Date of Award
12-2023
Document Type
Thesis
Degree Name
Master of Science (MS)
Department
Electrical Engineering and Computer Science
First Advisor
Naveed Mahmud, Ph.D.
Second Advisor
Ryan T. White, Ph.D.
Third Advisor
Samuel P. Kozaitis, Ph.D.
Fourth Advisor
Brian Lail, Ph.D.
Abstract
Running computer vision algorithms requires complex devices with lots of computing power, these types of devices are not well suited for space deployment. The harsh radiation environment and limited power budgets have hindered the ability of running advanced computer vision algorithms in space. This problem makes running an on-orbit servicing detection algorithm very difficult. This work proposes using a low powered FPGA to accelerate the computer vision algorithms that enable satellite component feature extraction. This work uses AMD/Xilinx’s Zynq SoC and DPU IP to run model inference. Experiments in this work centered around improving model post processing by creating implementations in Python, C++ and FPGA Fabric. The FPGA Fabric implementation included testing a synchronous and an asynchronous implementation. The final implementation shows that a model trained for satellite component feature extraction can run at 9 frames per second while running with an average power consumption of less than 9 Watts.
Recommended Citation
Ekblad, Andrew, "Accelerating Machine Learning Inference for Satellite Component Feature Extraction Using FPGAs." (2023). Theses and Dissertations. 1396.
https://repository.fit.edu/etd/1396
Included in
Artificial Intelligence and Robotics Commons, Digital Circuits Commons, Robotics Commons, Signal Processing Commons, Space Vehicles Commons, VLSI and Circuits, Embedded and Hardware Systems Commons