment (e.g. channel)

•prototyping – more convincing than “pure” simulations, not so

flexible, slow and expensive to develop complicated systems

– environment (channel) simulators still needed (approximations!), field tests expensive, repeatability?

ELECTRONICS

33

Y o u r s t r a t e g i c p a r t n e r

General hints (1) •good organization

– block diagrams, hierarchy, modularity, etc.– understand the system model better, explain

the problem to others– try to find independent (orthogonal) blocks!– careful testing & documentation (reports,

comments, etc.)

•always start from simple models, use idealizations, black boxes

– example: scalars instead of matrices– reduce idealizations step by step

•use analysis (decomposition) & synthesis (integration)

– proceed from optimal systems to approximations

– proceed from simple to more practical models

ELECTRONICS

34

Y o u r s t r a t e g i c p a r t n e r

General hints (2) •combine experiments & discussions•use simulations as much as possible after analysis

– prototypes as a “final proof” of the concept in a realistic environment

– case system, simulations, prototypes– seminars, colloquia, courses, discussion

groups– avoid blind man’s buff where everything

affects everything

•note the trade-off between concentration (specialization) & flexibility (generalization)

•proceed from general to specific systems and backwards

ELECTRONICS

35

Y o u r s t r a t e g i c p a r t n e r

How to make research more efficient •Advisors can help you in many ways

– collect essential literature (books, reviews, landmark papers), create an internal library

– make lists of big questions for basic research (potential master’s/doctoral thesis topics)

– make lists of new product ideas for applied research (potential EU/TEKES projects)

•Learn to know literature as well as research methods and communication skills

– language abilities, basic and applied research – be interested in other people’s problems

•We should work as a group in addition to as individual persons

– the own group should become one of the best in the world

– seminars are discussion forums, not only lectures

ELECTRONICS

36

Y o u r s t r a t e g i c p a r t n e r

Summary of research trade-offs

Encouragement

Criticism

History &roadmaps

Concentration

Creativity

Systematicwork

Flexibility

History &roadmaps

Experiments

Discussion

History &roadmaps

Details

Systems

History &roadmaps

Analysis

Synthesis

ELECTRONICS

37

Y o u r s t r a t e g i c p a r t n e r

8 Conclusions •basic elements in a research organization are people, ideas, funds and culture

•people work autonomously, tolerate ambiguity and show considerable initiative

•ideas are generated through a unique communication network

•generation of ideas is made easier by the characteristic spirit of a scientific community; scientists critically and impersonally evaluate scientific ideas and discoveries

•you need right problem, right timing and right approach

•get into the driver’s seat

ELECTRONICS

38

Y o u r s t r a t e g i c p a r t n e r

Big issues guiding our work

History &roadmaps

Fundamentallimits

Systemsengineering

ELECTRONICS

39

Y o u r s t r a t e g i c p a r t n e r

Phases of research

Review ofliterature

Experiments& discussions

Problem &hypothesis

Paper

Prototype

ELECTRONICS

40

Y o u r s t r a t e g i c p a r t n e r

References• R. K. Jain & H. C. Triandis, Management of research and

development organizations. 2nd ed. Wiley, 1997.• W. Benis & P. Biederman, Organizing genius. Addison

Wesley, 1998.• D. Sternberg, How to complete and survive a doctoral

thesis. St. Martin’s Press, 1981.• R. M. Felder & L. K. Silverman, “Learning and teaching

styles in engineering education,” Engineering Education, April 1988, pp. 674-681.

• H. B. Michaelson, How to write & publish engineering papers and reports. 3rd ed. Oryx Press, 1990.

• M. Davis, Scientific papers and presentations. Academic Press, 1997.

• Fundamental limits in Electrical Engineering (special issue). Proceedings of the IEEE, February 1981, pp. 147-282.

• R. N. Kostoff & R. R. Schaller, “Science and Technology Roadmaps,” IEEE Transactions on Engineering Management, May 2001, pp. 132 -143.

• J. N. Martin. Systems engineering guidebook: A process for developing systems and products. CRC Press, 1997.