使用最速下降算法提高极大似然估计算法的节点定位精度

Using steepest descent method to improve node localization accuracy of maximum likelihood estimation

阐述了极大似然估计算法用于无线传感器网络节点自定位的原理;阐述了最速下降算法求非线性方程组最优解的原理;提出在距离测量误差较大的情况下,使用最速下降算法优化极大似然估计算法所得的节点定位值,并通过模拟实验证实其可行性。实验结果表明,在无须多余通信代价的条件下,优化处理使定位精度得到很大提高,且算法收敛快,计算代价小,适用于无线传感器网络的节点自定位。

This paper expounded the localization principle of maximum likelihood estimation. Also, presented the principle of using steepest descent method to find an optimal solution for a system of nonlinear equations. Proposed steepest descent method to refine the initial node locations gotten by maximum likelihood estimation as the distance measurement error was large, and corresponding simulation experiment was done to testify the validity and advantages of such a disposal. The simulation results show that the refinement can improve the localization accuracy obviously with no more communication cost and small computation cost, fit to be utilized in wireless sensor networks.