Many-to-One Pursuit-Evasion Differential Game Method for UAVs Using Sequential Convex Programming

Aiming at the many-to-one pursuit-evasion game problem for fixed-wing unmanned aerial vehicles (UAVs) with complete kinematic constraints, this paper proposes a sequential convex programming-based differential game method for UAVs. The method decouples the coupled pursuit-evasion differential game problem into several differential game subproblems and transforms the non-convex escape objective function into an anti-capture constraint. Linearization is used to convexify the constraint, and then the penalty function method is leveraged to transform the constraint into a convex constraint penalty function. Consequently, the differential game subproblems are transformed into multiple sequence convex programming problems so that the approximate saddle-point solutions can be efficiently and quickly solved. The simulation experiment is conducted in a three-to-one pursuit-evasion game scenario, and the results demonstrate the effectiveness of the proposed method.

I. Introduction

With the development of unmanned technology, UAV swarm, as a new collaborative intelligence platform, have been widely used in the military field. Swarm gaming is one of the research areas that have attracted much attention, and its theory and methodology are the key factors for the victory of unmanned air combat in the future [1].