Potential field representation of environment model and its application to robot's force/position hybrid control

摘要

Recent biological studies on the spinalized frog suggested that the premotor circuitry of the frog is organized in a number of distinct vector field forms of modules, and the complex posture control patterns are generated through the vectorial summation of these modules. Functionally, each module correspondent to a "basis field", and the generation of a variety of patterns are exactly the same as to approximate the vector fields by linearly combining the basis fields. We consider the vector fields generated by artificial potential functions. Motivated by the biological studies, we focus our studies on robot manipulator's contact tasks on an uncertain environment. We apply the basis fields combination approach and extend it from the construction of robot's complex motion control patterns to the potential representation of the unknown environment model. We examine the effectiveness of this vector field representation of the environment model in robot's contact tasks.