Physics-informed Neural Networks for Robot Motion under Constraints
Keywords: Robotics, Motion Planning, Neural Fields, Implicit Neural Representation, Physics Informed Deep Learning
Abstract: Neural Motion Planners (NMPs) have emerged as a promising tool for solving robot tasks in complex environments. However, these methods often require expert data for learning, which limits their application to scenarios where data generation is time-consuming. Recent developments have also led to physics-informed deep learning models capable of representing complex Partial Differential Equations (PDEs). Inspired by these developments, we propose physics-informed neural motion planning methods for robot navigation and manipulation. Our framework represents a wave propagation model generating continuous arrival time to find path solutions informed by a nonlinear first-order PDE called the Eikonal equation. We evaluate our method in various robot navigation and manipulation tasks, including 3D robot navigation in the Gibson dataset, and 6-DOF robot manipulators constrained motion planning. Furthermore, the results show that our method exhibits high success rates and significantly lower computational times than the state-of-the-art methods, including NMPs that require training data from classical planners.
Submission Number: 7
Loading