Search

In the present fast-moving society, the Internet of Things (IoT) is transforming the way services are used in different industries. While it has many benefits, there are also considerable obstacles, especially in the areas of computing power, safety, and handling data. With the continuous evolution and importance of consumer electronics (CE) in smart cities, there is an increasing demand for sustainable and effective solutions to deal with challenges such as widespread sensing, advanced computing, prediction, monitoring, and data sharing. The artificial intelligence (AI) has emerged as a crucial component in the IoT environment, highlighting the need for energy-efficient CE in urban areas. The state of art methods are required to maximize resource usage and maintain high-quality services for smart systems in healthcare, transportation, AI-powered sensing (AIeS), and sustainable networks. The split learning is a technique for distributed deep learning, shows great potential as a solution for these CE applications. It can greatly reduce numerous obstacles linked with intelligent services in smart cities. The split learning enables the training of deep neural networks or split neural networks (SplitNN) using AIeS on various data sources. This method enables the secure and efficient processing of data without the requirement of directly sharing raw labeled data, which is crucial in industries like healthcare, finance, security, and surveillance where data privacy and security are vital. This guest editorial discusses and presents split learning methods in CE applications for smart cities. Using split learning, researchers and developers can develop creative solutions to address resource efficiency, data security, and service quality issues across different smart city sectors as presented further. As the IoT grows and changes, incorporating split learning into CE applications influences the platform for future smart cities.

Wireless sensor networks (WSNs), one of the urban sensing technologies, suffer from issues such as, data security, data offloading, bulk data storage, and network coverage which in turn affects urban sensing. Device-to-device (D2D) communication appears to be the most promising candidate for mitigating the problems faced by WSN. Enabling D2D communication in the mm-wave network enhances throughput and bandwidth. However, it suffers from signal degradation due to higher pathloss attenuation. Thus, the proposed scheme applies dynamic mode selection criteria for efficient D2D communication in a 5G mm-wave cellular network (CN). The radius of coverage of D2D users (D2Ds) is derived for switching of the modes, i.e., from inner to outer. The outer mode switching occurs when the D2Ds suffer from extensive pathloss attenuation or with increase in coverage radius. Furthermore, power optimization is done to obtain optimal D2D transmit power thereby, enhancing the data rate. This problem is solved by employing Lagrangian dual optimization technique. The performance metric, coverage probability is also derived which depicts the superiority of the proposed scheme in terms of quality-of-service (QoS) at a higher pathloss attenuation. Simulation results prove the efficacy of the proposed scheme. Finally, the proposed scheme is compared with the state-of-the-art schemes which further validates the superiority of the proposed scheme.

In the era of Industry 5.0, technology promises the rapid and smart growth of industries to improve the system performance substantially and be sustainable and more secure. Industry 5.0 enabled the conventional unmanned aerial vehicle (UAV) industries to represent a dynamic result that can offer various services irrespective of time, place, and in broader physical range. Hence, satellite systems of high performance having advanced pre-processing transmission capabilities are gaining considerable attention to increase overall system throughput. In enhancing the efficiency of interaction between humans and machines, gesture recognition in UAV-based industry 5.0 can perform a vital role as the transmission medium in end-user based applications with low latency periods. However, few smart consumer personalized gesture recognition (CPGR) systems are insufficient in the privacy preservation in UAV-based industry 5.0 applications. Also, privacy-preserving is time-consuming, especially for model training and real-time CPGR. Although wireless technologies with big-data schemes provide such applications but lack emotional attachments. This article proposes a gesture-aware system with pre-processing connected systems using a powerful emotion detection system by incorporating a conventional analysis-modification-synthesis framework used for industry 5.0 applications. This study demonstrates the effectiveness of different noise estimation techniques on the proposed Modulation domain Recognition (MDSR) approach for UAV-based Industry 5.0.