Font Size: a A A

Design And Realization Of Automatic Modulation Recognition Of Digital Communication Signals And All-Digital Demodulation

Posted on:2011-08-22Degree:MasterType:Thesis
Country:ChinaCandidate:T Q WangFull Text:PDF
GTID:2178360305971179Subject:Microelectronics and Solid State Electronics
Abstract/Summary:PDF Full Text Request
The automatic modulation recognition of communication signals and their corresponding modulation parameters estimation are very important in communications, which have important applications both in civil and military. It requires the receiver can recognize the modulation types correctly in the same hardware platform with poor prior knowledge, demodulate the digital modulated signals and get the modulation parameters, such as base-band signal, the frequency rate of containing. In recent years, a lot of ideas and methods have been put forward for modulation recognition. This paper mainly studies the automatic modulation recognition of six kinds of digital modulated signals and all-digital demodulation method.There are a lot of man-made factors in the early modulation recognition due to human intervention, as a result, the recognition results vary, and the species which can be identified are also very limited. In this paper, five parameters based on time series statistical characteristics have been analyzed. It can recognize the modulation type correctly of 2ASK, 4ASK,2FSK,4FSK,2PSK and 4PSK through doing some statistics of the signals' instantaneous amplitude, instantaneous frequency and instantaneous phase. By the use of Matlab programming, a variety of modulated signals data sample are generated, then their characteristic parameters are extracted to make a sample set, completing the classifier's training and testing.Considering of the merits and demerits of various classifiers comprehensively, this paper designed a layered network structure of MLP neural network classifier, which uses the BP algorithm, changes the decision threshold for each classification for all the characteristic parameters adaptively, making the success rate of recognition of the system improved significantly. The Matlab simulation has proved that the recognition rate of 2ASK is 91.67%, 4ASK 98.33%,2FSK 95.00%,4FSK 96.67%,2PSK and 4PSK 100%, the overall recognition rate of six kinds of modulated signals is 96.94%.Traditional demodulation system is single demodulation system for a particular modulation and specific bandwidth, so its application range is limited. In this paper, the six all-digital non-coherent demodulators has been designed considering the characteristics of the modulated signals under noise-free. Those demodulators could demodulate the signal, restore the base-band signal and the carrier frequency quickly according to the results of modulation recognition. The entire design uses Verilog HDL programming based on Quartusâ…ˇplatform,Finally, PCB design is achieved by the use of Altium Designer, and the A/D data acquisition module and the FPGA module is realized as well. Combined with laboratory DSP evaluation board, a software radio hardware platform is built, verifying the correctness and feasibility of modulation signal automatic modulation recognition and all-digital demodulation.
Keywords/Search Tags:characteristics parameters, neutral network classifier, non-coherent demodulation, software-defined radio
PDF Full Text Request
Related items