A Fast, Multi-Platform, High-Resolution SAR Echo Simulation Method Based on CPU-GPU Architecture

To improve the multi-platform applicability and simulation efficiency of high-resolution synthetic aperture radar (SAR) echo simulations, this paper proposes a fast, multi-platform, high-resolution SAR echo simulation method based on CPU-GPU architecture. First, the simulation process decouples the echo calculation from specific platform parameters by abstracting key modules such as system payload model, position-orientation model, and signal model into unified interfaces based on the SAR echo model, significantly enhancing the generality of the simulation algorithm. Second, the method adopts a “grid-stepping” strategy to design kernel functions for calculating the reflection coefficients of large-scale as well as high-resolution scenes. By using the CPU to manage the worker process pool and simulation task queue, the approach achieves dynamic scheduling of multi-GPU parallel computations, significantly improving GPU utilization and load balancing performance. Simulation results demonstrate that this method effectively simulates high-resolution SAR echo data for different platforms while offering high computational efficiency, ease of maintenance, and flexible scalability.