Signal Detection And Identification In Spectrum Sharing Communication System

Spectrum sharing communication system is an important solution to the problem ofthe low utilization rate of the static spectrum assignment policy, the basic idea of whichdetermines that the spectrum sensing is the core technology in this system. It demandsthat the signals in complicated wireless environment can not only be detected, but alsobe identified. Covariance detection algorithm is adopted, and the silence period is addedin the original cognitive system to achieve that. Due to the low performance ofcovariance algorithm in actual detection and the low utilization rate caused by thesilence period, this paper focuses on the high-performance detection based oncyclostationary and the characteristic detection based on PN(Pseudo random Noise)sequence to solve the above two problems, and also describes the implementation ofthese two algorithms based on the cognitive radio hardware platform. This paper alsodescribes spectrum sensing based on higher-order cyclic cumulants.First of all, this paper introduces the second-order cyclostationary and theengineering estimate of the cyclic spectrum density function. Proposal is tabled basedon CSD(Cyclic Spectral Density), and simulated analysis is made for the detectionperformance. The DSP implementation and actual testing verifies that there can be goodperformance in low SNR case.Secondly, as the cyclic spectrum detection in this system is unable to accuratelyidentify DTMB(Digital Television Terrestrial Multimedia Broadcasting) signals in thecondition where cognitive users’ signals exist. The author considers using parallelprocessing ability of FPGA, taking the priority for DTMB signals’ detection usingcorrelation detection based on PN sequence, to realize the function of identifying.Simulation verifies the effectiveness of this correlation detection, and FPGAimplementation of this module is finished.Finally, signal modulation recognition and OFDM sub-carrier estimation based onhigher-order cyclic cumulant is researched. The constructed classification characteristicsare verified to be effective by the simulation on BPSK, QPSK,16QAM three common digital signals. The sub-carrier estimation algorithm, which base on the fourth-ordercyclic cumulants, is also verified to be suitable for the adoption of16QAM modulatedOFDM signal.