使用滤波技术提高室内BLE指纹定位性能

Improving the Performance of Indoor BLE Fingerprint Localization by Filtering Technology

为了减小复杂的环境和多径效应对室内定位精度的影响,测试强干扰环境下滤波器对基于BLE信标的室内指纹库定位精度的提升。实现卡尔曼滤波和粒子滤波两种贝叶斯滤波技术,在办公环境下采集BLE数据,实测了不使用滤波技术、均值滤波、卡尔曼滤波、粒子滤波的定位精度,同时测试增加BLE信标数量对定位结果的影响。定位误差用累积分布函数、均方误差和绝对误差表示。实验表明使用滤波技术可以很大程度改善高干扰环境下的BLE指纹定位性能,其中粒子滤波的性能在各情况下均最佳。使用贝叶斯滤波器时BLE信标数量由4个提升为5个时,定位性能提升明显。

In order to reduce the influence of complex environment and multipath effect on indoor positioning accuracy,this paper measures the accuracy of indoor fingerprint database positioning based on Bluetooth Low Energy(BLE)beacon.Kalman Filter(KF)and Particle Filter(PF)techniques are implemented to improve positioning performance.The positioning accuracy without filtering technology,average filtering,KF and PF is measured in office environment.In addition,the effect of number of BLE beacons on positioning results is tested.The errors are compared in terms of cumulative distribution function curve,mean absolute error and root mean squared error.According to the experimental results,the positioning performance can be greatly improved by using filter technology,in which PF has the best performance in all cases.The performance improves significantly when the number of BLE beacons increases from 4 to 5 with Bayesian filter technology.