چكيده به لاتين
The fifth generation of mobile networks, boasting high speed, bandwidth, and low latency, facilitates efficient communication and supports billions of devices. Unmanned Aerial Vehicles (UAVs), initially developed for military applications, have now found utility in various sectors, including industry. The ease of mobility of UAVs has expanded their role as relays in fifth generation networks, particularly in environments where physical infrastructure is impractical. UAVs, equipped with sensors, contribute to data collection in the Internet of Things (IoT), creating a nexus for information gathering. The IoT aims to integrate billions of objects, enabling them to sense their surroundings, convey information to machines, and ultimately interact with each other and the environment. However, a significant challenge in the IoT lies in the limited processing power and storage capacity of its devices. Computational offloading methods, such as edge computing, fog computing, and cloud computing, have emerged as viable solutions to address this challenge. Leveraging the mobility of UAVs, they are now being employed as edge servers, marking a novel and expanding application. Nevertheless, using UAVs in the IoT introduces challenges related to security and data integrity. The mobility and limited energy of UAVs amplify technical issues, posing risks of data loss. Blockchain technology has emerged as a solution for secure and distributed data storage. Employing encryption methods and complex calculations, blockchain reduces susceptibility to value manipulation. Recent research has explored the integration of blockchain technology with UAVs, with a focus on computational offloading. However, many studies overlook end-users, especially IoT devices, and the dynamic selection of initial UAVs.
This study aims to propose a blockchain-based fifth-generation UAV-based computational offloading method. Addressing the impact of users’ mobility on transfer costs is crucial. The proposed method utilizes a recurrent neural network to predict IoT mobility and a genetic algorithm for dynamic UAV positioning and decision-making. The calculation of time and energy consumption is then performed. Implemented in a real environment, the evaluation results demonstrate an improvement of 63 percent in time and 66 percent in energy efficiency under varying numbers of IoT devices and UAVs.
