Radio Tomographic Imaging with Input Sensor Location Uncertainty

Abstract

The device-free localization (DFL) technique for target localization and tracking in a wireless sensor network is important in modern research. Radio tomographic imaging (RTI) is a DFL method that is widely used in today’s imagebased localization systems. In the RTI system, spatial loss fields (SLFs) represent the maps that indicate the degree of radio wave attenuation for each spatial location in the WSN due to obstacles. In the real-world RTI model, the data is always perturbed by uncertainty. Therefore, uncertainty in sensor node location leads to uncertainty in the input data of the RTI regression model. To address the sensor location uncertainty problem, this paper proposes a novel stochastic robust approximation (SRA) method for RTI (SRA-RTI). Simulation-based performance analysis shows that the proposed technique is robust against the uncertainty in the sensor node location.