Robot Dynamics Constraint for Inverse Kinematics

Inverse Kinematics is a fundamental tool in Cartesian/Operational Space control. Recent approaches make use of Quadratic Programming Optimization to obtain desired joint velocities or accelerations from Cartesian references. QP based IK also permits to specify constraints to affect the solution. Constraints are fundamental and necessary when working with real robotic hardware since they prevent possible damages: joint limits, self collision avoidance and joint velocity limits are examples of such constraints. In this work we present a constraint to take into account joint torque limits based on the robot dynamics and force/torque sensor measurements. Despite the robot dynamics can be naturally expressed at acceleration level, our main goal is to specify this constraint in a resolved motion rate control IK. For this reason we formulate it also at the velocity level to be used in any IK QP based scheme. Hence, this formulation allows to generate dynamically feasible motions of the robot even in simple IK velocity based schemes. We apply this constraint to our humanoid robot COMAN while performing a Cartesian task which requires high torques in some joints. The constraint is developed inside the OpenSoT library.

[1]  Olivier Stasse,et al.  A versatile Generalized Inverted Kinematics implementation for collaborative working humanoid robots: The Stack Of Tasks , 2009, ICAR.

[2]  Nikolaos G. Tsagarakis,et al.  OpenSoT: A whole-body control library for the compliant humanoid robot COMAN , 2015, 2015 IEEE International Conference on Robotics and Automation (ICRA).

[3]  François Keith,et al.  Dynamic Whole-Body Motion Generation Under Rigid Contacts and Other Unilateral Constraints , 2013, IEEE Transactions on Robotics.

[4]  Alessandro De Luca,et al.  Discrete-time redundancy resolution at the velocity level with acceleration/torque optimization properties , 2015, Robotics Auton. Syst..

[5]  Christian Kirches,et al.  qpOASES: a parametric active-set algorithm for quadratic programming , 2014, Math. Program. Comput..

[6]  Eiichi Yoshida,et al.  A Local Collision Avoidance Method for Non-strictly Convex Polyhedra , 2008, Robotics: Science and Systems.

[7]  Alexander Herzog,et al.  Momentum-based Balance Control for Torque-controlled Humanoids , 2013, ArXiv.

[8]  Pyung Hun Chang,et al.  The enhanced compact QP method for redundant manipulators using practical inequality constraints , 1998, Proceedings. 1998 IEEE International Conference on Robotics and Automation (Cat. No.98CH36146).

[9]  Olivier Stasse,et al.  Dancing Humanoid Robots: Systematic Use of OSID to Compute Dynamically Consistent Movements Following a Motion Capture Pattern , 2015, IEEE Robotics & Automation Magazine.

[10]  Nicolas Mansard,et al.  Whole-body motion integrating the capture point in the operational space inverse dynamics control , 2014, 2014 IEEE-RAS International Conference on Humanoid Robots.

[11]  Katsu Yamane Simulating and Generating Motions of Human Figures (Springer Tracts in Advanced Robotics, V. 9) , 2004 .