Research On Cascaded Space-time Processing Of Active Sonar

Space-time two-dimensional adaptive processing is a key technology in the field of radar, which use the pulse train of several same waveform(Cycle stationary signal) to realize the stable and full rank covariance matrix estimation by accumulating multiple snapshots(one cycle is one snapshot). In the sonar, such pulse train will take long time caused not only the reverberation time increases, and the echo signal is no longer stationary because of the target movement. Therefore, it is difficult to achieve underwater sonar detection by using pulse train in practice. There is a strong need to explore adaptive detection method of one single pulse signal. Space-time joint processing technique has good performance because of the large degree of freedom of weight vector. But at the same time, it needs high order inverse matrix calculation, which is hard to apply. The studys of dimension algorithm become a trend. This paper takes the research on short data space-time cascade adaptive processing for a given kind of non-stationary signal. The main work is as follows:1. The space detection methods of active sonar has already come to relatively maturity, this paper concentrates mainly on the performance enhancements of time domain detection. Adaptive MVDR processing model is established to detect analytical time series signals(Time Series MVDR, TSMVDR). It keeps the echo signal of unexpected times under the condition of distortionless, and effectively restrain the interference and interference signal by minimizing the output power. The corresponding algorithms(Algorithms of TSMVDR, TSMVDR-A) is presented by constructing analytic signal and approximate disposal through segment disposal. Based on TSMVDR and a MVDR model based on time-domain analytical signals (TAMVDR), the model of space-time cascade adaptive processing is built.2. Compared the performance with the space-time cascade adaptive processing. The results show that:i.When the number of hydrophone and order(time series) is insufficient, the the space-time joint processing has better performance than space-time cascade adaptive processing technique because the freedom of weight vector is flexible. ii.Regardless of the amount of calculation, the two techniques have the approximate performance using enough array elements and orders. The improvement of the gain is not clear as the number of weight vector increases. And it also leads to increasing of calculation. When the amount of calculation is limited, space-time cascade adaptive processing can get better performance in spatial and doppler dimension by using higher order of time. From this point, the space-time cascade processing has more extensive application value. The selection of model and algorithm should consider all kinds of factors in a practical application.3. Experimental verification:i. The simulation and experiment on the lake shows that the space-time cascade processing technique under the condition of high sampling rate is feasible by using single pulse. ii. The waveform designing of good space-time and Doppler detection precision is completed. which has good space-time and Doppler detection precision. the combination of frequency modulated signals with comb spectrum characteristic is the better plan. iii. Investigating the applicability of the technique. The technique requires that pulse envelope is constant, if it doesn’t meet this condition, but amplitude fluctuation is not big enough, the technique is also effective but performance degradation. Even so, it’s performance still is obviously better than conventional processing technique. And it’s also applicable to the convolution channel.