Authors:
Rinith Pakala
1
;
Niket Kathiriya
1
;
Hossein Haeri
2
;
Satya Maddipatla
3
;
Kshitij Jerath
2
;
Craig Beal
4
;
Sean Brennan
3
and
Cindy Chen
1
Affiliations:
1
Computer Science Department, University of Massachusetts Lowell, 220 Pawtucket St, Lowell, U.S.A.
;
2
Mechanical Engineering Department, University of Massachusetts Lowell, Lowell, U.S.A.
;
3
Mechanical Engineering Department, The Pennsylvania State University, University Park, U.S.A.
;
4
Mechanical Engineering Department, Bucknell University, Lewisburg, U.S.A.
Keyword(s):
Edge Computing, Middleware, Data Repository Server, Safety Critical Transmission.
Abstract:
The development of communication technologies in edge computing has fostered progress across various applications, particularly those involving vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communication. Enhanced infrastructure has improved data transmission network availability, promoting better connectivity and data collection from IoT devices. A notable IoT application is with the Intelligent Transportation System (ITS). IoT technology integration enables ITS to access a variety of data sources, including those pertaining to weather and road conditions. Real-time data on factors like temperature, humidity, precipitation, and friction contribute to improved decision-making models. Traditionally, these models are trained at the cloud level, which can lead to communication and computational delays. However, substantial advancements in cloud-to-edge computing have decreased communication relays and increased computational distribution, resulting in faster response time
s. Despite these benefits, the developments still largely depend on central cloud sources for computation due to restrictions in computational and storage capacity at the edge. This reliance leads to duplicated data transfers between edge servers and cloud application servers. Additionally, edge computing is further complicated by data models predominantly based on data heuristics. In this paper, we propose a system that streamlines edge computing by allowing computation at the edge, thus reducing latency in responding to requests across distributed networks. Our system is also designed to facilitate quick updates of predictions, ensuring vehicles receive more pertinent safety-critical model predictions. We will demonstrate the construction of our system for V2V and V2I applications, incorporating cloud-ware, middleware, and vehicle-ware levels.
(More)