Efficient collision control and auction-based resource allocation mechanism in dense LoRaWAN network via TCP using DRL technique

• Published in: International journal of information technology (Singapore. Online) 2024-10, Vol.16 (7), p.4039-4057
• Main Authors: Parween, Sultana, Hussain, Syed Zeeshan
• Format: Article
• Language: English
• Subjects: Artificial Intelligence; Computer Imaging; Computer Science; Image Processing and Computer Vision; Machine Learning; Original Research; Pattern Recognition and Graphics; Software Engineering; Vision
• ISSN: 2511-2104; 2511-2112
• DOI: 10.1007/s41870-024-01986-9

The Internet of Things (IoT) is basically about communication, which permits devices and circumstances to gather, share, and act on data. A transmission method called LoRa (Long Range) allows for the long-range transmission of data at a low delivery rate. The problems and the limitations of the IoT conception with a focus on LoRa technology are proposed in this study. This paper presents an analysis of the challenges and constraints associated with the Internet of Things (IoT) concept, with a specific emphasis on LoRa technology. A LoRaWAN (Long Range Wide Area Network) network consists of battery-operated devices that enable communication in both directions and is a form of Low Power Wide Area Network (LPWAN). However, various challenges present in IoT data transmission between sensor nodes due to its transmission rate that leads to high network collision. To overcome such challenges, Transmission Control Protocol (TCP) is used which is a transport layer protocol for efficient data transmission. However, existing work performs collision detection, collision avoidance, and resource allocation with limited parameters, which leads to inefficient collision avoidance. To leverage the existing issues, the proposed work adopts proper collision control and efficient resource allocation in the LoRaWAN network via TCP. In our proposed work, Multi-hop clustering is performed for LoRaWAN nodes and gateways using the Mini Batch K Means Clustering (MBKMC) algorithm to reduce the imbalance load among networks and computational complexities. Collision detection is performed using the Improved Gradient Boosting Decision Tree (IMPGBDT) algorithm. After that, collision avoidance is performed using the Osprey Optimization Algorithm (OOA) based on channel transmission signal strength. The channel is allocated to the demanded resource request using the Deep Q Network (DQN) algorithm. Here, resource allocation is performed based on the Profit Maximization Multi Round Auction (PMMRA) algorithm. We assess the proposed method using a variety of performance criteria.