一种基于多目标混沌PSO的机器人足球防守策略

Robot Soccer Defensive Strategy Based on Multi-objective Chaotic PSO

提出多目标混沌粒子群优化算法并应用于机器人足球防守策略之中。在对方多名队员进攻情况下,通过该策略引导我方队员防守跑位,并选取我方位置最佳队员对对方主攻队员进行截球,从而达到成功防守的目的。传统的防守策略,仅是根据具体环境采取应对策略,而基于多目标PSO的机器人足球防守策略通过粒子群优化算法的随机性能提高防守队员在动态比赛环境下的适应性,为了避免粒子群优化算法陷入局部极值,对粒子群的最优位置进行混沌优化以提高群体多样性。在FIRA仿真平台中,将加入多目标混沌PSO的机器人足球防守策略和传统防守策略相比较,实验结果表明基于多目标混沌PSO的机器人足球防守策略能较大程度地提高球队整体防守能力。

A robot soccer defensive strategy based on multi-objective chaotic particle swarm algorithm was proposed. When two or more opponents attacked, the proposed strategy could guide the defensive players move and defense. Meanwhile, in order to defense successfully, the proposed strategy selected the player with the best position to intercept the ball which was controlled by the main attacker. The traditional strategy could just take some coping strategies according to the different environment. However, the robot soccer defensive strategy based on multi-objective PSO increased the adaptability in the dynamic competition environment by the random characteristic of PSO. In order to avoid trapping in the local optimum, a chaotic optimization was combined in the proposed multi-objective PSO to improve the diversity of the population. To verify the effectiveness of the proposed algorithm in FIRA simulation platform, MO-CPSO was evaluated in the robot soccer competition, and the experiment results show that the robot soccer defensive strategy based on multi-objective chaotic PSO can improve the capacity of group defense than the traditional defensive strategy.