The extensive deployment of smart meters in millions of households provides a huge amount of individual electricity consumption data for demand side analysis at a fine granularity. Different from traditional aggregated system-level data, smart meter data is more irregular and unpredictable. As a result, probabilistic load forecasting, which can provide a better understanding of the uncertainty and volatility in future demand, is critical to constructing energy-efficient and reliable smart grids. In this paper, a recently developed technique called Bayesian deep learning is employed to solve this challenging problem. In particular, a novel multitask probabilistic load forecasting framework based on Bayesian deep learning is proposed to quantify the shared uncertainties across distinct customer groups while accounting for their differences. Further, a clustering-based pooling method is designed to increase the data diversity and volume for the framework. This not only addresses the problem of overfitting but also improves the predictive performance. Numerical results are presented which demonstrate that the proposed framework provides superior probabilistic forecasting accuracy over conventional methods.

Unmanned aerial vehicles (UAVs) can be deployed efficiently to provide high quality of service for Internet of Things (IoT). By using cooperative communication and relay technologies, a large swarm of UAVs can enlarge the effective coverage area of IoT services via multiple relay nodes. However, the low latency service requirement and the dynamic topology of UAV network bring in new challenges for the effective routing optimization among UAVs. In this paper, a layered UAV swarm network architecture is proposed and an optimal number of UAVs is analyzed. Furthermore, a low latency routing algorithm (LLRA) is designed based on the partial location information and the connectivity of the network architecture. Finally, the performance of the proposed LLRA is verified by numerical results, which can decrease the link average delay and improve the packet delivery ratio in contrast to traditional routing algorithms without layered architecture.

Non-intrusive load monitoring (NILM) is a technique that infers appliance-level energy consumption patterns and operation state changes based on feeder power signals. With the availability of fine-grained electric load profiles, there has been increasing interest in using this approach for demand-side energy management in smart grids. NILM is a multi-label classification problem due to the simultaneous operation of multiple appliances. Recently, deep learning based techniques have been shown to be a promising approach to solving this problem, but annotating the huge volume of load profile data with multiple active appliances for learning is very challenging and impractical. In this paper, a new semi-supervised multi-label deep learning based framework is proposed to address this problem with the goal of mitigating the reliance on large labeled datasets. Specifically, a temporal convolutional neural network (CNN) is used to automatically extract high-level load signatures for individual appliances. These signatures can be efficiently used to improve the feature representation capability of the framework. Case studies conducted on two open-access NILM datasets demonstrate the effectiveness and superiority of the proposed approach.

The space–air–ground-integrated network (SAGIN) will greatly promote the development of the Internet of Things (IoT). Green IoT will be an important part of SAGIN. Ambient backscatter communication (AmBC) is a potential solution for green SAGIN IoT. To improve the achievable sum rate (ASR) of the AmBC system, we propose a reconfigurable intelligent surface (RIS)-assisted AmBC system. In the single-backscatter device (BD) AmBC scenario, we first give the phase shifts that maximize the gain of the reflection link of the AmBC system, and then give the optimal reflection coefficient. The proposed scheme does not need to solve the convex semidefinite program (SDP) problem and has the characteristics of low computational complexity. In the multi-BD AmBC scenario, we first propose a multi-BD phase shifts initialization strategy to ensure the stability of the proposed scheme. Then, we give the optimal reflection coefficient and phase shifts based on the iterative method. Simulations show that the RIS-assisted AmBC scheme is superior to the non-RIS-assisted AmBC scheme.

Feature selection (FS) is a key factor for the performance of machine learning algorithms, as not all data and hence features are related to the various tasks. In this paper, we propose a novel scheme for convolutional FS for machine learning algorithms in computer vision. As not all the convolutional features are related to visual tracking, removing the unrelated ones will dramatically reduce the complexity and improve the algorithm performance. However, how to identify and select features related to the visual tracking task is still a challenge for machine learning algorithms. In the proposed scheme, a novel adaptive weights-objective function approach is established to evaluate and select the features. Furthermore, a quadratic programming method is introduced which improves the optimization efficiency. The experimental results demonstrate that our proposed scheme achieves superior performance compared to the state-of-art trackers on the challenging benchmarks in computer vision.

For object detection in computer vision, detection models trained by high-resolution images often fail to recognize or localize objects on low-resolution images. To tackle this problem, we propose a fully convolutional network named residual super-resolution single shot network (RSRSSN). RSRSSN consists of two sub-networks, super-resolution sub-network and detection sub-network. The super-resolution subnetwork in RSRSSN is achieved by stacking of identity residual blocks while the detection sub-network adopts the single shot multibox detector (SSD). Based on multi-task learning, we design a novel objective function called feature maps multibox loss to enforce low-resolution images to produce similar feature maps with their corresponding high-resolution ones. This information sharing mechanism is proved to be critical for solving the resolution mismatch problem in the experiments. A two-step training scheme is also proposed to train the RSRSSN in an end-to-end manner. Without any data augmentation, RSRSSN outperforms the SSD on both down-sampled PASCAL VOC and MS COCO in real-time object detection.