As a popular distributed machine learning framework, wireless federated edge learning (FEEL) can keep original data local, while uploading model training updates to protect privacy and prevent data silos. However, since wireless channels are usually unreliable, there is no guarantee that the model updates uploaded by local devices are correct, thus greatly degrading the performance of the wireless FEEL. Conventional retransmission schemes designed for wireless systems generally aim to maximize the system throughput or minimize the packet error rate, which is not suitable for the FEEL system. A novel retransmission scheme is proposed for the FEEL system to make a tradeoff between model training accuracy and retransmission latency. In the proposed scheme, a retransmission device selection criterion is first designed based on the channel condition, the number of local data, and the importance of model updates. In addition, we design the air interface signaling under this retransmission scheme to facilitate the implementation of the proposed scheme in practical scenarios. Finally, the effectiveness of the proposed retransmission scheme is validated through simulation experiments.

By periodically aggregating local learning updates from edge users, federated edge learning (FEEL) is envisioned as a promising means to reap the benefit of local rich data and protect users’ privacy. However, the scarce wireless communication resource greatly limits the number of participated users and is regarded as the main bottleneck which hinders the development of FEEL. To tackle this issue, we propose a user selection policy based on data importance for FEEL system. In order to quantify the data importance of each user, we first analyze the relationship between the loss decay and the squared norm of gradient. Then, we formulate a combinatorial optimization problem to maximize the learning efficiency by jointly considering user selection and communication resource allocation. By problem transformation and relaxation, the optimal user selection policy and resource allocation are derived, and a polynomial-time optimal algorithm is developed. Finally, we deploy two commonly used deep neural network (DNN) models for simulation. The results validate that our proposed algorithm has strong generalization ability and can attain higher learning efficiency compared with other traditional algorithms.

In recent years, high-resolution video has developed rapidly and widescreen smart devices have become popular. We present an Android application called WeWatch that enables high-resolution video to be shared across two mobile devices when they are close to each other. This concept has its inspiration in machine-to-machine connections in the Internet of Things (loT). We ensure that the two parts of the video are the same size over both screens and are synchronous. Further, a user can play, pause, or stop the video by moving one device a certain distance from the other. We decide on appropriate distances through experimentation. We implemented WeWatch on Android operating system and then optimize Watch so battery consumption is reduced. The user experience provided by WeWatch was evaluated by students through a questionnaire, and the reviews indicated that WeWatch does improve the viewing experience.

A forest fire is a severe threat to forest resources and human life. In this paper, we propose a forest-fire detection system that has an artificial neural network algorithm implemented in a wireless sensor network (WSN). The proposed detection system mitigates the threat of forest fires by provide accurate fire alarm with low maintenance cost. The accuracy is increased by the novel multicriteria detection, referred to as an alarm decision depends on multiple attributes of a forest fire. The multi-criteria detection is implemented by the artificial neural network algorithm. Meanwhile, we have developed a prototype of the proposed system consisting of the solar batter module, the fire detection module and the user interface module.

The Internet of Things (IoT) has received much attention over the past decade. With the rapid increase in the use of smart devices, we are now able to collect big data on a daily basis. The data we are gathering (and related problems) are becoming more complex and uncertain. Researchers have therefore turned to artificial intelligence (AI) to efficiently deal with the problems created by big data.

This special issue deals with the technology and applications of AI in the IoT and is a forum for scientists, engineers, broadcasters, manufacturers, software developers, and other related professionals to discuss related issues. The topics addressed in this special issue include current research progress, real-world applications, and security issues related to AI in IoT. The call-for-papers attracted a number of excellent submissions. After two-round reviews, five papers were selected for publication. These papers are organized in three groups.

The first group comprises one overview paper that outlines the technical progress of IoT. The second group comprises two papers addressing security issues in IoT. The last group comprises two papers that present some interesting real-world applications that will benefit daily life. The first paper,“I ²oT: Advanced Direction of the Internet of Things,”gives an excellent vision of how AI technologies can be combined with IoT. The author introduces the principle and conceptual model of intelligent IoT (I ²oT in short), which results from the integration of AI and IoT and is the most promising version of IoT. In the final section of the paper, the author makes recommendations for further study and standardization.

The wireless sensor network (WSN) is a key enabler of IoT because of its great sensing ability and ability to generate and process big data. Using AI to handle big data in a WSN is a critical research topic and deserves much effort. The next two papers,“An Instance-Learning-Based Intrusion-Detection System for Wireless Sensor Networks”and“Forest Fire Detection Using Artificial Neural Network Algorithm Implemented in Wireless Sensor Networks”fall within this scope. The former addresses the intrusion-detection issue in WSNs and presents an instance-learningbased intrusion-detection system (IL-IDS) to protect the network from routing attacks. By mining historical data (instances), critical rules about attacks can be created to help build a routing mechanism that is more robust to malicious behaviour. The latter paper deals with a more specific application, i.e., forest fire detection using an artificial neural network algorithm in a WSN. Forest fires threaten forest resources, human lives, and surrounding environments. The authors build a forest fire detection system that takes advantage of the unique features of a WSN, such as easy deployment, efficient data collection and environmental monitoring. An artificial neural network algorithm is designed to improve multi - criteria detection, which helps decrease the possibility of false alarms the system cost.

The last group comprises two papers about real-world applications of IoT:“We Watch: An Application for Watching Video Access Two Mobile Devices”and“A Parameter-Detection Algorithm for Moving Ships.”With the rapid development of wireless communications and embedded computing, IoT is no longer a concept but is gradually becoming a reality. One of the consequences of this trend is that people are surrounded by smart devices, which are changing almost every aspect of daily life. The former paper explores the blossoming of smart devices for a better viewing experience. It presents a unified platform based on Android where different devices can share screens. For instance, it allows a video to be played simultaneously on two devices that are close to each other. It provides a better way of watching videos by putting the screens of the two devices close together. However, the distance between the two screens needs to be accurately measured. This paper discusses a distancemeasuring mechanism based on Wi-Fi signal decay. By mining training data, the system can adaptively improve the measurement accuracy.

The vision-based technique is a general AI technique that involves abstracting information from dynamic or static pictures. It is essential for the fast approach of IoT. In the latter paper, the authors propose an algorithm for detecting the parameters of a moving ship in an inland river. Numerous different vision-based parameter-detection approaches have been used in traffic monitoring systems; however, few have been applied to waterway transport because of complexities such as rippling water and lack of calibration objects. The authors discuss interactive calibration without a reference as well as detection of a moving ship using an optimized visual foregrounddetection algorithm. This reduces the likelihood of false detection in dynamic water-based scenarios and improves the detection of ship size, speed and flow. The traffic parameter detection algorithm has been trialled in the Beijing - Hangzhou Grand Canal and has an accuracy of more than 90% for all parameters.

We thank all authors for their valuable contributions and we express our sincere gratitude to all the reviewers for their timely and insightful expert reviews. It is hoped that the contents in this special issue are informative and useful from the aspects of technology, standardization, and implementation.