Robot Hand-Eye Coordination Based on Fuzzy Logic

Hand-eye coordination of a robot system is a classical problem in robotic vision, which mathematically can be described as an ill-posed inverse mapping problem. Different techniques have been developed to solve this inverse mapping problem, but most of the reported work has a common limitation that is the relative position between the given robot and the camera must be fixed once the calibration is finished. However, very often in real-world applications this may not be realistic either due to the random disturbances to the system or due to the need of repositioning the camera or moving the robot base to a new position in order to perform a required task. In this case, usually a lengthy re-calibration process has to be carried out. In this paper, we describe our technique which eliminates the need of re-calibration for the vertical position change of the camera. Our technique is based on the integration of the utilization of a close form mathematical formulation and fuzzy logic. Using the fuzzy logic technique, we are able to deal the unknown amount of vertical changes of the camera position. Using image processing technique to estimate the distance change of the image land marks, we derived a fuzzy reasoning model which allows us to calculate the inverse mapping matrix. The algorithm and fuzzy reasoning system have been developed and implemented in a 6-degree-of-freedom robot system. Experiments were conducted and the result confirmed our design.