A Methodology to Evaluate Image Target Tracking Algorithms' Implementation in Different Hardware Architectures

A Methodology to Evaluate Image Target Tracking Algorithms' Implementation in Different Hardware Architectures

Store separation flight tests are fundamental to ensure the safe separation between the store and the aircraft. However, this activity is considered high-risk, high-cost, and can take days and a large number of devices to be executed. Therefore, any reduction of time or equipment brings advantages i...

Full description

Saved in:
Bibliographic Details
Published in:IEEE/AIAA Digital Avionics Systems Conference pp. 1 - 8
Main Authors: de Paulo, Rodrigo Carvalho, Duarte, Gabriel C., Leite, Nelson Paiva Oliveira, Loubach, Denis S.
Format: Conference Proceeding
Language:English
Published: IEEE 29-09-2024
Subjects:
Avionics systems
Computer architecture
Cyber-physical systems
Embedded systems
Flight tests
Graphics processing units
Hardware
Hardware architectures
Object tracking
Performance evaluation
Proposals
Real-time systems
Store Separation Trajectory
Target tracking
Tensors
Tracking algorithms
Trajectory
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Store separation flight tests are fundamental to ensure the safe separation between the store and the aircraft. However, this activity is considered high-risk, high-cost, and can take days and a large number of devices to be executed. Therefore, any reduction of time or equipment brings advantages in terms of both efficiency and cost. This paper proposes a methodology to evaluate the performance of different tracking algorithms in fixed hardware architectures, such as CPU and GPU, as well as in reconfigurable hardware architectures, i.e. FPGA. Our methodology takes into account important features such as computer performance metrics, object tracking precision and effectiveness, and flight test environment. Our proposal allows a comparison between architecture performance and effectiveness while executing the tracking algorithms with real-time constraints.
ISSN:2155-7209
DOI:10.1109/DASC62030.2024.10749072