F-RRT: An Efficient Algorithm for Semi-Constrained Path Planning Problems
Guillaume de Mathelin de Papigny, Franco Ivan Gassibe, Vincent Padois
AI summary
Problem
Industrial tasks often require end-effectors to follow a path within flexible tolerances, but traditional sampling-based planners struggle with the high dimensionality and non-differentiable constraints of these semi-constrained manifolds.
Approach
The method reparameterizes the planning problem into a lower-dimensional tolerance space combining path progress and task tolerances, using a modified RRT planner guided by a Quadratic Programming solver to map steps to valid joint configurations.
Key results
- Introduction of F-RRT, a novel sampling-based planner for semi-constrained tasks
- Reparameterization of planning into a lower-dimensional tolerance space to bypass high-dimensional configuration sampling
- Integration of a QP-based inverse kinematics solver ensuring joint continuity, limit compliance, and singularity handling
- Demonstrated superior speed and robustness in complex, cluttered industrial environments compared to existing methods
Why it matters
Provides a practical, low-hyperparameter solution for automating complex industrial tasks like welding and gluing, directly benefiting manufacturing and robotics engineers.
Abstract
This paper addresses the challenging problem of Semi-Constrained End-Effector Path Planning for robotic manip- ulators. This problem arises when complex specifications restrict the end-effector’s motion during the execution of industrial tasks. Traditional path planning algorithms often struggle with such problems due to the difficulty of exploring the robot’s valid configuration space, or constrained manifold, under these condi- tions. In this work, we propose a novel sampling-based approach that efficiently navigates the constrained manifold by exploring an alternative space representing the end-effector’s degrees of freedom, such as process-related tolerances, throughout the task. This method retains the simplicity of sampling-based techniques. Building on this approach, we introduce the F-RRT algorithm, an adaptation of the renowned RRT planner [1]. F-RRT demon- strates enhanced speed and robustness compared to existing solutions, particularly in complex and cluttered environments.