Algorithms of time compression and analysis of formed

Algorithms of time compression and analysis of formed patterns in autonomous adaptive control systems Mazur Yuri, Zhdanov Alexander Lebedev Institute of Precision Mechanics and Computer Engineering, Moscow, Russia Autonomous Adaptive Control Lab (AAC Lab) http: //www. ipmce. ru http: //www. aac-lab. com

Structure and functions of «nervous system» Autonomous Adaptive Control (AAC) (only main elements shown) “Body” of controlled Object “Nervous system” Sensors Recognition and pattern formation subsystem Subsystem of knowledge base generation Memory of recognized patterns Emotions apparatus Environment Knowledge base Actuators Decision making subsystem 2

Memory module for recognized patterns Fixed size of recognized patterns` memory Adding patterns to the memory Deleting patterns from the memory TN+2 TN TN+1 Trash TN+4 TN+3 3

Problems of recognized patterns` module and solutions Problems: • Unreasonable using of memory (saving of all input data). • Absence of information's compression. Solutions: • Compression of information in the time direction. • Introduction of importance characteristic of every pattern for control system (CS) and using of this characteristic for compression. Reason: • Storing of maximally important information for CS helps to increase further analysis's efficiency of this information and deduction of new regularities, which, probably, accelerates learning process and increase control quality. 4

Time-based compression of information TN+1 Initial patterns sequence TN TN+2 TN+3 TN+7 TN+5 TN+4 TN+6 TN+8 Needs 9 memory cells T Adding patterns to the memory Compression of data on the basis of average values Compressed patterns sequence Needs 4 memory cells T Adding patterns to the memory 5

Consideration of information importance for control system in compression process In the AAC method role of information importance (pattern/event) is played by emotional estimation of pattern, because function depending from emotional estimation is used as criterion of control quality. 6

Using of emotional estimation for compression of information New pattern ? Emotions intervals Level for all patterns all Level 0 for patterns with low emotions 1≤ |S| <5 Level 1 for patterns with medium emotions 5 ≤ |S| <9 Level 2 for patterns with high emotions 9 ≤ |S| <13. . . 9 ≤ |S| <13 5 ≤ |S| <9 1≤ |S| <5 |S| <1 7

Features of the developed algorithm • Averaging function is used for attributes values of compressed patterns. • Updating of compressed patterns is happened on each time point. • Fixed and preliminary determined memory size is used for intermediate calculations. • Level of compression depends on pattern importance for control system and amount of time passed from the moment of pattern recognition. 8

Computer implementation of the algorithm Input sequence Emotions Level for all patterns Level 0 Level 1 9

Thank you for your attention! Any questions? 10