Ambient Backscatter Communication System Empowered by Matching Game and Machine Learning for Enabling Massive IoT Over 6G HetNets
Amany A. Khalifa,Ahmed M. Abd El-Haleem,Mahmoud M. Elmesalawy,Hesham M. Z. Badr
DOI: https://doi.org/10.1109/access.2024.3388720
IF: 3.9
2024-04-24
IEEE Access
Abstract:Ambient backscatter communication (ABC) is considered as a promising paradigm for meeting the 6G massive Internet of Things (IoT) requirements which is expected to revolutionize our world. In this paper, a new multimode matching game and machine learning-based IoT ambient backscatter communication scheme is proposed to maximize the ABC system rate and capacity over the LTE and Wi-Fi multi-RAT heterogeneous network, thereby supporting the 6G green massive IoT communication. The proposed algorithm is designed to support different rate and capacity requirements for different massive Machine Type Communication (mMTC) use cases such as sensor networks, smart grid, agriculture and low data rate Ultra Reliable Low Latency Communication (URLLC) use cases such as tactile interaction. The proposed optimization algorithm runs into two phases, the first one is a matching game-based algorithm that selects the optimum association between the IoT tags and the primary users (PU) downlink signals from a specific base station which maximizes the IoT tags rate while minimizing the resulting interference to the PU. Each IoT tag can ride the PU downlink signal using one of three different riding modes according to the required IoT ABC system rate and capacity, whereas mode 1 allows multiple IoT tags to ride the whole PU downlink signal resource blocks, in mode 2 each IoT tag can ride only one subcarrier from the PU downlink signal resource blocks, while in mode 3 multiple IoT tags can ride the same subcarrier from the PU downlink signal resource blocks. In addition, unmanned aerial vehicles (UAVs) flying HetNodes equipped with LTE and Wi-Fi receivers are used as backscatter receivers to receive the IoT tags uplink backscattered signals, so the second optimization phase is formulated to maximize the total sum rate of the ABC system by dividing its service area into clusters using the enhanced unsupervised k-means algorithm, also the enhanced k-means algorithm finds the optimum location of each cluster's serving UAV flying HetNode that maximizes the channels gain between the IoT tags and the serving UAV flying HetNode in order to maximize the total system rate. The system model was implemented within the MATLAB environment where simulations across the various scenarios are conducted to assess the effectiveness of the proposed algorithm. Simulation results and the performance analysis demonstrated that the proposed algorithm can support the required rate for the most mMTC and low data rate URLLC IoT applications with average IoT tag rates in the range of 15 Kbps to 115 Kbps, and outperforms the algorithm-free riding technique in the case of massive IoT applications. The proposed mode 2 (first enhanced mode) achieves the best performance in terms of the average IoT tags rate and the total system rate with the lowest interference to the primary system users, on the other hand, mode 3 (second enhanced mode) improves the system capacity with maximum IoT tags satisfaction ratio. The capacity and satisfaction ratio of the proposed mode 3 outperforms mode 1 by 300 % and 138% respectively and outperforms mode 2 by 2,000 % and 420 % respectively. The proposed algorithm reduces the interference power to the PUs on the average by relative to the algorithm-free riding technique. From the result, we can conclude that the proposed algorithm supports different IoT applications and achieves the required data rates with minimal effect on the primary system keeping the PU's data rate within the required range compared to the algorithm-free riding technique with the cost of higher time complexity.
computer science, information systems,telecommunications,engineering, electrical & electronic