A time domain eigen value method for robust indoor localization

The hazardous nature of the indoor environment and the rapid growth of commercial indoor positioning systems have placed a significant emphasis on developing robust localization techniques. Our work focuses on developing super resolution techniques that can provide accurate time delay estimates under line-of-sight (LOS) conditions and generation of location information rich fingerprints that can be utilized for localization under non LOS conditions. First a detailed behavioral analysis of the subspace separation based super resolution algorithms is presented. Then we examine the newly introduced time domain eigen-value (TD-EV) method which effectively combines the time domain multiple signal classification (TD-MUSIC) and the frequency domain eigen-value (FD-EV) algorithms. This is done to secure the bandwidth versatility, superior path resolvability, and noise immunity of TD-MUSIC algorithm and FD-EV's ability to resurface underestimated local peaks submerged beneath the noise floor under constrained conditions. This makes TD-EV a prime candidate for accurate time delay estimation under severe multi-path and noise conditions prevalent in indoor environments. Additionally, these attributes provide a location information rich fingerprint from the resultant pseudo-spectrum output of our method for location based fingerprinting techniques.