Planning reliably through noise in new environments using neural operators
Planning under Noise with Neural Operators
Machine LearningRobotics
Summary
Planning paths for robots or agents can be hard when their moves are noisy or uncertain, and when the environment is new or complex. The authors created a method called Planning Under Noise (PUN) that learns how noise affects value maps for navigation, helping agents plan reliably even in unseen spaces. They improved how boundaries like walls reflect paths, letting the system explore better, and used a special sampling method extended to higher dimensions. Their approach is faster and more successful than other planners in tests involving noise and complex spaces.
What this means in practice
- •For robotics engineers: Plan reliable paths for robots in noisy environments with varying or unknown geometries, improving success and speed over existing methods.
- •For autonomous vehicle developers: Enable faster and more robust navigation in uncertain and cluttered scenarios by predicting noise-aware navigation fields without costly computations at runtime.
Authors
Michael Chen, Hong Chul Nam, Xi Deng, Anima Anandkumar
Abstract
Planning under uncertain control execution noise requires reliable feedback policies across diverse environments. We introduce Planning Under Noise (PUN), a neural operator that learns noise-aware value fields for planning in unseen geometries. Building on an established screened-Poisson formulation of stochastic optimal control, we address two challenges: specifying boundary conditions that support navigation across varying geometries and efficiently computing fields with reliable gradients. Zero-Dirichlet obstacle boundaries terminate sampled paths at walls, weakening the goal signal along long routes and making it harder to estimate reliably. We instead impose reflecting Neumann conditions on obstacles and a Dirichlet condition at the goal, allowing sampled paths to continue exploring free space until they reach the target. To solve the resulting PDE without a configuration-space mesh, we extend a Monte Carlo (MC) solver, Walk on Stars (WoSt), to higher dimensions. However, local MC fluctuations can reverse estimated gradient directions and cause planning failures, making direct planning require costly sampling. PUN amortizes this computation by learning geometry-to-field mappings from noisy, partially converged Monte Carlo labels, avoiding expensive convergence for every training example. At inference, test-time scaling uses bounded local search guided by the predicted global field. We evaluate PUN on planning tasks across dimensions. Across four noise levels, PUN matches or outperforms all compared planners, achieving 95-100% success on the 4D replanning benchmark while maintaining the lowest runtime. At the highest tested noise level in 4D, PUN achieves 10.7-60.6% relative gains in success rate over other methods while running 3.29-32.15 times faster.