Keisuke Ando

Papers

Total Citations

H-Index

About

Keisuke Ando is a leading researcher in multi-agent systems and robotic soccer, with a primary focus on advancing decision-making in the RoboCupSoccer Simulation 2D (RSS2D) domain. His most notable contribution is the development and analysis of the Expected Possession Value (EPV) framework, a sophisticated metric that quantifies the probability of a team maintaining ball control in dynamic, adversarial environments. By modeling the constant motion of players and the ball, Ando’s work enables autonomous agents to evaluate the strategic worth of actions in real time—transforming raw positional data into actionable intelligence for passing, dribbling, and positioning. This breakthrough directly addresses the core challenge of RSS2D, where agents must continuously adapt to fluid game states. While his seminal 2023 paper has garnered initial citations, its impact is growing as the EPV approach becomes a foundational tool for optimizing team coordination and long-term planning in competitive multi-robot settings. Ando’s research bridges reinforcement learning and sports analytics, offering a principled method for assessing possession dynamics that has implications beyond robotics, including in human soccer strategy and autonomous vehicle coordination.