Teardrop Risk Field-Based Spatiotemporal Path Planning Method for Mobile Robots in Dynamic Environments

Area of Science:

• Robotics
• Artificial Intelligence
• Autonomous Systems

Background:

• Autonomous robots require efficient and safe navigation in dynamic environments.
• Traditional methods often fail to account for obstacle velocities, leading to suboptimal or unsafe path planning.
• Existing approaches like ORCA, MPPI, and ST-RRT* have limitations in handling dynamic threats.

Purpose of the Study:

• To introduce a novel spatiotemporal RRT framework, TRF-ST-RRT, for enhanced autonomous robot navigation.
• To develop a velocity-based "teardrop" risk field (TRF) for anisotropic safety corridor construction.
• To improve computational efficiency and safety in dynamic environments compared to existing methods.

Main Methods:

• Proposed TRF-ST-RRT framework integrating a velocity-based teardrop risk field (TRF).
• Implemented a three-layer probabilistic hybrid sampling strategy for accelerated 3D spatiotemporal search.
• Utilized line-of-sight path simplification, temporal reconstruction, and Bézier smoothing for trajectory executability.

Main Results:

• TRF-ST-RRT demonstrated substantial reductions in planning time and iteration counts.
• Achieved notable improvements in overall driving efficiency.
• Showcased enhanced dynamic evasion success rates compared to state-of-the-art baselines.

Conclusions:

• The proposed TRF-ST-RRT framework offers a significant advancement in safe and efficient autonomous navigation.
• The velocity-based teardrop risk field effectively handles dynamic obstacles, ensuring robust spatial buffers.
• This approach provides a promising solution for real-time robotic navigation in complex, dynamic environments.