2D仿真机器人足球中Agent定位设计

Agent positioning design in simulation robot soccer 2D

为了模拟真实情况,通常在2D仿真机器人足球的Server里人为地加入了一些随机噪声[1-2],使得球员获得的定位信息不够准确,甚至是随机的,从而导致球员会做出一些错误的决策。针对该问题,本文提出一种解决方法——半逆推量化公式法。从Server中的量化噪声公式入手,进行数学的分析与证明和仿真实验的验证,从而得出该有效的方法,并将它写入代码来提高球员的定位能力,以使得传球、截球、射门等能够更为精准。即球员获得的信息更为准确,那么球员做出的决策判断就更为合理。

In order to simulate real situation, some random noise was added to the server of Simulation Robot Soccer 2D which made positioning information players obtained not exact enough even stochastic, resulting in players to make a wrong decision. For this problem, a solution： semi-inverse quantitative formulas method was proposed in this paper. Starting from quantitative noise formulas in the server, this effective method could be obtained after mathematical analysis and proof and verification in simulation experiment. Coding this method to improve players location capability would bring more exact passes, intercepts and shoots. Namely the more exact information players got the more rational decision and judgmnent players made.