Characterizing Multiple Wireless Sensor Networks for Large-Scale Radio Tomography

Radio Tomographic Imaging (RTI) is an emerging Device-Free Passive Localization (DFPL) technology that uses a collection of cheap wireless transceivers to form a Wireless Sensor Network (WSN). Unlike device-based active localization, DFPL does not require a target of interest to be wearing any kind of device. The basic concept of RTI utilizes the changes in Received Signal Strength (RSS) between the links of each transceiver to create an attenuation image of the area. This image can then be used for target detection, tracking, and localization. Each transceiver in the WSN must transmit sequentially to prevent collisions. This is not a problem when the number of transceivers in the WSN are small. However, large-scale RTI with a large number of transceivers suffer from high computational complexity, low frame rates, and physical distance limitations on the range of the transceivers. The goal of this research is to determine the applicability and characterize the feasibility of using multiple WSNs to address the limitations with a large-scale RTI network. The concept to this new variant of RTI, called Multiple-Networks RTI (mnRTI), is to divide the transceivers into multiple WSNs as opposed to using one WSN. Analytical, simulated, and experimental data are computed, collected, and compared between a RTI network with one WSN to a mnRTI network with two WSNs. The WSN(s) comprise a total of 70 wireless transceivers covering an area of no more than 19 ft x 16 ft. Simulated and experimental results are presented from a series of stationary and moving target data collection. Preliminary results demonstrate multiple WSNs can potentially provide similar or better results than the traditional RTI method with one WSN. Multiple WSNs have higher frame rates and lower computational complexity. Also, position estimation accuracy are comparable, if not better, than the traditional RTI method with one WSN