Fuzzy-Behavior Synthesis, Coordination, and Evolution in an Adaptive Behavior Hierarchy

Abstract

Autonomous control and navigation of mobile robotic vehicles are fundamental enabling technologies for automation in a variety of operating domains ranging from industrial environments to remote planetary surfaces. The engineering problem to be solved generally consists of achieving real-time sensor-based motion control among obstacles in the environment while performing useful tasks throughout its accessible regions. In many instances, mobile robots are required to do so using limited resources (e.g. power, computation, sensors, etc.) that are resident on-board the vehicle. Traditional approaches have been based on functional decomposition of tasks, which employed computationally intensive planning algorithms and explicit pre-determined world models. The resulting serial execution of sensing, modeling, planning and acting produced intelligent behavior, but at the great expense of real-time performance.