Adaptive Control of a Flexible-Link Robotic Manipulator with Unknown Payload Dynamics

For many applications, flexible-link robot arms may handle payloads that are not simple rigid bodies. In space applications, the RMS (remote manipulator system) will be manipulating satellites that may contain fuel or have flexible appendages. High-performance control schemes for flexible-link manipulators require end-point feedback. If the payload dynamics are not accounted for in the control design, degraded performance or instability are possible [1]. This paper presents technology that extends the concept of end-point control to handle payloads with unknown internal dynamics. High-performance control is merged with an innovative identification algorithm in a self-tuning regulator approach. The identification of the payload is done using recently developed subspace-fitting techniques. These techniques allow real-time determination of the order of the payload dynamics. Sufficient excitation problems are addressed by performing the identification closed loop. Experimental results are presented.