Singularity Tolerant Robotic Manipulation for Autonomous Field Tracking in an Unknown Environment

This paper presents a comprehensive motion control strategy for the autonomous operation of robotic manipulators combined with the sensor-driven recursive estimation of an unknown field of interest. The spatial distribution of the environmental phenomenon is modeled by a radial basis function (RBF) network and their weight parameters are estimated by a recursive least square (RLS) method using the collective measurements from the on-board sensors mounted to the manipulator. The asymptotic tracking has been simultaneously achieved by the control law based on the gradient of the estimated field. Since the target location cannot be known a priori, the motion controller has to be designed in explicit consideration of tolerating the singular configuration of the manipulator kinematics. By using the null space decomposition of the task space for the Jacobian near the singularities, a systematic method is suggested to command the task space control law in spite of the singular configurations. Simulation results using the three link planar robot are presented to support the main ideas.