AI Movement Design for Robots, Autonomous Vehicles & Gaming
SUTD Researchers Train AI with Reinforcement Learning to Defeat Champion Street Fighter Players
Tags: Singapore University of Technology and Design, Singapore, Transportation & Automotive, Electronics & Robotics
SUTD researchers developed a reinforcement learning algorithm that trains AI to outperform champion Street Fighter players by generating complex movement strategies. The AI’s approach uses millions of trial movements to refine its skills, demonstrating success against in-built game AIs and showcasing high energy efficiency with minimal hardware consumption. Potential applications include training for autonomous vehicles, collaborative robots, and high-level gaming scenarios. This technology could transform fields requiring precise movement design by accelerating AI's strategic learning and adaptability. The study points toward future competitive applications in video games and real-world movement systems.
IP Type or Form Factor: Software & Algorithm
TRL: 4 - minimum viable product built in lab
Industry or Tech Area: Automobiles Autonomous; Robotics
