焦聚NAO机器人点球大战——一种基于概率模型的最优设计算法实现

Focus on NAO robot penalty kick——An implementation of optimal design algorithm based on probability model

NAO机器人作为robocub标准平台组的比赛机器人,使用于各大学间的机器人足球比赛。通过大量比赛视频获取数据,将数据写入数据库进行分类总结,集中分析点球命中率低下原因。运用正态分布函数计算不同角度点球命中概率,根据概率密度分布图比较得到各区域进球概论,并用三角函数算出最优射门角度。基于java语言,对点球过程及射门角度进行优化设计,并以封装形式实现整个程序流程。基于python语言,利用单线程实现点球过程中助跑、转身和射门动作,利用多线程实现多名球员共同庆祝动作。使点球流程更加逼真,拥有显示比赛中的足球魅力。最后利用Choregraphe程序仿真机器人点球过程,缩小助跑角度的同时将进球概率最大化。

During the University campaign, Robocub standard platform group robots has powered by NAO robots. Through a large number of robocub standard platform group game, we concentrate to analyze the cause of low penalty scoring rate. The normal probability is used to calculate the penalty scoring probability of different angles. According to the probability density distribution, the results of the regional goals are obtained. And the optimal shooting angle is calculated by trigonometric function. The penalty process and shooting angle are optimized by java language. At the same time through the encapsulate method, It achieve the entire penalty process. Based on the python language, using of single-threaded penalty process help to achieve running, turning and shooting actions, and using of multi-threaded to help players to accomplish celebrating actions. Above of that make the penalty process more realistic, and find back more charm of football. Lastly, Choregraphe program is used to simulate the robot's penalty process, which minimizes the run-up angle and maximizes the scoring probability.