@Article{cmc.2026.087414,
AUTHOR = {Ahmed Mazin Jalal, Muhammad Asshad, Amjed A. Ahmed, Nidal A. Al-Dmour, Mohammad Ahmed Alomari, AbdulGuddoos S. A. Gaid, Omar Almomani, Taher M. Ghazal},
TITLE = {Forward Federated Learning for Intelligent Sensing and Navigation in 6G-Supported Internet of Vehicles},
JOURNAL = {Computers, Materials \& Continua},
VOLUME = {},
YEAR = {},
NUMBER = {},
PAGES = {{pages}},
URL = {http://www.techscience.com/cmc/online/detail/28274},
ISSN = {1546-2226},
ABSTRACT = {The Internet of Vehicles (IoV) is moving towards sixth-generation (6G) communication technologies to enable intelligent transportation to transmit data at ultra-low latency and high speed. Effective vehicle navigation and environmental sensing are key to providing instantaneous decision-making in dynamic driving conditions. This paper introduces a Forward Federated Learning (FFL)-based navigation aid system for 6G-enabled IoVs to provide precise sensing and navigation assistance amid changing environmental conditions. The suggested framework collects environmental data, such as neighboring vehicles, road signs, collision distance, and road conditions, using distributed sensing devices. Perceived data is used to aid navigation and driving with greater accuracy. FFL is used to check consistency across sensing intervals and to ensure effective actuation performance. Moreover, the framework ensures the sensing-to-actuation latency requirement is verified by leveraging the ultra-low latency of 6G communications. The forward learning process facilitates a smooth transfer between vehicles and roadside units (RSUs), enabling effective information exchange in mobility. To minimize computational load, information that is outdated relative to surrounding vehicles, traffic density, and traffic lights is removed from the learning process. The suggested framework, in turn, balances dynamic sensing and navigation processes dynamically via federated learning evaluation and 6G communication facilities. The experimental findings indicate that the developed approach reduces communication delay by 11.242% and computational overhead by 10.84% in the maximum vehicle-speed scenarios, making the method applicable in intelligent vehicular settings.},
DOI = {10.32604/cmc.2026.087414}
}