基于隐马尔可夫模型的机器人足球赛场态势评估

摘要

For robot soccer stadium situation assessment problem, a novel model of robot soccer stadium situation assessment based on the Hidden Markov Model ( HMM) is proposed. First, seven key factors through the characteristics of game environment as the inputs of HMM are abstracted. And then the parameters of HMM using the improved Bauw-Welch algorithm based on the existing experimental database are optimized. Finally, the result of the model is got. The experiment result shows model can satisfy real-time and accuracy demands and provide timely and reliable references for the change of the corresponding strategy for current soccer situation.%针对机器人足球赛场态势评估问题,提出一种基于隐马尔可夫模型(HMM)的机器人足球赛场态势评估模型.通过对赛场信息进行提取和量化,抽象出7个相关态势因素,作为HMM模型的输入数据.通过采用改进的Bauw-Welch算法优化HMM的参数,从而获得赛场态势评估结果.实验结果表明,所提出的方法能较好地满足赛场实时性和准确性的要求,可以为球场上对应策略的变更提供及时、可靠的依据.