基于卡尔曼滤波的仿真机器人定位方法  

Simulation Robot Location Method Based On Kalman Filter

作　　者：张青[1] 李龙澍[1] 刘跃[2] 

机构地区：[1]安徽大学计算机科学与技术学院,安徽合肥230601 [2]合肥市公安局网络安全保卫支队,安徽合肥230001

出　　处：《计算机仿真》2013年第12期317-320,共4页Computer Simulation

基　　金：安徽省自然科学基金(090412054);安徽高等学校省级自然科学基金:(KJ2011Z020)

摘　　要：机器人准确定位问题是RoboCup3D仿真比赛中的一个关键部分。由于视觉信息存在误差,所以能否准确定位,直接关系到机器人能否准确快速地完成任务。采用RoboCup3D仿真平台进行比赛,已从当最初的3V3发展为11V11,无论是球队的进攻配合还是防守站位都必不可少。采用的策略都基于机器人在球场上的位置,使得机器人的定位问题显得尤为重要。为此提出一种改进预测模型的卡尔曼滤波定位方法,通过对机器人位置预测模型的改进,提高卡尔曼滤波的收敛性。在3D仿真平台上测试,定位效果有了明显的提升。Accurate localization of robot is a key part of RoboCup 3D simulation game. In consequence of visual information error, the accurate localization directly affects whether the robot can complete the task quickly and accurately. With the RoboCup 3D simulation platform, which has been developed from 3 vs 3 to 11 vs 11, both the attack coordination and the defense position are essential in the game. These strategies are based on the position of the robot on the court, which makes the location particularly important. Consequently, an improved prediction model Kalman filter method was proposed to get a better convergence by improving the prediction model of the robot position. The experimental result on simulation platform shows that the location can be obviously improved.

关 键 词：机器人足球仿真 定位 位置预测 卡尔曼滤波 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]