Robust wifi localization using received signal strength

We present an investigation into the problem of robust wifi localization using received signal strength. Various RF based probabilistic wifi localization model has been developed that uses fingerprinting for location determination. But often they are too complicated and dependent on user's information and infrastructural information. We proposed three models which also takes into account that access points (APs) are randomly distributed, for developing the robust model which is simple, faster, generic, independent of user's information and can also withstand the infrastructural change. These models are naive Bayes classifier and it's modification . We used two techniques sampling Boolean matrix for utilizing the temporal spatial constraints respectively. To the best of our knowledge it is the first time anyone has used the boolean matrix to utilize the spatial constraints without the user information. Use of sampling gives us flexibility of varying sampling period to obtain delicate balance between latency and robustness. We employed confusion matrix and ranking statistics to analyse the performance of the classifier. Implementation of method shows that we can obtain accuracy up to 92.4% and 98.6% respectively for rank 1 and rank 2 classifier with a sampling period of 2s.