Belief-Conditioned One-Step Diffusion: Real-Time Trajectory Planning with Just-Enough Sensing

2025-08-16

AI Summary

In partially observable environments, robotic perception incurs excessive energy consumption due to redundant sensor usage. Method: This paper proposes a belief-conditioned single-step diffusion planning framework that jointly optimizes trajectory generation and minimal sensor subset selection to maintain state uncertainty at the task-required minimum. Diffusion models are directly employed to capture the coupling between state uncertainty and sensor selection; denoised trajectory distributions serve as differentiable, calibrated surrogates for localization error—bypassing traditional covariance propagation and enabling end-to-end closed-loop perception-motion co-planning. The diffusion process is conditioned jointly on pose belief rasterization and sensor masks, integrated with a soft actor-critic algorithm for real-time, short-horizon decision-making. Results: Evaluated on autonomous surface vehicles in sea trials, the method achieves 10-ms single-step inference latency, preserves target arrival rate, and significantly reduces sensing energy consumption.

Technology Category

• Intelligent Robots: State Estimation
• Planning, Routing, and Scheduling: Planning with Language Models
• Computer Vision: Diffusion Models for Vision

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Robots equipped with rich sensor suites can localize reliably in partially-observable environments, but powering every sensor continuously is wasteful and often infeasible. Belief-space planners address this by propagating pose-belief covariance through analytic models and switching sensors heuristically--a brittle, runtime-expensive approach. Data-driven approaches--including diffusion models--learn multi-modal trajectories from demonstrations, but presuppose an accurate, always-on state estimate. We address the largely open problem: for a given task in a mapped environment, which extit{minimal sensor subset} must be active at each location to maintain state uncertainty extit{just low enough} to complete the task? Our key insight is that when a diffusion planner is explicitly conditioned on a pose-belief raster and a sensor mask, the spread of its denoising trajectories yields a calibrated, differentiable proxy for the expected localisation error. Building on this insight, we present Belief-Conditioned One-Step Diffusion (B-COD), the first planner that, in a 10 ms forward pass, returns a short-horizon trajectory, per-waypoint aleatoric variances, and a proxy for localisation error--eliminating external covariance rollouts. We show that this single proxy suffices for a soft-actor-critic to choose sensors online, optimising energy while bounding pose-covariance growth. We deploy B-COD in real-time marine trials on an unmanned surface vehicle and show that it reduces sensing energy consumption while matching the goal-reach performance of an always-on baseline.

Problem

• Minimizing sensor usage while maintaining low state uncertainty
• Real-time trajectory planning with pose-belief and sensor constraints
• Energy-efficient sensing without compromising localization accuracy

Innovation

• Diffusion planner conditioned on pose-belief raster
• Real-time trajectory planning in 10 ms
• Soft-actor-critic optimizes sensor energy online