Evolving a behavior-based control architecture- From simulations to the real world

Genetic programming makes it possible to automatically search the space of possible programs. First we evolved a behavior-based control architecture using computer simulations. Then we replicated one of the experiments with a service robot, showing that Koza's classic experiment of evolving a control architecture can be transfered to the real world with a change of representation. The use of a service robot necessitates safety measures which are also explained. Results are reported for the experiments using computer simulations and with the real robot. 1 MOTIVATION Many mobile robots are carefully programmed by hand. Apart from repetitive tasks which may simply be recorded, programmingrobots to complete tasks in arbitrary environments is a di cultprocess which usually takes a long time. The di culty of the task is due in part to unforeseeable interactions between the robot and the environment. In addition, a handwritten programmaybehave very di erent in realityfrom the way it was...

• Type: Conference
• Year: 1999
• Where: GECCO
• Authors: Marc Ebner, Andreas Zell
• Keywords: Behavior-based Control Architecture, GECCO 1999, Optimization, Real World, Service Robot