Application of Extended Game in Multi-UUV Pursuit-Escape Task

Abstract

Abstract With the extensive deployment of unmanned systems in various fields, the requirements for unmanned underwater vehicles (UUV) are becoming higher and higher. A single UUV can no longer meet the task requirements, and the cooperative control of multiple UUVs is required. In this paper, facing the hunting-escape task, establishing the kinematics and observation communication model of UUV, the multi-UUV pursuit task is completed based on this model. The multi-UUV maneuvering strategy set, payoff function and Nash equilibrium solution are given by using dynamic extended form game. In the game task, the pursuit UUVs decompose many-to-one game toone-to-one, and the escape UUV calculates the highest income among the multiple UUVs, that is, in the process of the task, the pursuit party simplifies the task to one-to-one interception, and the other is the best strategy for many-to-one selection. Finally, the pursuit and escape simulation experiments are carried out to verify the effectiveness and superiority of the algorithm. From the perspective of effectiveness, after 29 games, the pursuit UUV successfully encircles the escape UUV. The effectiveness of the algorithm is verified from the encircling path, the change of the distance between the pursuit UUV and the breaching party’s UUV, and the final situation. The effectiveness of UUV controller is analyzed from lateral, longitudinal velocity and heading control. In terms of the superiority of the algorithm, when the pursuit party chooses the maneuver strategy given by the interception and breakthrough theory, the pursuit party fails after the same time and strategy selection steps, and the escape succeeds. The superiority of the algorithm can be obtained from the comparison of the simulation results.