Efficient Swept Volume-Based Trajectory Generation for Arbitrary-Shaped Ground Robot Navigation

TL;DR

This work tackles the challenge of safely navigating arbitrarily shaped ground robots in cluttered environments with efficient continuous collision avoidance (CCA). It introduces a coarse-to-fine framework comprising topology path generation, SE(2) motion sequence generation, and subproblem trajectory optimization that leverages SVSDF for SE(2) segments and simpler R^2 planning elsewhere. Key contributions include geometry-aware topology shortening, SE(2) motion sequence construction with high-/low-risk region classification, and a MINCO-based back-end that splices optimized SE(2) and R^2 segments into a continuous trajectory. Extensive benchmarks and real-world experiments show substantial gains in planning speed and CCA success rate compared with state-of-the-art baselines, demonstrating practical applicability for real-time autonomous navigation of complex-shaped robots.

Abstract

Navigating an arbitrary-shaped ground robot safely in cluttered environments remains a challenging problem. The existing trajectory planners that account for the robot's physical geometry severely suffer from the intractable runtime. To achieve both computational efficiency and Continuous Collision Avoidance (CCA) of arbitrary-shaped ground robot planning, we proposed a novel coarse-to-fine navigation framework that significantly accelerates planning. In the first stage, a sampling-based method selectively generates distinct topological paths that guarantee a minimum inflated margin. In the second stage, a geometry-aware front-end strategy is designed to discretize these topologies into full-state robot motion sequences while concurrently partitioning the paths into SE(2) sub-problems and simpler R2 sub-problems for back-end optimization. In the final stage, an SVSDF-based optimizer generates trajectories tailored to these sub-problems and seamlessly splices them into a continuous final motion plan. Extensive benchmark comparisons show that the proposed method is one to several orders of magnitude faster than the cutting-edge methods in runtime while maintaining a high planning success rate and ensuring CCA.

Figures (11)

• Figure 1: A T-Shaped delivery robot navigating a cluttered indoor environment in the real-world experiment. A whole-body trajectory is generated by the proposed framework to ensure precise continuous collision avoidance.
• Figure 2: The overview of our proposed planning framework.
• Figure 3: (a) Actual experiment robot with dimension marked, where the inscribed circumference of the robot is shown. (b) The robot body frame ESDF is built inside the robot geometry. The sum of the gradient (green arrow) can be used to avoid the obstacle.
• Figure 4: A detoured and long path is shortened based on the robot's actual geometry. (a) At each discretized point, the obstacle that blocked its visibility to the last point is pushed to generate a new waypoint and stored with a safe orientation (green T-shaped). (b) The final simplified topological path (green line).
• Figure 5: A T-shaped robot detects a collision with the environment, where the robot kernel discretizes 18 gridded configurations at 20-degree clockwise rotational increments.