摘要
室内定位中信号传播环境多为非视距、视距混合环境,在不同环境下的测量数据统计特性不同,因此需要进行非视距鉴别。经典的非视距鉴别方法计算复杂,不适合工程应用,针对此问题提出利用信号强度和首径估计值相结合进行非视距鉴别,鉴别率可达90%。基于超宽带的定位系统在非视距环境下的测量数据含有非高斯有色噪声,传统的卡尔曼滤波要求测量噪声为白噪声,否则滤波效果变差。针对此问题,提出一种基于现代时间序列分析理论的自适应滤波方法,用于处理测量误差。分析了误差特性,建立了误差模型并推导了滤波公式。经实验分析可得,该方法适合处理动态数据,可减小30%~40%的测量误差。
Most of the signal propagation environments are made up of non-line of sight and visual environment in the indoor positioning. In different environments,the statistical properties of the measurement data are different,so it is necessary to perform non visual range identification. The classical non-line of sight identification method is computationally complex,and is not suitable for engineering applications. In view of this problem,we propose using the combination of signal strength and the first path to identify the non line of sight. The identification rate can reach 90%.We found that ultra-wide-band positioning system in the non-line of sight environment where measurement data contained the cost of Gauss colored noise. The traditional Kalman filter required that the noise be white noise. Otherwise the filtering effect was poor. To solve this problem,a new method based on the theory of modern time series analysis was proposed. The error characteristic was analyzed,and the error model was established. By experimental analysis,our method can reduce the random error of 30% ~ 40%.
出处
《计算机应用与软件》
2017年第10期157-162,184,共7页
Computer Applications and Software
