I. Introduction
With the increasing focus and demand of intelligence of robots, multiagent games have been widely studied in recent years, among which the research of pursuit-evasion (PE) game is gaining momentum [1]. In PE game, the pursuers attempt to capture the evaders while the evaders want to escape or delay the capture. This scenario is widely found in nature, where the predator strives to chase its food and the prey struggles for surviving [2], [3], [4]. A more important application of the PE game lies in the military, where there is a tremendous demand for pursuit strategies to attack the target and evasion strategies to avoid some damage [5], [6]. Considering different settings, researchers have focused on various problems derived from the PE game, such as target-attacker-defender (TAD) problems [7], reach-avoid games [8], perimeter defense problems [9], and games in a region with [10] or without [11] boundaries.