基于DRSS模型的相对误差估计及凸优化混合定位方法研究

Research on the DRSS Model-Based Hybrid Localization Method Using Relative Error Estimation and Convex Optimization

基于接收信号强度差(Difference of Received Signal Strength, DRSS)的定位模型具有节省能量、带宽和时间的优点,并且在定位过程中隐藏了发射机的传输方式,非常有益于机密监视或军事应用。然而DRSS模型具有较高的非凸性,在定位求解时比较困难,本文提出了一种改进的定位方法——相对误差及凸优化混合定位方法。首先借助相对误差方法构建最小化问题,然后借助半正定规划和二阶锥规划对模型进行近似求解。为了验证所提方法的有效性,引入均方根误差(Root Mean Square Error, RMSE)作为估计方法精度的评判标准,通过对比本文所提方法以及现有四种方法(A-BLUE、U-BLUE、LARE-SDP、SOCP)的RMSE,研究结果发现本文提出方法的RMSE值最低,并且更加贴近理论误差的CRLB下界。The positioning model based on Difference of Received Signal Strength (DRSS) has the advantages of saving energy, bandwidth, and time, and hides the transmission mode of the transmitter during the positioning process, which is very beneficial for confidential monitoring or military applications. However, the DRSS model has high nonconvexity and is difficult to solve in localization. This paper proposes an improved localization method—a hybrid localization method of relative error and convex optimization. Firstly, the minimization problem is constructed using the relative error method, and then the model is approximately solved using semi positive definite programming and second-order cone programming. In order to verify the effectiveness of the proposed method, Root Mean Square Error (RMSE) was introduced as the evaluation criterion for the accuracy of the estimation method. By comparing the RMSE of the proposed method with four existing methods (A-BLUE, U-BLUE, LARE-SDP, SOCP), the research results showed that the RMSE value of the proposed method was the lowest and closer to the CRLB lower bound of the theoretical error.