Guidance, Navigation, and Control System Simulations via Graphics Processor Unit

摘要

Monte Carlo simulation is crucial in the design, development, and execution of a guided projectile program. Graphics processing units (GPUs) are powerful parallel computing devices that are increasingly being used for general purpose (GP) computing. This technical report details the use of GPUs for Monte Carlo simulations with the goal of aiding the Guidance, Navigation, and Control (GN&C) engineer during the design phase of a guided weapon. This report provides a brief overview of GP GPU computing, a basic six-degree-of- freedom projectile dynamic model, and the implementation of a GPU. Run-time performance comparisons are performed between serial Monte Carlo simulations performed on a central processing unit (CPU) and parallel simulations performed on a GPU. The results show that for large numbers of trajectories, significant runtime reductions are possible for Monte Carlo simulations performed on the GPU in comparison to simulations performed serially on the CPU.