The Research On Modulation Signal Identification In Software Radio

Any communication signals must inform the modulation mode and parameter of the signal to receive and demodulation. The aim of this thesis is that it can analyze the modulation types and estimate the modulation parameters of the communication signals from the received signals, Given that any modulation infomation is unknown, which is traveling in multi-signals and noise environment. This is very important for analyzing and processing signal. The importance of the thesis is that given modulation mode and bandwidth of traditional singularity communication system would no longer adapt the request of presently multi-modulates and multi-serves commucation system. With rapid development of communication technology, the system and modulation mannerof communication signals became more and more complicated and various. Modulation identification would be more effect in the development of software radio.The main work in this paper can be summarized as follows:1. This paper introduces the architecture of software radio and related theory, some key technologies of software radio are researched, then dissussed the algorithm of modulation and demodulation.2. The common model of modulation signals is discussed. This part expatiates the frequency spectrum of analog signals AM,PM,FM,DSB,USB and LSB,it shows clearly the characteristic of analog signals. Analyse the instantaneous extent, instantaneous phase and instantaneous frequency of digital signals 2ASK,4ASK,2FSK,4FSK,2PSK and 4PSK. Lay a foundation for the after identification. At last expatiates the main character parameters based on instantaneous characters.3. To avoid the fixed limitation of decision-theoretic, the paper introduce the identification algorithm based on higher-orde cumulants (HOC) of digital signals.the algorithm shows great effect in identifying five different digital modulated signals 2ASK/2PSK,4ASK,4PSK,2FSK and 4FSK. Simulations show that the algorithm is able to avoid the influence of gauss white noise, the character parameters would be used in the backward study in neural network identification algorithm.4. In this paper we propose an automatic modulation recognition system to recognize seven digital signal classes as 2ASK,4ASK,2FSK,4FSK,2PSK,4PSK and 16QAM using decision-theoretic based feature set addition to statistical pattern based feature set with Back Propagation (BP) neural network. In order to verify the performance of the system, we carry out a large amount of emulation experiment from the neural networks. Computer simulation results show that the robustness and practicability of this recognition method.