Energy-Efficient GaN-Accelerated Stream Processing Framework for Smart City IoT Applications

Akey Sungheetha; R. Rajesh Sharma; John Blake; Sheila Mahapatra; Nilanga Abeysinghe; Komal Parashar

doi:10.1007/978-3-032-14038-8_30

Energy-Efficient GaN-Accelerated Stream Processing Framework for Smart City IoT Applications

Akey Sungheetha^*
, R. Rajesh Sharma
, John Blake
, Sheila Mahapatra
, Nilanga Abeysinghe
, Komal Parashar

^*Corresponding author for this work

Alliance University
SLIIT
Chitkara University, Punjab

Research output: Chapter in Book/Published conference output › Conference publication

Abstract

This paper presents a novel energy-efficient framework integrating Gallium Nitride (GaN) hardware accelerators with real-time stream processing for sustainable smart city IoT applications. The proposed system combines high-speed modulation characteristics of scaled GaN laser diodes operating at 2.4 GHz with reconfigurable Multiple-In Multiple-Out (MIMO) antenna arrays to enable efficient processing of intensive data streams from urban IoT sensors. The hybrid architecture leverages both edge and cloud computing paradigms, achieving statistically significant improvements of 47.7% in energy efficiency (95% CI: 44.2–51.3%, p<0.001) and 68.0% latency reduction (95% CI: 65.1–70.9%, p<0.001) compared to traditional approaches. The system incorporates resonant-cavity light-emitting diode technology for high-bandwidth data transmission and employs machine learning-based adaptive stream processing algorithms optimized for urban infrastructure monitoring. Experimental validation across 14 international deployments in Singapore, Barcelona, Toronto, and Dubai demonstrates consistent performance improvements while maintaining 99.8% system reliability and processing throughput of 8.7 Gbps.

Original language	English
Title of host publication	Computer Vision and Robotics
Subtitle of host publication	Proceedings of CVR 2025, Volume 3
Editors	Harish Sharma, Abhishek Bhatt, Chirag Modi, Andries Engelbrecht
Pages	381-394
Number of pages	14
ISBN (Electronic)	9783032140388
DOIs	https://doi.org/10.1007/978-3-032-14038-8_30
Publication status	E-pub ahead of print - 31 Jan 2026
Event	5th International Conference on Computer Vision and Robotics, CVR 2025 - Goa, India Duration: 25 Apr 2025 → 26 Apr 2025

Publication series

Name	Lecture Notes in Networks and Systems (LNNS)
Volume	1770
ISSN (Print)	2367-3370
ISSN (Electronic)	2367-3389

Conference

Conference	5th International Conference on Computer Vision and Robotics, CVR 2025
Country/Territory	India
City	Goa
Period	25/04/25 → 26/04/25

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

SDG 7 Affordable and Clean Energy
SDG 11 Sustainable Cities and Communities

Keywords

Edge Computing
Energy Efficiency
GaN Technology
IoT Analytics
Smart Cities
Stream Processing

Access to Document

10.1007/978-3-032-14038-8_30

Cite this

Sungheetha, A., Rajesh Sharma, R., Blake, J., Mahapatra, S., Abeysinghe, N., & Parashar, K. (2026). Energy-Efficient GaN-Accelerated Stream Processing Framework for Smart City IoT Applications. In H. Sharma, A. Bhatt, C. Modi, & A. Engelbrecht (Eds.), Computer Vision and Robotics: Proceedings of CVR 2025, Volume 3 (pp. 381-394). (Lecture Notes in Networks and Systems (LNNS); Vol. 1770). Advance online publication. https://doi.org/10.1007/978-3-032-14038-8_30

@inproceedings{ae5e519c4df14c728cf87b9749e18dbc,

title = "Energy-Efficient GaN-Accelerated Stream Processing Framework for Smart City IoT Applications",

abstract = "This paper presents a novel energy-efficient framework integrating Gallium Nitride (GaN) hardware accelerators with real-time stream processing for sustainable smart city IoT applications. The proposed system combines high-speed modulation characteristics of scaled GaN laser diodes operating at 2.4 GHz with reconfigurable Multiple-In Multiple-Out (MIMO) antenna arrays to enable efficient processing of intensive data streams from urban IoT sensors. The hybrid architecture leverages both edge and cloud computing paradigms, achieving statistically significant improvements of 47.7\% in energy efficiency (95\% CI: 44.2–51.3\%, p<0.001) and 68.0\% latency reduction (95\% CI: 65.1–70.9\%, p<0.001) compared to traditional approaches. The system incorporates resonant-cavity light-emitting diode technology for high-bandwidth data transmission and employs machine learning-based adaptive stream processing algorithms optimized for urban infrastructure monitoring. Experimental validation across 14 international deployments in Singapore, Barcelona, Toronto, and Dubai demonstrates consistent performance improvements while maintaining 99.8\% system reliability and processing throughput of 8.7 Gbps.",

keywords = "Edge Computing, Energy Efficiency, GaN Technology, IoT Analytics, Smart Cities, Stream Processing",

author = "Akey Sungheetha and \{Rajesh Sharma\}, R. and John Blake and Sheila Mahapatra and Nilanga Abeysinghe and Komal Parashar",

year = "2026",

month = jan,

day = "31",

doi = "10.1007/978-3-032-14038-8\_30",

language = "English",

isbn = "9783032140371",

series = "Lecture Notes in Networks and Systems (LNNS)",

pages = "381--394",

editor = "Harish Sharma and Abhishek Bhatt and Chirag Modi and Andries Engelbrecht",

booktitle = "Computer Vision and Robotics",

note = "5th International Conference on Computer Vision and Robotics, CVR 2025 ; Conference date: 25-04-2025 Through 26-04-2025",

}

Sungheetha, A, Rajesh Sharma, R, Blake, J, Mahapatra, S, Abeysinghe, N & Parashar, K 2026, Energy-Efficient GaN-Accelerated Stream Processing Framework for Smart City IoT Applications. in H Sharma, A Bhatt, C Modi & A Engelbrecht (eds), Computer Vision and Robotics: Proceedings of CVR 2025, Volume 3. Lecture Notes in Networks and Systems (LNNS), vol. 1770, pp. 381-394, 5th International Conference on Computer Vision and Robotics, CVR 2025, Goa, India, 25/04/25. https://doi.org/10.1007/978-3-032-14038-8_30

Energy-Efficient GaN-Accelerated Stream Processing Framework for Smart City IoT Applications. / Sungheetha, Akey; Rajesh Sharma, R.; Blake, John et al.
Computer Vision and Robotics: Proceedings of CVR 2025, Volume 3. ed. / Harish Sharma; Abhishek Bhatt; Chirag Modi; Andries Engelbrecht. 2026. p. 381-394 (Lecture Notes in Networks and Systems (LNNS); Vol. 1770).

Research output: Chapter in Book/Published conference output › Conference publication

TY - GEN

T1 - Energy-Efficient GaN-Accelerated Stream Processing Framework for Smart City IoT Applications

AU - Sungheetha, Akey

AU - Rajesh Sharma, R.

AU - Blake, John

AU - Mahapatra, Sheila

AU - Abeysinghe, Nilanga

AU - Parashar, Komal

PY - 2026/1/31

Y1 - 2026/1/31

N2 - This paper presents a novel energy-efficient framework integrating Gallium Nitride (GaN) hardware accelerators with real-time stream processing for sustainable smart city IoT applications. The proposed system combines high-speed modulation characteristics of scaled GaN laser diodes operating at 2.4 GHz with reconfigurable Multiple-In Multiple-Out (MIMO) antenna arrays to enable efficient processing of intensive data streams from urban IoT sensors. The hybrid architecture leverages both edge and cloud computing paradigms, achieving statistically significant improvements of 47.7% in energy efficiency (95% CI: 44.2–51.3%, p<0.001) and 68.0% latency reduction (95% CI: 65.1–70.9%, p<0.001) compared to traditional approaches. The system incorporates resonant-cavity light-emitting diode technology for high-bandwidth data transmission and employs machine learning-based adaptive stream processing algorithms optimized for urban infrastructure monitoring. Experimental validation across 14 international deployments in Singapore, Barcelona, Toronto, and Dubai demonstrates consistent performance improvements while maintaining 99.8% system reliability and processing throughput of 8.7 Gbps.

AB - This paper presents a novel energy-efficient framework integrating Gallium Nitride (GaN) hardware accelerators with real-time stream processing for sustainable smart city IoT applications. The proposed system combines high-speed modulation characteristics of scaled GaN laser diodes operating at 2.4 GHz with reconfigurable Multiple-In Multiple-Out (MIMO) antenna arrays to enable efficient processing of intensive data streams from urban IoT sensors. The hybrid architecture leverages both edge and cloud computing paradigms, achieving statistically significant improvements of 47.7% in energy efficiency (95% CI: 44.2–51.3%, p<0.001) and 68.0% latency reduction (95% CI: 65.1–70.9%, p<0.001) compared to traditional approaches. The system incorporates resonant-cavity light-emitting diode technology for high-bandwidth data transmission and employs machine learning-based adaptive stream processing algorithms optimized for urban infrastructure monitoring. Experimental validation across 14 international deployments in Singapore, Barcelona, Toronto, and Dubai demonstrates consistent performance improvements while maintaining 99.8% system reliability and processing throughput of 8.7 Gbps.

KW - Edge Computing

KW - Energy Efficiency

KW - GaN Technology

KW - IoT Analytics

KW - Smart Cities

KW - Stream Processing

UR - https://link.springer.com/chapter/10.1007/978-3-032-14038-8_30

UR - https://www.scopus.com/pages/publications/105029426674

U2 - 10.1007/978-3-032-14038-8_30

DO - 10.1007/978-3-032-14038-8_30

M3 - Conference publication

AN - SCOPUS:105029426674

SN - 9783032140371

T3 - Lecture Notes in Networks and Systems (LNNS)

SP - 381

EP - 394

BT - Computer Vision and Robotics

A2 - Sharma, Harish

A2 - Bhatt, Abhishek

A2 - Modi, Chirag

A2 - Engelbrecht, Andries

T2 - 5th International Conference on Computer Vision and Robotics, CVR 2025

Y2 - 25 April 2025 through 26 April 2025

ER -

Sungheetha A, Rajesh Sharma R, Blake J, Mahapatra S, Abeysinghe N, Parashar K. Energy-Efficient GaN-Accelerated Stream Processing Framework for Smart City IoT Applications. In Sharma H, Bhatt A, Modi C, Engelbrecht A, editors, Computer Vision and Robotics: Proceedings of CVR 2025, Volume 3. 2026. p. 381-394. (Lecture Notes in Networks and Systems (LNNS)). Epub 2026 Jan 31. doi: 10.1007/978-3-032-14038-8_30