RSS based indoor localization with limited deployment load

One major bottleneck in the practical implementation of received signal strength (RSS) based indoor localization systems is the extensive deployment load required to construct radio maps through fingerprinting. Several works aimed to employ radio propagation models as alternative to fingerprinting but the different sources of inaccuracies in the generation of these models result in high localization errors. In this paper, we propose an indoor localization scheme that can be directly deployed and employed without building a full radio map of the indoor environment. The proposed scheme employs the information from a radio propagation simulator and limited number of calibration measurements to perform direct localization using manifold alignment. For moving users, we exploit the correlation of their reported observations to improve the localization accuracy. The online performance evaluation shows that our algorithm achieves localization errors in the order of 2.5 to 3 m with as low as 15% - 30 % of the complete fingerprinting load.