Safety barrier functions and multi-camera tracking for human-robot shared environment

A new vision in human-robot collaboration has allowed to place robots nearby human operators, working close to each other in industrial environments. As a consequence, human safety has become a dominant issue, together with production efficiency. In this paper we propose an optimization-based control algorithm that allows robots to avoid obstacles (like human operators) while minimizing the difference between the nominal acceleration input and the commanded one. Control Barrier Functions are exploited to build safety barriers around each robot link, to guarantee collision-free trajectories along the whole robot body. Human accelerations and velocities are computed by means of a bank of Kalman filters. To solve obstruction problems, two RGB-D cameras are used and the measured skeleton data are processed and merged using the mentioned bank of Kalman filters. The algorithm is implemented on an Universal Robots UR5 in order to validate the proposed approach. (C) 2019 Elsevier B.V. All rights reserved.