雷达终端高精度快速浮点算法研究

High Accuracy Fast Floating-point Algorithm on Radar Terminal System

针对雷达终端处理过程中,浮点运算在现场可编程门阵列(FPGA)中实现较为耗时、资源占用较高的问题,首先在分析对比当前主要浮点处理方法的基础上,选择CORDIC算法进行浮点处理,并对算法原理进行了详细的介绍,通过仿真,分析了位宽及迭代次数对精度的影响。然后,以arctan函数为例详细介绍了其在FPGA中实现的方法,结果显示arctan流水运算速度可达200 MHz,资源占用率较低、实时性强、精度较高且易于调整。

In radar terminal system, floating-point operation when implemented in FPGA is time-consuming and with high resource u- tilization. After comparing current floating-point processing methods, CORDIC algorithm with high processing performance is select- ed. The theory of CORDIC algorithm is introduced in depth and the influence of bit wide and the iteration numbers to the accuracy is discussed by simulating. Finally, the implementation of aretan function in FPGA as an example is presented in detail. The results show that the speed is up to 210 MHz, the resource utilization is lower and the accuracy is higher which is also easy adjusted.