Hybrid Classical/RL Local Planner for Ground Robot Navigation

TL;DR

This work tackles local planning for ground robot navigation by merging a classical Dynamics Window Approach (DWA) with a learning-based SACPlanner in a parallel framework. A heuristic, clearance-based switch selects the planner that yields the best response to current surroundings, avoiding the need for trained switching networks. Empirical results on a live robot show the hybrid planner reduces navigation time by about 26% and eliminates collisions across challenging scenarios, while maintaining smooth motion. The approach delivers practical benefits by combining obstacle-avoidance prowess with smooth, efficient trajectory tracking, using a lightweight switching mechanism within a ROS-based pipeline. The work thus advances robust, real-time local planning for dynamic environments without additional neural-network complexity for switching.

Abstract

Local planning is an optimization process within a mobile robot navigation stack that searches for the best velocity vector, given the robot and environment state. Depending on how the optimization criteria and constraints are defined, some planners may be better than others in specific situations. We consider two conceptually different planners. The first planner explores the velocity space in real-time and has superior path-tracking and motion smoothness performance. The second planner was trained using reinforcement learning methods to produce the best velocity based on its training $"$experience$"$. It is better at avoiding dynamic obstacles but at the expense of motion smoothness. We propose a simple yet effective meta-reasoning approach that takes advantage of both approaches by switching between planners based on the surroundings. We demonstrate the superiority of our hybrid planner, both qualitatively and quantitatively, over the individual planners on a live robot in different scenarios, achieving an improvement of 26% in the navigation time.

Figures (8)

• Figure 1: Confronting an obstacle in a series of arc-motions.
• Figure 2: ROS framework and the architecture of our hybrid local planner.
• Figure 3: Waypoint generation.
• Figure 4: Clearance detection.
• Figure 5: Dummy training environment (left) and the associated polar costmap (right).