Live SAR imaging from small UAVs is an emerging field. On-board processing of the radar data requires high-performance and energy-efficient platforms. One candidate for this are Transport Triggered Architecture (TTA) processors. We implement Backprojection and Backprojection Autofocus on a TTA processor specially adapted for this task using OpenCL. The resulting implementation is compared to other platforms in terms of energy efficiency. We find that the TTA is on-par with embedded GPUs and surpasses other OpenCL-based platforms. It is outperformed only by a dedicated FPGA implementation.
Funding Information: Niklas Rother thanks T. Fiedler for his help with the ASIC synthesis. Tampere University authors thank Academy of Finland (decisions #331344 and #353199)
10.15488/15965
%0 Generic
%1 rother2023synthetic
%A Rother, Niklas
%A Mätzner, Leonard
%A Jääskeläinen, Pekka
%A Leppänen, Topi
%A Schleusner, Jens Karsten
%A Blume, Holger Christoph
%B International Radar Conference 2023
%D 2023
%K Energy-efficiency SAR TTA backprojection myown
%R 10.15488/15965
%T Synthetic Aperture Radar Algorithms on Transport Triggered Architecture Processors using OpenCL
%X Live SAR imaging from small UAVs is an emerging field. On-board processing of the radar data requires high-performance and energy-efficient platforms. One candidate for this are Transport Triggered Architecture (TTA) processors. We implement Backprojection and Backprojection Autofocus on a TTA processor specially adapted for this task using OpenCL. The resulting implementation is compared to other platforms in terms of energy efficiency. We find that the TTA is on-par with embedded GPUs and surpasses other OpenCL-based platforms. It is outperformed only by a dedicated FPGA implementation.
%@ 9781665482783
@conference{rother2023synthetic,
abstract = {Live SAR imaging from small UAVs is an emerging field. On-board processing of the radar data requires high-performance and energy-efficient platforms. One candidate for this are Transport Triggered Architecture (TTA) processors. We implement Backprojection and Backprojection Autofocus on a TTA processor specially adapted for this task using OpenCL. The resulting implementation is compared to other platforms in terms of energy efficiency. We find that the TTA is on-par with embedded GPUs and surpasses other OpenCL-based platforms. It is outperformed only by a dedicated FPGA implementation.},
added-at = {2024-02-05T16:20:37.000+0100},
author = {Rother, Niklas and Mätzner, Leonard and Jääskeläinen, Pekka and Leppänen, Topi and Schleusner, Jens Karsten and Blume, Holger Christoph},
biburl = {https://www.bibsonomy.org/bibtex/2efaa6b8b769ca8f3b20e6fd70af18d58/fabcho},
booktitle = {International Radar Conference 2023},
comment = {Funding Information: Niklas Rother thanks T. Fiedler for his help with the ASIC synthesis. Tampere University authors thank Academy of Finland (decisions #331344 and #353199)
10.15488/15965},
doi = {10.15488/15965},
interhash = {a2f4d2e881af78ca0d12f5fb2a3d1ae3},
intrahash = {efaa6b8b769ca8f3b20e6fd70af18d58},
isbn = {9781665482783},
keywords = {Energy-efficiency SAR TTA backprojection myown},
month = dec,
timestamp = {2024-03-05T15:39:40.000+0100},
title = {Synthetic Aperture Radar Algorithms on Transport Triggered Architecture Processors using OpenCL},
year = 2023
}