基于多传感器数据融合的机器人里程计设计与实现  

Design and Implementation of Robot Odometer Based on Multi-Sensor Data Fusion

作　　者：余翀[1] 高翔[2] 邱其文[2] 

机构地区：[1]复旦大学信息科学与工程学院,上海200433 [2]南京邮电大学自动化学院,南京210046

出　　处：《传感技术学报》2012年第1期53-57,共5页Chinese Journal of Sensors and Actuators

摘　　要：设计并实现了一种基于FPGA和多光电鼠标的高精度机器人里程计。多个PS/2光电鼠标传感器测量位移数据,利用FPGA解析PS/2协议并完成数据融合,得出高精度机器人里程计结果。针对传统Luo一致性数据融合算法的缺陷进行改进,并通过归一化特征值加权法得到每个传感器测量值被系统综合支持的程度,完成多传感器测量数据融合。实验结果表明:该数据算法计算步骤固定,方便在FPGA上实现;该里程计在有异常数据干扰情况下,能够达到较高的测量精度。An high-precision odometer by using FPGA and multi-optical-mouse sensors was designed and implemented. Firstly, PS/2 optical mouse sensors were used to measure the displacement data;then PS/2 protocol was parsed and multi-sensor data fusion algorithm was implemented by FPGA to obtain the results of high-precision robot odometer. On the base of improving the classical data fusion algorithm of consistency proposed by Luo, by using the normalized eigenvalue weighted value to obtain the level of support for each sensor integrating all sensor measurements in the system, the process of muhi-sensor data fusion is completed. Experimental results show that the algorithm steps are fixed and easy to implement by FPGA. In the interference of abnormal data, the robot odometer can meet the requirements of high accuracy.

关 键 词：导航定位技术 机器人里程计 数据融合 多传感器 光电鼠标 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]