One of the bottleneck problems of software radar is real-time signal processing.In order to improve the real-time performance of software radar signal processing,take advantage of graphics processors unit(GPU) computation ability for radar signal processing hardware acceleration.This paper designs implementation strategy of radar signal processing in CPU-GPU system and optimizes radar signal processing algorithm on GPU parallel computing characteristics.As the results show,contrast with the CPU computing platform,GPU computation can achieve more than 20 times speedup,and can execute radar signal processing of the entire process in real-time,reflect the good engineering value and application prospect.%软件雷达实现的瓶颈问题之一是信号的实时处理.为提高软件雷达信号处理的实时性,利用图形处理器(GPU)的并行运算能力进行雷达信号处理的硬件加速.设计雷达信号处理在CPU-GPU系统中的执行策略.针对GPU并行计算特点对雷达信号处理算法进行优化.实验结果表明,通过与同期中央处理器(CPU)运算平台比较,GPU运算可实现20倍以上的加速比,并且可以实时完成雷达信号处理的整个流程,体现出良好的工程价值与应用前景.
展开▼
