A Hybrid Evolutionary Approach for Multi Robot Coordinated Planning at Intersections

TL;DR

Multi-robot coordination at intersections is framed as a non-differentiable, large-scale optimization problem. The authors propose RADES, a gradient-free Rank-based Differential Evolution with a Successful Archive, to optimize lattice-roadmap parameters $\mathbf{x}=(b_i,h_i)$ and minimize $\sum_{i=1}^{N} L(P_i)$ over directed graphs $G_i=(V_i,E_i)$. Key contributions include the lattice-roadmap formulation, the archive-enhanced RADES algorithm, and empirical results across ten intersection scenarios showing competitive convergence and feasibility vs. seven DE-based baselines. The approach offers a scalable, sampling-efficient method for generating time-parameterized, collision-free trajectories in constrained, real-world environments.

Abstract

Coordinated multi-robot motion planning at intersections is key for safe mobility in roads, factories and warehouses. The rapidly exploring random tree (RRT) algorithms are popular in multi-robot motion planning. However, generating the graph configuration space and searching in the composite tensor configuration space is computationally expensive for large number of sample points. In this paper, we propose a new evolutionary-based algorithm using a parametric lattice-based configuration and the discrete-based RRT for collision-free multi-robot planning at intersections. Our computational experiments using complex planning intersection scenarios have shown the feasibility and the superiority of the proposed algorithm compared to seven other related approaches. Our results offer new sampling and representation mechanisms to render optimization-based approaches for multi-robot navigation.

Figures (8)

• Figure 1: Overview of a cross intersection scenario with 7 robots (vehicles): (a) initial configuration of robots, and (b) the origin (with numbers), the destination (red color) and the ideal trajectory configurations with arrows.
• Figure 2: Basic approach for lattice-based roadmap configuration.
• Figure 3: Statistical comparisons derived from pair-wise Wilcoxon tests at 5% significance level.
• Figure 4: Number of cases of outperformance in blue color, equal performance in cyan color, and underperformance in yellow color derived from pair-wise Wilcoxon tests at 5% significance level. The bars in blue/cyan/orange color show the number of times in which the algorithm is significantly better/similar/worse compared to other algorithms.
• Figure 5: Examples of the average convergence performance over independent runs.