Autonomous target tracking of UAVs based on low-power neural network hardware

Wei Yang; Zhanpeng Jin; Clare Thiem; Bryant Wysocki; Dan Shen; Genshe Chen

doi:10.1117/12.2054049

Autonomous target tracking of UAVs based on low-power neural network hardware

Wei Yang
, Zhanpeng Jin
, Clare Thiem
, Bryant Wysocki
, Dan Shen
, Genshe Chen

Department of Computer Science and Engineering

State University of New York Binghamton University
Air Force Research Laboratory
Intelligent Fusion Technology, Inc.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

5 Scopus citations

Abstract

Detecting and identifying targets in unmanned aerial vehicle (UAV) images and videos have been challenging problems due to various types of image distortion. Moreover, the significantly high processing overhead of existing image/video processing techniques and the limited computing resources available on UAVs force most of the processing tasks to be performed by the ground control station (GCS) in an off-line manner. In order to achieve fast and autonomous target identification on UAVs, it is thus imperative to investigate novel processing paradigms that can fulfill the real-time processing requirements, while fitting the size, weight, and power (SWaP) constrained environment. In this paper, we present a new autonomous target identification approach on UAVs, leveraging the emerging neuromorphic hardware which is capable of massively parallel pattern recognition processing and demands only a limited level of power consumption. A proof-of-concept prototype was developed based on a micro-UAV platform (Parrot AR Drone) and the CogniMemTMneural network chip, for processing the video data acquired from a UAV camera on the y. The aim of this study was to demonstrate the feasibility and potential of incorporating emerging neuromorphic hardware into next-generation UAVs and their superior performance and power advantages towards the real-time, autonomous target tracking.

Original language	English
Title of host publication	Machine Intelligence and Bio-inspired Computation
Subtitle of host publication	Theory and Applications VIII
Publisher	SPIE
ISBN (Print)	9781628410563
DOIs	https://doi.org/10.1117/12.2054049
State	Published - 2014
Event	Machine Intelligence and Bio-inspired Computation: Theory and Applications VIII - Baltimore, MD, United States Duration: May 8 2014 → May 9 2014

Publication series

Name	Proceedings of SPIE - The International Society for Optical Engineering
Volume	9119
ISSN (Print)	0277-786X
ISSN (Electronic)	1996-756X

Conference

Conference	Machine Intelligence and Bio-inspired Computation: Theory and Applications VIII
Country/Territory	United States
City	Baltimore, MD
Period	05/8/14 → 05/9/14

Keywords

Neural network
Neuromorphic hardware
Object tracking
UAVs

Access to Document

10.1117/12.2054049

Cite this

Yang, W., Jin, Z., Thiem, C., Wysocki, B., Shen, D., & Chen, G. (2014). Autonomous target tracking of UAVs based on low-power neural network hardware. In Machine Intelligence and Bio-inspired Computation: Theory and Applications VIII Article 91190P (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 9119). SPIE. https://doi.org/10.1117/12.2054049

@inproceedings{9815b982512048eab507d29481b5a2a5,

title = "Autonomous target tracking of UAVs based on low-power neural network hardware",

abstract = "Detecting and identifying targets in unmanned aerial vehicle (UAV) images and videos have been challenging problems due to various types of image distortion. Moreover, the significantly high processing overhead of existing image/video processing techniques and the limited computing resources available on UAVs force most of the processing tasks to be performed by the ground control station (GCS) in an off-line manner. In order to achieve fast and autonomous target identification on UAVs, it is thus imperative to investigate novel processing paradigms that can fulfill the real-time processing requirements, while fitting the size, weight, and power (SWaP) constrained environment. In this paper, we present a new autonomous target identification approach on UAVs, leveraging the emerging neuromorphic hardware which is capable of massively parallel pattern recognition processing and demands only a limited level of power consumption. A proof-of-concept prototype was developed based on a micro-UAV platform (Parrot AR Drone) and the CogniMemTMneural network chip, for processing the video data acquired from a UAV camera on the y. The aim of this study was to demonstrate the feasibility and potential of incorporating emerging neuromorphic hardware into next-generation UAVs and their superior performance and power advantages towards the real-time, autonomous target tracking.",

keywords = "Neural network, Neuromorphic hardware, Object tracking, UAVs",

author = "Wei Yang and Zhanpeng Jin and Clare Thiem and Bryant Wysocki and Dan Shen and Genshe Chen",

year = "2014",

doi = "10.1117/12.2054049",

language = "English",

isbn = "9781628410563",

series = "Proceedings of SPIE - The International Society for Optical Engineering",

publisher = "SPIE",

booktitle = "Machine Intelligence and Bio-inspired Computation",

address = "United States",

note = "Machine Intelligence and Bio-inspired Computation: Theory and Applications VIII ; Conference date: 08-05-2014 Through 09-05-2014",

}

Yang, W, Jin, Z, Thiem, C, Wysocki, B, Shen, D & Chen, G 2014, Autonomous target tracking of UAVs based on low-power neural network hardware. in Machine Intelligence and Bio-inspired Computation: Theory and Applications VIII., 91190P, Proceedings of SPIE - The International Society for Optical Engineering, vol. 9119, SPIE, Machine Intelligence and Bio-inspired Computation: Theory and Applications VIII, Baltimore, MD, United States, 05/8/14. https://doi.org/10.1117/12.2054049

Autonomous target tracking of UAVs based on low-power neural network hardware. / Yang, Wei; Jin, Zhanpeng; Thiem, Clare et al.
Machine Intelligence and Bio-inspired Computation: Theory and Applications VIII. SPIE, 2014. 91190P (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 9119).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Autonomous target tracking of UAVs based on low-power neural network hardware

AU - Yang, Wei

AU - Jin, Zhanpeng

AU - Thiem, Clare

AU - Wysocki, Bryant

AU - Shen, Dan

AU - Chen, Genshe

PY - 2014

Y1 - 2014

N2 - Detecting and identifying targets in unmanned aerial vehicle (UAV) images and videos have been challenging problems due to various types of image distortion. Moreover, the significantly high processing overhead of existing image/video processing techniques and the limited computing resources available on UAVs force most of the processing tasks to be performed by the ground control station (GCS) in an off-line manner. In order to achieve fast and autonomous target identification on UAVs, it is thus imperative to investigate novel processing paradigms that can fulfill the real-time processing requirements, while fitting the size, weight, and power (SWaP) constrained environment. In this paper, we present a new autonomous target identification approach on UAVs, leveraging the emerging neuromorphic hardware which is capable of massively parallel pattern recognition processing and demands only a limited level of power consumption. A proof-of-concept prototype was developed based on a micro-UAV platform (Parrot AR Drone) and the CogniMemTMneural network chip, for processing the video data acquired from a UAV camera on the y. The aim of this study was to demonstrate the feasibility and potential of incorporating emerging neuromorphic hardware into next-generation UAVs and their superior performance and power advantages towards the real-time, autonomous target tracking.

AB - Detecting and identifying targets in unmanned aerial vehicle (UAV) images and videos have been challenging problems due to various types of image distortion. Moreover, the significantly high processing overhead of existing image/video processing techniques and the limited computing resources available on UAVs force most of the processing tasks to be performed by the ground control station (GCS) in an off-line manner. In order to achieve fast and autonomous target identification on UAVs, it is thus imperative to investigate novel processing paradigms that can fulfill the real-time processing requirements, while fitting the size, weight, and power (SWaP) constrained environment. In this paper, we present a new autonomous target identification approach on UAVs, leveraging the emerging neuromorphic hardware which is capable of massively parallel pattern recognition processing and demands only a limited level of power consumption. A proof-of-concept prototype was developed based on a micro-UAV platform (Parrot AR Drone) and the CogniMemTMneural network chip, for processing the video data acquired from a UAV camera on the y. The aim of this study was to demonstrate the feasibility and potential of incorporating emerging neuromorphic hardware into next-generation UAVs and their superior performance and power advantages towards the real-time, autonomous target tracking.

KW - Neural network

KW - Neuromorphic hardware

KW - Object tracking

KW - UAVs

UR - https://www.scopus.com/pages/publications/84906351375

U2 - 10.1117/12.2054049

DO - 10.1117/12.2054049

M3 - Conference contribution

AN - SCOPUS:84906351375

SN - 9781628410563

T3 - Proceedings of SPIE - The International Society for Optical Engineering

BT - Machine Intelligence and Bio-inspired Computation

PB - SPIE

T2 - Machine Intelligence and Bio-inspired Computation: Theory and Applications VIII

Y2 - 8 May 2014 through 9 May 2014

ER -

Autonomous target tracking of UAVs based on low-power neural network hardware

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this