臺灣博碩士論文加值系統

摘要
一個稱為方波的學習策略以運動的精確度來換取簡化的指令且擷取了運動軌跡的相似性，已經在先前被提出。然而，我們考慮到的除了運動的軌跡還有運動的速度需要被學會。我們延續了這個工作而提出了另一個方法稱為三角波法來實現速度的追蹤與指令的簡化。當原來複雜的指令在容許精確度的運動下，被簡化成簡單的形式且用簡單結構及較少記憶空間的的學習機制來儲存與操作。此外，這個學習機制也可以從調整指令的大小來改變其速度，所以簡化的指令提供了一組具有不同距離及速度的運動，而毋須重算系統的動力學。以網求揮拍動作為主的實驗報告證實了我們所提出的學習機制履行了精確度與簡化指令的交換與速度的追蹤。

A learning scheme, called rectangular-wave method, has been proposed previously.
It trades motion accuracy for motion command simplification and captures the similarity
of position trajectory, (eg., human signature). However, we consider not only the position but
also the velocity trajectories should be learnt.
As a continuing work, we propose another method to implement motion command
simplification and velocity tracking. We refer it as the triangular-wave method.
When the original complex motion commands capable of accurate motion tracking are simplified into simple forms,
the simplified motion commands can then be stored and manipulated by using
learning mechanisms with simple structures and scanty memory resources, and they can be
fast and smoothly executed. In addition, this learning scheme can also perform motion
command scaling in speed, so that simplified motion commands be provided for a number
of similar motions of different movement distances and velocities without system dynamics re-calculation.
Experiments based on tennis swing motions are reported that demonstrate the
effectiveness of the proposed learning scheme in implementing these accuracy-simplification and velocity tracking.

Chinese Abstract............................................................................................................................................i
English Abstract.............................................................................................................................................ii
Acknowledgements.......................................................................................................................................iii
Contents.........................................................................................................................................................iv
List of Figures................................................................................................................................................vi
1 Introduction..................................................................................................................................................1
2 Biological Backgrounds
2.1 Human Motor Control..............................................................................................................................3
2.2Eguilibrium-Point Hypothesis....................................................................................................................5
2.3Speed-Sensitive Movement Strategy..........................................................................................................6
2.4Speed-Insensitive Movement Strategy........................................................................................................7
2.5Speed-Insensitive Movement Strategy as a Default....................................................................................8
3 Learning Scheme for Motion Command Simplification and Scaling...........................................................9
3.1Motion Command Simplification...............................................................................................................9
3.1.1 Rectangular-Wave Method...................................................................................................................10
3.1.2 Triangular-wave Method.......................................................................................................................15
3.2 Motion Command Scaling........................................................................................................................20
3.2.1 Scaling in Size........................................................................................................................................21
3.2.2 Scaling in Speed.....................................................................................................................................24
4. Experiment...................................................................................................................................................27
.
4.1 Result Discriptions....................................................................................................................................30
4.2 Discussions................................................................................................................................................36
4.2.1 Performance Evaluation of Using the Triangular-Wave and the Rectangular-wave Methods.................36
4.2.2 Scaling in Size and Speed........................................................................................................................38
..
5 Conclusion
5.1 Future Work...............................................................................................................................................48
A Learning Scheme for Accurate Motion Tracking..........................................................................................49
A.1 Trajectory Mapping....................................................................................................................................50
A.2 FNN Implementation..................................................................................................................................52
A.2.1 Fussy Neural Network.............................................................................................................................53
Bibliography.......................................................................................................................................................58

[1] H.R. Berenji and P. Khedkar, ``Learning and Tuning Fuzzy Logic
Controllers Through Reinforcements,IEEE Trans. on Neural Networks,Vol. 3(5), pp. 724-740, 1992.
[2] J.J. Craig, Introduction to Robotics, Addison-Wesley, 1989.
[3] P.M. Fitts, ``The Information Capacity of the Human Motor Program in
Controlling the Amplitude of Movement,'' J. Experimental Psychology, Vol. 47,pp. 381-394, 1954.
[4]T. Flash, ``The Control of Hand Equilibrium trajectories in
Multi-Joint Arm Movements,'' Vol. 57, pp. 257-274, 1987.
[5] D.M. Corcos, G.L. Gottlieb , and G.C. Agarwal , "Organizing Principles for Single
Joint Movements-(1) Speed Insensitive Strategy ," vol 62,no2,pp.341-357, Auguest 1989.
[6] G.L. Gottlieb ,D.M. Corcos, and G.C. Agarwal , "Organizing Principles for Single
Joint Movements-(2) Speed Sensitive Strategy ," vol 62,no2, pp.358-370, Auguest 1989
[7] G.L. Gottlieb ,D.M. Corcos, G.C. Agarwal ,and M.L. Latash, "Organizing Principles for Single
Joint Movements-(3) Speed Insensitive Strategy as a Default ," {Journal of Neurophysiology.} vol 63,no3, pp.625-636 March 1990.
[8] Z. Hasan, ``Optimal Movement Trajectories and Joint Stiffness
in Unperturbed, Initially Loaded Movements," {\sl J. Biological Cybernetics,} Vol.53, pp. 373-382, 1986.
[9] J.C. Houk and W.Z. Rymer, ``Neural Control of Muscle Length and Tension,'' sl Handbook of Physiology:
The Nervous System II, Bethesda, MD, American Physiol. Soc.,Sec. 1, Vol. II, Chap. 8, pp. 257-323, 1981.
[10]M. Kawato, K. Furukawa, and R. Suzuki, ``A Hierarchical Neural-Network
Model for Control and Learning of Voluntary Movement,'' Biological Vol. 57, pp. 169-185, 1987.
[11] S.L. Lehman, ``Input Identification Depends on Model Complexity,'' in Multiple Muscle Systems, Winters and Woo, Ed., Springer-Verlag, New York, pp. 94-100, 1990.
[12] S.F. Liu, "Robot Motion Command Simplification and Scaling Using Learning Mechanisms, " S.M. thesis, Dept. Cont. Eng.,
Chiao-Tung University, Taiwan, R.O.C. 1997.
[13] A. Polit and E. Bizzi, ``Characteristics of Motor Programs Underlying
Arm Movements in Monkeys,'' J. Neurophysiology, Vol. 42(1), pp. 183-194, 1979.
[14]R. Plamondon and F. Maarse, ``An Evaluation of Motor Models of Handwriting,'' IEEE Trans. Systems, Man, and Cybernetics, Vol. 19(5), pp.1060-1072, 1989.
[15]T.D. Sanger, ``Neural Network Learning Control of Robot Manipulators Using Gradually Increasing Task Difficulty,''
IEEE Trans. Robotics and Automation,Vol. 10(3), pp. 323-333, 1994.
[16] R.A. Schmidt, Motor Control and Learning: A Behavioral Emphasis,Second Edition, Human Kinetics Publishers, Champaign, Illinois, 1988.
[17] T. Shibata and T. Fukuda, ``Hierarchical Intelligent Control for Robotic Motion,'' IEEE Trans. Neural Network, Vol. 5(5), pp. 823-832, 1994.
[18] M. Takegaki and S. Arimoto, ``A New Feedback Method for Dynamic Control of Manipulators," ASME J. Dynamic Systems, Measurement,and Control, Vol. 103(2), pp. 119-125, 1981.
[19]K.Y. Young and J.K. Wang, ``Similarity Analysis for Robot Motions Using and FNN Learning Algorithm,'' IEEE Conf. on Decision and Control, pp. 2523-2528, 1997.
[20] C.H. Wu, K.Y. Young, K.S. Hwang, and S. Lehman, ``Voluntary Movements for Robotic Control,'' IEEE Control Systems Magazine, Vol. 12(1), pp. 8-14, 1992.
[21] R.R. Yager, ``An Alternative Procedure for the Calculation of Fuzzy Logic Controller Values," J. Jap. Soc. Fuzzy Technol. (SOFT), Vol.3(4), pp. 736-746, 1991.
[22] B.H. Yang and H. Asada, ``Progressive Learning and Its Application to Robot Impedance Learning,'' IEEE Trans. on Neural Networks, Vol. 7(4), pp. 941-952,1996.
[23] K.Y. Young and C.C. Fan, ``An Approach to Simplify the Learning Space for Robot Learning Control,'' Fuzzy Sets and Systems, Vol. 95(1), pp. 23-38, 1998.
[24] J.M. Hollerbach, ``Dynamic Scaling of Manipulator Trajectories,'' ASME J. Dynamic Systems, Measurement, and Control,
Vol. 106, pp. 102-106, 1984.
[25] K.Y. Young, J.F. Lee, and H.J. Jou, ``Robot Learning Schemes That Trade Motion Accuracy for Command Simplification,'' accepted by Fuzzy Sets and systems.