Wireless Transmitter Identification Using Multicore Path Network

Abstract

Due to the widespread use of wireless communication network, the criminals easily access the Internet through these distributed points for illegal activities. However the identity of the software layer is easily falsified. If the hardware characteristics can be analyzed from the wireless transmitter, it will greatly improve the accuracy of the wireless transmitters. This paper proposes a framework for identifying wireless transmitters using multicore path network (MPN). The nonlinear power amplifier (PA) model uses volterra series to describe the nonlinear behavior of the wireless transmitters. In the proposed MPN, cyclic spectrum features are extracted as the network input. The framework not only reuses features by residual branch path but also explores new features by dense connection path. Meanwhile the MPN merges different scales through different convolution kernels. Through simulation results, we demonstrate that the proposed scheme can be superior than recent methods and has moderate computational complexity.