Study Of An Approach For Indoor Positioning By RFID Based On RSSI

Positioning technologies can be divided into outdoor location and indoor location according to the circumstances of the target. Currently satellite positioning is the major approach for outdoor location. GPS has been most widely used for its high accuracy and high real-time quality. Nevertheless, satellite positioning has been not able to meet the requirements for indoor positioning owing to the complexity of indoor electromagnetic environment and the shielding of buildings against signal, which straightly leads to a research enthusiasm of indoor location. Nowadays the primary techniques for indoor positioning include ultrasound, Zigbee, infrared ray and RFID, etc. And RFID has been drawing people’s wider attention with its advantage of lower cost and communication without contact or insight propagation. LANDMARC takes the values of RSSI (Received Signal Strength Indicator) on readers as the representation of distance, and creatively adopts reference electro-tags to assist positioning by active electro-tags on the idea of neighbors, overcoming electromagnetic interference and multipath effect. However, the accuracy can’t fully satisfy the needs of application. To solve the problem of a lack of precision with LANDMARC, this paper has done some probe as following:1. First, we do some simulation of T nearest neighbor algorithm and add coefficient-distance noise using the model of large-scale-path logarithm signal degradation. Then make experiments on the platform of CC2430/2431for Zigbee. The results suggest that T nearest neighbor algorithm has a certain anti-jamming capability, and what’s more, it proves the validity of the model of large-scale-path logarithm signal degradation.2. We continually change the number of reference electro-tags, and find that the accuracy merely gets higher in local area as the reference electro-tags become more and more.3. After analysis of geometry, we lay the16reference electro-tags in triangular grids instead of rectangular ones. Experiments have shown that this new layout presents more excellent performance than traditional ones.4. This paper proposes an improved method called Observation Mensuration Deviation-offset (OMD) by choosing several Observation Mensuration test points in the room before hand. And the experimental results manipulated on the platform of CC2430/2431prove that our approach generates a mean error of100tests of5.71%lower than T nearest neighbor algorithm.5. To compare the performance of these positioning algorithms directly, this paper has developed a graphically simulation system for indoor positioning and tracing by RFID based on Windows.