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

Gokul Puthumanaillam, Aditya Penumarti, Manav Vora, Paulo Padrao, Jose Fuentes, Leonardo Bobadilla, Jane Shin, Melkior Ornik
Proceedings of The 9th Conference on Robot Learning, PMLR 305:68-92, 2025.

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 *minimal sensor subset* must be active at each location to maintain state uncertainty *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 localization 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 localization 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.

Cite this Paper

BibTeX
@InProceedings{pmlr-v305-puthumanaillam25a, title = {Belief-Conditioned One-Step Diffusion: Real-Time Trajectory Planning with Just-Enough Sensing}, author = {Puthumanaillam, Gokul and Penumarti, Aditya and Vora, Manav and Padrao, Paulo and Fuentes, Jose and Bobadilla, Leonardo and Shin, Jane and Ornik, Melkior}, booktitle = {Proceedings of The 9th Conference on Robot Learning}, pages = {68--92}, year = {2025}, editor = {Lim, Joseph and Song, Shuran and Park, Hae-Won}, volume = {305}, series = {Proceedings of Machine Learning Research}, month = {27--30 Sep}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v305/main/assets/puthumanaillam25a/puthumanaillam25a.pdf}, url = {https://proceedings.mlr.press/v305/puthumanaillam25a.html}, 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 *minimal sensor subset* must be active at each location to maintain state uncertainty *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 localization 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 localization 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.} }
Endnote
%0 Conference Paper %T Belief-Conditioned One-Step Diffusion: Real-Time Trajectory Planning with Just-Enough Sensing %A Gokul Puthumanaillam %A Aditya Penumarti %A Manav Vora %A Paulo Padrao %A Jose Fuentes %A Leonardo Bobadilla %A Jane Shin %A Melkior Ornik %B Proceedings of The 9th Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2025 %E Joseph Lim %E Shuran Song %E Hae-Won Park %F pmlr-v305-puthumanaillam25a %I PMLR %P 68--92 %U https://proceedings.mlr.press/v305/puthumanaillam25a.html %V 305 %X 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 *minimal sensor subset* must be active at each location to maintain state uncertainty *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 localization 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 localization 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.
APA
Puthumanaillam, G., Penumarti, A., Vora, M., Padrao, P., Fuentes, J., Bobadilla, L., Shin, J. & Ornik, M.. (2025). Belief-Conditioned One-Step Diffusion: Real-Time Trajectory Planning with Just-Enough Sensing. Proceedings of The 9th Conference on Robot Learning, in Proceedings of Machine Learning Research 305:68-92 Available from https://proceedings.mlr.press/v305/puthumanaillam25a.html.

Related Material