To Short-range Millimeter-wave Radar Target Recognition

As the high-resolution radar range as a much more refined target structure and shape information, therefore, based on a dimensional range of radar target identification has great theoretical and practical significance. This paper studys target identification for a typical tank. In the paper,there are:Firstly,it introduces the millimeter wave stepped-frequency pulse radar working principle,introduced the definition of the radar signal and spectrum analysis, discusses stepping frequency radar waveform parameter selection principles, and designes consistent with this requirement of the radar waveform.Secondly.it introduces the objectives of multiple scattering center model, and then the scattering points in a typical model simulation tanks out of the tank's multi-angular one-dimensional range profile, also analyzes the characteristics of high-resolution radar ground clutter, simulating the environment of a particular grass clutter, and wavelet analysis for radar clutter suppression, get good simulation results.Thirdly, the range profiles of target is decomposed using a wavelet to extract characteristic information under noise-free condition based on wavelet transform and neural network, then the characteristic information of target will be used as feature template for training neural network In the case of superposition of clutter and noise, through pretreatment (to clutter, noise reduction), the characteristic information of targe is extracted from the target of one-dimensional target as information for target identification.Finally, according to the aforementioned theory and method, hardware circuit was designed based on DSP chip (TMS320F2812).