隨機角色分配策略於機器人足球賽之模擬研究

Abstract

本論文主要探討足球機器人致勝策略以及路徑設計。利用了基本移動與進階移動來規劃出機器人的避障路徑，並且描述每個足球員角色所執行的任務。另外提出一個隨機角色分配機制，利用機器人與足球在場上的位置，計算出足球員與球的距離以及相對於目標點的角度，並且依據以上位置資訊給予不同的權值計算規則來分配角色。 最後利用FIRA(Federation of International Robot-soccer Association)機器人足球賽軟體Robot Soccer v1.5a 模擬器來執行策略實驗。並且實驗中結果驗證加入角色分配機制對於五對五比賽中，可以提高攻守效率來達到較短的進球時間以及較佳的勝率。

This thesis mainly discusses the wining strategies of soccer robots anddesigns the corresponding routes of robots. We use basic movement and advanced movement to draw up obstacle avoidance, and describe the tasks performed by the role of each soccer player. On the other hand, we propose a dynamic role playing strategy by the corresponding position between soccer players and the ball, and give them different computing rules to assign the roles by above position information. Finally, we use the simulation software, FIRA(Federation of International Robot-soccer Association) Robot Soccer v1.5a, to run the experiment. By the experiment result, it shows that the teams added dynamic role playing strategy have high performance of offensive or defensive, and get shorter competition time and have better rates of scoring.