SBAMP: Sampling Based Adaptive Motion Planning

TL;DR

SBAMP addresses the challenge of real-time, dynamic navigation by integrating global RRT* planning with an online Lyapunov-stable SEDS-inspired local controller, enabling continuous, stable trajectory following without pretraining data. The framework refits a Lyapunov-stable vector field for each RRT* segment and interleaves fast local updates with slower global replanning under a formal average-dwell-time stability guarantee, preserving global optimality while ensuring reactivity. Empirical validation on RoboRacer and in simulation demonstrates rapid disturbance recovery, robust obstacle avoidance, and resilient operation under extreme perturbations, outperforming standard RRT* in terms of replanning frequency and safety. This approach offers a scalable, data-free pathway to stable, real-time motion planning in dynamic, unstructured environments, with potential extensions to MPC/MPPI and higher-dimensional manipulation tasks.

Abstract

Autonomous robotic systems must navigate complex, dynamic environments in real time, often facing unpredictable obstacles and rapidly changing conditions. Traditional sampling-based methods, such as RRT*, excel at generating collision-free paths but struggle to adapt to sudden changes without extensive replanning. Conversely, learning-based dynamical systems, such as the Stable Estimator of Dynamical Systems (SEDS), offer smooth, adaptive trajectory tracking but typically rely on pre-collected demonstration data, limiting their generalization to novel scenarios. This paper introduces Sampling-Based Adaptive Motion Planning (SBAMP), a novel framework that overcomes these limitations by integrating RRT* for global path planning with a SEDS-based local controller for continuous, adaptive trajectory adjustment. Our approach requires no pre-trained datasets and ensures smooth transitions between planned waypoints, maintaining stability through Lyapunov-based guarantees. We validate SBAMP in both simulated environments and real hardware using the RoboRacer platform, demonstrating superior performance in dynamic obstacle scenarios, rapid recovery from perturbations, and robust handling of sharp turns. Experimental results highlight SBAMP's ability to adapt in real time without sacrificing global path optimality, providing a scalable solution for dynamic, unstructured environments.

Figures (13)

• Figure 1: Flowchart of the SBAMP theoretical framework. When New Path Available? is true, the SEDS generator refits the dynamical system to the latest RRT$^*$ segment; otherwise, the existing SEDS velocity command is executed.
• Figure 2: ROS2 architecture for SBAMP on RoboRacer. Laser scan and odometry feed the occupancy_grid_node; next_waypoint_node and rrt_node produce $\tau$; sbamp_node generates control; visualization_node is launched if enabled.
• Figure 3: RoboRacer simulation environment with SBAMP-generated vector field and RRT* path.
• Figure 4: F1/10 hardware platform used for real‐world validation.
• Figure 5: Replanning Frequency vs. Lateral Perturbation for RRT* and SBAMP