Research On Signal Detection And Parameters Estimation Algorithms For Space Objects Surveillance Using General Survey Radar

The near earth space is a crucial area for the national security as well as an important resource for the economy development. As the degradation of operating environment for the spacecrafts, it is extremely urgent to realize exhaustive space objects surveillance. As the basic instrument for wide-field space objects detection, the general survey type radar plays a very important role in space surveillance system.Some key techniques in signal processing and data processing of the"wide resident beam"radar, which is a kind of general survey radar, are studied in this dissertation. The main studies and the contributions can be summarized as follows.Firstly, the motion characteristics of space objects with respect to the survey station are studied. The analytical expressions of motion characteristics are derived, and the effects caused by the eccentricity are also discussed. For the application of"wide resident beam"radar, the movement characteristics calculation for space objects with known orbit elements is obtained and the model of radar echo signal is founded. From the point of view of output SNR (signal noise ratio) gain, the performance and applicability of FFT (fast Fourier transform), which is a common moving target detection technique, are analyzed.Secondly, the space objects detection and parameters estimation are studied. At the beginning, the principle of LFM (linear frequency modulated) signals detection and parameters estimation based on FrFT (fractional Fourier transform) are summarized. Then, the detection performance and estimation accuracy are analyzed theoretically where LFM signal with WGN (white Gaussian noise) is processed by discrete type DFrFT (discrete fractional Fourier transform). The factors which cause the error in"peak searching"are analyzed when DFrFT is applyed. The results show that the resolution of DFrFT in FrF (fractional Fourier) domain is the dominating one. For solving the finite FrF resolution which is determined by the time-bandwidth product of the signal, a novel approach using DFrFT interpolation is proposed. Using DFrFT interpolation, the method breaks through the FrF resolution and gets higher accuracy of parameters estimation. The results of simulation validate the effectiveness of this approach. The rotation angleαin time-frequency plane affects the"peak searching"as well as the computation complexity, therefore, the rotation range and interval of angleαare also analyzed.Subsequently, the DOA (direction of arrival) estimation of space objects with antenna array is studied. As an implementation of spatial filtering, the spatial matrix filter (SMF) improves the performance of DOA estimation of multiple sources through keeping the signals in spatial passband (sector-of-interest) and attenuating the noise and interference in spatial stopband (outside the sector-of-interest). To eliminate the effects caused by SMF, a novel approach based on the fourth-order cumulant (FOC) of the data snapshots is proposed. As the fourth order cumulant of Gaussian noise is zero, this approach is capable of improving the performance of DOA estimation in lower SNR or CGN (colored Gaussian noise) situations. Meanwhile, by sliding the passband of SMF, the approach makes it possible to estimate DOAs of multiple sources with a number more than the array elements. Moreover, the approach is rewritten in an iterative form to calibrate the channel gain and phase error of antenna array. Results of simulation verify the approach's validity.Finally, the techniques are investigated to solve the problem of direction data association when multi-targets cross the radar FOV (field of view) synchronously. Towards the position determination problem of"wide resident beam"radar, the model for multi-receiver data association is established at first. Then, the expression of the range rate of space objects with respect to the receiver is derived. Under the assumption that the orbit of space objects is circular, the calculation of inclination using the Doppler measurements is derived. Considering the difference of DI (Doppler inclination) and effects caused by false and miss alarm, the cost function for direction data association is founded. This approach just use the measurements acquired by two receivers to achieve direction data association, so there is no need to set up other receivers for acquiring redundant direction data. At last, the effects caused by the eccentricity and the measurement error on DI calculation are analyzed. The feasibility and validity of this approach is validated by the simulation.