The Internet of Things (IoT) has been attracting people to its capability to deal with smart applications. However, with the development of IoT, there are more attacks to threaten IoT systems. Especially, the distributed denial of service (DDoS) attack can lead to mighty destruction to IoT servers, causing the whole IoT network to be out-of-service. Hitherto, given the lack of a common standard for defining the IoT-driven DDoS (IoT-DDoS) attack, the DDoS defense system for IoT is developed without accurate guidance. Additionally, defense against multi-layer IoT-DDoS attacks is rarely covered by previous works. To address these issues, a deep reinforcement learning-based multi-layer IoT-DDoS defense system (DRL-MLDS) is proposed with the reward metrics in compliance with IEEE P2668 – the first of its kind. In addition, to provide a resilient blocking time configuration for false-positive samples, a new power-law-based blocking time mechanism is developed to cooperate with the DRL-MLDS. The outcome reveals that the DRL-MLDS can reach the same accuracy level (i.e., more than 96%) as previous works under single protocol-based IoT DDoS attack, as well as providing around 97% defense accuracy on multi-layer IoT-DDoS attack, which was rarely discussed in previous works. Additionally, by applying the IEEE P2668-compliant reward metrics, the applicability index (ADex) of DRL-MLDS can be improved from 3.2 to 4.4, fulfilling the recommendation of ADex (e.g., $\mathbf {>}3.5$ ) toward IoT best practices. The DRL-MLDS can be extended to Metaverse design and applications.

The Received signal strength (RSS) fingerprint-based approaches are widely used for indoor location-based services. The emerging Long Range Wide Area Network (LoRaWAN) is a cost-effective solution for indoor latency-tolerant location-based services attributed to its long-range property. In general, there are serious RSS fluctuations due to fadings along the communication path, thus significantly jeopardizing the localization accuracy. To overcome the challenge, we propose the extreme RSS to stabilize the fingerprint database and formulate boundary autocorrelation to downsize tremendously the searching complexity and thus proliferating localization accuracy. In essence, the RSS fluctuations are modeled as a Bernoulli random process so that the RSS stability can be estimated by a newly defined fluctuation analytic function. To mitigate the impact of the perturbative fluctuation, the extreme RSS is further defined to cultivate a highly stable and robust fingerprint database which withstand environmental dynamics. In addition, boundary autocorrelation is developed to measure and compare the similarity between the measured RSS values versus the prestored fingerprint database. RSS values with low autocorrelation coefficients are eradicated from the typically lengthy searching. The downsized complexity significantly improves the localization accuracy. Experiments were carried out and the results revealed that the proposed method achieved sub-10-meter localization accuracy in indoor environments. Such accuracy is encouraging and superior in contemporary LoRaWAN measurements.

Urban trees provide essential ecosystem services on regulating temperature and humidity, filtering urban pollutants, and improving air quality. However, the increasing number of urban trees put pressure on maintenance and public safety. The total compensatory value of the trees, consisting of inspection, maintenance, and settlement of tree damages, is more than $2 trillion USD. At this point in time, there is no known research on manifesting guidance on automated tree health assessment. The Internet-of-Things (IoT) proliferates the deployment of wireless sensors and networks. A concept of the IoT trees is raised to implement various sensors on the trees for automated health monitoring and assessment. In this paper, an urban tree health index (UTHI) is first developed to indicate the health of urban IoT trees. The index will facilitate preventive measures on urban trees. To construct the indexing model, seven (7) dynamic (time-series) features and seven (7) static features are extracted to explore the ambient effects on urban tree health. Afterward, a heterogeneous neural network (HNN) for UTHI modeling is developed to adopt the heterogeneous feature structure. In HNN, the dynamic features are analyzed in the gated recurrent unit (GRU) layer and the static features are analyzed in a hidden layer. The novel fusion layer then aggregates the outputs computed from those layers and further explores unseen correlations among all features. The experimental result verifies that the HNN-based modeling achieves high accuracy and model fitness with the error rate of less than 5%. In addition, the HNN achieves 34% to 66% improvement of accuracy in comparison with the other machine learning algorithms. The supremacy of the developed model is that all indexing features can be predefined or monitored by the IoT sensors, thus rendering an automated and economic urban tree management.

As one of the most common species on Earth, tree plays an essential role in protecting ecological environment. Without trees large loss will occur in all aspects, such as soil erosion and climate warming, which will also harm human activities. In this case, it is vital to keep tree healthy growth to prevent the potential hazard to human. A Narrow Band-Internet of Things (NB-IoT) based tree health monitoring system is proposed in this study. It could provide effective guidance and monitoring for arborists to keep trees healthy. In this system, data collection of four various environmental features, which are thought as crucial factors in tree health, is achieved by relative sensors. Since NB-IoT has the advantages of low-power consummation and large coverage area, it is adopted to provide the transmission means for collected data and forward them to the sever for further analysis in this study. As an efficient method solving classification problems with small amount of data, K-nearest neighbors (KNN) suits this study well and is adopted. A tree health monitoring system is established finally, in which the data of environmental features are used as input to gain the health state of trees as output and the accuracy of this system could achieve 87.3%.

Improving driving safety is the key to reducing traffic accidents caused by human factors such as cognitive errors, judgment errors, slow emergency response, etc. Studies have shown these human factors are highly correlated with drivers stress levels. Hence, various driver stress detection (DSD) schemes have been developed to improve driving safety. However, existing approaches face great difficulties in achieving high detection accuracy and practicability. To address this challenge, an Ensembled Multiscale Classifier (EMC) is proposed to realize the DSD and further reduce traffic accidents. In EMC, the stress level is classified into three categories, namely Low-Stress Level (LSL), Mid-Stress Level (MSL), and High-Stress Level (HSL). Fiducial features and non-fiducial features are considered to reach a balance between detection accuracy and implementation practicability. Specifically, fiducial features are extracted directly from ECG signals and analyzed by the neural network with backpropagation (NNBP). For non-fiducial features, they are extracted from the transformed ECG signals and analyzed by the 1-D convolutional neural network (1-D CNN). The outputs of the NNBP and 1-D CNN are coordinated by an ensembled decision-making layer, thereby deriving a probabilistic prediction of drivers stress levels. Experimental results reveal the developed method has high model fitness and 95.9% detection accuracy.