Research Analyzer
← Back ICRA 2026

Multi-Robot Trajectory Planning Via Constrained Bayesian Optimization and Local Cost Map Learning with STL-Based Conflict Resolution

Sourav Raxit, Abdullah Al Redwan Newaz, Jose Fuentes, Paulo Padrao, Ana Cavalcanti, Leonardo Bobadilla

PDF

AI summary

Key figure (auto-extracted from paper)
A two-stage framework combining constrained Bayesian optimization and STL-based conflict resolution enables scalable, safe, and specification-compliant multi-robot trajectory planning validated in simulation and real-world experiments.
Multi-robot planning Signal Temporal Logic Bayesian Optimization Conflict-Based Search Kinodynamic constraints Autonomous vehicles

Problem

Multi-robot motion planning under Signal Temporal Logic specifications faces scalability bottlenecks in exact solvers and excessive sampling requirements in conventional methods, hindering adaptability to dynamic, real-world conditions.

Approach

The method decouples planning into a single-robot constrained Bayesian Optimization tree search (cBOT) for efficient local cost map learning and a multi-robot STL-enhanced Kinodynamic Conflict-Based Search (STL-KCBS) that uses temporal logic monitors for robust conflict detection and resolution.

Key results

  • cBOT generates shorter collision-free trajectories using fewer samples than RRT-based planners
  • STL-KCBS ensures formal STL specification satisfaction with probabilistic completeness
  • Benchmarking demonstrates superior trajectory efficiency and safety over existing baselines
  • Real-world autonomous surface vehicle experiments validate robustness in uncertain environments

Why it matters

Provides a scalable, formally verified planning framework that bridges theoretical multi-robot coordination with practical deployment in complex, dynamic environments.

Abstract

We address multi-robot motion planning under Signal Temporal Logic (STL) specifications with kinodynamic constraints. Exact approaches face scalability bottlenecks and limited adaptability, while conventional sampling-based meth- ods require excessive samples to construct optimal trajectories. We propose a two-stage framework integrating sampling-based online learning with formal STL reasoning. At the single-robot level, our constrained Bayesian Optimization-based Tree search (cBOT) planner uses a Gaussian process as a surrogate model to learn local cost maps and feasibility constraints, generating shorter collision-free trajectories with fewer samples. At the multi-robot level, our STL-enhanced Kinodynamic Conflict- Based Search (STL-KCBS) algorithm incorporates STL moni- toring into conflict detection and resolution, ensuring specifica- tion satisfaction while maintaining scalability and probabilistic completeness. Benchmarking demonstrates improved trajectory efficiency and safety over existing methods. Real-world exper- iments with autonomous surface vehicles validate robustness and practical applicability in uncertain environments. The STL- cBOT Planner will be released as an open-source package, and videos of real-world and simulated experiments are available at https://stlbot.github.io/.

Index terms

Path Planning for Multiple Mobile Robots or Agents Integrated Planning and Learning Motion and Path Planning

Related papers