Introduction To Intelligent Control M Yamakita Dept of

Introduction To Intelligent Control M. Yamakita Dept. of Mechanical and Control Systems Eng. Tokyo Inst. Of Tech. 1/30/2022 1

Controlled system becomes more and more complex. It is almost impossible to represent mathematical differential and difference equation representation of the systems. Emergent technology is needed ! Intelligent Control 1/30/2022 2

Trends from ’ 60 Artificial Intelligence (AI) Crisp Logic Fuzzy Logic Symbolic Representation Non-Symbolic Representation (ANN) Control Theory Classical Control Theory （PID) 1/30/2022 Modern Control Theory Robust Control Theory Adaptive Control Theory Hybrid System Control 3

What’s Intelligent Control ? AI Information Processing Formal Language Planning Scheduling Management OR k y or on em ati ck M miz dba ti ee cs Op n. F ami Dy Dyn ac Intelligent Control db s e c e i nt F n. nam me on y D Dy age ati an in M oord C CONTROL (G. Saridis, 1979) Intelligent Control ≠ Fuzzy Control 1/30/2022 ( 4

Structure Of Intelligent Control １． Hierarchical Intelligent Control (Albus, Saridis) 2. Reactive Intelligent Control (Brooks) (Subsumption Architecture) 1/30/2022 5

Hierarchical Intelligent Control (Saridis) PRECISION INTELLIGENCE Organization Level ORGANIZER DISPACHER 1/30/2022 MOTION COORDINATOR VISION COORDINATOR PLANNING COORDINATOR MOTION CONTROLLER VISION CONTROLLER COMMUNICATION CONTROLLER ACTUATORS VISION HARDWARE NETWORK HARDWARE … Coordination Level … Execution Level … 6

Reactive Intelligent Control (Brooks) Modify the World RESET Create Maps R INPUT I Discover New Area Suppressor Inhibitor BEHAVIORAL MODULE S OUTPUT Avoid Collisions Move Around 1/30/2022 7

Supporting Technologies １． Extensions of conventional control technologies Robust optimal control Adaptive control Learning control 2. New technologies FAN(Fuzzy, AI, and Neural network) technology (Fukuda) Soft computing (Zadeh) 1/30/2022 8

Dynamical System Representation (State Space Representation) 1/30/2022 9

Robust Optimal Control A nominal system Set of uncertain systems Model set of uncertain systems 1/30/2022 10

Adaptive Control 1/30/2022 11

Symbolic System Representation (Rule Based Representation) Area 3 ? Area 2 Area 1 Classical AI, Automaton etc. 1/30/2022 12

Crisp Logic vs. Fuzzy Logic Tall Mr. A （ Mrs. B １８０ｃｍ １７０ｃｍ ？ １７０ｃｍ １６０ｃｍ Mr. C 1/30/2022 Short 13

When we describe real world symbolically, there always exist ‘gray zone’ state. It is very difficult to describe the gray zone property by conventional crisp logic. Or, we must define undesirably many categories. Fuzzy Logic 1/30/2022 14

Introduction of membership functions Degree of property 100% 50% ) 160 Short 1/30/2022 170 180 Tall x Height 15

Perceptron O 1 O 2 B A B C A--B （A is connected to B) B--C C--A Triangle A C D A--B B--C C--D NOT Triangle Human easily recognize O 2 as triangle ! 1/30/2022 16

B A C B Mimic the brain function ! A 1/30/2022 D C 17

Adjustable Weights No Hidden Layer 1/30/2022 Activation Function （Rosenbratto Type Perceptron） 18

Multi Layered Neural Network Adjustable Weights Activation Function . . Generalized delta rule, Back-propagation algorithm (Amari, Rumelhalt) 1/30/2022 19

References 1. M. M. Gupta, N. k. Sinha：Intelligent Control Systems, IEEE Press. (1996) 2. K. Furuta et. ：Intelligent Control, Corona Pub. (1988) （in Japanese) 3. B. Widrow, E. Walach: Adaptive Inverse Control, Prentice Hall (1996) 1/30/2022 20