Real-time trajectory resolution for dual robot machining

Abstract

A real-time trajectory planner is presented. A two-pronged approach provides alternative trajectories in response to disturbances encountered during a two-manipulator machining application. First, based on the current torque demand, a weighted pseudo-inverse technique is used to reduce the joint accelerations contributing the most to the saturated joint torques. Second, the null space of the relative Jacobian is exploited to minimize the compliance of the system. Influence coefficients are used to share the null space between the cost function and penalty functions.