基于三次样条插值的无线信号强度衰减模型

为提高指纹识别定位技术在无线网络中的应用效率,解决无线信号强度衰减模型计算效率低、构建复杂以及参数难确定等问题,提出使用较少参数实现精确的无线信号强度传播模型.该模型是一种经验模型,其由距离或信号强度的三次样条插值函数来调整经典衰减模型中的衰减因子,无需增加额外参数,计算简单.实验结果显示：在室内走廊环境中,该模型以10%的样本数据进行训练后,无线信号强度预测误差不超过0.4 dBm,由该模型生成校准数据,利用指纹识别法可获得平均误差1.3 m以内的定位精度.

Wireless signal strength propagation model base on cubic spline interpolation

An accurate wireless signal strength propagation model with less parameters was proposed to increase the application efficiency of fingerprint based positioning approaches in wireless networks, and to address the problems of wireless signal attenuation models, e.g. low computation efficiency, build complexity, and difficulty of parameter setting. The proposed model is an empirical model, which changes the attenuation factor n of the classical attenuation model, in terms of a cubic interpolation spline function of distance or received signal strength indicator (RSSI). The proposed model does not introduce any new parameter and is easy to compute. Experimental results indicated that in an indoor corridor environment, by trained with 10% samples, the proposed model estimated the wireless signal strength with an average error less than 0.4 dBm. Employing the calibration data generated by the proposed model, the fingerprint based positioning approach achieved a mean error less than 1.3 m．