The recent deployment of Automatic Identification System (AIS) enables real-time monitoring of vessel movements, generating massive volumes of navigational data. To extract meaningful behavioral patterns, trajectory clustering has become a key technique for knowledge discovery and decision-making in maritime scenarios, supporting tasks like route optimization, anomaly detection, and situational awareness. However, existing trajectory representation learning methods in maritime scenarios struggle to jointly capture spatial structural and motion dynamics. The non-uniform sampling and varying trajectory lengths also reduce the efficiency of traditional distance-based similarity calculations. Additionally, clustering methods based on pairwise similarity focus on local relationships and struggle to capture global structural relationships among trajectories. To address the aforementioned challenges, this study introduces a generalizable trajectory clustering framework. First, the raw vessel trajectory data are remapped onto two-dimensional grid to construct multi-channel trajectory images, effectively preserving their intrinsic spatiotemporal features. Second, we develop a feature enhance convolutional autoencoder (FE-CAE) that collaboratively models multiple attention modules to integrate spatial structural information and motion semantics in trajectory images, thereby learning discriminative trajectory representations and transforming trajectory similarity in the feature space, enabling fast, efficient and scalable similarity estimation. Finally, we construct trajectory network based on pairwise similarity, in which trajectories with similar navigation behaviors are tightly connected, and apply community detection to identify distinct navigational behavior patterns, thereby establishing a unified paradigm that bridges representation learning and network modeling. Extensive experiments conducted on large-scale AIS datasets from multiple real-world maritime water areas, and demonstrate that the proposed method outperforms widely adopted methods in both clustering efficiency and robustness. Therefore, underscoring the practical significance and wide applicability of the proposed trajectory clustering framework in advancing intelligent vessel traffic service.

This paper investigates the distributed adaptive event-triggered least-distance formation control problem for nonlinear multiagent systems (MASs) over switching digraphs via noncooperative game theory. First, distributed event-triggered estimators incorporating a switching-activated communication strategy are proposed to estimate the moving target and all agents’ decisions while reducing inter-agent communication. Based on the designed distributed estimators, a distributed time-varying Nash equilibrium (NE) seeking algorithm is established such that all agents’ decisions asymptotically reach the NE solution. Since high-order derivatives of the proposed distributed estimator states do not exist due to digraphs switching and event-triggering, the backstepping control design cannot be implemented. To overcome this difficulty, three-stage cascade filters are designed to provide sufficiently smooth signals. Then, based on the developed three-stage cascade filters, an adaptive event-triggered least-distance formation controller is proposed by the backstepping control technique. It is proved that the constructed formation control method can ensure that all agents achieve the least-distance formation, i.e., all agents’ outputs asymptotically reach a desired shape while minimizing the overall distance to the moving target. Finally, a simulation example on nonholonomic mobile robots is provided to illustrate the validity of the developed theoretical results.

This article investigates the adaptive neural network (NN) cooperative optimal output-feedback control problem for nonlinear multi-agent (NMA) systems under switching topology. Since the optimization point of global cost function is unknown, an optimal signal generator and high-order filter are established to estimate optimization point and high-order derivatives of optimal virtual signals. Meanwhile, a NN state observer is designed to estimate unmeasurable states. Based on the NN state observer and high-order filter, an adaptive NN cooperative optimal output-feedback control scheme is proposed by backstepping control design technology. It is shown that the proposed adaptive NN cooperative optimal output-feedback control method ensures that all signals of systems are bounded and the global cost function is minimized. In addition, we apply the proposed cooperative optimal control approach to unmanned surface vehicle systems, the simulation results verify its effectiveness.

Autonomous underwater vehicles (AUV) are popular robots that can independently operate complex tasks in underwater environment. Since AUV not only has highly moving characteristics, but also provides efficiently communicating supports, it is necessary to consider the energy balance of path planning for collision avoidance and communication capacity for all agents, so as to achieve largest endurance during tasks processing. Therefore, this paper focuses on the multi-objective optimization problem (MOP) in AUV case, and includes a new policy of weighted prioritized experience repay (WPER) to improve deep Q-learning network (DQN), which can obtain optimal costs balance between path planning and communication capacity. The proposed WPER policy enhances the experience pool of DQN, so that the AUV can distinguish the importance of empirical samples, improve the sampling efficiency and the training accuracy. In simulation experiment, we construct a three-dimensional underwater environment containing AUV, obstacles and serviced agents, and examine several comparison methods such as Q-learning and its variants to validate the performance of proposed DQN using WPER. Furthermore, we also design scenarios of simulating situation to investigate the capability of proposed method, and the results reveal that our method can effectively obtain MOP solution of shorter path length and larger communication capacity without collision.

In this article, we study the distributed fuzzy formation control problem for a class of strict-feedback nonlinear multiagent systems (NMASs) under communication delays and jointly connected switching topology. Since the communication between agents is affected by time-varying delay and some agents cannot access the leader's information under jointly connected switching topology, a communication-delay-related distributed formation observer is designed to estimate the leader's information and simultaneously mitigate the effects of communication delays. By using fuzzy logic systems to approximate the unknown functions, the controlled uncertain NMASs are transformed into the strict-feedback parameterized NMASs. Then, based on the designed communication-delay-related distributed formation observer and the backstepping control design theory, a fuzzy adaptive formation control algorithm is proposed. By constructing the Lyapunov functions, it is proved that the designed communication-delay-related distributed formation observer errors converge to zero exponentially and the proposed distributed fuzzy formation control algorithm can ensure that the closed-loop systems are semi-globally uniformly ultimately bounded, with the formation tracking errors converging to an adjustable neighborhood around zero. Finally, we apply the distributed fuzzy formation control scheme to marine surface vehicles (MSV), the simulation results and comparisons with the previous control methods verify its effectiveness.

The adaptive neural network (NN) exact-optimal consensus fault tolerant control (FTC) problem is investigated for uncertain high-order nonlinear multiagent systems (NMASs) with intermittent actuator faults. NNs are utilized to model unknown agents, and an adaptive NN state observer with asymptotical property is established. Since the optimization point is not directly known to the agents, the optimal signal generator is formulated to estimate it. Based on the designed NN state observer and optimal signal generator, an adaptive NN exact-optimal consensus output-feedback FTC scheme is proposed by using the backstepping control technology. It is proved that the controlled NMAS is asymptotically stable, and the observer errors and the tracking errors between the outputs and optimization point asymptotically converge to zero. Finally, we apply the proposed adaptive NN exact-optimal consensus output-feedback FTC approach to multiple marine surface vehicles (MSVs), and the simulation and comparison results verify its effectiveness.

In this article, a fuzzy optimal event-triggered dynamic positioning control approach with aQ-learning value iteration (VI) algorithm is developed for unmanned surface vehicles (USVs) systems. The USV systems are first modeled by Takagi-Sugeno (T-S) fuzzy systems. To reduce the communication resources and controller update times, an event-triggered mechanism is designed via employing the sampled augmented systems states and triggered control input signals. Based on the developed event-triggered mechanism and Bellman optimality theory, a fuzzy optimal event-triggered control (ETC) approach is presented. Since solution of optimal control policy reduces to algebraic Riccati equations (AREs), its analytical solution is difficult to solve directly. Then, to search its approximation solution, a VI algorithm is formulated. By rigorous proof, the proposed optimal ETC scheme can assure that the USVs systems are asymptotically stable and theQ-learning algorithm is convergent. Finally, the simulation and comparisons results with previous optimal controllers verify the feasibility of the presented optimal ETC scheme.

This article proposes a new adaptive fuzzy distributed output-feedback optimal control methodology for high-order strict-feedback nonlinear multiagent systems (MASs), which is composed of a fuzzy feedforward output-feedback distributed controller and a fuzzy error feedback correction distributed optimal controller. The fuzzy feedforward distributed controller is designed by backstepping design technique and fuzzy state observer, which can solve unmeasured states problem of the agents and transform the high-order strict-feedback nonlinear MASs into a linearizable feedback nonlinear MASs, while the fuzzy error feedback correction distributed controller is designed for the error nonlinear MASs based on adaptive dynamic programming and multiplayer differential graphical game, which can find a Nash equilibrium solution of the muti-player differential graphical game. It is proved that the formulated adaptive distributed fuzzy optimal control scheme can achieve the global optimal control objective, and make the controlled MASs stable, the followers output follow the leaders output. The simulation results on ship autopilot systems demonstrate its effectiveness.

In this paper, the collision avoidance process of underwater vehicles is considered, and an improved potential function is proposed to reduce the energy consumption of underwater vehicles when approaching the target position. Aiming at the negative gradient of the potential function in the controller, the potential function based on target position is proposed to prevent underwater vehicles from oscillating near the target position due to the excessive negative gradient of the potential function. Then, the Lyapunov function is used to prove that the potential function can achieve the purpose of underwater vehicles collision avoidance [1], and the improved potential function is compared with the general potential function during collision avoidance. Finally, two potential functions are compared by simulation, which shows the advantage of improving the potential energy function in reducing unnecessary energy consumption.

This article studies the fuzzy adaptive event-triggered (ET) consensus control issue of nonlinear multiagent systems (NMASs) under jointly connected switching networks. Since the leader and its high-order derivatives are unknown under jointly connected switching networks, a novel distributed ET reference generator equipped with an ET mechanism is constructed to estimate them. Meanwhile, the continuous information transmission among agents is avoided and the network channel utilization is optimized. Subsequently, fuzzy logic systems (FLSs) are employed to approximate unknown dynamics, and a fuzzy adaptive ET consensus control algorithm only using intermittent communication is designed by backstepping control methodology. It is demonstrated that all the closed-loop signals are semi-globally uniformly ultimately bounded (SGUUB), with the tracking errors converging to a small neighborhood around zero. Finally, we apply the developed fuzzy adaptive ET consensus control algorithm to unmanned surface vehicles (USVs), and the simulation results verify the effectiveness of the proposed ET consensus control algorithm.

This paper investigates the observer-based fuzzy event-triggered consensus fault-tolerant control (FTC) problem for nonlinear multiagent systems (MASs) with jointly connected switching topologies and actuator faults. Since a part of agents cannot receive information from their neighbors and leader under switching topologies, a distributed observer is designed to estimate unknown leader. At the same time, to avoid continuous information transmission and enhance the efficiency of network resources utilization among agents, an event-triggered communication mechanism is constructed to schedule inter-agent communication. Meanwhile, a fuzzy state observer is formulated to estimate the unmeasured states of the agents. Based on the distributed event-triggered observer and fuzzy state observer, an output-feedback event-triggered fuzzy FTC scheme is proposed by backstepping recursive control design. It is demonstrated that all signals of the controlled MASs are semi-globally uniformly ultimately bounded (SGUUB), consensus tracking errors converge to a small neighborhood of zero, and continuous communication between agents is avoided. Finally, simulation results on marine surface vehicles (MSVs) testify the advantages and effectiveness of the theoretical results.

Artificial intelligence and Automatic Identification Systems (AIS) play pivotal roles in intelligent maritime navigation for the modern maritime industry. Many artificial intelligence maritime applications based on AIS data have dramatically benefited traditional operations and managements in the field of maritime industry, and also provided state-of-the-art predictive analytics for vessel collisions and route optimization. However, the problem of modeling the interactions of vessels in complex waters still needs to be adequately addressed. In this paper, we focus on using spatiotemporal AIS data to model and forecast multiple vessel trajectories amid dynamic interaction patterns, and we propose a forecast model based on a novel neural network, namely a spatiotemporal multi-graph fusion network (STMGF-Net). The innovative STMGF-Net comprises three crucial modules. First, a Spatiotemporal graph construction module generates interaction graphs of various navigation modes, such as motions, risks, and attributes of vessels, Second, a multi-mode fusion module embeds and fuses the above interaction graphs into STMGF-Net. Finally, squeeze-and-excitation and temporal convolutional networks are introduced as Squeeze-and-excitation temporal convolutional modules to enhance the overall efficiency of the model significantly. Overall, the STMGF-Net can recognize complex spatiotemporal interaction patterns among neighboring vessels so as to capture and integrate these interaction features for achieving high-precision prediction performance in intelligent maritime navigation. In numerical experiments, three water areas of Zhoushan Islands, Yangshan Waters, and Yangtze River Waters are used as training and testing datasets. The results show that STMGF-Net improved prediction errors of average and final distance with increase of 49.637% and 50.622% than classic and state-of-art graph neural networks.

In this article, the adaptive fuzzy dynamic event-triggered output feedback resilient consensus control issue is investigated for nonlinear multiagent systems (MASs) subject to denial-of-service (DoS) attacks. Fuzzy logic systems (FLSs) are employed to model uncertain agents, and a state observer is constructed to estimate unmeasurable states. An event-triggered distributed resilient observer is designed to save the communication resources between agents, and estimate the unknown leader and its high-order derivatives in case of the communication topology being interrupted by DoS attacks. By the designed state observer and distributed resilient observer, a dynamic event-triggered resilient consensus control method is presented. It is proved that the controlled MASs are stable, and the followers can track the leader under DoS attacks. Moreover, the Zeno behavior can be excluded. Finally, we apply the developed resilient consensus control algorithm to multiple unmanned surface vehicles (USVs), the simulation results verify its effectiveness.

With the continuous development of artificial intelligence, the application of unmanned surface vessels (USVs) in the field of intelligent maritime vessels is becoming increasingly mature. Significant achievements have been made in several areas such as ocean environmental monitoring, maritime rescue, and maritime security. To further enhance the autonomous exploration capability, navigation safety, and operational efficiency of USV in route planning, this paper proposes an improved rapidly-exploring random tree* (RRT*) algorithm for USV path planning in known environments. Through simulation experiments, it has been verified that the proposed RRT* can achieve superior results in USV path planning. The proposed RRT* algorithm imposes restrictions on the sampling point range, building upon the original RRT* algorithm, resulting in more precise path planning. By comparing with the classical RRT algorithm and other traditional RRT algorithms, our approach significantly reduces the required distance for path planning while ensuring the safety of ship navigation. However, this improvement also comes with an increase in the time cost as it enhances the quality of the planned path.

Accurate prediction of vessel traffic flow plays a significant role in the field of modern intelligent transportation system. In order to enhance the prediction accuracy of vessel traffic flow, this paper combines genetic algorithm (GA) and Back Propagation neural network (BPNN) to build a prediction model. Based on the vessel traffic flow data of The Wuhan Yangtze River Bridge, the simulation experiments were carried out from 2013 to 2018. The average relative error of BPNN optimized by GA is 4.03%, which is better than the average relative error of direct BPNN prediction is 5.57%. The results show that accuracy of the prediction model using BPNN with GA optimization is higher than the traditional BPNN. The BPNN optimized by GA has achieved more ideal results in the forecast of vessel traffic flow. This paper provides the theoretical basis for the relevant decision-making of the water safety authorities so as to guarantee the water traffic safety.

This paper proposes a novel approach of in-store route data obtained by RFID for customer segmentation based on process mining (PM). The PM technique can dig out the hidden information of customers' in-store shopping route deeply. Firstly, we use PM to customize and decompose consumers' in-store movement paths. Then, we divide the fish area into patterns, and define these patterns as passing area, staying area and entering area respectively. Finally, we decompose the customer’s shopping path into several sub-processes according to the customer’s moving sequences obtained by RFID technology. In the experiment, we use the K-mediods algorithm based on Levenshtein distance to classify and analyze the characteristics of consumer moving path process logs, and investigate correlation analysis of different types of consumers obtained by analyzing the consumers' in-store movement paths through PM technology will be carried out.

In this article, we use the method of machine learning to model the consumer behavior, in order to identify the consumers with high volume purchases, and make a quantitative analysis of their behavior patterns. Firstly, radio frequency identification (RFID) sensors are installed in physical supermarkets for data sampling, tracking the shopping behavior of consumers in supermarkets, combining them with point of sale (POS) data at checkout, and transferring them to consumer behavior data set. Then, based on POS data, we mark consumers as high volume, low volume and none volume customer groups with historical purchase volume as target variables. Based on RFID data, stay time, shopping background, flow direction and number of return visits are extracted as explanatory variables. In the experiment, a classification model of consumer behavior is established by using long short-term memory based neural network (LSTM-NN). Numerical experiments show that LSTM-NN has higher recognition accuracy than logistic regression and support vector machine, which are improved by 5.26% and 6.97% respectively.

In this paper, we propose a new method for voice feature extraction by using a hierarchical clustering approach of inverse discrete cosine transform (IDCT) cepstrum coefficient. Since the IDCT cepstrum coefficient is transformed into the feature vector based on Euclidean metric, we call the proposed voice feature as “E-vector”. Comparing with other voice features, e.g. Mel frequency cepstrum coefficient (MFCC) and histogram of DCT cepstrum coefficients (HDCC), the E-vector can represent the dynamic characteristics and interactional relationships of the voice. In the numerical experiments, we use the speech files based on the voice of 630 people in TIMIT corpus, and employ Gaussian mixture model to compare the recognition accuracy of E-vector with that of MFCC and HDCC. The results show that E-vector outperformed other voice features on personal identification, and presented higher extensiveness on voice feature extraction.

Customer relationship management (CRM) is one of the most important recommendation systems to manage customer groups and understand potential relationships. CRM systems use many communication approaches such as telephone, email and social media to collect mass of customer data, which can help the companies (managers) to increase sales growth and enhance customer retention. Since 2000s, the advancements of RFID technology brought a new perspective on customer in-store behavior. Based on such RFID data, this paper describes a novel analytic approach applying social network analysis (SNA) into customer relationship network (CRN) and presents three contributions. First, to be different from existing studies, we attempt to construct the CRN based the shopping path (visiting patterns), which is recognized as a data stream of digitally encoded coherent signals. The CRN can help us to understand the customer relationships from the network perspective. Second, we employ network clustering (also called community detection or node classification) method to extract customer groups by maximizing the modularity. The modularity provides us a structural and relational understanding of customer groups. Third, we use within-module agree and participant coefficient to measure how a customer well-connected to other customers in the CRN; and identify such customers as "hub" customers by a topological process.

Consumers classification is one of the most important task in the retail sector. RFID (Radio Frequency IDentification) - A wireless non-contact technology is made easier to classify the consumers’ in-store behavior, recently. This paper presents an extraction of consumer purchasing behavior using statistical learning theory SVM (Support Vector Machine). In this research, we present our recent investigation outcome on the consumers shopping behavior in a Japanese supermarket using RFID data. We observe that it is possible to express the individual difference of consumers how are they spending time (we call it stay time in this paper) on shopping in a certain area of the supermarket. The contribution of this research is in two folds: we employ a SVM model on dealing with the RFID data of the consumer in-store behaviour firstly, as compared with other forecast model such as linear regression analysis and bayesian network, SVM provides a significant improvement in the forecasting accuracy of purchase behaviour (from 81.49% to 88.18%). Secondly, the kernel trick is adopted inside the SVM theory to choose the appropriate kernel for consumer purchasing behavior extraction.

This paper represents our recent studies about the prediction of purchase behavior and an advancement of in-store behavior with respect to RFID technology. In contrast to prior innovators in this research field, this paper has paid special attention to stay time spent on shopping in a target area rather than the whole supermarket, which can serve us to interpret the decision process of purchasing one product or a series of products in a much more intuitive and precise measurement. Also, we develop an integrated model to combine purchase behavior and in-store behavior. A probabilistic graphical model - bayesian network is employed to demonstrate a quantitative analysis process of purchase behavior decision over stay time. In order to distinguish purchase intention among different customers, an attitudinal factor - purchase background of customer is introduced in this paper to build bayesian network. As bayesian network can only deal with the discrete variables, a clustering algorithm is applied to discretize the continuous variables. In the experiments, the optimal cluster number of stay time and purchase background is examined for maximizing the performance evaluation with higher accuracy, and the results also show bayesian network has a better accuracy than other typical prediction models. Finally, we investigate the sensitivity and specificity of purchase behavior predicted by our proposal in adjustment of decision threshold, and use ROC (Receiver Operating Characteristic) curve to determine the optimal decision threshold which can maximize the classification accuracy of models.