Indian Journal of Science and Technology
Year: 2023, Volume: 16, Issue: 16, Pages: 1230-1240
Impana Appaji1*, P Raviraj2
1Research Scholar, GSSSIETW Affiliated to VTU Belagavi, Mysuru, India and Senior Analyst at
Infosys, Mysuru, India
2Professor at Department of Computer Science and Engineering, GSSSIETW Affiliated to VTU Belagavi, Mysuru, India
Email: [email protected]
Received Date:01 September 2022, Accepted Date:04 April 2023, Published Date:25 April 2023
Objectives: To develop a computational framework which is capable of analyzing the realistic scenario with better decision-making of traffic management using enhanced learning-based model. Methods: A discrete baseline architecture is designed for proposed traffic model considering road network and properties of vehicles. A specific set of logical condition is formulated for constructing assumptions required for studies followed by formulating traffic environment. A reinforcement learning scheme is applied in order to obtain state attributes, action attributes, and reward attributes followed by subjecting all the attribute information to Long-Short Term Memory Attention network. The outcome of the model is inform of decision towards proper vehicular communication. The implementation is carried out by two dataset viz. Hangzou data and New York data. The prime parameter for the evaluation is average travel time while the comparison is carried out with multiple standard dataset. The simulated implementation of the proposed scheme is carried out on Hangzou simulation set up where topology for Internet-of-Vehicle (IoV) is executed on top of it. Findings: The study outcome exhibited significant improvement in average travel time of emergency as well as normal vehicles assessed with respect to various existing dataset of 6x6 uniflow, 6x6 biflow, Newyotk, and Hangzhou. The study outcome also exhibited Newyork to show approximately 85% of reduced travel time compared to 6x6 uniflow, 6x6 biflow, and Hangzou set up. This outcome was also found in agreement with power consumption where Hangzou set up was shown to offer approximately 96% of reduced power consumption in contrast to Newyork set up. Novelty: The proposed study contributes towards yielding a generalized assessment framework for traffic management which is capable of evaluating average travel time and power consumption unlike any existing system in cost effective manner.
Keywords: Vehicular communication System; Internetofthings; Reinforcement Learning; Decision Making; Long Short Term Memory; Graph attention Networks
© 2023 Appaji & Raviraj. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Published By Indian Society for Education and Environment (iSee)
Subscribe now for latest articles and news.