Deviation Analysis For Approximate Maximum Likelihood Localization Algorithms

Approximate maximum likelihood localization Algorithm (AML) based on TDOA put forward by Yiu-Tong Chan estimates the tag accurate position through establishing a maximum likelihood function and solving its extreme value. However, if the hyperbolas of the TDOA equations have two intersections, AML will get two solutions that fit the equations. In this case, the solution chosen with the AML suggestion may not be optimal. Through simulation, we prove that AML may choose non-optimal solution when hyperbolas have two intersections. Further, this article analyzes the cause of appearing two solutions and the areas where exist two solutions. From the simulations and analysis, deviation areas are related to the sensors' arranging positions. In order to avoid deviations caused by not selecting the optimal solution, some suggestions for sensors application arrangement are given.