基于龙格-库塔的弹道微分方程解算的FPGA实现

The FPGA Design for the Solution of Ballistic Differential Equation Based on Runge-Kutta

摘要:

针对火控计算中求解弹道微分方程组的速度、精确度以及资源配置间的冲突，设计了一种在FPGA硬件平台上实现、基于龙格-库塔的微分方程组求解器。通过在方程组之间以及方程内的计算步骤中采用多通道二级并行的方式实现算法硬件加速。实验结果表明，对于精度要求高、计算复杂度大的微分方程组，FPGA硬件实现的运算性能较传统的软件实现方式加速性能明显。

Abstract:

In order to solve the computing problem of the ballistic differential equations: the conflict between the calculating speed, accuracy and the resource consumption, the design put forwards a kind of FPGA hardware differential equation solver based on Runge-Kutta. The design realizes the hardware acceleration by using parallel processing to deal with the data pre-processing and the numerical calculation. It also use dual channel to calculate the internal function result. The experiment shows that, for the equation of higher accuracy and complexity, the FPGA hardware implementation has a better accelerating performance than traditional software implementation.