无线电工程

为了实现高精度室内定位，在超宽带(Ultra-Wideband, UWB)定位中运用天牛须搜索(Beetle Antennae Search, BAS)算法，将三维定位的非线性方程组求解问题转化为最优化问题，在行人航位推算(Pedestrian Dead Reckoning, PDR)定位中采用基于时间周期性的峰值检测法与自适应步长估计算法减少伪波峰对步态检测的干扰，以提高2种定位技术的定位精度和可靠性。采用基于PDR航向角动态改变过程噪声Q值的偏移卡尔曼滤波法来识别UWB信号传播情况，从而实现利用UWB定位修正PDR航向角累积误差，利用PDR定位修正UWB非视距(Non-Line-of-Sight, NLOS)定位误差。搭建一套室内定位的实验演示系统进行验证，测试结果表明，所提算法可以有效降低视距(Line-of-Sight, LOS)和NLOS情况下UWB定位误差。特别是在NLOS情况下，融合定位算法比单一UWB定位算法的定位精度提升了约68%,平均定位误差达到0.137 m。

To achieve high precision indoor positioning, the Beetle Antennae Search(BAS) is used in Ultra-Wideband(UWB) positioning, transforming the solution of nonlinear equations of three-dimensional positioning into an optimization problem.The time-periodic peak detection method and self-adaptive step size estimation algorithm are adopted in Pedestrian Dead Reckoning(PDR) positioning to reduce the interference of pseudo peaks on step detection, which will improve the positioning accuracy and reliability of both techniques.The biased Kalman filter whose process noise Q value is dynamically changed with the PDR heading angle is used to identify the signal transmission conditions, as a result, the cumulative error of PDR heading angle is corrected by UWB positioning and the error of UWB positioning in Non-Line-of-Sight(NLOS) condition is corrected by PDR positioning.An experimental indoor positioning system is built for verification, and the testing results show that the proposed fusion method can effectively reduce the UWB positioning error under both Line-of-Sight(LOS) and NLOS conditions.Especially in NLOS condition, as compared with single UWB positioning method, the positioning accuracy of fusion positioning algorithm is improved by about 68%,and the average positioning error is about 0.137 m.