Self-Organizing Neural Network Controllers for Mobile Robots
Ralf Salomon · 2007
: This paper proposes a new self-organizing control architecture for mobile robots. It consists of a controller and a value system. The controller uses visual sensors to determine the motor commands, whereas the value system uses stimuli from proprioceptive sensors. This design is justified by (1) both components are significantly decoupled by using different sensory modalities, (2) the feedback of proprioceptive sensory patterns is omnipresent in biological systems and has been widely neglected in control systems, and (3) proprioceptive sensors operate more reliably and can be used more efficiently than visual sensors, such as a CCD camera. Experiments with the Khepera robot show that by using proprioceptive sensors, the control architecture can adapt to different environments and can yield very robust behavior with respect to sensor failures. INTRODUCTION One goal of research on autonomous agents is to build robots that are able to behave without human control in dynam...