AI-Empowered Interference Mitigation Technique for M2M Networks

Abstract

Machine-to-machine, or M2M, communication has reshaped how industries function as a whole. Its advancement has fueled the growth of the Industrial Internet of Things (IIoT), transforming factories into smart, networked environments. M2M enables devices and machines to connect with one another in real-time, resulting in swifter data exchange, realtime surveillance, and optimized operations. M2M communication, on the other hand, relies primarily on wireless networks, which are prone to randomness and uncertainty. These networks are prone to noise and interference, which reduces the efficiency of communication and degrades performance. This may have a major effect on the dependability and resilience of smart industrial systems. Interference mitigation techniques are employed to improve the performance of M2M communication. In this paper, we propose a hybrid AI-empowered interference technique. A virtual setup of the cellular network is simulated in MATLAB. The channel gain data generated from the simulation is fed as a dataset to the clustering algorithm DBSCAN (Density-Based Spatial Clustering of Applications with Noise). Furthermore, the best data cluster is fed as the initial population to the genetic algorithm (GA) to enhance the performance. Through each iteration of the GA process, the solution with the best fitness is chosen as the initial population for the next iteration. Ultimately, we have the population with the best fitness function as the solution. The GA employed in the proposed framework performs well by converging at a throughput of 21.2 bps and gradually decreasing execution time over the generations.