Real-Time Robot Trajectory Generation with Python

Design and performance measurements of a framework for external real-time trajectory generation for industrial robots is presented. The framework is implemented entirely in Python. It serves as a proof of concept for performing real-time trajectory generation in Python, from a PC with connection to the motion controller in an industrial robot controller. Robotic applications requiring advanced, custom trajectory generation, and a high level of integration with sensors and other external systems, may benefit from the efficiency of Python in terms of reduced development time, lower code complexity, and a large amount of accessible software technologies. The presented framework, dubbed PyMoCo, supplies a set of simple trajectory generators, which are comparable to those found in contemporary industrial robot controllers. Designing and implementing new trajectory generators and integrating or extending the included trajectory generators is central to the design of PyMoCo. Laboratory applications involving real-time sensor- and vision-based robot control has demonstrated the usability of PyMoCo as a motion control framework and Python as a robotics application platform. For robotics applications with a control frequency not exceeding a couple of hundred Hz, computation deadlines no shorter than some couples of milliseconds and jitter tolerance at the order of a millisecond, PyMoCo may be considered a feasible and flexible framework for testing and prototype development.