Machine Learning-Based Communication Failure Identification Scheme for Directional Industrial IoT Networks

In industrial systems, massive content, such as high-quality video and a large amount of sensing data, should be exchanged between Industrial Internet of Things (IIoT) devices under strict deadlines. The use of millimeter-wave (mmWave) frequencies of 28 and 60 GHz can satisfy the requirements of IIoT by providing a high data rate. In the mmWave band, it is necessary to use a directional antenna owing to its short wavelength. Consequently, directional links are vulnerable to adverse effects, such as deafness problems, where a communicating node cannot receive signals from other transmitting nodes. To alleviate the deafness problem, in this article, we propose a machine learning-based communication failure identification scheme for reliable device-to-device (D2D) communication in the mmWave band. The proposed scheme determines the type of network failure (deafness/interference) according to the IIoT device's state parameters. Based on the identification scheme, we additionally propose ML-DMAC to improve the throughput and minimize the deafness duration of D2D communication. The performance evaluation shows that the proposed ML-DMAC outperforms existing schemes in aggregate throughput and deafness duration by approximately 31% and 88%, respectively.