Optimized Tree-Structured LoRa Networks for Energy-Efficient Data Aggregation

Employing Convolutional Neural Networks (CNNs) to predict traffic accidents presents a variety of significant benefits that could transform the industry. Their remarkable ability to effectively extract spatial features and automatically learn them greatly boosts hybrid models’ performance, achieving an impressive accuracy rate of 95% when integrated with XGBoost. This allows CNNs to proficiently process and gain insights from complex data, leading to marked improvements in predictive accuracy for traffic accident forecasting and safety management. By consolidating information from multiple sources, CNNs support the development of innovative approaches to accident prevention within a broader traffic management context. Looking ahead, there is considerable potential for further research, which could focus on developing sophisticated CNN architectures and utilizing more pertinent input data from conventional traffic accident prediction sources—ultimately enhancing road safety and traffic management efficiency. In this research, we provide a novel tree-structure-based information forwarding routing technique. To develop energy-efficient adaptive information forwarding in a decentralized manner, we modify wireless sensors in this study to choose better neighbour LoRa nodes.According to the suggested procedure, a wireless sensor first gathers the state and action data of its nearby LoRa node, including the distance and direction of information transmission, the amount of energy left, and the energy used for the information transmission action. For the future generation of IoT, where mesh topology is primarily used in current networks, LoRa technology is a possible alternative. This paper describes a LoRabased tree network and communication protocol that accounts for energy constraints. When employing LoRa technology, the nodes self-configure according to the physical link’s behaviour and forwarding, where mesh nodes are used to gather information when scalable. Analytical, computational, and experimental results show that the suggested approaches achieve energy efficiency and produce notable gains in energy performance. These findings provide helpful starting points for optimizing large-scale IoT networks. According to the experimental results, this approach also enhanced the lifetime of the entire LoRa network and reduced energy consumption in LoRa nodes.