A Real-Time Convolution Algorithm and Architecture with Applications in SAR Processing
FYSISCH EN ELEKTRONISCH LAB TNO THE HAGUE (NETHERLANDS)
Pagination or Media Count:
In this report we propose an algorithm that maps a large one- dimensional linear convolution problem onto a small two-dimensional linear convolution problem. We show that this property can be derived from a signal flow representation of a sectioned convolution. The introduction of short length FFT processor elements into the signal flow graph results in a structure that can be efficiently implemented in a dedicated hardware architecture. Although the performance of the architecture is not optimal in terms of computational complexity, the performance in terms of data rate, input data sequence lengths and amount of required hardware is high. The architecture design is illustrated with a configuration suitable for real-time on-board airborne synthetic aperture radar SAR processing for the Phased Array Universal Sar PHARUS, which is currently under development at TNO-FEL.
- Computer Programming and Software
- Active and Passive Radar Detection and Equipment