Multi-Agent Path Planning in Complex Environments using Gaussian Belief Propagation with Global Path Finding

TL;DR

This work tackles multi-agent path planning in complex environments by augmenting Gaussian Belief Propagation (GBP) with a tracking factor that enforces adherence to global paths. It implements two global-path integration strategies—waypoint tracking (WT) as a baseline and path tracking (PT) with a dedicated tracking factor—and evaluates them using two global planners: RRT* and a lane-structured SP. Key contributions include introducing the tracking factor, defining its measurement and Jacobian, and validating the approach in simulation with metrics for collisions and path deviation; results show substantial reductions in path deviation (up to 28% in single-agent and 16% in multi-agent) and, when paired with structured planning, elimination of collisions in some scenarios. The findings suggest that tight global-path adherence via the tracking factor, particularly with structured global planning, improves coordination and safety in complex, potentially communication-challenged environments, with practical implications for scalable multi-robot systems.

Abstract

Multi-agent path planning is a critical challenge in robotics, requiring agents to navigate complex environments while avoiding collisions and optimizing travel efficiency. This work addresses the limitations of existing approaches by combining Gaussian belief propagation with path integration and introducing a novel tracking factor to ensure strict adherence to global paths. The proposed method is tested with two different global path-planning approaches: rapidly exploring random trees and a structured planner, which leverages predefined lane structures to improve coordination. A simulation environment was developed to validate the proposed method across diverse scenarios, each posing unique challenges in navigation and communication. Simulation results demonstrate that the tracking factor reduces path deviation by 28% in single-agent and 16% in multi-agent scenarios, highlighting its effectiveness in improving multi-agent coordination, especially when combined with structured global planning.

Figures (6)

• Figure 1: Illustration of the Complex environment with multiple crossings and bends. Larger spheres represent agents, while the chains denote variables in agents' factor graphs and the solid lines indicate agents' traveled path.
• Figure 2: Illustration of the tracking factor concept. The tracking factor measurement is shown in orange, the state variables is shown in blue, and the gray line is the global path connecting two waypoints.
• Figure 3: The path deviation visualized between waypoint tracking (blue), path tracking (green), and desired path (black).
• Figure 4: Illustration of the two environments used in the experiments: (a) junction environment, (b) complex environment.
• Figure 5: Results from the junction environment. There is a slight increase in inter-robot collisions when using path tracking instead of waypoint. The number of collisions with the environment is the same for WT and PT. Average over 5 runs for each failure rate with 1200 robots.