Simultaneous obstacle avoidance and trajectory tracking for multi-robot systems via acceleration-level motion planning
TL;DR
It is demonstrated that strict collision-free guarantees are achieved for arbitrarily large robot teams without excessive conservatism and the integrated strategy yields the lowest tracking deviation and the shortest total intervention time among all tested deadlock-handling approaches.
Abstract
Considerable attention has been directed toward collision-free trajectory tracking in multi-robot systems, for which CBFs have been widely adopted as a safety-critical tool. However, existing CBF methods are constrained by their dependence on maximum braking estimation. Moreover, effective deadlock resolution is often treated as an afterthought, resulting in frequent external interventions and degraded tracking accuracy. To overcome these limitations, a novel CBF is proposed, by which the reliance on maximum braking estimation is entirely eliminated. Furthermore, an auxiliary velocity vector is refined and directly integrated into the trajectory tracking controller for deadlock resolution. Through quantitative comparative experiments, it is demonstrated that strict collision-free guarantees are achieved for arbitrarily large robot teams without excessive conservatism. It is also shown that the integrated strategy yields the lowest tracking deviation and the shortest total intervention time among all tested deadlock-handling approaches. Consequently, the proposed framework is established as a practically viable solution for safe and efficient multi-robot trajectory tracking, wherein both safety constraints and deadlock resolution are addressed in a unified manner.