CogniPlan: Uncertainty-Guided Path Planning with Conditional Generative Layout Prediction

Yizhuo Wang, Haodong He, Jingsong Liang, Yuhong Cao, Ritabrata Chakraborty, Guillaume Adrien Sartoretti
Proceedings of The 9th Conference on Robot Learning, PMLR 305:1382-1396, 2025.

Abstract

Path planning in unknown environments is a crucial yet inherently challenging capability for mobile robots, which primarily encompasses two coupled tasks: autonomous exploration and point-goal navigation. In both cases, the robot must perceive the environment, update its belief, and accurately estimate potential information gain on-the-fly to guide planning. In this work, we propose CogniPlan, a novel path planning framework that leverages multiple plausible layouts predicted by a conditional generative inpainting model, mirroring how humans rely on cognitive maps during navigation. These predictions, based on the partially observed map and a set of layout conditioning vectors, enable our planner to reason effectively under uncertainty. We demonstrate strong synergy between generative image-based layout prediction and graph-attention-based path planning, allowing CogniPlan to combine the scalability of graph representations with the fidelity and predictiveness of occupancy maps, yielding notable performance gains in both exploration and navigation. We extensively evaluate CogniPlan on two datasets (hundreds of maps and realistic floor plans), consistently outperforming state-of-the-art planners. We further deploy it in a high-fidelity simulator and on hardware, showcasing its high-quality path planning and real-world applicability.

Cite this Paper

BibTeX
@InProceedings{pmlr-v305-wang25d, title = {CogniPlan: Uncertainty-Guided Path Planning with Conditional Generative Layout Prediction}, author = {Wang, Yizhuo and He, Haodong and Liang, Jingsong and Cao, Yuhong and Chakraborty, Ritabrata and Sartoretti, Guillaume Adrien}, booktitle = {Proceedings of The 9th Conference on Robot Learning}, pages = {1382--1396}, 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/wang25d/wang25d.pdf}, url = {https://proceedings.mlr.press/v305/wang25d.html}, abstract = {Path planning in unknown environments is a crucial yet inherently challenging capability for mobile robots, which primarily encompasses two coupled tasks: autonomous exploration and point-goal navigation. In both cases, the robot must perceive the environment, update its belief, and accurately estimate potential information gain on-the-fly to guide planning. In this work, we propose CogniPlan, a novel path planning framework that leverages multiple plausible layouts predicted by a conditional generative inpainting model, mirroring how humans rely on cognitive maps during navigation. These predictions, based on the partially observed map and a set of layout conditioning vectors, enable our planner to reason effectively under uncertainty. We demonstrate strong synergy between generative image-based layout prediction and graph-attention-based path planning, allowing CogniPlan to combine the scalability of graph representations with the fidelity and predictiveness of occupancy maps, yielding notable performance gains in both exploration and navigation. We extensively evaluate CogniPlan on two datasets (hundreds of maps and realistic floor plans), consistently outperforming state-of-the-art planners. We further deploy it in a high-fidelity simulator and on hardware, showcasing its high-quality path planning and real-world applicability.} }
Endnote
%0 Conference Paper %T CogniPlan: Uncertainty-Guided Path Planning with Conditional Generative Layout Prediction %A Yizhuo Wang %A Haodong He %A Jingsong Liang %A Yuhong Cao %A Ritabrata Chakraborty %A Guillaume Adrien Sartoretti %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-wang25d %I PMLR %P 1382--1396 %U https://proceedings.mlr.press/v305/wang25d.html %V 305 %X Path planning in unknown environments is a crucial yet inherently challenging capability for mobile robots, which primarily encompasses two coupled tasks: autonomous exploration and point-goal navigation. In both cases, the robot must perceive the environment, update its belief, and accurately estimate potential information gain on-the-fly to guide planning. In this work, we propose CogniPlan, a novel path planning framework that leverages multiple plausible layouts predicted by a conditional generative inpainting model, mirroring how humans rely on cognitive maps during navigation. These predictions, based on the partially observed map and a set of layout conditioning vectors, enable our planner to reason effectively under uncertainty. We demonstrate strong synergy between generative image-based layout prediction and graph-attention-based path planning, allowing CogniPlan to combine the scalability of graph representations with the fidelity and predictiveness of occupancy maps, yielding notable performance gains in both exploration and navigation. We extensively evaluate CogniPlan on two datasets (hundreds of maps and realistic floor plans), consistently outperforming state-of-the-art planners. We further deploy it in a high-fidelity simulator and on hardware, showcasing its high-quality path planning and real-world applicability.
APA
Wang, Y., He, H., Liang, J., Cao, Y., Chakraborty, R. & Sartoretti, G.A.. (2025). CogniPlan: Uncertainty-Guided Path Planning with Conditional Generative Layout Prediction. Proceedings of The 9th Conference on Robot Learning, in Proceedings of Machine Learning Research 305:1382-1396 Available from https://proceedings.mlr.press/v305/wang25d.html.

Related Material