Automatic Generation of Ice Hockey Defensive Motion via Coverage Control and Control Barrier Functions

Abstract

A successful defensive strategy in ice hockey games is often designed empirically by an experienced professional. The majority of previous work on automating the strategy focuses on analyzing spatial data to decide the most optimal formation and action but cannot generalize the system to real games with real-time capabilities. We propose a novel control logic for generating real-time ice hockey defensive motion based on a control barrier function (CBF) and coverage control to extend our antecessors' logic that succeeds in duplicating ideal formations for specific scenes. To this end, we first present an ellipsoidal CBF to overcome the drawbacks of the existing line-based CBF of our antecessors. We also tune and add a novel density function to reflect real specifications more precisely than the previous work. The control logic is then demonstrated through simulations with offensive motion in real games. It is confirmed that the present logic generates valid defensive movements without specification to these specific scenes. We further exemplify that the logic generates proper motion under ice hockey's man-to-man and zone defense strategies and their intermediate strategies by tuning the logic. This would contribute to reducing the efforts of the practitioners to educate ice hockey players.