Research map / 3D vision / 3D generation
Gallant: Voxel Grid-based Humanoid Locomotion and Local-navigation across 3-D Constrained Terrains
Qingwei Ben, Botian Xu, Kailin Li, Feiyu Jia, Wentao Zhang, Jingping Wang, Jingbo Wang, Dahua Lin, Jiangmiao Pang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| simulate LiDAR returns from dynamic geometry such as the robot's own links (ablation) | Gallant · Ceiling terrain (simulation) | success rate (%): 28.4 -> 97.1 | Improved |
| simulate LiDAR returns from dynamic geometry such as the robot's own links (ablation) | Gallant · Platform terrain (simulation) | success rate (%): 22.16 -> 96.1 | Improved |
| use z-grouped 2D CNN instead of standard 3D CNN (ablation) | Gallant · Pile terrain (simulation) | success rate (%): 65.3 -> 82.1 | Improved |
| use z-grouped 2D CNN instead of standard 3D CNN (ablation) | Gallant · Ceiling terrain (simulation) | success rate (%): 97.5 -> 97.1 | No clear effect |
| use dense 2D CNN instead of sparse 3D CNN (ablation) | Gallant · Pile terrain (simulation) | success rate (%): 52.4 -> 82.1 | Improved |
| use dense 2D CNN instead of sparse 2D CNN (ablation) | Gallant · Pile terrain (simulation) | success rate (%): 57.6 -> 82.1 | Improved |
| add voxel grid to actor instead of height map only (ablation) | Gallant · Ceiling terrain (simulation) | success rate (%): 5.3 -> 97.1 | Improved |
| add voxel grid to actor instead of height map only (ablation) | Gallant · Pile terrain (simulation) | success rate (%): 86.2 -> 82.1 | Hurt |
| add height map as privileged critic input (ablation) | Gallant · Pile terrain (simulation) | success rate (%): 72.3 -> 82.1 | Improved |
| use 5 cm voxels instead of 10 cm (ablation) | Gallant · Platform terrain (simulation) | success rate (%): 75.5 -> 96.1 | Improved |
| use 5 cm voxels instead of 2.5 cm (ablation) | Gallant · Ceiling terrain (simulation) | success rate (%): 13.3 -> 97.1 | Improved |
| use voxel grid instead of elevation map | Gallant · real-world Ceiling and Door | success rate | Improved |
Closest papers by meaning
- Let Humanoids Hike! Integrative Skill Development on Complex Trails (CVPR 2025 · similarity 0.86)
- Hybrid Internal Model: Learning Agile Legged Locomotion with Simulated Robot Response (ICLR 2024 · similarity 0.84)
- Humanoid Horizon: Extending Task Horizon in Whole-Body Loco-Manipulation via Parallel Training, Dynamic Starting, and Reward Gating (NeurIPS 2026 · similarity 0.83)
- Image2Sim: Scaling Embodied Navigation via Generative Neural Simulator (NeurIPS 2026 · similarity 0.83)
- Dynam3D: Dynamic Layered 3D Tokens Empower VLM for Vision-and-Language Navigation (NeurIPS 2025 · similarity 0.82)
- Affordances-Oriented Planning Using Foundation Models for Continuous Vision-Language Navigation (AAAI 2025 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
