TRACKING HIGH-SPEED PROJECTILES WITH AN EVENT-BASED PIPELINE
Loading...
Authors
Hashimoto, Andrea
Gutierrez, Yolanda E.
Subjects
computer science
neural network
event-based
Dynamic Vision Sensor
STICK
neural network
event-based
Dynamic Vision Sensor
STICK
Advisors
Monaco, Vinnie
Date of Issue
2022-03
Date
Publisher
Monterey, CA; Naval Postgraduate School
Language
Abstract
The potential accuracy and reliability of event-based pipelines in light of their low-resource and lightweight physical demands make them promising candidates for critical systems with strict environmental constraints. The research done for this paper is intended to expand on event-based pipelines as an optimal means of tracking high-speed projectiles in real time. Time intervals between spikes in a neural network can be implemented in such a way that linear mathematical functions are predictable, as shown by Xavier Lagorce and Ryad Benosman's 2015 paper, "STICK: Spike Time Interval Computational Kernel, A Framework for General Purpose Computation using Neuron, Precise Timing, Delays and Synchrony." However, there has yet to be research on predicting non-linear functions with this method. In this work, an event-based sensor is used to gather high-speed projectile data, which is then processed to determine the optimal parameters for the ballistic equations. The specific spiking neural network is designed and integrated for further implementation in STICK. While smaller components of the ballistic functions are still necessary for the complete functionality of a STICK implementation to be applied to trajectories, this work provides proof of concept that the combination of these two technologies has the capability to allow for trajectory tracking without the current operational cost, constraints, and larger scale requirements of other current tracking techniques.
Type
Thesis
Description
Series/Report No
Department
Computer Science (CS)
Computer Science (CS)
Organization
Identifiers
NPS Report Number
Sponsors
National Reconnaissance Office, Virginia, 20151
Funder
Format
Citation
Distribution Statement
Approved for public release. Distribution is unlimited.
Rights
Copyright is reserved by the copyright owner
This publication is a work of the U.S. Government as defined in Title 17, United States Code, Section 101. Copyright protection is not available for this work in the United States.
This publication is a work of the U.S. Government as defined in Title 17, United States Code, Section 101. Copyright protection is not available for this work in the United States.