Abstract: | The coordination betwen the sensor and motor systems is an essential feature in autonomous intelligent systems. This thesis investigates the evolutionary approach to sensorimotor control using learner classifier systems. A simple classifier system is used to solve the problem of coordinating a pair of eyes and an arm in order to catch an object.First, an analysis of the previous approaches based on neural networks is done. Then a review of what a simple classifier system is, as well as the specific implementation of the simple classifier system to solve this problem. Afterwards ana analysis of the results is presented. Finally a review of the advantages and disadvantages of this approach comparing it with previous ones is offered. Results have shown that classifier systems are a promising tool - which solve this sensorimotor coordination problem - further work needs to be done to determine the limitations of this approach.
|