DocumentCode :
3256401
Title :
Evolving neural network controllers
Author :
Salama, Rameri ; Hingston, Philip
Author_Institution :
Dept. of Comput. Sci., Western Australia Univ., Nedlands, WA, Australia
Volume :
2
fYear :
1995
fDate :
29 Nov-1 Dec 1995
Firstpage :
579
Abstract :
An emerging design paradigm uses evolutionary processes to search for optima in design space. The evolutionary technique has the advantage of being a declarative paradigm; the user specifies the task, and a genetic algorithm searches for an optimum solution. Normal techniques require the definition of the controller, and this is computationally expensive. We use a genetic algorithm to design a neural network-based controller for a hexapod robot. The robot must perform the task of moving from a start position to a goal position, under varying degrees of simulated instrument and sensor noise. The findings show that it is possible to embed a degree of noise tolerance into the solution. This is useful in situations where the environment of the robot may change over time
Keywords :
control system synthesis; genetic algorithms; intelligent control; legged locomotion; neurocontrollers; noise; optimal control; search problems; changing environment; declarative paradigm; design paradigm; design space optima searching; evolutionary technique; genetic algorithm; hexapod robot; mobile robot; neural network controllers; noise tolerance; simulated instrument noise; simulated sensor noise; user-specified tasks; Algorithm design and analysis; Control systems; Genetic algorithms; Genetic programming; Neural networks; Orbital robotics; Robot control; Robot sensing systems; Vehicles; Working environment noise;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Evolutionary Computation, 1995., IEEE International Conference on
Conference_Location :
Perth, WA
Print_ISBN :
0-7803-2759-4
Type :
conf
DOI :
10.1109/ICEC.1995.487448
Filename :
487448
Link To Document :
بازگشت