@brazovayeye

Evolution of a control architecture for a mobile robot

. Proceedings of the Second International Conference on Evolvable Systems: From Biology to Hardware (ICES 98), volume 1478 of LNCS, page 303--310. Lausanne, Switzerland, Springer Verlag, (23-25 September 1998)

Abstract

Programming a robot to perform a desired task in an unknown environment is a difficult task. Due to unexpected interactions between the environment and the robot many iterations of program development are often required. Using genetic programming the robots themselves may search the space of possible programs. In an experiment which was conducted over a period of two months we evolved a behavior-based control architecture for a large sized mobile robot, a Real World Interface B21. This is the first time that a large mobile robot was used in evolutionary robotics using tree-based genetic programming. In addition, our architecture uses conditional statements to build up a hierarchical reactive control structure. Sonar sensors are used to sense the environment. Because the robot is able to exert a considerable force if it crashes into an object, safety measures had to be taken to automatically monitor the run.

Links and resources

Tags

community

  • @brazovayeye
  • @dblp
@brazovayeye's tags highlighted