Fast Simulation For Heterogeneous Clutter Based On GPU

To provide more realistic clutter data for airborne radar system simulation in order to evaluate the performance of radar in a specific scene more accurately, this thesis uses the digital elevation model to simulate the clutter in the coastal area, which can effectively reflect the clutter of the actual terrain. However, a large number of digital elevation data need to be selected to ensure the accuracy, which leads to slow simulation speed. Aiming at the problem of simulation speed the clutter simulation process is optimized: the power of scattering clutter units are calculated by the same steps and a large part of the work can be implemented by using parallel processing. The graphics processing unit can complete parallel processing function to accelerate it. This paper explores the method of modeling airborne radar clutter in coastal area, the process of inhomogeneous clutter simulation based on graphics processing unit, and the simulation results are analyzed and summarized. The final analysis describes how the inhomogeneous clutter affects the digital radar signal processing result of each module.The main content of the thesis is as follows:1. By means of modeling airborne radar clutter model based on digital elevation model in coastal areas, the clutter power spectrum is simulated in these areas. The simulation process includes the methods of using digital elevation model, airborne radar carrier parameters and locations to calculate the scattering unit parameters, and how to use these parameters to simulate power spectrum: judging each scattering unit belongs to which range-Doppler unit, summing the power of scattering clutter unit in the same range-Doppler unit, completing inhomogeneous clutter power spectrum simulation.2. Due to the demand of real-time and the situation of low simulation speed, taking the characteristics of the GPU parallel computing into account, it is useful to combine parallel programming and memory hardware acceleration to clutter simulation for improving the means of simulating the clutter, putting forward ideas to optimize simulation algorithm and implementing fast simulation of inhomogeneous clutter. By comparative analysis and improvement of the scale optimization algorithm, the simulation speed increases significantly after the parallel simulation algorithm.3. The target detection effects in real environment of inhomogeneous clutter is researched in this paper. In addition, the target echo in pulse compression, moving target indication, moving target detection and constant false alarm rate detection are simulated respectively. In the signal processing simulation in line with the premise of target information, comparing the two methods which are based on statistical model and digital elevation model respectively, we found that in specific situations based on digital elevation model inhomogeneous clutter will affect the signal processing system.