With the continuous increase of maritime activities, maritime Internet of Things (M-IoT) devices are also growing explosively, which puts forward higher requirements for maritime networks with limited resources and complex environment. As the basis of 6G edge intelligence, edge computing technology can be adopted to deal with massive heterogeneous ocean data effectively to meet the needs of different maritime application scenarios. To improve the execution efficiency of maritime edge computing under the constraints of delay and energy consumption, we propose an edge computing network architecture for M-IoT based on the space-air-ground-sea integrated communication network of 6G, and put forward an intelligent task offloading algorithm based on Reinforcement Learning(RL), which migrates heavy computation tasks to suitable servers by means of cloud-edge-end cooperation. Simulation results show the superiority of the algorithm in time delay and energy efficiency.
This article considers the group consistency of second-order MAS with directly connected spanning tree communication topology. Because the MAS is divided into several groups, we proposed a group consistency control method based on intermittent control, and the range of parameters is given when the system achieves consensus. The protocol can realize periodic control and reduce the working hours of the controller in period. Furthermore, the group consistency of MAS is turn to the stability analysis of error, and a group consistency protocol of MAS with time-delays is designed. Finally, two examples are used for verify the theory.
Current Industrial Internet faces serious threats where attackers propagate malicious flows, resulting in communication failures in the Industrial Internet. In this work, we propose a practical and novel method to detect malicious traffic attack in real time with high accuracy. Our primary idea is to capture network flow, extract adequate network flow features, construct a long short-term memory (LSTM) based deep learning model, and identify the property of the corresponding network flow. Whether the network suffers attack or not is then determined according to the detection results. The corresponding prototype is also implemented in the Industrial Internet which is equipped with Software Defined Networking (SDN). Experimental results validate that the proposed method is effective in defending against malicious traffic attack in real-world network.
Intrusion detection is a traditional practice of security experts, however, there are several issues which still need to be tackled. Therefore, in this paper, after highlighting these issues, we present an architecture for a hybrid Intrusion Detection System (IDS) for an adaptive and incremental detection of both known and unknown attacks. The IDS is composed of supervised and unsupervised modules, namely, a Deep Neural Network (DNN) and the K-Nearest Neighbors (KNN) algorithm, respectively. The proposed system is near-autonomous since the intervention of the expert is minimized through the active learning (AL) approach. A query strategy for the labeling process is presented, it aims at teaching the supervised module to detect unknown attacks and improve the detection of the already-known attacks. This teaching is achieved through sliding windows (SW) in an incremental fashion where the DNN is retrained when the data is available over time, thus rendering the IDS adaptive to cope with the evolutionary aspect of the network traffic. A set of experiments was conducted on the CICIDS2017 dataset in order to evaluate the performance of the IDS, promising results were obtained.
Particle filtering is an indispensable method for non-Gaussian state estimation, but it has some problems, such as particle degradation and requiring a large number of particles to ensure accuracy. Biota intelligence algorithms led by Cuckoo (CS) and Firefly (FA) have achieved certain results after introducing particle filtering, respectively. This paper respectively in the two kinds of bionic algorithm convergence factor and adaptive step length and random mobile innovation, seized the cuckoo algorithm (CS) in the construction of the initial value and the firefly algorithm (FA) in the iteration convergence advantages, using the improved after the update mechanism of cuckoo algorithm optimizing the initial population, and will be updated after optimization way of firefly algorithm combined with particle filter. Experimental results show that this method can ensure the diversity of particles and greatly reduce the number of particles needed for prediction while improving the filtering accuracy.
With the development of renewable energy technology, and combined with the traditional fuel energy as an energy system, which integrated with load equipment to form a ship-harbour energy system. In order to ensure the safe operation of the energy system, we take a limitation on the carbon dioxide emissions of the ship during sailing voyage to reduce the waste of resources caused by redundant energy generation. Therefore, based on the above situations, an energy management model is proposed in this paper. In addition, considering the above problem, a reliable and efficient distributed economic optimal scheduling strategy is constructed. The simulation results show that the method can provide high-precision multi-energy complementary optimal scheduling scheme for the ship-harbour energy system under the premise of convergence.
This work designs a robust composite disturbance cancellation scheme for the unmanned surface vessel through incorporating the robust disturbance observer with the vectorial backstepping technique. The ocean disturbances are expressed by the first-order Markov process with the unknown amplitude matrix. By designing a robust disturbance observer, the on-line estimations of disturbances are obtained and they are applied to the time-varying unknown disturbance compensation by the backstepping technique. The backstepping control law is derived on the basis of compound interference monitor, which determines the actual position and diffraction angle of unmanned surface vessels to track the required reference path randomly and accurately. Ensure that the closed-loop control signal is globally limited. At last, the effectiveness of the control strategy is confirmed through the simulation study upon an unmanned surface vessel.
Nowadays unmanned surface vessel (USV) automatic collision avoidance technology in a dynamic environment is getting increasingly concern in the intelligent navigation field. And taking USV's dynamic characteristics, COLREGs (International Regulations for Avoiding Collisions at sea), and the complexity of the marine environment into account is a trend of USV automatic collision avoidance technology. Based on the traditional potential field approach and COLREGs, this improved potential field model designed the collision avoidance algorithm integrated with rules, considered collision action timing, steering angle amplitude, and resume timing. More actually, the novel algorithm divided the collision avoidance process into three stages: track keeping, collision avoidance, and course resumption. At last, the simulation results show the feasibility and availability of collision avoidance.
With the develop of Internet of Things(loT), maritime IoT(MIoT) has been widely used in sea. It integrate the maritime information, and realize the monitoring and systematic management of the massive ocean data. The MIoT is the application of IoT technology in the ocean. Compared with traditional information and computing technologies in the ocean, the MIoT is still in the primary stage. In this paper, the existing literature of MIoT is classified and summarized, also including its applications situation and future development.
Under the denial-of-service (DoS) attacks, we will design a adaptive secure tracking control for a kind of second-order uncertain multi-agent systems finally realizes the leader-following consensus problem. Combining the algebraic graph theory and the commonly used Lyapunov function method, an adaptive secure algorithm is proposed. Compared with the results of the existing second-order multi-agent consensus problem, we can ensure that the second-order unmodeled nonlinear multiagent system can achieve leadership tracking consensus in the case of DoS attacks. Finally, taking the automatic road system model as an example, the validity of the protocol is verified.
This paper studies the nonlinear autonomous platoon control via vehicular networks. Firstly, a nonlinear autonomous platoon model is established by considering the influence of consecutive cyber-attack. Secondly, an event-driven framework is constructed based on the model, and a decentralized resilient controller design method is proposed to ensure that the platoon can still achieve safe driving in the case of consecutive cyber-attack. Finally, the effectiveness of the proposed algorithm is analyzed and compared in Matlab. Keywords-nonlinear autonomous platoon, event-driven, consecutive cyber-attack, resilient controller
Traffic light detection algorithm, as an important algorithm in automatic driving and assisted driving tasks, is the focus of research. Because the traffic light is in the complex environment, it brings many difficulties to the detection algorithm. Traditional algorithms are based on color and geometric features and have many barriers. Therefore, a traffic light detection algorithm based on deep learning is proposed in paper. With YOLOv5 as the algorithm core and K-means to generate anchor, 63.3%AP and 143FPS detection speed are obtained on traffic lights in BDD100K data set respectively. The method in this paper achieves good results in speed and accuracy in an end-to-end manner.
This study aimed to build up a machine learning model through resnet34 framework with strong generalization ability and an auxiliary diagnosis system to assist physicians to classify pathological sections from multiple parts of human beings. Usually, a pathological section can be distinguished between tumor, stroma, complex, lymph, debris, mucosa, adipose, and empty. U-net, resnet18, and resnet34 were employed to build the models. Accuracy was applied to access the performance of models since our data sets are homogeneous. Resnet34, selected to be our final framework, achieved the best performance with the highest accuracy between 95.0% and 97.2% with training on the data sets, adjusting options, and adding a fully connected layer. Since it is not possible to deploy the required machine learning environment on every physician's computer, we deployed an online recognition system based on the models, presented as a website, for physicians to use.
A three-dimensional (3D) lidar is the main sensing module of an unmanned surface vehicle (USV). The interference of water clutter will reduce the energy efficiency of target detection and affect autonomous navigation's obstacle avoidance function. Based on 3D lidar, this study proposes a surface target DBSCAN-VoxelNet joint detection algorithm. The proposed algorithm employs a noise density clustering approach (DBSCAN) to filter surface clutter interference; a depth neural network VoxelNet is employed to divide surface sparse point cloud data into voxels, and the results are input into a Hash table for efficient query; the feature tensor is extracted through the feature learning layer, and the tensor is input into the convolution layer to obtain the global target information, resulting in high-precision target detection. The experimental results reveal that the proposed joint detection algorithm performs well in suppressing clutter in the water area , with a mean average precision (mAP) of 82.4%, which effectively enhances surface target detection accuracy.
For the unmanned surface vehicle (USV) autonomous navigation and obstacle avoidance in the inland waters and coastal areas, it's essential to understand the water scene and divide the navigable area by taking the sea-sky-line and shoreline as the reference. Owing to the noise interference such as water surface reflection, it is not easy to detect the sea-sky-line and shoreline. In order to segment the water scene more accurately, the gradient image is firstly generated by Sobel operator, and then the contour of sea-sky-line and shoreline is enhanced by superposition with the image eliminated by sea surface reflection. Considering the local features in each image partition, superpixels are generated by multi-scale morphological gradient reconstruction (MMGR) and watershed algorithm. Finally, the superpixels is aggregated by fuzzy c-means (FCM), to get water scene segmentation. Experimental results show that the algorithm proposed (SEFCM) is superior to other similar algorithms in the accuracy of water scene segmentation, and more robust to illumination interference.
Industrial Internet is widely used in the production field. As the openness of networks increases, industrial networks facing increasing security risks. Information and communication technologies are now available for most industrial manufacturing. This industry-oriented evolution has driven the emergence of cloud systems, the Internet of Things (IoT), Big Data, and Industry 4.0. However, new technologies are always accompanied by security vulnerabilities, which often expose unpredictable risks. Industrial safety has become one of the most essential and challenging requirements. In this article, we highlight the serious challenges facing Industry 4.0, introduce industrial security issues and present the current awareness of security within the industry. In this paper, we propose solutions for the anomaly detection and defense of the industrial Internet based on the demand characteristics of network security, the main types of intrusions and their vulnerability characteristics. The main work is as follows: This paper first analyzes the basic network security issues, including the network security needs, the security threats and the solutions. Secondly, the security requirements of the industrial Internet are analyzed with the characteristics of industrial sites. Then, the threats and attacks on the network are analyzed, i.e., system-related threats and process-related threats; finally, the current research status is introduced from the perspective of network protection, and the research angle of this paper, i.e., network anomaly detection and network defense, is proposed in conjunction with relevant standards. This paper proposes a software-defined network (SDN)-based industrial Internet security gateway for the security protection of the industrial Internet. Since there are some known types of attacks in the industrial network, in order to fully exploit the effective information, we combine the ExtratreesClassifier to enhance the detection rate of anomaly detection. In order to verify the effectiveness of the algorithm, this paper simulates an industrial network attack, using the acquired training data for testing. The test data are industrial network traffic datasets, and the experimental results show that the algorithm is suitable for anomaly detection in industrial networks.
This paper studies the data-driven prescribed performance control (PPC) problem about the discrete-time nonlinear multi agent system with fixed communication topologies in tracking error constraints. The new transformation error strategy is constructed by using equivalent dynamic linearization techniques, and a controller with the predetermined performance is designed to make the tracking error converge to a predetermined area. This control scheme can effectively ensure fast convergence of tracking error and achieve a certain convergence rate. In addition, better tracking performance can be achieved by adjusting the appropriate parameters. The effectiveness of the design will be verified by a simulation example.
Maritime Internet of Things (MIoT) is regarded as one of the important paradigms to realize the Smart Ocean. With the increasing number of vessels, it is significant to ensure navigation safety and prevent traffic accidents. Due to the advantages of fast mobility and flexible deployment, Unmanned Aerial Vehicles (UAV) can be employed to detect and collect data packets from Sensor Nodes (SNs). However, it is extreme challenging for one energy-constrained UAV to complete the data collection task while updating information in time. In this paper, we present a Multi-UA V-Assisted MioT (MUA-MIOT) architecture for collecting navigational data packets which can provide important information support for navigation safety. In addition, the optimal deployment for MUA-MIoT is investigated to achieve data packet collection and information detection. UAVis used as a data collection point to collect data packets from sensor nodes by hovering over them. The Integer Linear Programming (ILP) problem is formulated with the objective of minimizing energy consumption in the MUA-MIoT. Considering a series of constraints such as reliability and connectivity, the optimal deployment is given and then its feasibility and scalability is evaluated in simulations.
Separable nonlinear models (SNLMs) adopt a linear combination of nonlinear functions, which is often used in the field of system identification, machine learning and signal processing. In this paper, we studied the performance of gradient-based algorithms in identifying separable nonlinear models. We put forward a gradient descent-based variable projection (GD-VP) algorithm which taking advantage of the particular structure of SNLM. In each iteration, the algorithm eliminates the linear parameters of the model, then updates the nonlinear parameters through the gradient descent (GD) algorithm. To improve the convergence rate of GD algorithm, an accelerated GD-VP algorithm is derived by employing the Aitken acceleration technology. Numerical experiments shows the efficiency of the proposed algorithm.
This paper studies the problem of event-based output feedback controller design for linear systems. A new approach for establishing event-triggered (ET) controller via output feedback are presented. Based on the designed online iteration algorithm, the gains of the ET controllers are updated at different trigger instants to stabilize the system. By establishing discontinuous Lyapunov function (LF), it is proved that the constructed controllers can ensure the global uniform ultimate boundedness (uubs) of the systems. Moreover, it is shown that the proposed ET control approach achieves better system performance than the existing ones with fixed gains. Finally, an example is employed to verify the effectiveness and advantages of the presented technique.
