Continuous Reachability Task Transition Using Control Barrier Functions⁎This work was supported in part by the National Science Foundation under award #1749357. C. Santoyo was supported by the NSF Graduate Research Fellowship Program under Grant No. DGE1650044. Issue 2 (2020)

• Record Type: Journal Article
• Authors: Srinivasan, Mohit; Santoyo, Cesar; Coogan, Samuel
• Is Part Of: IFAC-PapersOnLine. Volume 53:Issue 2(2020)
• Page Start: 9696
• Page End: 9701
• Publication Date: 2020
• Subjects: Mobile robots -- Autonomous robotic systems -- Robotics technology; Automatic control -- Periodicals; 629.805
• DOI: 10.1016/j.ifacol.2020.12.2621
• Languages: English
• ISSNs: 2405-8963

Abstract

Abstract: In this paper, a method to achieve smooth transitions between sequential reachability tasks for a continuous time mobile robotic system is presented. Control barrier functions provide formal guarantees of forward invariance of safe sets and finite-time reachability and are able to enforce task execution. Barrier functions used in quadratic programs result in implementation of controllers with real-time performance guarantees. Existing approaches for multi-objective task execution using control barrier functions leverage discretely switched, sequential quadratic programs to achieve successive tasks. However, discrete switching can lead to control input discontinuities which can affect a robot's performance. Hence, we propose a method which ensures continuous transitions between sequential quadratic programs. In particular, a time varying component to the barrier function constraint is introduced which allows for a smooth transition between objectives. Robotic implementation results are also provided.