Effect of clutter topology on multi-hop localizer placement

Accuracy in range-based localization systems can degrade rapidly in the presence of clutter in the environment. This is due to the incidence of Non-Line-of-Sight (NLOS) distance measurements between the anchors and an unlocalized node. While a large corpus of research work dealt with the scenario where the NLOS distances form a minority of the total distances to anchors, there has been not much research done to handle the situation where a majority or even all distance measurements to anchors are NLOS in nature. In our previous work, we showed that using localizers in a cluttered environment can improve the localization accuracy of a target node even when all the distance measurements are NLOS. Instead of NLOS bias, these techniques suffer residual multi-hop error, which is caused due to the distance overestimate when a multi-hop chain is used instead of the straight-line distance. In this paper, we analyze the effect of clutter topology on the multi-hop error. We use machine learning techniques to estimate the aggregate forms of the multi-hop error for a given clutter topology when only characteristic features of the clutter topology are provided.