Manipulator inverse dynamics computation on FPGA for reconfigurable applications

摘要

Reconfigurability and flexibility are often desired characteristics in an automatic system whose capability needs to be changed from time to time to meet the need of its operation. In this study, a Hardware Description Language (HDL) software core for joint torques/forces computation for an n-joint general purpose manipulator is designed. This HDL core can be used in a reconfigurable robot control design. The inverse dynamics computation determines required joint torques for creating desired accelerations. Our study results in a state machine with forty-seven states. A complete torque computation can be finished in 167 clock cycles in microseconds if Xilinx XC2V4000 Field Programmable Gate Array (FPGA) device is used. This manipulator-structure independent core design can be integrated with other components on FPGAs. There could be a significant reduction in the load on the microprocessor by shifting heavy-load computation to FPGAs.