The development of autonomous driving has brought with it requirements for intelligence, safety, and stability. One example this is the need to construct effective forms interactive cognition between pedestrians vehicles in dynamic, complex, uncertain environments. Pedestrian action detection a form that fundamental success technologies. Specifically, detect pedestrians, recognize their limb movements, understand meaning actions before making appropriate decisions response. In survey, we...

Federated learning can effectively protect local data privacy in 5G-V2X environment and ensure protection Internet of vehicles environment. The advantages low delay 5G network should be better utilized the vehicle-road cooperative system. But existing asynchronous federated obtains a model through different node training completes update global central server. There are problems such as single point failure, leakage, deviation aggregation parameters. In response to above problems, we...

Recently the research of vehicle detection is mainly through machine learning, but it still has low accuracy problem. With study researchers, using deep learning methods becomes hot. In this paper, a selective search method and target model based on Fast R-CNN are used to detect vehicle. The strategy optimizes by preprocessing sample image new network structure. Firstly, experiment uses public KITTI data set self-collected BUU-T2Y set, respectively, for training validation test. Secondly,...

Combinatorial optimization problems (COPs) are a class of NP-hard with great practical significance. Traditional approaches for COPs suffer from high computational time and reliance on expert knowledge, machine learning (ML) methods, as powerful tools have been used to overcome these problems. In this review, the in energy areas series modern ML approaches, i.e., interdisciplinary COPs, areas, mainly investigated. Recent works solving using sorted out firstly by methods which include...

Automatic cloud detection in remote sensing images is of great significance. Deep-learning-based methods can achieve with high accuracy; however, network training heavily relies on a large number labels. Manually labelling pixel-wise level and non-cloud annotations for many laborious requires expert-level knowledge. Different types satellite cannot share set data, due to the difference spectral range spatial resolution between them. Hence, labelled samples each upcoming image are required...

Photovoltaic (PV) power generation has brought about enormous economic and environmental benefits, promoting sustainable development. However, due to the intermittency volatility of PV power, high penetration rate may pose challenges planning operation systems. Accurate forecasting is crucial for safe stable grid. This paper proposes a short-term method using K-means clustering, ensemble learning (EL), feature rise-dimensional (FRD) approach, quantile regression (QR) improve accuracy...

We know little about the intensity and determinants of interorganisational collaboration within homeless network. This study describes characteristics relationships (along with variables predicting their degree collaboration) 68 organisations such a network in Montreal (Quebec, Canada).Data were collected primarily through self-administered questionnaire. Descriptive analyses conducted followed by social multivariate analyses.The has high density (50.5%) decentralised structure maintains...

The imperative for swift and intelligent decision making in production scheduling has intensified recent years. Deep reinforcement learning, akin to human cognitive processes, heralded advancements complex found applicability the domain. Yet, its deployment industrial settings is marred by large state spaces, protracted training times, challenging convergence, necessitating a more efficacious approach. Addressing these concerns, this paper introduces an innovative, accelerated deep learning...

To improve the revenue of dairy farms, cow estrus must be accurately monitored to track mating time. Narrow Band Internet Things (NB-IoT) is considered as a promising technology realize cost-effective detection system attributing its wide coverage and low power consumption. increase success rate real-time detection, machine learning based algorithms have been applied extract patterns from data. However, due lack multivariate time series data, most previous studies do not consider using...

Recently GaN-based high electron mobility transistors (HEMTs) have revealed the superior properties of a breakdown field and saturation velocity. Reduction gate leakage current is one key issues to be solved for their further improvement. This paper reports that an Al layer as thin 3 nm was inserted between conventional Ni/Au Schottky contact n-GaN epilayers, behaviour Al/Ni/Au investigated under various annealing conditions by current-voltage (I–V) measurements. A non-linear fitting method...

Due to the intermittency and fluctuation of photovoltaic (PV) output power, a high proportion grid-connected PV power generation systems has significant impact on systems. Accurate forecasting can alleviate uncertainty is great significance for stable operation scheduling Therefore, in this study, feature rise-dimensional (FRD) two-layer ensemble learning (TLEL) model short-term deterministic probability proposed. First， based eXtreme Gradient Boosting (XGBoost), Random Forest (RF),...

High-precision cloud detection is a key step in the processing of remote sensing imagery. However, existing methods struggle to extract high-accuracy pixels, especially for images thin and fragmented clouds or those over high-brightness surfaces. In this study, we developed new model by combining models Fully Convolutional Network-8 sample (FCN-8s) U-network (U-net) (based on three visible bands) take full advantage spectral spatial information. proposed Network Ensembling Learning (FCNEL)...

This paper investigates the behaviour of reverse-bias leakage current Schottky diode with a thin Al inserting layer inserted between Al0.245Ga0.755N/GaN heterostructure and Ni/Au contact in temperature range 25–350 °C. It compares without Aluminium layer. The experimental results show that minimum point I–V curve drifts to minus voltage, increase increasing, returns 0 point. dependence gate-leakage currents novelty traditional are studied. introduces interface states metal Al0.245Ga0.755N....

To address the issue of building energy consumption optimization, a novel prediction method based on Random Forest (RF) and Auto Regressive Moving Average (ARMA) algorithm is presented in this paper. Considering was affected by factors involving equipment usage, personnel information, climate conditions, random forest introduced to build forecasting models historical data. It includes working-day mode non-working day according operating characteristics. switching challenge between working...

In multi-label tasks, sufficient and class-balanced label is usually hard to obtain, which makes it challenging train a good classifier. this paper, we consider the problem of learning from imbalanced incomplete supervision, where only small subset labeled data available distribution highly imbalanced. This setting importance commonly appears in variety real applications. For instance, considering ride-sharing liability judgment task, disputes due reasons, however, expensive manually...

Autonomous driving has to deal with human-vehicle interaction, in which one of the key tasks is detect pedestrians. In this paper, HOG, a classical algorithm pedestrian detection field used for extracting features and SVM classifier training. The feature obtained through training testing using INRIA dataset data acquired by autonomous vehicles. Meanwhile, we design visualization system better application on vehicles detecting pedestrian. This detects image input users calls that been trained...

Thanks to rapid development of Virtual Reality technologies, the research is not only limited in military training and scientific visualization realm, it has been expanded into more multidisciplinary areas, such as education, archaeology, conservation culture heritage, etc. Many heritage objects can be found museum, among them, historical mock-up represents a city with its precious value. The objective this paper propose prototype future interface that uses VR technology visualizing...

Target detection has a wide range of applications in many areas life, and it is also research hotspot the field unmanned driving. Urban roads are complex changeable, especially at intersections, which have always been difficult key part pilotless technology. Traffic policemen intersections link, but there few existing algorithms, speed generally slow. Aiming this problem, paper proposes real-time method traffic police based on YOLOv3 network.The YOLO network robust capable quickly completing...

Nowadays, the anomaly detection of aluminum electrolysis cell is a big problem in industry. The unbalanced time series samples common industrial applications. number under normal conditions much larger than that abnormal conditions. In electrolytic industry, this even more serious, it very difficult to find production because experts do not have clear criterion judge abnormalities. traditional machine learning algorithms, such as support vector (SVM) and convolutional neural network (CNN),...