Reactive Symbolic Planning and Control in Dynamic Adversarial Environments

Matching-based capture strategies for 3D heterogeneous multiplayer reach-avoid differential games

“…HJ methods [6] intrinsically can deal with general obstacle environments, but they require the state space discretization (thus suffer from loss of accuracy), and are limited to systems up to about five states in practice [5]. The same scalability issue exists when approximating the game as a finite-state model and then synthesizing the strategies using LTL algorithms [32]. Efficient algorithms have been proposed with restrictions, such as discrete-time dynamics and linear-quadratic approximations [1].…”