Piscataway: Institute of Electrical and Electronics Engineers
摘要:
This paper addresses the problem of visual servo control for a five-degree-of-freedom robot manipulator to perform pick-and-place tasks. A new hybrid switched reactive-based visual servo control structure, inspired from a sensing-and-reaction behavior without the inverse interaction matrix computing, is proposed to handle this problem efficiently. The proposed structure is similar to the logic-based approaches, but the requirements of the fuzzy modeling and/or the fuzzy-rule base learning are omitted. To achieve this, a novel hardware-aware reactive function approach is presented to directly map an image position error vector to a desired end-effector velocity vector under the consideration of hardware limitations. This approach helps to simplify the implementation of visual servo systems with improved reliability. Moreover, the proposed control system is a hybrid switching controller consisting of both image-based and position-based reactive planning schemes, which allows improving the robustness and effectiveness of the visual servo system. Experimental results validate the performance of the proposed visual servo control method in a realistic scenario setting.
