A Novel QP-Based Kinematic Redundancy Resolution Method With Joint Constraints Satisfaction
Authors:
Open Access
- 14 April 2022
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 10, 41023-41037
Abstract
Kinematically redundant manipulators are advantageous for their increased dexterity and ability to fulfill some secondary requirements along with their primary task to follow the prescribed trajectory. The redundancy results in a non-trivial inverse kinematics problem (IK). Standard methods of redundancy resolution are based on the pseudoinverse of the Jacobian matrix of the manipulator. The excessive degrees of freedom are utilized to perform secondary tasks that are projected onto the null space of the Jacobian matrix. However, the joint constraints satisfaction cannot be easily ensured in this way—even though methods that account for the joint position limits are known, the constraints for joint velocities and especially accelerations are not straightforward to include. In this paper, a novel redundancy resolution method based on a less common quadratic programming (QP) approach is described. The proposed velocity-level IK method allows fulfilment of the joint constraints at the position, velocity, and acceleration levels. In the derived formulas, accelerations instead of the usual velocities are used—the discretized joint state equations allow the use of joint accelerations as decision variables in the QP problem. The developed algorithm is investigated in a series of numerical tests in which the kinematics of the KUKA LWR4+ redundant 7-DOF manipulator is exploited. The newly elaborated QP-based IK method is firstly compared with the classic pseudoinverse-based approach and then tested for its ability to keep the joint accelerations within the prescribed bounds. The prospects of the proposed approach are discussed in the concluding section.Keywords
Funding Information
- National Science Centre, Poland (2018/29/B/ST8/00374)
- Warsaw University of Technology (WUT) Internal Grant for Research in Automation, Electronic, and Electrical Engineering
This publication has 49 references indexed in Scilit:
- Closed-Loop Inverse Kinematics for Redundant Robots: Comparative Assessment and Two EnhancementsIEEE/ASME Transactions on Mechatronics, 2014
- qpOASES: a parametric active-set algorithm for quadratic programmingMathematical Programming Computation, 2014
- Least-Squares-Based Reaction Control of Space ManipulatorsJournal of Guidance, Control, and Dynamics, 2012
- An online active set strategy to overcome the limitations of explicit MPCInternational Journal of Robust and Nonlinear Control, 2007
- Selectively Damped Least Squares for Inverse KinematicsJournal of Graphics Tools, 2005
- An updated set of basic linear algebra subprograms (BLAS)ACM Transactions on Mathematical Software, 2002
- Singularity-robust task-priority redundancy resolution for real-time kinematic control of robot manipulatorsIEEE Transactions on Robotics and Automation, 1997
- A weighted least-norm solution based scheme for avoiding joint limits for redundant joint manipulatorsIEEE Transactions on Robotics and Automation, 1995
- Redundancy resolution through local optimization: A reviewJournal of Robotic Systems, 1989
- Obstacle Avoidance for Kinematically Redundant Manipulators in Dynamically Varying EnvironmentsThe International Journal of Robotics Research, 1985