Safe Flocking in Spite of Actuator Faults
The safe flocking problem requires a collection of N mobile agents to (a) converge to and maintain an equi-spaced lattice formation, (b) arrive at a destination, and (c) always maintain a minimum safe separation. Safe flocking in Euclidean spaces is a well-studied and difficult coordination problem. Motivated by real-world deployment of multi-agent systems, this paper studies one-dimensional safe flocking, where agents are afflicted by actuator faults. An actuator fault is a new type of failure that causes an affected agent to be stuck moving with an arbitrary velocity. In this setting, first, a self-stabilizing solution for the problem is presented. This relies on a failure detector for actuator faults. Next, it is shown that certain actuator faults cannot be detected, while others may require O(N) time for detection. Finally, a simple failure detector that achieves the latter bound is presented. Several simulation results are presented for illustrating the effects of failures on the progress towards flocking.
KeywordsMobile Agent Detection Time Failure Detector Reachable State Actuator Fault
Unable to display preview. Download preview PDF.
- 6.Franceschelli, M., Egerstedt, M., Giua, A.: Motion probes for fault detection and recovery in networked control systems. In: American Control Conference 2008, pp. 4358–4363 (2008)Google Scholar
- 9.Johnson, T.: Fault-Tolerant Distributed Cyber-Physical Systems: Two Case Studies. Master’s thesis, Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign, Urbana, IL 61801 (May 2010)Google Scholar
- 10.Johnson, T., Mitra, S.: Safe and stabilizing distributed flocking in spite of actuator faults. Tech. Rep. UILU-ENG-10-2204 (CRHC-10-02), University of Illinois at Urbana-Champaign, Urbana, IL (May 2010)Google Scholar
- 13.Shaw, E.: Fish in schools. Natural History 84(8), 40–45 (1975)Google Scholar