# Agile and Generalized Legged Locomotion via Attention-Based Neural Map Encoding

arXiv: https://arxiv.org/abs/2601.08485 · HTML: https://mlsanity.com/paper/2601.08485/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 13 Jan 2026.

## Built on by (most cited first, 5 of 5)

- [SceneBot: Contact-Prompted General Humanoid Whole Body Tracking with Scene-Interaction](https://mlsanity.com/paper/2606.27581/) (2026)
- [Physics-Informed Reinforcement Learning of Spatial Density Velocity Potentials for Map-Free Racing](https://mlsanity.com/paper/2604.09499/) (2026)
- [Global-Local Attention Decomposition for Terrain Encoding in Humanoid Perceptive Locomotion](https://mlsanity.com/paper/2606.00637/) (2026)
- [Fauna Sprout: A lightweight, approachable, developer-ready humanoid robot](https://mlsanity.com/paper/2601.18963/) (2026)
- [GigaBrain-WBC-0.5: A Behavior World Model for Robust Humanoid Whole-Body Tracking with Environment Interaction](https://mlsanity.com/paper/2608.18234/) (2026)

## Abstract (arXiv)

Achieving agile and generalized legged locomotion across terrains requires tight integration of perception and control, especially under occlusions and sparse footholds. Existing methods have demonstrated agility on parkour courses but often rely on end-to-end sensorimotor models with limited generalization and interpretability. By contrast, methods targeting generalized locomotion typically exhibit limited agility and struggle with visual occlusions. We introduce a unified reinforcement learning (RL) framework for agile and generalized locomotion that incorporates a novel attention-based map encoder in the control policy. This encoder extracts local and global mapping features and uses attention mechanisms to focus on salient regions, producing an interpretable and generalized embedding for RL-based control. We further propose a learning-based mapping pipeline that provides fast, uncertainty-aware terrain representations robust to noise and occlusions, serving as policy inputs. It uses neural networks to convert depth observations into local elevations with uncertainties, and fuses them with odometry. The pipeline also integrates with parallel simulation so that we can train controllers with online mapping, aiding sim-to-real transfer. We validate our framework with the proposed mapping pipeline on a quadruped and a biped robot, and the resulting controllers demonstrate strong agility and generalization to unseen terrains in simulation and in real-world experiments.