Journal
IEEE TRANSACTIONS ON GREEN COMMUNICATIONS AND NETWORKING
Volume 6, Issue 3, Pages 1591-1603Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TGCN.2022.3148293
Keywords
Blockchains; Training; Computational modeling; Smart cities; Metadata; Computer architecture; Smart contracts; Smart cities; green; intelligence networking; non-fungible token (NFT)
Categories
Funding
- Natural Science Foundation of Beijing [4212004, L201002]
- BUPT Excellent Ph.D.
- Students Foundation [CX2021208]
Ask authors/readers for more resources
Advanced communication and AI technologies play a crucial role in the development of green smart cities. Connected and automated vehicles utilize intelligence networking and AI modeling for automatic decision-making, compensating for the lack of individual vehicle experience. However, the energy consumption of model training and intelligence networking is significant. This paper proposes a non-fungible token-based green intelligence networking scheme, using metadata to tokenize and describe intelligence, thereby reducing energy consumption.
Advanced communication and artificial intelligence (AI) technologies facilitate the development of green smart cities. Connected and automated vehicles (CAVs) are crucial components, which are aware of the environment by collecting data from sensors and modeling through AI to realize automatic decision-making. Intelligence networking enables each CAV to train appropriate models locally to learn how to drive in different environments, which can make up for the lack of experience of single-vehicle. However, model training and intelligence networking consume a lot of energy, calling on high-energy-efficient solutions. In this paper, we propose a non-fungible token (NFT)-based green intelligence networking scheme (NGIN) for CAVs in smart cities. We use NFT to tokenize and describe intelligence by metadata, enabling applications to network intelligence efficiently, thereby reducing energy consumption. We present the architecture, modules, and efficient mechanisms of distributed intelligence networking. Moreover, we formulate the core problem as a discrete Markov decision process (MDP) and adopt the quantum-inspired reinforcement learning (QRL) algorithm to solve it. Also, the convergence rate and performance are evaluated. Simulation results demonstrate the effectiveness of the proposed scheme.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available