Research On The Spatial Spectrum Estimation For Wideband Signal

The main task of spatial spectrum estimation is to get the spatial parameters of signals or position of information source, which is also the problem in those fields of radar, telecommunication and sonar, etc. It can theoretically greatly improve estimated angle accuracy, angular resolution and some other parameters' accuracy of spatial signals in system processing bandwidth and has extremely expansive future prospects in those fields.With the development of telecommunication techniques, wideband signals such as frequency hopping signal, spread spectrum signals and chirp signals have more and more application. In addition, some signals are wideband in their nature, such as voice signals and earthquake waves. So it has significant meaning in both theory and practice.It's explored in this paper that wideband signal focusing transform, spatial spectrum estimation algorithms based on array manifold transformation, high resolution spatial spectrum estimation, fourth cumulative value algorithm and wideband Chirp signal spatial spectrum estimation according to the characteristic of large bandwidth-duration product for wideband signals. It's presented the basic rationale of algorithms, their performance and simulation verification for DOA. The main research is focusing on content as follows:1 .It's proposed the combination of the ideas of spatial re-sampling and beam forming which is put into practice by methods of digital filter or FFT. Beam forming spatial spectrum estimation is achieved by those two methods and their performance is also compared by simulation.2.It's proposed a modified rotated signal subspace (RSS) wideband direction finding algorithm which not only reduces the operation of focusing matrix but also improves the precision of DOA and resolution.3.It's proposed a wideband direction finding algorithm based on quasi array manifold transform and a generalized array manifold interpolating wideband signal spatial spectrum estimation algorithm. The effective of algorithms above is proved by simulation experiments.4.It's proposed a wideband signal spatial spectrum estimation algorithm which includes focusing and beam space domain preprocessing and it could not only greatly reduce operation of characteristics structure class methods but also make the threshold of signal to noise ratio better and improve the performance of estimation.5.It's proposed a fourth order cumulant slicing algorithm in the wideband signal spatial spectrum estimation which has the same function in the sense of blind Gaussian but much less operation.6.It's presented a time domain wideband source DOA algorithm which constructs a new time domain data model by computing the continuous wavelet cross spectrum between referenced elements of array and the others. Time frequency correlation matrix is constructed in main energy gathering zone by choosing time frequency points which is instead of traditional array correlation matrix. Then the DOA estimation could be realized by characteristic decomposing. The algorithm is put into practice in both time domain and frequency domain which make the best of the energy gathering property of time frequency distribution. It not only has been realized the separation of non-stationary signal in time-frequency domain, but also improved the signal to noise ratio and the algorithm robustness to noise.