Fillet-based RRT*: A Rapid Convergence Implementation of RRT* for Curvature Constrained Vehicles

James Swedeen,Greg Droge,Randall Christensen

DOI: https://doi.org/10.48550/arXiv.2302.11648

2023-02-23

Abstract:Rapidly exploring random trees (RRTs) have proven effective in quickly finding feasible solutions to complex motion planning problems. RRT* is an extension of the RRT algorithm that provides probabilistic asymptotic optimality guarantees when using straight-line motion primitives. This work provides extensions to RRT and RRT* that employ fillets as motion primitives, allowing path curvature constraints to be considered when planning. Two fillets are developed, an arc-based fillet that uses circular arcs to generate paths that respect maximum curvature constraints and a spline-based fillet that uses Bezier curves to additionally respect curvature continuity requirements. Planning with these fillets is shown to far exceed the performance of RRT* using Dubin's path motion primitives, approaching the performance of planning with straight-line path primitives. Path sampling heuristics are also introduced to accelerate convergence for nonholonomic motion planning. Comparisons to established RRT* approaches are made using the Open Motion Planning Library (OMPL).

Robotics,Systems and Control

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to improve the convergence speed and performance of the RRT* algorithm while considering the vehicle path curvature constraints in path planning. Specifically, although the traditional RRT* algorithm can provide asymptotically optimal solutions in probability, it performs poorly and has a slow convergence speed when dealing with non - holonomic motion constraints (such as maximum curvature limits). To solve these problems, the paper proposes a "fillet" method based on circular arcs and Bezier curves as the basic motion unit for path planning to naturally consider the path curvature constraints, and accelerates the convergence speed of the algorithm by introducing intelligent sampling and path optimization techniques.

The main contributions of the paper include:
1. **Developed two fillet methods**: One is a fillet based on circular arcs, and the other is a fillet based on Bezier curves. These two methods are respectively used to generate paths that satisfy the maximum curvature constraint and the curvature continuity requirement.
2. **Improved the RRT* algorithm**: By introducing intelligent sampling and path optimization techniques, the convergence speed of the algorithm in non - holonomic motion planning is accelerated.
3. **Combined multiple sampling and path optimization techniques**: Combining the ideas of Informed RRT* and Smart RRT*, a new Smart and Informed RRT* (SI - RRT*) algorithm is developed, which reduces the need for parameter adjustment and improves the robustness and efficiency of the algorithm.

These improvements make the fillet - based RRT* algorithm far outperform the RRT* algorithm using Dubin path basic motions in terms of performance when dealing with path curvature constraints, and close to the planning performance using straight - line path basic motions.