The discussion begins with a review of Sheridan's paradigm for supervisory control of telerobots. Here the machine intelligence is distributed between the operator site and the plant site, and there are a number of local sensing and control loops as well as remote monitoring and command loops. For operation of a complex process plant, this control philosophy must be expanded and extended in a number of areas.
After description of some recent research on operation of a robot-assisted biochemical analysis laboratory planned for Space Station Freedom, an extended architecture is proposed and several topics for continued research are outlined. These include, but are not limited to: distributed supervisory control by intelligent agents, design and function of the telerobots, use of neural networks for sensing and preprocessing, smart components, advanced artificial intelligence, and communication requirements.
Interested in reading the full paper? (14 pages, 737,474 bytes, pdf)