PhysicsFC: Learning User-Controlled Skills for a Physics-Based Football Player Controller

TL;DR

PhysicsFC presents a hierarchical framework for learning user-controllable, physics-based football skills by training separate skill policies (Dribble, Trap, Move, Kick) that output latent vectors for a physics-based motion embedding model (CALM). The approach introduces Skill Transition-Based Initialization (STI) to enable smooth transitions between skills and Data-Embedded Goal-Conditioned Latent Guidance (DEGCL) to preserve motion diversity during Move. Quantitative evaluations and interactive demos (including 11v11 gameplay) demonstrate improved transition speed, goal accuracy, and naturalness of ball handling under user control. The work advances realistic, interactive football simulations by tightly coupling high-level user intent with low-level physics-based motion generation, enabling responsive, controllable gameplay with complex skill repertoires.

Abstract

We propose PhysicsFC, a method for controlling physically simulated football player characters to perform a variety of football skills--such as dribbling, trapping, moving, and kicking--based on user input, while seamlessly transitioning between these skills. Our skill-specific policies, which generate latent variables for each football skill, are trained using an existing physics-based motion embedding model that serves as a foundation for reproducing football motions. Key features include a tailored reward design for the Dribble policy, a two-phase reward structure combined with projectile dynamics-based initialization for the Trap policy, and a Data-Embedded Goal-Conditioned Latent Guidance (DEGCL) method for the Move policy. Using the trained skill policies, the proposed football player finite state machine (PhysicsFC FSM) allows users to interactively control the character. To ensure smooth and agile transitions between skill policies, as defined in the FSM, we introduce the Skill Transition-Based Initialization (STI), which is applied during the training of each skill policy. We develop several interactive scenarios to showcase PhysicsFC's effectiveness, including competitive trapping and dribbling, give-and-go plays, and 11v11 football games, where multiple PhysicsFC agents produce natural and controllable physics-based football player behaviors. Quantitative evaluations further validate the performance of individual skill policies and the transitions between them, using the presented metrics and experimental designs.

Figures (14)

• Figure 1: PhysicsFC FSM.
• Figure 2: Example of Dribble Policy Training with Skill Transition-Based State Initialization (STI): (a) Numerous episodes are simulated using trained skill policies, and the character and ball states are stored in STI buffers for each skill. (b) During Dribble policy training, half of the episodes are initialized with states randomly sampled from the Trap STI buffer, while the other half are initialized from the Move STI buffer. Through these episodes, the Dribble policy learns to initiate dribbling quickly in various situations, both while moving and immediately after trapping.
• Figure 3: Foot collision mesh.
• Figure 4: Visualization of Dribble reward.
• Figure 5: Visualization of Trap reward.