Algorithm-specific Optimizations for On-Board Real-Time Backprojection on FPGA

Abstract:
This paper details a design optimization on a hardware accelerator for an on-board real-time SAR imaging system using the Backprojection algorithm, focusing on algorithm-specific approximations intended to reduce the overhead introduced by intensive functions such as the square root, sine, and cosine. This paper revisits existing approximation methods, such as polynomial approximation and Chebyshev polynomials, and compares their performance on the algorithm. Results demonstrate that the algorithm-specific approximations perform better than traditional methods while using up to 88% fewer LUTs for the trigonometric approximations and up to 94% fewer LUTs for the square root approximation and maintaining the image quality.