Research map / Reinforcement learning / Visual and representation RL
Push-and-Step: From RL-Based Balance Recovery to Physical Simulation of Dense Crowds
Alexis Jensen, Pei Xu, Ioannis Karamouzas, Charles Pontonnier, Julien Pettré
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 |
|---|---|---|---|
| add imitation reward r_imit (ablation) | pi_pretrain · Pretraining ablation | heading deviation: 44.81° -> 5.93° | Improved |
| add imitation reward r_imit (ablation) | pi_pretrain · Pretraining ablation | total foot sliding: 24cm -> 22cm | No clear effect |
| add imitation reward r_imit (ablation) | pi_pretrain · Pretraining ablation | total kinetic energy: 2381J -> 933J | Improved |
| add motion quality reward r_quality (ablation) | pi_pretrain · Pretraining ablation | total foot sliding: 49cm -> 22cm | Improved |
| add motion quality reward r_quality (ablation) | pi_pretrain · Pretraining ablation | total kinetic energy: 1444J -> 933J | Improved |
| add motion quality reward r_quality (ablation) | pi_pretrain · Pretraining ablation | heading deviation: 13.19° -> 5.93° | Improved |
| add balance reward r_balance (ablation) | pi_pretrain · 80 push trials | average max push withstood: 21 N -> 230 N | Improved |
| add hand placement reward r_hand (ablation) | pi_adapt · Adaptation ablation | final hand height: 1.16m -> 0.81m | Improved |
| add hand placement reward r_hand (ablation) | pi_adapt · Adaptation ablation | transmitted impulse: 75Ns -> 40Ns | Improved |
| add social contact reward r_social (ablation) | pi_adapt · Adaptation ablation | transmitted impulse: 74Ns -> 40Ns | Improved |
| add social contact reward r_social (ablation) | pi_adapt · Adaptation ablation | max hand height: 0.81m -> 0.85m | Mixed |
| add social contact reward r_social (ablation) | pi_adapt · Adaptation ablation | final hand height: 0.81m -> 0.81m | No clear effect |
Closest papers by meaning
- RESCUE: Crowd Evacuation Simulation via Controlling SDM-United Characters (ICCV 2025 · similarity 0.84)
- ELHINN: Unifying Dense Crowd Simulation Across Scales via Eulerian–Lagrangian Hydrodynamics (ECCV 2026 · similarity 0.84)
- PAMotion: Physics-Aware Motion Generation for Full-Body Interaction with Multiple Objects (CVPR 2026 · similarity 0.83)
- Stabilizing Reinforcement Learning in Differentiable Multiphysics Simulation (ICLR 2025 · similarity 0.82)
- Social Physics Informed Diffusion Model for Crowd Simulation (AAAI 2024 · similarity 0.82)
- Continuous Locomotive Crowd Behavior Generation (CVPR 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.
