Baseband Psk, Qam Signal Modulation Subclass Of Automatic Identification Study

In recent years, MPSK and MQAM modulation has been widely used in the field of wireless communications such as radio, satellite and cellular mobile communication, especially BPSK, UQPSK, QPSK, OQPSK, 8PSK and pi/4-QPSK modulation types are frequently used in satellite communication. However, at this stage, there is still no unified system to identify these two sub-set modulation, and many existing modulation recognition algorithms have not taken the impact brought about by shape-filtering in actual communication into account. This dissertation is formed around the algorithm and design of modulation recognition scheme for PSK, QAM sub-set signals. Main content is as follows:(1) The basic characteristics of PSK and QAM sub-set signals have been analyzed based on the constellation diagram. Two clustering methods-subsraction and mountain clustering, which are often used in modulation classification techniques based on constellation diagram, have been simulated and compared. A method to amend the threshold of substraction cluster after differential treatment for two groups of signals (OQPSK and QPSK, 8PSK and pi/4-QPSK) with the same constellation diagram has been proposed and proved by simulation.(2) Four typical SNR estimation algorithms have been analyzed and compared through simulation respectively from the aspects such as affects of data length, range of SNR estimated and shape-filtering on the algorithm and signal types which the algorithm is suited for. The simulation results show that, SNR estimation algorithm based on auto-correlation matrix eigendecomposition has a stable performance, and does not require any prior knowledge. In the modulation type unknown circumstances, it can properly estimate SNR, so as to prepare for the following modulation recognition.(3) Theoretical high-order cumulants'values of PSK, QAM signals are deduced, a modulation classification method has been proposed to recognize 16 square QAM, M-ary star QAMs ( M = 8,16,32), MPSK ( M = 2,4,8), OQPSK, UQPSK and pi/4-QPSK signals. Simulation results show that the method doesn't need SNR as prior knowledge, and can reach high classification accuracy. (4) The impact of shape-filtering on high-order cumulants algorithm has been discussed, and a modulation classification method which can be applied to PSK, QAM modulation type after shape-filtering was summed. Then, the impact of the input data, roll-off factor to the average correct recognition rate (CRR) was analyzed through simulation. Results demonstrate that when roll-off factorsβ< 0.5the method can reach high correct recognition rate at low SNR.(5) The modulation classification algorithms based on the maximum likelihood for BPSK and UQPSK signals, relative phase difference for 8PSK and pi/4-QPSK signals and detecting offset timing for OQPSK and QPSK signals have been analyzed. Then, the impact of shape-filtering and signal to noise ratio estimation error on algorithm based on maximum likelihood was dicussed. Finally, A Comprehensive modulation identification scheme for BPSK, UQPSK, QPSK, OQPSK, 8PSK and pi/4-QPSK signals was designed. The simulation results show that the scheme can classify these signals in a large SNR range.