Safe human dual-robot interaction based on control barrier functions and cooperation functions

Abstract

Nowadays, humans are allowed to work side-by-side with a robot team, which consists of dual robots in most cases. To ensure human safety, the motion of each robot should be reactive to and compliant with humans via human-in-the-loop control. Furthermore, when the robots conduct a cooperative task, the reactive and compliant motion of each robot must fulfill the constraints imposed by the cooperation with the other robot. It is challenging to guarantee human safety and robot cooperation simultaneously, especially in a decentralized architecture. This letter presents a decentralized control framework that guarantees both human safety and robot cooperation in human dual-robot interaction. First, the high-order time-varying control barrier functions (HO-TV-CBFs) are defined to represent human safety, based on which a safety control set is formulated to guarantee human safety. Second, the cooperation functions are introduced to abstract the cooperation constraints of different cooperative tasks of dual robots, which are further guaranteed by a cooperation control set. Then, a centralized control framework based on the safety and cooperation control sets is constructed. Finally, the centralized control framework is decentralized while maintaining the same performance. The proposed control framework is verified via experiments.