Singularity Avoidance with Application to Online Trajectory Optimization for Serial Manipulators

Manipulability maximization for inverse kinematics is done, e.g., in Dufour and Suleiman (2017). Several important tasks in robotics require compliance in A potential function on the torque level, as an additive the robot's end-effector including handling tasks, such as impedance, based on the manipulability measure is proposed the peg-in-hole task, see, e.g., Park et al. (2017) and Song in Ott (2008) for singularity avoidance. Due to the et al. (2021), or more recently tasks in physical humanrobot complexity introduced by maximizing the manipulability interaction (pHRI), see, e.g., Sharifi et al. (2022) measure, an optimization approach using a dynamic neural and Li et al. (2018). To this end, control concepts enabling network is introduced in Jin et al. (2017) for tracking compliance in the end-effector, e.g., prescribing a specific control including the consideration of joint velocity limits.