Intelligent Systems Techniques Research Articles

Dynamic spatial aware graph transformer for spatiotemporal traffic flow forecasting

Accurately predicting traffic flow is a crucial upstream technique in intelligent transportation systems for future travel plans, the efficiency of urban transport, and the regulation of transport departments, etc. The mainstream spatiotemporal graph convolutional neural networks are usually based on prior knowledge to predefine adjacency matrix graphs for spatial dependencies of the road network. However, modeling spatial correlation statically limits these models to accurately predict traffic flow, since the spatial correlations of road segments change over time. To address these issues, we propose a spatiotemporal gated transformer network with a graph latent information learning structure, termed GL-STGTN, for spatiotemporal traffic flow forecasting. First, we propose a graph latent information learning structure to dynamically learn the spatial dependencies for road network conditions from global and local learning perspectives. Second, we design a spatiotemporal gated transformer network (STGTN) block, which consists of a localized geographically aware block to extract the local embedding of spatial correlations and a temporal-aware enlarged block to extract local temporal dependencies. The learned spatial and temporal feature embeddings are further aggregated in a spatial multi-head attention module and a temporal multi-head attention module, respectively. In the end, a spatiotemporal fusion layer fuses the spatial and temporal information from the stacked STGTN blocks. Experiments on two public real-world benchmark datasets show that our model outperforms six state-of-the-art models for multi-step traffic flow forecasting.

Exploring the design of reward functions in deep reinforcement learning-based vehicle velocity control algorithms

ABSTRACT The application of deep reinforcement learning (DRL) techniques in intelligent transportation systems garners significant attention. In this field, reward function design is a crucial factor for DRL performance. Current research predominantly relies on a trial-and-error approach for designing reward functions, lacking mathematical support and necessitating extensive empirical experimentation. Our research uses vehicle velocity control as a case study, build training and test sets, and develop a DRL framework for speed control. This framework examines both single-objective and multi-objective optimization in reward function designs. In single-objective optimization, we introduce “expected optimal velocity” as an optimization objective and analyze how different reward functions affect performance, providing a mathematical perspective on optimizing reward functions. In multi-objective optimization, we propose a reward function design paradigm and validate its effectiveness. Our findings offer a versatile framework and theoretical guidance for developing and optimizing reward functions in DRL, particularly for intelligent transportation systems.

Artificial Intelligence (AI) Techniques for Intelligent Control Systems in Mechanical Engineering

In mechanical engineering, control systems are essential to the dependable and effective operation of mechanical equipment and processes. The advantages of integrating AI into control systems include energy savings, greater product quality, increased process effectiveness, and adaptive and predictive control. Various AI approaches, including machine learning methods like decision trees, neural networks, and support vector machines, are extensively studied for system identification, modelling, and adaptive control. AI techniques enhance these applications' control performance, energy efficiency, problem discovery, and maintenance. The methodology uses metaheuristic techniques for creating intelligent control systems, such as simulated annealing, ant colony optimisation, and harmony search. With the help of these algorithms, the solution space is efficiently searched for ideal control strategies while avoiding local optima. The importance of evaluating the learned control algorithm on a different dataset or running experiments to determine its performance is also underlined. Performance assessment benchmarks for tracking error, settling time, overshoot, and energy efficiency are provided in the results section. This paper investigates the use of AI techniques in mechanical engineering intelligent control systems. The paper also offers areas for additional research in intelligent control systems and identifies research directions for future advancements.

Joint Multi-Domain Resource Allocation and Trajectory Optimization in UAV-Assisted Maritime IoT Networks

The integration of Maritime Internet of Things (M-IoT) technology and unmanned aerial/surface vehicles (UAVs/USVs) has been emerging as a promising navigational information technique in intelligent ocean systems. In this article, we consider the UAV-assisted M-IoT network where USVs offload computation-intensive maritime tasks via non-orthogonal multiple access (NOMA) to the UAV equipped with the mobile-edge computing (MEC) server subject to the UAV mobility. To improve the energy efficiency of offloading transmission and workload computation, we focus on minimizing the total energy consumption by jointly optimizing the USVs’ offloaded workload, transmit power, computation resource allocation, as well as the UAV trajectory subject to the USVs’ latency requirements. Despite the nature of mixed discrete and non-convex programming of the formulated problem, we exploit the vertical decomposition and propose a two-layered algorithm for solving it efficiently. Specifically, the top-layered algorithm is proposed to solve the problem of optimizing the UAV trajectory based on the idea of deep reinforcement learning (DRL), and the underlying algorithm is proposed to optimize the underlying multidomain resource allocation problem based on the idea of the Lagrangian multiplier method. Numerical results are provided to validate the effectiveness of our proposed algorithms as well as the performance advantage of NOMA-enabled computation offloading in terms of overall energy consumption.

Data fusion for ITS: A systematic literature review

In recent years, the development of intelligent transportation systems (ITS) has involved the input of various kinds of heterogeneous data in real time and from multiple sources, which presents several additional challenges. Studies on Data Fusion (DF) have delivered significant enhancements in ITS and demonstrated a substantial impact on its evolution. This paper introduces a systematic literature review on recent data fusion methods and extracts the main issues and challenges of using these techniques in intelligent transportation systems (ITS). It endeavors to identify and discuss the multi-sensor data sources and properties used for various traffic domains, including autonomous vehicles, detection models, driving assistance, traffic prediction, Vehicular communication, Localization, and management systems. Moreover, it attempts to associate abstractions of observation-level fusion, feature-level fusion, and decision-level fusion with different methods to better understand how DF is used in ITS applications. Consequently, the main objective of this paper is to review DF methods used for ITS studies to extract its trendy challenges. The review outcomes are (i) a description of the current Data fusion methods that adopt multi-sensor sources of heterogeneous data under different evaluation strategies, (ii) identifying several research gaps, current challenges, and new research trends.

Bioinspired Scene Classification by Deep Active Learning With Remote Sensing Applications

Accurately classifying sceneries with different spatial configurations is an indispensable technique in computer vision and intelligent systems, for example, scene parsing, robot motion planning, and autonomous driving. Remarkable performance has been achieved by the deep recognition models in the past decade. As far as we know, however, these deep architectures are incapable of explicitly encoding the human visual perception, that is, the sequence of gaze movements and the subsequent cognitive processes. In this article, a biologically inspired deep model is proposed for scene classification, where the human gaze behaviors are robustly discovered and represented by a unified deep active learning (UDAL) framework. More specifically, to characterize objects' components with varied sizes, an objectness measure is employed to decompose each scenery into a set of semantically aware object patches. To represent each region at a low level, a local-global feature fusion scheme is developed which optimally integrates multimodal features by automatically calculating each feature's weight. To mimic the human visual perception of various sceneries, we develop the UDAL that hierarchically represents the human gaze behavior by recognizing semantically important regions within the scenery. Importantly, UDAL combines the semantically salient region detection and the deep gaze shifting path (GSP) representation learning into a principled framework, where only the partial semantic tags are required. Meanwhile, by incorporating the sparsity penalty, the contaminated/redundant low-level regional features can be intelligently avoided. Finally, the learned deep GSP features from the entire scene images are integrated to form an image kernel machine, which is subsequently fed into a kernel SVM to classify different sceneries. Experimental evaluations on six well-known scenery sets (including remote sensing images) have shown the competitiveness of our approach.

The Architecture of an Intelligent Technical Support System for Electricity Spot Market

With the accelerating pace of China's electricity market reform, the construction of the electricity spot market has been put on the agenda. However, as the number and scope of market participants gradually expand, the market-oriented transaction power continues to rise, and the intraprovincial and interprovincial transaction varieties are increasingly abundant. How to design an intelligent and powerful system that can meet the performance requirements of high concurrency and high-frequency transactions in the future market is a major problem in power reform. Based on the research of theoretical research results, this paper builds the front-end interaction platform of the southern spot electricity system based on the regional center to provide a data declaration interface for market users, including market management, market declaration, market release, market evaluation, intelligent analysis, front-end data interface, security protection, and other functional modules, provide declaration information and some market evaluation results to the southern regional spot power system platform, and obtain clearance results and published information from this platform.

Application of Intelligent Systems in Volt-VAr Centralized Control in Modern Distribution Systems of Electrical Energy

Centralized control of voltage magnitude and reactive power (Volt-VAr) is a highly complex combinatorial problem that seeks to determine the optimal adjustment of a set of control variables such as active and reactive power generation of distributed generators (DGs), modules in operation of capacitor banks, and voltage regulator taps; these with the purpose of ensuring an optimal operation of distribution systems. Looking for tools that allow real-time automation of this type of control, this study applies different intelligent system (ISs) techniques, such as decision trees, artificial neural networks, and support vector machines. Voltage magnitudes at nodes, current flow magnitudes in the circuits, and active and reactive power injections at the nodes at different grid points were used as input data. Training was performed from available measurements and actions recorded at the system control center. The tests were performed in a 42-bus distribution test system demonstrating the efficiency and robustness of the proposed solution techniques when compared with the results of a conventional mathematical model.

AI@TSS- Intelligent technical support scam detection system

Technical Support Scam (TSS) is a cybercrime that not only elicits the trust of a user but swindles their property. To detect TSS attacks in malicious samples, an intelligent TSS-aware system, called the AI@TSS system, was proposed in the study. The proposed AI@TSS system was built by the LightGBM algorithm with different kinds of features. In the study, 8263 malicious web page samples and 8263 TSS web page samples were collected and 42 features are defined for modeling the AI@TSS system. The performance analysis results demonstrate that the AI@TSS system reaches the accuracy and precision of 98% and 100%, respectively. The comparison shows superior to the existing detection methods, with 2.16% and 3.48% improvement in accuracy and precision.

Artificial Neural Network (ANN) based Soil Temperature model of Highly Plastic Clay

ABSTRACT Artificial neural networks (ANNs) are one of the popular methods of artificial intelligence that seek to follow the human mind function and nervous system with its successful application increased in many areas of engineering. The current study is focused to develop an ANN-based predictive model of the soil temperature of Yazoo clay using based on field instrumentation data. To provide an acceptable dataset for developing the predictive model, the investigation was carried out in six instrumented slopes within the 25 miles (40.2 km) radius from metropolitan Jackson in Mississippi. The six selected slopes were instrumented with soil moisture sensors, automated rain gauge, air, and soil temperature sensors starting from mid-August 2018. Volumetric moisture content, precipitation, air, and soil temperature values at 1.5 m (5 ft) depth at the crest of the six slopes were collected using automated data loggers and observed for more than seventeen months. The established database consisting of 13650 datasets was implemented for ANN intelligent system and multiple-degree Fourier series non-linear regression technique for predicting the hourly soil temperature. The total hourly natural rainfall and time, average previous soil temperature, and average hourly air temperature were set to be the inputs of the model and the hourly soil temperature was set to be the output of the model. These datasets were used as the training data and validated with each target slope. Sensitivity analysis was also conducted, and the most influential input parameters on the data output were determined. In this study, the change of soil temperature with atmospheric temperature was investigated, and a predictor model was developed by adopting the Levenberg-Marquardt (LM) algorithm method and the Tan-sigmoid transfer function. The developed ANNs model showed an excellent fit with the observed field values.

INTELLIGENT LIQUID LEVEL CONTROL OF A COUPLED NONLINEAR THREE TANK SYSTEM SUBJECTED TO VARIABLE FLOW PARAMETERS

In this paper, an intelligent control system technique is proposed to model and control of a nonlinear coupled three tank system. Two pumps fed the tank 1 and tank 2 and a fractional flow of these two pumps fed tank 3. The main aim of this paper is to make a set point tracking experiments of the tanks level using a nonlinear autoregressive moving average L-2 (NARMA L-2) and neural network predictive controllers. The proposed controllers are designed with the same neural network architecture and algorithm. Comparison of the system with the proposed controllers for tracking a step and random level set points for a fixed and variable flow parameter and some good results have been obtained.

Artificial Intelligence Technique and Wireless Sensor Networks in Energy Management System for Secure Power Optimization

The modern basic building blocks of a control system consist of data acquisition, dispensation of data by the system operators and the remote control of system devices. However, the physical controls, technical examinations and deductions were originally implemented to aid the process and control of power system design. The complexity of the power system keeps increasing due the technical improvements, diversity and dynamic requirements. Artificial intelligence is the science of automating intelligent activities presently attainable by individuals. Intelligent system techniques may be of excessive benefit in the application of area power system controls. Whereas smart grid can be measured as a modern electric power grid structure for better productivity and dependability via automatic control, excessive power converters, modern communications setup, sensing and metering equipment, and modern energy management techniques established on the optimization of demand, energy and network accessibility,and so on. The enormous depiction of the entire transmission grid, in the perspective of smart grids, is quite unclear; and in Nigeria no studies have been put on ground in order for the existing network to be turn into a smart grid. In this research work emphasis is placed on generation and transmission stations; power optimization using artificial intelligent techniques and wireless sensor networks for power control management system.

Projects for Predictive Maintenance and Operation between COPET G&T and Gnarus Institute

HIGHLIGHTS Monitoring system with analysis through intelligent system techniques. Predictive assessment of the operating conditions of circuit breakers. Monitoring of lightning arresters in the face of atmospheric discharges. Process of detection and localization of internal faults in SF6 substations.

DBScan and WrapDBScan methods applying for intellectual variance analysis in employee’s moving

This article discusses the development of mathematical support and software for detecting anomalous behavior of users based on biometric characteristics of their behavior analysis. One of the challenges in intelligent UBA systems is acquisition of useful information from a large volumes of unstructured, unmatched data. Methods and algorithms of intelligent data processing and machine learning used in UBA/DSS systems help to work on a task of solving problems of data analysis of different directivities. It is proposed an application of machine learning methods in implementation of mobile UBA system. There was formed the list of the most significant factors submitted to the input of the analyzing methods during the study. Two approaches of detecting abnormal user behavior have been proposed. The application of machine learning techniques in intelligent UBA systems will make it possible to predict information risks and insider excfiltration of these organizations in advance.

Parallel Swarm Intelligent Motion Planning with Energy‐Balanced for Multirobot in Obstacle Environment

Multirobot motion planning is always one of the critical techniques in edge intelligent systems, which involve a variety of algorithms, such as map modeling, path search, and trajectory optimization and smoothing. To overcome the slow running speed and imbalance of energy consumption, a swarm intelligence solution based on parallel computing is proposed to plan motion paths for multirobot with many task nodes in a complex scene that have multiple irregularly‐shaped obstacles, which objective is to find a smooth trajectory under the constraints of the shortest total distance and the energy‐balanced consumption for all robots to travel between nodes. In a practical scenario, the imbalance of task allocation will inevitably lead to some robots stopping on the way. Thus, we firstly model a gridded scene as a weighted MTSP (multitraveling salesman problem) in which the weights are the energies of obstacle constraints and path length. Then, a hybridization of particle swarm and ant colony optimization (GPSO‐AC) based on a platform of Compute Unified Device Architecture (CUDA) is presented to find the optimal path for the weighted MTSPs. Next, we improve the A∗ algorithm to generate a weighted obstacle avoidance path on the gridded map, but there are still many sharp turns on it. Therefore, an improved smooth grid path algorithm is proposed by integrating the dynamic constraints in this paper to optimize the trajectory smoothly, to be more in line with the law of robot motion, which can more realistically simulate the multirobot in a real scene. Finally, experimental comparisons with other methods on the designed platform of GPUs demonstrate the applicability of the proposed algorithm in different scenarios, and our method strikes a good balance between energy consumption and optimality, with significantly faster and better performance than other considered approaches, and the effects of the adjustment coefficient q on the performance of the algorithm are also discussed in the experiments.

Unmanned Aerial Vehicle Video-Based Target Tracking Algorithm Using Sparse Representation

Target tracking based on unmanned aerial vehicle (UAV) video is a significant technique in intelligent urban surveillance systems for smart city applications, such as smart transportation, road traffic monitoring, inspection of stolen vehicle, etc. In this paper, a computer vision-based target tracking algorithm aiming at locating UAV-captured targets, like pedestrian and vehicle, is proposed using sparse representation theory. First of all, each target candidate is sparsely represented in the subspace spanned by a joint dictionary. Then, the sparse representation coefficient is further constrained by an $L_{2}$ regularization based on temporal consistency. To cope with the partial occlusion appearing in UAV videos, a Markov random field (MRF)-based binary support vector with contiguous occlusion constraint is introduced to our sparse representation model. For long-term tracking, the particle filter framework along with a dynamic template update scheme is designed. Both qualitative and quantitative experiments implemented on visible (Vis) and infrared (IR) UAV videos prove that the presented tracker can achieve better performances in terms of precision rate and success rate when compared with other state-of-the-art trackers.

Human action recognition based on convolutional neural network and spatial pyramid representation

Detecting and recognizing human action in natural scenarios, such as indoor and outdoor, is a significant technique in computer vision and intelligent systems, which is widely applied in video surveillance, pedestrian tracking and human-computer interaction. Conventional approaches have been proposed based on various features and achieved impressive performance. However, these methods failed to cope with partial occlusion and changes of posture. In order to address these limitations, we propose a novel human action recognition method. More specifically, in order to capture image spatial composition, we leverage a three-level spatial pyramid feature extraction scheme, where each pyramid is encoded by local features. Thereafter, regions generated by a proposal algorithm are fed into a dual-aggregation net for deep representation extraction. Afterwards, both local features and deep features are fused to describe each image. To describe human action category, we design a metric CXQDA based on Cosine measure and Cross-view Quadratic Discriminant Analysis (XQDA) to calculate the similarity among different action categories. Experimental results demonstrate that our proposed method can effectively cope with object scale variations, partial occlusion and achieve competitive performance.

Scene categorization towards urban tunnel traffic by image quality assessment

Scene categorization is an indispensable technique in intelligent systems, such as scene parsing, video surveillance or autonomous driving. Considering traffic analysis under big data, in this paper, we propose scene categorization towards urban tunnel traffic based on image quality assessment. Specifically, the dataset is obtained through analyzing urban tunnel traffic incidents from 2016 to 2018. And we classify the traffic accidents in the big data environment. Then, the vehicles in the surveillance videos are extracted using conventional detector. The spatial information of vehicles in the image reflects the traffic situation. In order to encode such important information, we leverage the information clustering algorithm based on information entropy for image classification. Afterward, we establish a quality evaluation model based on each clustered images. The trained image quality assessment model will guide tunnel traffic classification and event analysis. The experimental results show the correct rate is more than 90%, and the overall detection effect is better than the k-modes algorithm and the Ng’k-modes algorithm.

Scene Categorization by Deeply Learning Gaze Behavior in a Semisupervised Context.

Accurately recognizing different categories of sceneries with sophisticated spatial configurations is a useful technique in computer vision and intelligent systems, e.g., scene understanding and autonomous driving. Competitive accuracies have been observed by the deep recognition models recently. Nevertheless, these deep architectures cannot explicitly characterize human visual perception, that is, the sequence of gaze allocation and the subsequent cognitive processes when viewing each scenery. In this paper, a novel spatially aware aggregation network is proposed for scene categorization, where the human gaze behavior is discovered in a semisupervised setting. In particular, as semantically labeling a large quantity of scene images is labor-intensive, a semisupervised and structure-preserved non-negative matrix factorization (NMF) is proposed to detect a set of visually/semantically salient regions from each scenery. Afterward, the gaze shifting path (GSP) is engineered to characterize the process of humans perceiving each scene picture. To deeply describe each GSP, a novel spatially aware CNN termed SA-Net is developed. It accepts input regions with various shapes and statistically aggregates all the salient regions along each GSP. Finally, the learned deep GSP features from the entire scene images are fused into an image kernel, which is subsequently integrated into a kernel SVM to categorize different sceneries. Comparative experiments on six scene image sets have shown the advantage of our method.

Fuzzy overall equipment effectiveness and line performance measurement using artificial neural network

Purpose The overall equipment effectiveness (OEE) is a powerful metric in production as well as one of the methods in evaluating function for measuring productivity in the production process. In the existing method, measuring OEE is based on three main elements consisting availability, performance and quality. The purpose of this paper is to evaluate the recognized metrics of production: OEE and overall line effectiveness (OLE) by using smart systems techniques. Design/methodology/approach In this paper, to improve the calculative methods and productivity with three methods: measuring OEE using Mamdani fuzzy inference systems (FIS), measuring OEE using Sugeno FIS, and measuring OLE using FIS and artificial neural networks (ANNs) are proposed. Findings The proposed methodologies aim to decrease some weaknesses of OEE and OLE methods by exploiting intelligent system techniques, such as FIS and ANNs. In particular, this research will solve the following issues that occur in manual and automatic data gathering. This technique is an effective way of measuring OEE and OLE with regard to different weights of losses as well as difference in the weight of the machines. In addition, it allows the operator’s knowledge to take a part in the measurement using uncertain input and output with implementation of linguistic terms. The presented method is the details and capabilities of those methods in various tested scenarios, and the results have been fully analyzed. Originality/value In relation to other methodologies, it allows the operator’s knowledge to take part in the measurement using uncertain input and output with implementation of linguistic terms. The presented method is the details and capabilities of those methods in various tested scenarios, and the results have been fully analyzed.