ShuttleEnv: An Interactive Data-Driven RL Environment for Badminton Strategy Modeling

Ang Li; Xinyang Gong; Bozhou Chen; Yunlong Lu; Jiaming Ji; Yongyi Wang; Yaodong Yang; Wenxin Li

cs.AIMar 18, 2026

ShuttleEnv: An Interactive Data-Driven RL Environment for Badminton Strategy Modeling

Authors: Ang Li, Xinyang Gong, Bozhou Chen, Yunlong Lu, Jiaming Ji, Yongyi Wang, Yaodong Yang, Wenxin Li

Organizations: School of Computer Science, Peking University · Institute for Artificial Intelligence, Peking University

Abstract

We present ShuttleEnv, an interactive and data-driven simulation environment for badminton, designed to support reinforcement learning and strategic behavior analysis in fast-paced adversarial sports. The environment is grounded in elite-player match data and employs explicit probabilistic models to simulate rally-level dynamics, enabling realistic and interpretable agent-opponent interactions without relying on physics-based simulation. In this demonstration, we showcase multiple trained agents within ShuttleEnv and provide live, step-by-step visualization of badminton rallies, allowing attendees to explore different play styles, observe emergent strategies, and interactively analyze decision-making behaviors. ShuttleEnv serves as a reusable platform for research, visualization, and demonstration of intelligent agents in sports AI. Our ShuttleEnv demo video URL: https://drive.google.com/file/d/1hTR4P16U27H2O0-w316bR73pxE2ucczX/view

Figures & tables

Figure 1: The workflow of our data-driven badminton environment. The RL agent interacts with the environment in a standard loop. Inside the environment, the outcome of an action $a_{t}$ is determined by a two-stage probabilistic process, driven by two pre-trained models ( $M_{\text{succ}}$ and $M_{\text{ret}}$ ) which leverage the elite-level badminton match data. The agent receives a new observation $o_{t+1}$ generated from the action history window to make the next decision and a sparse reward $R_{t}$ at the end of a rally.

	Stroke Type				Landing Zone
Player	Top-1	Top-2	Top-3	Top-4	Top-1	Top-2	Top-3	Top-4
Lin Dan	0.473	0.664	0.785	0.880	0.580	0.763	0.854	0.898
Lee Chong Wei	0.438	0.689	0.816	0.871	0.594	0.796	0.884	0.924

Table 1: Top-

k

accuracy for next-action prediction.

Figure 2: Representative frame from the visualization demo. The rendering overlays rally context (active player, action metadata, and score) on a 3D court scene. The visualization highlights how different shot selections and target areas lead to distinct rally trajectories and outcomes, supporting qualitative inspection of tactical decision-making.

Game-Playing Agents

Game-playing agents use search, learning, and reasoning to choose actions in games with strategic, uncertain, or changing environments.

See all 194 papers

ShuttleArena: Interpretable Self-Play in Physics-Based Badminton
Aug 26, 2026Peize DingGame-Playing Agents Self-Play RL
Columbia University New York, NY 10027, USA
Humanoid Badminton: Learning Dynamic Racket Skills from Limited Human Motion Data
Sep 25, 2026Jingzhi Cui, Zhexiong Wang, Bangjie Xu +9Robot Skill Learning Robotics
Tsinghua University · Hong Kong Embodied AI Lab · Zhejiang University +3
Humanoid Whole-Body Badminton via an Annealed Reinforcement Learning Curriculum
Nov 14, 2025Chenhao Liu, Leyun Jiang, Ningyuan Tian +4RL for Robotics Reinforcement Learning

Aug 26, 2026J/K move · Enter open · S save

ShuttleArena: Interpretable Self-Play in Physics-Based Badminton

Peize Ding

Columbia University New York, NY 10027, USA

Game-Playing Agents Self-Play RL RL Benchmarks Interpretability in RL

Open paper AskPDF arXiv

Badminton is a compact but challenging domain for game AI: a player must choose a physically feasible shuttle trajectory, anticipate the opponent's interception, and recover to a court position whose value depends on the opponent's next response. The central challenge is that shot selection and recovery are not separable: the best recovery depends on the shot-induced opponent response, while the value of the shot depends on whether the hitter can cover the reply. This paper presents ShuttleArena, a physics-based singles badminton self-play environment that couples continuous shuttle flight, player interception, structured shot generation, and post-shot recovery. The policy uses role-conditioned outputs: a masked interception choice on receiver turns and a factorized hitter action over shot azimuth, shot elevation, shot speed, and recovery target, enabling interpretable tactical probes. Episodes are single rallies rather than full scored games, and training uses Proximal Policy Optimization (PPO) self-play against a staged checkpoint opponent pool with sparse terminal rally-outcome rewards and a factor-specific recovery update. Evaluation with frozen checkpoint play, controlled tactical probes, recovery ablations, qualitative rollouts, and a human-data sanity check shows competitive improvement together with interpretable opponent-conditioned changes in shot geometry and recovery behavior. The learned policies produce recognizable badminton-like structure while also reflecting the abstractions of the simulator, and the recovery intervention shows that learned recovery behavior is competitively important. These results suggest that physics-based racket sports are a useful testbed for interactive digital entertainment AI because they require agents to coordinate execution, positioning, and opponent-relative tactical value.

Abstract

Game-Playing Agents

Explore similar work

ShuttleArena: Interpretable Self-Play in Physics-Based Badminton

Humanoid Badminton: Learning Dynamic Racket Skills from Limited Human Motion Data

Humanoid Whole-Body Badminton via an Annealed Reinforcement Learning Curriculum