This research will investigate the problem of precision payload tracking from mobile robots. Payload tracking denotes maintaining alignment between a payload and a target while both payload and target may be moving. For instance, a teleoperated mobile robot on the moon will need to constantly align its communication antenna with an antenna on the Earth to enable high bandwidth communication, and hence effective teleoperation. Tracking is important for many other applications including cooperative manipulation, mobile surveying and reconnaissance, optical communication, and active vision. In the case of moderate speeds across rough terrain, tracking demands high slew rates and large motion ranges due to vehicle motion disturbances. This differs from satellite antenna tracking or telescope pointing where the need is for very high precision but incurred motion rates are small. Moreover, mobile robots are commonly mass and power limited. Attaining tracking stability requirements coupled with large articulation ranges, high slew rates, low mass and power makes the problem of payload tracking from mobile robots challenging.
This research emphasizes the use of the complete robot system (mechanism, planning and control) to achieve precision tracking. This is important because an independent fine pointing device such as a gimbal- without cooperation from locomotion, suspension and isolation devices- may not be able provide the large torques, angular excursions, and disturbance rejection needed in rough terrain. As the terrain roughness and corresponding excursions and disturbances increase, so does the need for robot systems to functionally cooperate to achieve payload tracking.
The proposed approach is to develop systemic design for tracking through configuration, analysis and simulations, followed by testing and demonstration. The tracking problem is formulated in a control-theory framework to allow use of standard control techniques for analysis. A dynamic simulation software package, SIMKITTM is used to evaluate tracking performance of various configurations. Data from a payload tracking testbed (lunar-rover prototype or other) will be used for verifying and tuning the simulator and analytic models.
The research is limited to wheeled robots moving at moderate speeds in rough but not severe terrains. The application area is precision antenna tracking from roving mobile robots to enable high data rate, low power telemetry. The example used is the case of providing video telemetry from a lunar rover to an Earth station. The expected contributions from the work include tools to model and evaluate payload tracking, metrics for measuring tracking performance, control theory formulation for design, and guidelines for developing mobile robot configurations suitable for precision payload tracking.
email: deepak@ri.cmu.edu
Last Modified:
Thursday, 27-Mar-1997 17:03:26 EST