Safe Human-Robot Coetaneousness Through Model Predictive Control Barrier Functions and Motion Distributions

Abstract

Future real-world applications will consist of robots and human workers collaborating with each other in a shared environment to increase productivity. In such scenarios, it is necessary to guarantee the safety of humans while maintaining precise control of the robots performing tasks. Probabilistic movement primitives (ProMPs) are a powerful tool for defining a distribution of trajectories for dynamic systems. However, they have been solely used for determining robot trajectories. In this paper, we utilize ProMPs to predict the probabilistic motion of humans in the environment. To achieve this, we propose a combination of model predictive control (MPC) and control barrier functions (CBFs) to guide a robot along a predefined trajectory while guaranteeing it always maintains a desired distance from a human worker motion distribution defined by a ProMP. A case study is provided to demonstrate the efficacy of our methods.