In the practical scenario of construction sites with extremely complicated working environment and numerous personnel, it is challenging to detect safety helmet wearing (SHW) in real time on premise ensuring high precision performance. this paper, a novel SHW detection model basis improved YOLOv3 (named CSYOLOv3) presented heighten capability target site. Firstly, backbone network darknet53 by applying cross stage partial (CSPNet), which reduces calculation cost improves training speed....

Automated social behaviour analysis of mice has become an increasingly popular research area in behavioural neuroscience. Recently, pose information (i.e., locations keypoints or skeleton) been used to interpret behaviours mice. Nevertheless, effective encoding and decoding interaction underlying the rarely investigated existing methods. In particular, it is challenging model complex interactions between due highly deformable body shapes ambiguous movement patterns. To deal with modelling...

Industrial Wireless Sensor Networks (IWSNs), a novel technique in the field of industrial control, can greatly reduce cost measurement and as well improve productive efficiency. Different from (WSNs) non-industrial areas, IWSNs has high requirements for reliability, especially large-scale industry application. As network architecture great influences on performance IWSNs, this paper discusses node placement problem IWSNs. Considering reliability requirements, setup energy balance model is...

Since Modbus TCP is capable of the stable transmission while 5G has advantage low latency, they can be applied in different industrial situations, forming wired/wireless heterogeneous networks. However, and have protocol standards, so that cannot exchange information directly. To solve this problem, paper to propose a conversion mechanism system design method between 5G. First, communication rate format are analyzed, their mapping presented. Second, hardware software architecture gateway...

To address the problems of traditional network intrusion detection models with long training and time, difficult hyperparameter selection, poor performance due to data imbalance, this paper proposes an method named AB-LightGBM, which utilizes adaptive synthesis (ADASYN) oversampling technique Bayesian optimized light gradient boosting machine (LightGBM). In order solve problem rate a small number attacks caused by imbalanced data, ADASYN algorithm is used increase samples. addition, accuracy...

Automated analysis of mouse behaviours is crucial for many applications in neuroscience. However, quantifying from videos or images remains a challenging problem, where pose estimation plays an important role describing behaviours. Although deep learning based methods have made promising advances human estimation, they cannot be directly applied to mice due different physiological natures. Particularly, since body highly deformable, it challenge accurately locate keypoints on the body. In...

Industrial Wireless Sensor Networks (IWSNs), a novel technique in industry control, can greatly reduce the cost of measurement and control improve productive efficiency. Different from (WSNs) non-industrial applications, communication reliability IWSNs has to be guaranteed as real-time field data need transmitted system through IWSNs. Obviously, network architecture significant influence on performance IWSNs, therefore this paper investigates optimal node placement problem ensure cost. To...

This paper proposes a biogeography-based optimization (BBO) method augmented with Kalman filter, which is called KFBBO, for PID parameter tuning in wireless networked learning control system (WNLCS). Because of unreliable transmission data and commands networks, the noisy prone to errors, results poor performance by conventional real-world applications. BBO as new evolutionary proposed solve this problem dynamically optimizing parameters. network environment noisy, we also use filter...

Clustering is a concept in data mining, which divides set into different classes or clusters according to specific standard, making the similarity of objects same cluster as large possible. by fast search and find density peaks (DPC) novel clustering algorithm based on density. It simple novel, only requiring fewer parameters achieve better effect, without requirement for iterative solution. And it has expandability can detect any shape. However, DPC still some defects, such employs clear...

In this paper, a new backoff algorithm based on the dynamic modulating parameters was presented after analysis of binary exponential (BEB) and multiplicative increase linear decrease (MILD) IEEE 802.11 protocol. The reflected slot utilization to transmission probability in wireless channel, making average filter every twice sending changing priority re-node network modulate collision window nodes, which reduced effectively improved use channel. It time delay greatly. simulation results show...

According to different data sampling rates which are needed in machinery condition monitoring systems, one dual-mode wireless sensor network structure is proposed this paper. For the change status of machinery, faster need high-rate intermittent acquisition and lower low-rate continuous acquisition, so configured run dual-model satisfy two type acquisitions. The design implementation kind nodes shared gateway, transmitter real industrial ethernet, introduced detail, some experience result...

Wireless networked control system (WINCS), based on distributed systems of wireless network, integrate technology, computer network technology and communication technology. The delay controlled plant transmission may have bad impacts performance. Aiming at solving the problem, this paper analyzes existing problems solutions for WINCS firstly. Then, components simulation software true time toolbox function each module are detailed. Thirdly, a is constructed platform, which DC servo ZigBee...

Hybrid fieldbus network integrating wireless networks with existing wired fieldbuses has become new a research direction in industrial automation systems. In comparison to fieldbuses, the hybrid wired/wireless different system architecture, data transmission mechanism, communication protocol, etc. This leads challenges that need be addressed. paper proposes which consists of control (WICN), PROFIBUS-DP (Process Field Bus-Decentralized Periphery) network, and MODBUS/TCP (Mod Bus/Transmission...

Point cloud based retrieval for place recognition is still a challenging problem due to pronounced variations in scene appearance across different environmental contexts. Addressing the limitations of existing point cloud-based methods complex warehousing environments, which involve downsampling and inefficient utilization input data, resulting information loss. This paper introduces novel network model on Kernel Convolution (KPConv). approach leverages both intensity data as initial inputs,...

In this paper, a novel distributed online anomaly detection method in resource-constrained WSNs was proposed. Firstly, the spatiotemporal correlation existing sensed data exploited and series of single detectors were built each deployment sensor node based on ensemble learning theory. Secondly, these trained broadcasted to member nodes cluster, combining with its detector, initial detector built. Thirdly, considering resources-constrained WSNs, pruning biogeographical optimization (BBO)...