Applying an Evolvable Hardware Approach to Autonomous Robot Design

Microprocessors have long been the central controller for many autonomous robots. However, as a processor is a general-purpose device, it is possible that if a control system for a robot were designed specifically for a certain task, improved performance would result. However, such an approach would be impractical for anything but the most simplistic goals, assuming that a human were the designer. One possible solution is to approach the problem using Evolvable Hardware (EHW). This method uses Genetic Algorithms to apply evolutionary forces to the development process. In this research, the EHW approach will be applied to design a control system for a robot that has increasingly complex tasks to perform. This design process will then be compared with control systems developed by a human designer, to determine whether the EHW approach can benefit robot designs.