Answers are grounded in this paper and cite their sources.
Addressing the real-time collision-free trajectory generation problem for large-scale multi-agent systems in dynamic interactive environments, this paper proposes a hierarchical event-triggered distributed model predictive control algorithm based on control barrier functions (HCBF–HET–DMPC). The algorithm constructs a hybrid CBF collision-avoidance strategy that integrates hard constraints at the first prediction step with soft constraints over the prediction horizon, thereby enforcing an unrelaxed safety constraint at the first prediction step and providing soft collision-avoidance guidance over the remaining prediction horizon. A risk-aware hierarchical event-triggered mechanism (ETM) dynamically switches solution modes on demand to prevent redundant computation. Built on a decoupled decentralised architecture, the algorithm uses distributed local optimisation based on shared predicted-trajectory information. Simulation results demonstrate that in large-scale, densely coupled multi-agent scenarios with up to 300 agents, HCBF–HET–DMPC substantially improves trajectory feasibility and the global agent completion rate. By avoiding unnecessary optimisation updates, the ETM preserves collision-avoidance performance while reducing optimiser activation and total computation time by more than 30%. The algorithm thus achieves a favourable balance among collision-avoidance safety, optimisation feasibility and computational efficiency for large-scale swarms.
No takes yet. Share an insight, caveat, or question.
Wang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: