Online Robot-Object Synchronization With Geometric Constraints and Limits on Velocity, Acceleration, and Jerk

This letter presents a new approach to online synchronization of a robot to the motion of an object. This is done by an iterative trajectory planning approach. Trajectory planning is divided into geometric path planning and subsequent time optimal parameterization of the geometric path. The geometric path planning approach can handle position constraints in joint space as well as in task space and results in a sequential quadratic program in two variables for each joint. The time optimal path parameterization respecting joint velocity, acceleration, and jerk constraints is formulated as a linear program. However, the overall optimization problem is nonlinear. Only few iterations of this nonlinear optimization problem can be performed online. Therefore, a method for obtaining a time scaling function is presented such that the robot and the object are synchronized exactly. The synchronization task is executed online and experimental results are shown for a 6-DOF manipulator.