摘要
针对RSSI方式对移动节点进行导航定位时会引入较大的模型误差的问题,提出了一种高效的硬件分组粒子滤波算法对节点间RSSI信号进行滤波预处理,确保RSSI值和节点间距离呈单调函数关系。直接利用RSSI信息和网络拓扑结构实现机器人定位,避免将RSSI值转换为几何距离时带来的模型误差。将滤波处理和导航计算分散到多个信标节点上执行以提高算法实时性,在导航实施之前无需精确已知各信标节点的几何坐标,免去了人工部署信标节点的步骤,可适合网络拓扑结构或环境参数频繁变化的场合,该方法在现场实验中导航控制精度可达到0.6 m。
A group improved particle filter algorithm was designed to preprocess the RSSI （received singal strength indicator） signal which includes plenty of background noise. In order to avoid the model error, the coordinate space which was quantized by using RSSI value was used to describe the robot＇ s state and target position. The navigation system consists of some beacon nodes, and each of them is a distributed control unit. Then the navigation control center gathers the control information from each beacon nodes and calculates the final outputs for mobile robot. The coordination of beacon node is not needed to know before navigation, and it allow for adaptation to dynamic or unknown scenarios. The experimental indicate that an accuracy of 0.6 m is attainable with this method.
出处
《电机与控制学报》
EI
CSCD
北大核心
2008年第6期717-722,共6页
Electric Machines and Control
基金
国家自然科学基金(60572010)
黑龙江省自然科学基金(F2007-08)
关键词
无线传感器网络
粒子滤波
自主导航
接收信号强度
wireless sensor network
particle filters
autonomous navigation
received singal trength indications
