推算定位算法在拦截策略中的研究与应用

Research and Application of DR Algorithm in Heading-off Strategy

机器人足球比赛中,研究有效的拦截对方球员的进攻和防守是一项关键的战术,以往开发的拦截策略中存在"漏拦"和"错拦"等现象,造成拦截成功率很低。将推算定位(Dead Reckoning,DR)算法用到拦截策略的研究中,有效估算出对方球员和球的下一步动向,实施有效拦截,提高了比赛过程中拦截判断的精度。对DR算法中的阈值的确定进行了分析和改进,提出一种判据函数,有效地解决动态门选取的问题;针对算法执行和网络的延时给予了延时补偿,提高系统的实时性,使拦截能够快速准确地实现。利用DR算法推算出赛场球和球员的运动趋势,指导赛场策略的生成并和队员的动作与位置的部署。实验数据表明,方法具有很强的优越性。

In robot soccer competition, effective heading - off against the opponent＇ s aggression or defense is a key tactics. Dead Reckoning （DR） algorithm is applied to research heading- off policy, to estimate the trends of the ball and the opponent robots, and then implement effective heading - off action. The threshold value setting in DR al- gorithm is analyzed and improved. A criterion function is put forward to solve the threshold value - selection problem, and time -lapse equalization attributed to calculation and network has been solved. DR algorithm is used to reckon the moving trends of the ball and the players, to support the decision - making system with guidance on their actions and locations. Experiments data indicates that the method above is effective.