Research On The Methods Of Highmaneuvering Weak Targets Detection

Extending integration time can increase signal energy of targets, and improve the radar detection performance to targets. However, with the increasing of long-time coherent integration, the target echo signal will produce range migration, Doppler deviation, diffusion and Doppler ambiguity phenomenons obviously to the high speed and high maneuvering targets. For this type of target, if not compensate range migration and Doppler migration effectively, so targets signal’s energy can be accumulated great. This thesis focuses on the actual problems of Space-based radar to high maneuvering weak targets detection, and introduces several effective long-time target energy coherent accumulation methods, and studies how to super-resolution estimate the speed of different targets in multi-target environment.Firstly, this thesis introduces the Chirp signal and pulse compression principle, and establishes the echo signal model of the high maneuvering weak targets, and analyzes range migration and Doppler migration problems because of the targets making relative movement.Secondly, this thesis introduces the Maximum Likelihood method, by constructing a matching function and design appropriate Doppler modulated frequency search interval and center frequency search interval can effectively eliminate range migration and Doppler migration. Analyzes the method of based on first-order Keystone, it does not need the prior information, can one-time correct the phenomenon of multiple targets range migration, but can’t solve the two range curvatures and Doppler frequency migration problem. Aiming at this problem, introduces the method based on two-order Keystone transform and matched filtering combination, this method can effectively solve the problem of energy diffusion because of targets doing maneuvering movement, and can also effectively do coherent integration to targets energy on occurring Doppler ambiguity.Thirdly, this thesis also introduces the method of two-dimensional frequency domain matched filtering, which makes target echo after pulse compression convert to the range Doppler two-dimensional frequency domain, and do matched filtering processing through constructing the compensation function. This method requiresonly one dimension search for acceleration estimation, the range walk caused by the target’s radial velocity and the range bend and Doppler diffusion caused by radial acceleration can all be eliminated well. Finally, this thesis studies a method of multiple targets speed super resolution estimation of basing on subspace projection, the method do first-order Keystone transform on the distance frequency-azimuth time domain data firstly, thus correct linear envelope walk of target signal, and make use of before and after the spatial smoothing method to solve coherent processing for target distance unit data in the distance time domain and azimuth time domain, then do feature decomposition for the correlation matrix obtained by using before and after spatial smoothing, and make the obtained noise subspace project toward the search steering vector with respect to target speed, finally according to the different peak to make super resolution estimation for the velocity of the target parameters. Simulation results show that this method can make effective estimation for the radial velocity corresponding to multiple targets in the same range unit, so as to improve the capability of Space-based Radar to detect multiple target of the difference of radial velocity of the smaller.