Compressive Sensing Based Fingerprinter Positioning with MCA-based Pre-processing

The sparse nature of location finding in the spatial domain makes it possible to exploit the compressive sensing (CS) theory for wireless location. CS-based location algorithm can largely reduce the number of measurements while achieve a high level of localization accuracy, which makes the CS-based solution very attractive for indoor positioning. In this paper, a novel CS-based fingerprinting location algorithm with minor component analysis (MCA) is proposed by us. MCA theory is firstly introduced into CS to solve the coherence of Received Signal Strength (RSS) measurements in wireless location scenario. MCA-based pre-processing can better satisfy the restricted isometry property (RIP) condition by finding the minor components of RSS measurements. Analytical studies and simulations are provided to indicate that the proposed novel method using MAC significantly outperforms that with orthogonalization pre-processing, and also has lower complexity.