Improvement of RFID based location fingerprint technique for indoor environment

In this paper, the improvement of an indoor localization using a fingerprint technique is proposed. The radio frequency identification (RFID) system is utilized. In general, the RFID localization can be categorized into two main types: one is the reader localization in which the reader locations will be estimated and one is the tag localization in which the tag locations will be estimation. The selection of these two types depends on the applications. In this paper, the reader localization is considered because of the lower system cost. For the reader localization, although a large number of tags is used, the tag cost is much cheaper than the reader cost. The passive tags are employed as references attached to the ceiling at known locations and the reader carrying on the vehicle of interest is the target to be localized. The basic principle of the fingerprint technique is to find the location of the target by comparing its signal (or information) pattern with a previously recoded database of known signal (or information)-location data. Therefore, there are two main steps to estimate the target location: (i) construction of database containing the signals (or information) and their corresponding locations, and (ii) estimation of the target location. In this paper, firstly, the detected tags found by the reader at each location of interest are collected and from now on the detected tags are called a fingerprint. Secondly, the location of the reader can be estimated using three proposed methods using the intersection between the detected tags and fingerprints. The effectiveness of each method is verified by experiment data. The best result of location estimation error among three proposed methods is less than 35 cm.