WiFi与惯导融合的渐消因子扩展卡尔曼滤波实时定位算法

A Real-time Positioning Algorithm Using Fading Factor Kalman Filter Based on WiFi and Inertial Fusion

针对无线信号强度易受干扰,基于RSSI指纹库室内定位技术的定位结果常出现跳跃不稳定现象,提出并实现一种WiFi与惯导融合的渐消因子扩展卡尔曼滤波实时定位算法。该方法基于加速度数据进行多重约束波峰-波谷检测实现自适应步态识别,根据室内几何布局特征划分矢量域修正方向传感器数据确定其航向角,获得行人位移参数。然后建立基于渐消因子扩展卡尔曼滤波融合模型,实现最终位置估计。实验结果表明该算法可以有效抑制无线定位的跳跃或堆积现象,进而增强室内定位稳健性与可靠性,平均定位精度在2 m左右。

Due to the indoor positioning errors produced by the unsteadiness of location technique, a newpositioning technology by fusing inertial measuring unit IMU） and WiFi wireless signals with fading-factor-based extended Kalman f i lter （ EKF） is proposed. A multijDle restrictions for peak-valley detection is developed on acceler-ation for real-time step recognition. Then the paper utilizes the feature of indoor environment to amend the orientation for getting a correct heading angle. Final ly, this paper proposes a fading- factor-based EK F fu s io n model based on displacement constraintwith WiFi and inertial sensors positioning techniques for user’ s location estimation. The experimenrithm can effectively suppress the unsteadiness of jump or centralization, and enhance the indoor location robustness and rel ia bi l i-ty. The average positioning accuracy is about two meters.