The Internet of Things (IoT) is a futuristic technology that enables smart devices to be networked, creating intelligent ecosystems. Nevertheless, one should design intelligent networks to be autonomous and adaptive, allowing them to handle changing environments. This research paper examines recent mechanisms for enhancing adaptability and autonomy in IoT-based networks. Through the use of artificial intelligence, machine learning, and edge computing techniques, we present a new architecture that enhances decision-making and resource utilisation in real-time IoT networks. Our proposed architecture is based on real-time data processing, adaptive communication, and self-adaptive algorithms with the view of offering scalable and stable networks. Statistics employed here are from real IoT applications executed in real industrial and household settings, say sensor readings (motion, temperature, humidity), network performance metrics (latency, bandwidth), and system performance statistics (response time, energy consumption), with over 500,000 samples collected from 200+ IoT nodes over six months. Our experimental findings, utilising Python and Matplotlib, demonstrate improved network performance, reliability, and resource utilisation compared to conventional methods. The study provides valuable insights into the development of future autonomous IoT networks that can adapt effectively to novel situations and offer seamless integration across diverse environments.
Maheswari et al. (Sat,) studied this question.