A Framework for Accelerating Radar Simulation on Heterogeneous Parallel Platforms
In the context of aircrew training on flight simulators, the simulation of the radar system requires a large amount of calculation which needs to be performed in real-time. Mapping the virtual environment (landmass, vegetation, buildings, roads, etc.), modeling energy propagation and performing the post-processing are among the most intensive tasks. Most of these are currently done on a computers Central Processing Unit (CPU) in a sequential fashion. The Graphics Processing Unit (GPU) as the core of the modern graphic cards offers high computation capability usable not only for visualisation tasks but also to perform massively parallelized mathematical operations. By leveraging additional computational power from such GPUs in combination with multi-core CPUs, we intend to increase the performance and the fidelity of radar simulation for surveillance and high-resolution modes.