|Abstract:||The main alternative to traditional robots in recent years have been behaviour-based robots, which challenge the traditional assumption of functional decomposition in behavioural decomposition. They deliver real-time performance in a dynamic world. However, behaviour-based robots have become hard to design due to the increasing complexity of interactions between different behavioural modules, as the number of required behavioural modules increases. One way out of this problem is to try and automate the design process. Evolutionary approaches use ideas borrowed from evolution in order to solve problems in highly complex search space. If some object fitness function can be derived for any given architecture, thee is the possibility of automatic evolutionary approach, called genetic programming, to evolve a behaviour-based robot automatically. In addition, this project also intends to simultaneously co-evolve a body with the control system of a robot.