Gallant: Voxel Grid-based Humanoid Locomotion and Local-navigation across 3D Constrained Terrains

2025-11-18

AI Summary

Existing methods rely on depth or elevation maps, providing only local 2D planar perception—insufficient for robust navigation of humanoid robots in complex 3D constrained environments (e.g., multi-level stairs, narrow passages, lateral obstacles). To address this, we propose a voxel-grid-based full 3D environmental representation framework. For the first time, we voxelize LiDAR point clouds and apply a z-axis grouped 2D CNN for end-to-end perception–control co-learning. Integrated with a high-fidelity LiDAR simulation system, our approach enables globally consistent 3D structural modeling and policy optimization. Experiments demonstrate near-100% success rates on challenging tasks including stair climbing and platform ascent—significantly surpassing conventional ground-centric perception paradigms. Our method establishes a scalable, embodied intelligence navigation paradigm for complex 3D terrains.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Computer Vision: 3D Computer Vision
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

Robust humanoid locomotion requires accurate and globally consistent perception of the surrounding 3D environment. However, existing perception modules, mainly based on depth images or elevation maps, offer only partial and locally flattened views of the environment, failing to capture the full 3D structure. This paper presents Gallant, a voxel-grid-based framework for humanoid locomotion and local navigation in 3D constrained terrains. It leverages voxelized LiDAR data as a lightweight and structured perceptual representation, and employs a z-grouped 2D CNN to map this representation to the control policy, enabling fully end-to-end optimization. A high-fidelity LiDAR simulation that dynamically generates realistic observations is developed to support scalable, LiDAR-based training and ensure sim-to-real consistency. Experimental results show that Gallant's broader perceptual coverage facilitates the use of a single policy that goes beyond the limitations of previous methods confined to ground-level obstacles, extending to lateral clutter, overhead constraints, multi-level structures, and narrow passages. Gallant also firstly achieves near 100% success rates in challenging scenarios such as stair climbing and stepping onto elevated platforms through improved end-to-end optimization.

Problem

• Achieving robust humanoid locomotion across complex 3D constrained terrains
• Overcoming limitations of partial environmental perception in existing methods
• Enabling navigation through lateral clutter, overhead constraints and narrow passages

Innovation

• Voxel grid-based framework for humanoid locomotion
• Z-grouped 2D CNN mapping perception to control policy
• High-fidelity LiDAR simulation enabling sim-to-real transfer