联合TDOA-AOA无线传感器网络半定规划定位算法研究

Joint TDOA-AOA Source Localization in Wireless Sensor Networks Using Semidefinite Programming

在无线传感器网络定位中,TDOA和AOA联合定位可有效利用多种位置信息提高定位精度。由于传统联合加权最小二乘(WLS)的目标函数非线性,在应用于无线传感器网络定位时,会产生多个局部最优解。因此,针对该问题本文将约束加权最小二乘问题转化为二次约束二次规划问题,之后通过引入半定松弛(SDR)方法将联合定位问题转换为低复杂度的半定规划问题(SDP),进而寻找全局最优解。并且针对实际应用中参考节点带误差的情形分析和推导了定位算法。与已有算法相比,提出的算法在参考节点无误差和有误差时都有更高的精度。此外,提出的SDP算法还能够实现只有两个参考节点下的目标定位。

Joint TDOA-AOA localization is able to improve location estimation accuracy through utilizing all the available information in wireless sensor networks（ WSNs）. Due to the non-linearity of the equations,joint weighted least square（ WLS） may converge to local optimum solution when applied to wireless sensor networks localization.Therefore,by applying semidefinite relaxation（ SDR） to the quadratically constrained quadratic program problem which is derived from WLS method,we transformed the joint localization problem into a low complexity semidefinite programming（ SDP） problem to find the global optimum solution. Localization in the presence of sensor position errors has also been studied. Simulation results show that,compared with the existing approaches,the proposed approach has higher accuracy whether with or without sensor position errors. Furthermore,the proposed SDP method still has good performance with only two sensors offering location information.