Anytime Computation Algorithms for Stochastically Parametric Approach-Evasion Differential Games

Erich Mueller,Sze Zheng Yong,Minghui Zhu,Emilio Frazzoli
DOI: https://doi.org/10.1109/iros.2013.6696902
2013-01-01
Abstract:We consider an approach-evasion differential game where the inputs of one of the players are upper bounded by a random variable. The game enjoys the order preserving property where a larger relaxation of the random variable induces a smaller value function. Two numerical computation algorithms are proposed to asymptotically recover the expected value function. The performance of the proposed algorithms is compared via a stochastically parametric homicidal chauffeur game. The algorithms are also applied to the scenario of merging lanes in urban transportation.
What problem does this paper attempt to address?