基于核岭回归的自适应蓝牙定位方法

Adaptive Bluetooth location method based on kernel ridge regression

针对室内高精度定位需求和蓝牙信号强度动态变化特征,提出了一种基于核岭回归（KRR）的定位方法,只需利用蓝牙锚节点之间的信号强度及其物理坐标信息,学习蓝牙信号强度与物理坐标的回归模型,并能在线动态更新模型参数,实现自适应免标定定位。实验结果表明,KRR方法对信号强度的动态变化具有较好的适应性和鲁棒性,平均定位误差为1.25m,相比信号—距离映射方法（SDM）能取得更高的定位精度;实验也验证了有效的滤波处理能进一步改善定位效果。

In order to meet the requirement of high precise indoor location and overcome the difficulty of the dynamic characteristics of the Bluetooth RSSI,this paper proposed a location method based on kernel ridge regression（ KRR） . It was adaptive and calibration-free,which made use of the information of RSSI from Bluetooth beacons and each beacon’s physical coordinates,and learned the regression model between them. Meanwhile,it could dynamically update the model parameters. Experimental results show that the method is robust and adaptive to the RSSI dynamic changes,and the average location error is measured to be 1. 25 m,which achieves higher location precision than the signal-distance map（ SDM） method. Moreover,the experiment shows that the efficient filter method can further improve the precision.