Controller design for human-assist robots with force filtering process

This paper describes a controller design method for robots supporting humans in real environments. Such robots must deal with various interactions between robots and the human environment. Furthermore, they also must autonomously classify the interactions from sensor information and select their actions according to the tasks and situations. This study introduces the idea of a spatial filter, a filter in the spatial domain, to construct such a control system. The method decomposes sensor information into function modes, spatial information essential for control. The force filtering process based on function modes makes the controller design simple and explicit. Some experimental results are presented in order to verify the validity of the proposed method.