TGH-Planner: Topology-Guided Hierarchical Planner for Nonholonomic Robots in 2-D Unknown and Complex Environments

While motion planning has witnessed significant advances, challenges remain in rapidly generating safe and feasible trajectories in unknown and complex environments, particularly under nonholonomic dynamics constraints. To address these issues, we propose the topology-guided hierarchical (TGH) planner, which combines a global path planner for efficient topological exploration with a local trajectory optimizer to enhance feasibility and safety. The global path planner integrates topological path set, a memory-based structure that preserves and reuses historical topological paths to accelerate replanning and stabilize global guidance. To enable global topological maintenance, we propose a computationally efficient homotopy equivalence check algorithm, which speeds up the identification of topologically distinct paths. Guided by the global path, the local optimizer conducts B-spline-based trajectory optimization, incorporating an angular velocity constraint for nonholonomic kinematics, and an unknown obstacle risk to proactively mitigate potential collision risks. Experimental results demonstrate that TGH-Planner outperforms existing methods, achieving a 100% navigation success rate, a 4.2–7.5× speedup in replanning, a 4%–33% increase in average velocity, and a 2%–35% reduction in execution time. Real-world deployment further confirms its effectiveness for the autonomous navigation of nonholonomic robots in complex, unknown 2-D environments.