Sensor based terrain guidance of distributed cooperative mobile robots

Abstract

Navigation in outdoor terrain is difficult due to lack of easily and uniquely identifiable landmarks. This problem is further complicated for a system with multiple robots navigating a common terrain. The paper describes a field-capable system for navigation, obstacle avoidance, simulated visual training of mobile robots, group world perception modeling using visual feedback from multiple robots, and fusion of sonar range data with vision information for the purpose of terrain learning. A neural network approach is proposed for fusion of the robots visual feedback. In this approach, each mobile robot is presumed to be equipped with one camera and sonar sensor. In the proposed technique, self-localization of the robots and localization of obstacles are performed based on the visual and sonar feedback from the neural network. Computer simulation of the technique is done with FMCell simulation software an interactive graphical simulation environment. Results of simulation runs illustrating the capabilities of this technique are provided. The technique provides a better and simplified approach for visual servoing of a multi-agent system.