A New Wide-band Digital Receiver

Wide-band digital receiver is the import part of the new generation of the detection system for radar signal. The high-speed A/D converter is placed near the RF end to cover a wide survey band, then the data rate converter is used to reduce the data rate in the following process of dealing with data and meanwhile to keep the intact information of target signal and finally deal with signals of low speed rate with programmable devices such as Digital Signal Processor (DSP ) Field Programmable gates array(FPGA)etc. Its advantage lies in its capability of receiving, dealing with different signals and using different signals processing methods by changing the software on the programmable devices. However, there is a great challenge that the mismatch between the high-speed data flow of the wide-band A/D converter and the processing capability of the general DSP when considering the realization of wide-band digital receiver. In this dissertation band-pass sampling and efficient digital down conversion are suggested to resolve the gap. Since data rate can only be converted on the base of obtaining the target signal's frequency, In this dissertation how to estimate the signal frequency with certain precision by using quick calculation under the conditions of certain signal to noise rate and short-length data when adopting the data output of the ADC is mainly discussed.In chapter 3, two efficient methods of converting data rate in the high probability digital receivers are introduced. One is the improved sampling method through linear interpolation combined with band-pass filtering based upon the research of the theorems of band-pass sampling while the other is wide-band efficient digital down conversion which put decimation operation before mixing frequency and filtering operation in general down conversion by utilizing polyphase structure.Short data instantaneous frequency estimation is one of the key technologies to realize wide-band digital receiver. Data collected in electronic detecting environment can be presumed as sinusoidal signal in short time term, which can be divided into complex sampling and real sampling according to different way of sampling. Kay's method (phase difference weighting method)is appropriate when dealing with single complex sinusoidal signal but high SNRthreshold is needed. In chapter 4, a detailed analysis of threshold of Kay is given and many methods are proposed to reduce the threshold. Since when Kay's method applies in the case of real signal, quadrature sampling must be made first, and so calculation becomes much more. In this dissertation two quick algorithms to estimate frequency of single real sinusoidal signal with short-length data are proposed, one based on the extension of the Prony's method, and the other on linear fitting. Hardware realization of the two algorithms could be easy and the run time satisfies the need of projects.In chapter 5, a algorithm estimating frequency with cubic interpolation is proposed. In this algorithm, only four data samples after DFT are used to estimate frequency, and the accuracy of estimation can meet the requirement under the short data condition, which have advantages such as the few increasing amount of calculation, dealing with multi-signal, and low requirement of signal to noise rate. Moreover, a further study is conducted using the cubic interpolation to estimate frequency with six samples in this dissertation.In chapter 6, the realizations of fast frequency estimation are studied for two/multi signals. The fast frequency estimation method, dealing with two signals directly, is proposed basing the Prony's method. Two technologies, spectral channelization and spectral analysis with zero crossing that could be both applied to multi-signals, are also introduced. The mentioned above interpolation algorithms can also be applied to estimate the frequencies of the multi-signals, the base of these algorithms is picking up the peak of periodogram. However, if multi-signals arrive at the same time, the peak of signals may not be found easily.