PyRoboCOP: Python']Python-Based Robotic Control and Optimization Package for Manipulation and Collision Avoidance

Contacts are central to most manipulation tasks as they provide additional dexterity to robots to perform challenging tasks. However, frictional contacts leads to complex complementarity constraints. Planning in the presence of contacts requires robust handling of these constraints to find feasible solutions. This paper presents PyRoboCOP which is a lightweight Python-based package for control and optimization of robotic systems described by nonlinear Differential Algebraic Equations (DAEs). In particular, the proposed optimization package can handle systems with contacts that are described by complementarity constraints. We also present a general framework for specifying obstacle avoidance constraints using complementarity constraints. The package performs direct transcription of the DAEs into a set of nonlinear equations by performing orthogonal collocation on finite elements. The resulting optimization problem belongs to the class of Mathematical Programs with Complementarity Constraints (MPCCs). MPCCs fail to satisfy commonly assumed constraint qualifications and require special handling of the complementarity constraints in order for NonLinear Program (NLP) solvers to solve them effectively. PyRoboCOP provides automatic reformulation of the complementarity constraints that enables NLP solvers to perform optimization of robotic systems. The package is interfaced with for obtaining sparse derivatives by automatic differentiation and for performing optimization. We provide extensive numerical examples for various different robotic systems with collision avoidance as well as contact constraints represented using complementarity constraints. We provide comparisons with other open source optimization packages like and . The code is open sourced and available at https://github.com/merlresearch/PyRoboCOP. Note to Practitioners-PyRoboCOP is intended to be an easy-to-use software package written in Python which can be used for optimization, estimation and control for a large class of robotic systems. Including, in particular, contact-rich applications to deal with complex scenarios that arise when making and breaking contacts during a task. Typical problems that can be solved with our work are trajectory and control sequence optimization, parameter estimation. To make the proposed software package easier for practitioners, the paper provides access to the package and a large number of example problems. Furthermore, the package also provides a guide describing the details of all the methods a user might have to implement for their own system. Compared to some of the other packages, PyRoboCOP works with NumPy object arrays which is the native computing package in Python. We believe that this will make it much easier to learn and use compared to some of the other optimal control packages.