# LEGO: Leveraging Experience in Roadmap Generation for Sampling-Based Planning

arXiv: https://arxiv.org/abs/1907.09574 · HTML: https://mlsanity.com/paper/1907.09574/

## Key facts

- Cited by 19 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 22 Jul 2019.

## Built on by (most cited first, 19 of 19)

- [Motion Policy Networks](https://mlsanity.com/paper/2210.12209/) (2022)
- [NTFields: Neural Time Fields for Physics-Informed Robot Motion Planning](https://mlsanity.com/paper/2210.00120/) (2022)
- [Neuro-algorithmic Policies enable Fast Combinatorial Generalization](https://mlsanity.com/paper/2102.07456/) (2021)
- [PRESTO: Fast Motion Planning Using Diffusion Models Based on Key-Configuration Environment Representation](https://mlsanity.com/paper/2409.16012/) (2024)
- [Motion Planning Transformers: A Motion Planning Framework for Mobile Robots](https://mlsanity.com/paper/2106.02791/) (2021)
- [Deep Reactive Policy: Learning Reactive Manipulator Motion Planning for Dynamic Environments](https://mlsanity.com/paper/2509.06953/) (2025)
- [Progressive Learning for Physics-informed Neural Motion Planning](https://mlsanity.com/paper/2306.00616/) (2023)
- [Learning Sampling Dictionaries for Efficient and Generalizable Robot Motion Planning with Transformers](https://mlsanity.com/paper/2306.00851/) (2023)
- [CTRMs: Learning to Construct Cooperative Timed Roadmaps for Multi-agent Path Planning in Continuous Spaces](https://mlsanity.com/paper/2201.09467/) (2022)
- [A Survey on the Integration of Machine Learning with Sampling-based Motion Planning](https://mlsanity.com/paper/2211.08368/) (2022)
- [Physics-informed Temporal Difference Metric Learning for Robot Motion Planning](https://mlsanity.com/paper/2505.05691/) (2025)
- [Planning from Pixels in Environments with Combinatorially Hard Search Spaces](https://mlsanity.com/paper/2110.06149/) (2021)
- [Using Deep Learning to Bootstrap Abstractions for Hierarchical Robot Planning](https://mlsanity.com/paper/2202.00907/) (2022)
- [Approximating Constraint Manifolds Using Generative Models for Sampling-Based Constrained Motion Planning](https://mlsanity.com/paper/2204.06791/) (2022)
- [Robot Motion Planning as Video Prediction: A Spatio-Temporal Neural Network-based Motion Planner](https://mlsanity.com/paper/2208.11287/) (2022)
- [Transformer-Enhanced Motion Planner: Attention-Guided Sampling for State-Specific Decision Making](https://mlsanity.com/paper/2404.19403/) (2024)
- [Toward Generalist Neural Motion Planners for Robotic Manipulators: Challenges and Opportunities](https://mlsanity.com/paper/2603.24318/) (2026)
- [Learning Continuous Cost-to-Go Functions for Non-holonomic Systems](https://mlsanity.com/paper/2103.11168/) (2021)
- [SIL-RRT*: Learning Sampling Distribution through Self Imitation Learning](https://mlsanity.com/paper/2411.17293/) (2024)

## Abstract (arXiv)

We consider the problem of leveraging prior experience to generate roadmaps in sampling-based motion planning. A desirable roadmap is one that is sparse, allowing for fast search, with nodes spread out at key locations such that a low-cost feasible path exists. An increasingly popular approach is to learn a distribution of nodes that would produce such a roadmap. State-of-the-art is to train a conditional variational auto-encoder (CVAE) on the prior dataset with the shortest paths as target input. While this is quite effective on many problems, we show it can fail in the face of complex obstacle configurations or mismatch between training and testing. We present an algorithm LEGO that addresses these issues by training the CVAE with target samples that satisfy two important criteria. Firstly, these samples belong only to bottleneck regions along near-optimal paths that are otherwise difficult-to-sample with a uniform sampler. Secondly, these samples are spread out across diverse regions to maximize the likelihood of a feasible path existing. We formally define these properties and prove performance guarantees for LEGO. We extensively evaluate LEGO on a range of planning problems, including robot arm planning, and report significant gains over heuristics as well as learned baselines.