Measurement-Based Access Point Deployment Mechanism for Indoor Localization

In this paper, we study how to deploy new access points (AP) to achieve improved accuracy for WiFi- based indoor localization systems. Existing mechanisms in this aspect are typically simulation based and further they do not consider how to use pre-existing APs in target environment for achieving high localization performance. To overcome these issues, in this paper, we propose a measurement-based AP deployment mechanism (MAPD). MAPD takes advantage of those pre-existing APs to identify candidate positions with poor localization accuracy for deploying new APs. We then collect the fingerprints for all possible AP deployment layouts via over- deployment of APs, one at each candidate position. Finally, we present a greedy search algorithm to identify m positions out of the n candidate positions (m≤n) while minimizing the location error. Experimental results demonstrate that the localization errors can be largely reduced: Mean error distance can be reduced by 0.56 meter (26%) and 0.17 meter (10%) as compared with the case without deploying new APs and previous work, respectively; Moreover, the maximum location error can be reduced by 1.53 meter (27%) and 0.51 meter (11%), respectively.