Using advanced deep learning (DL) algorithms for forecasting significant wave height of coastal sea waves over a relatively short period can generate important information on its impact and behaviour. This is vital for prior planning and decision making for events such as search and rescue and wave surges along the coastal environment. Short-term 24 h forecasting could provide adequate time for relevant groups to take precautionary action. This study uses features of ocean waves such as zero up crossing wave period (Tz), peak energy wave period (Tp), sea surface temperature (SST) and significant lags for significant wave height (Hs) forecasting. The dataset was collected from 2014 to 2019 at 30 min intervals along the coastal regions of major cities in Queensland, Australia. The novelty of this study is the development and application of a highly accurate hybrid Boruta random forest (BRF)–ensemble empirical mode decomposition (EEMD)–bidirectional long short-term memory (BiLSTM) algorithm to predict significant wave height (Hs). The EEMD–BiLSTM model outperforms all other models with a higher Pearson’s correlation (R) value of 0.9961 (BiLSTM—0.991, EEMD-support vector regression (SVR)—0.9852, SVR—0.9801) and comparatively lower relative mean square error (RMSE) of 0.0214 (BiLSTM—0.0248, EEMD-SVR—0.043, SVR—0.0507) for Cairns and similarly a higher Pearson’s correlation (R) value of 0.9965 (BiLSTM—0.9903, EEMD–SVR—0.9953, SVR—0.9935) and comparatively lower RMSE of 0.0413 (BiLSTM—0.075, EEMD-SVR—0.0481, SVR—0.057) for Gold Coast.

Some M2M applications such as event monitoring involve a group of devices in a vicinity that act in a co-ordinated manner. An LTE network can exploit the correlated traffic characteristics of such devices by proactively assigning resources to devices based upon the activity of neighboring devices in the same group. This can reduce latency compared to waiting for each device in the group to request resources reactively per the standard LTE protocol. In this paper, we specify a new low complexity predictive resource allocation algorithm, known as the one way algorithm, for use with delay sensitive event based M2M applications in the LTE uplink. This algorithm requires minimal incremental processing power and memory resources at the eNodeB, yet can reduce the mean uplink latency below the minimum possible value for a non-predictive resource allocation algorithm. We develop mathematical models for the probability of a prediction, the probability of a successful prediction, the probability of an unsuccessful prediction, resource usage/ wastage probabilities, and mean uplink latency. The validity of these models is demonstrated by comparison with the results from a simulation. The models can be used offline by network operators or online in real time by the eNodeB scheduler to optimize performance.

More and more devices, such as Bluetooth and IEEE 802.15.4 devices forming Wireless Personal Area Networks (WPANs) and IEEE 802.11 devices constituting Wireless Local Area Networks (WLANs), share the 2.4 GHz Industrial, Scientific and Medical (ISM) band in the realm of the Internet of Things (IoT) and Smart Cities. However, the coexistence of these devices could pose a real challenge—co-channel interference that would severely compromise network performances. Although the coexistence issues has been partially discussed elsewhere in some articles, there is no single review that fully summarises and compares recent research outcomes and challenges of IEEE 802.15.4 networks, Bluetooth and WLANs together. In this work, we revisit and provide a comprehensive review on the coexistence and interference mitigation for those three types of networks. We summarize the strengths and weaknesses of the current methodologies, analysis and simulation models in terms of numerous important metrics such as the packet reception ratio, latency, scalability and energy efficiency. We discover that although Bluetooth and IEEE 802.15.4 networks are both WPANs, they show quite different performances in the presence of WLANs. IEEE 802.15.4 networks are adversely impacted by WLANs, whereas WLANs are interfered by Bluetooth. When IEEE 802.15.4 networks and Bluetooth co-locate, they are unlikely to harm each other. Finally, we also discuss the future research trends and challenges especially Deep-Learning and Reinforcement-Learning-based approaches to detecting and mitigating the co-channel interference caused by WPANs and WLANs.

LTE is a candidate wide area communications network for the Smart Grid and can enable applications such as AMI, Demand Response and WAMS. We compare the uplink performance of the LTE FDD and TDD modes for a typical Smart Grid scenario involving a large number of devices sending small to medium size packets to understand the advantages and disadvantages of these two modes. An OPNET simulation model is employed to facilitate realistic comparisons based upon latency and channel utilization. We demonstrate that there is a critical packet size above which there is a step increase in uplink latency due to the nature of the LTE uplink resource scheduling process. It is shown that FDD leads to better uplink performance in terms of latency, while TDD can provide greater flexibility when the split between uplink and downlink data is asymmetrical (as it is expected to be in a Smart Grid environment). It is also demonstrated that the capacity of both FDD and TDD systems in terms of the number of serviced devices is control channel (PDCCH) limited for small infrequent packets, but TDD has the advantage that the capacity remains data channel (PUSCH) limited for smaller packet sizes and lower data burst rates than an FDD system.

The use of Artificial Intelligence (AI) to detect defects such as concrete cracks in civil and transport infrastructure has the potential to make inspections less expensive, quicker, safer and more objective by reducing the need for on-site human labour. One deployment scenario involves using a drone to carry an embedded device and camera, with the device making localised predictions at the edge about the existence of defects using a trained convolutional neural network (CNN) for image classification. In this paper, we trained six CNNs, namely Resnet18, Resnet50, GoogLeNet, MobileNetV2, MobileNetV3-Small and MobileNetV3-Large, using transfer learning technology to classify images of concrete structures as containing a crack or not. To enhance the model’s robustness, the original dataset, comprising 3000 images of concrete structures, was augmented using salt and pepper noise, as well as motion blur, separately. The results show that Resnet50 generally provides the highest validation accuracy (96% with the original dataset and a batch size of 16) and the highest validation F1-score (95% with the original dataset and a batch size of 16). The trained model was then deployed on an Nvidia Jetson Nano device for real-time inference, demonstrating its capability to accurately detect cracks in both laboratory and field settings. This study highlights the potential of using transfer learning on Edge AI devices for Structural Health Monitoring, providing a cost-effective and efficient solution for automated crack detection in concrete structures.