Performance evaluation of modified-hybrid radio tomographic imaging for human localization in outdoor environment

Abstract

The accuracy is the main factor in developing the localization system. However, due to the environmental noise and interference, the accuracy of localization has been affected. Since Device-free Localization has a multipath problem, thus Radio Frequency Tomography (RTI) has been introduced. This approach is used to localize the human position. This approach offers great potential in monitoring activities especially in perimeter surveillance application. Conventionally, RTI uses Linear Back Projection algorithm (LBP) to reconstruct the tomographic image. However, this algorithm suffers with the ill-posed problem caused by back-projection and the smearing effect due to the overlapping image. This leads to a low-quality tomographic image projection. To improve the quality tomographic image, several regularization approaches has been introduced by other researchers. However, because the target occupies only a small amount of space compared to the entire area monitored, the resulting image is blurred with noise. Therefore, this paper proposed a Modified Hybrid Radio Tomographic Imaging (HRTI-M). Through this proposed method, the area smeared on the RTI image has been reduced. Moreover, the average error of the reconstruction area has been able to be reduced from more than 3% to less than 1%.