International Journal of Systems, Control and Communications (39 papers in press)

Regular Issues

• A systems and control approach to early employment prediction based on deep learning of educational behaviour  
  by Shi Ming, Ge Bowei, Wang Jianhong, Hao Lin, Wei Chenyue, Jia Xiaoqi, Ren Jiaxin 
  Abstract: Early employability prediction is critical for career guidance and refining educational strategies in todays competitive job market. This study employs deep neural networks (DNN) and comprehensive learning behaviour data to develop a robust predictive model. Data spanning academic performance, study habits, and extracurricular activities underwent rigorous filtering to ensure quality and relevance. An optimised DNN achieved a prediction accuracy of 92.4%, outperforming logistic regression (81.7%) and effectively identifying complex nonlinear data patterns. Statistical modelling and graphical visualisation confirmed the models stability and interpretability. The experimental design ensured reproducibility and strong quality control. In addition to predicting employability, the system provides students with personalised career guidance and offers valuable recommendations to educators and policymakers. This approach supports evidence-based decisions in education and human resources, aligning academic achievements with job market demands and fostering long-term success through cutting-edge deep learning methodologies.
  Keywords: deep neural networks; DNN; student employment prediction; learning behaviour analysis; educational data mining; EDM; employability skills.
  DOI: 10.1504/IJSCC.2025.10073514
   
  
• Recent advances in intelligent system design for crypto trading bots: a control and communication perspective  
  by Wang Yun 
  Abstract: More people are into AI-fuelled crypto trading bots these days. Simple need to automate their trade plans and profit from the erratic nature of the crypto market. This exploration covers some trading bots and demarcates them. These include rule-focused bots, those hooked on machine learning, sentiment analysis bots, arbitrage-minded bots, and portfolio-managing bots. Knowing the pros and cons of each type of bot allows traders to make better decisions in various sensed market climates. Additionally, this research into AI-dependent crypto trading robots will be able to highlight areas of future investigation and development. One such area involves refining the accuracy and reliability of sentiment analysis and machine learning methods. Other topics for research include new arbitrage and risk management strategies. Finally, it gives a fresh look at AI crypto trading bots for aggressive traders who would want to maximise the opportunities that this highly volatile crypto market will afford them.
  Keywords: AI-powered trading bots; machine learning in finance; cryptocurrency trading automation; sentiment analysis bots; arbitrage trading bots; portfolio management bots; rule-based trading strategies; performance evaluation metrics.
  DOI: 10.1504/IJSCC.2025.10073515
   
  
• Digital media interaction control technology based on tactile feedback mechanism  
  by Yizuo Wang, Xinlong Zhou 
  Abstract: In digital media interaction scenarios, the interaction control is difficult to achieve ideal results due to factors such as sensor accuracy, data processing speed, and user interface friendliness. Therefore, a research on digital media interaction control technology based on tactile feedback mechanism is proposed. Firstly, based on the NLT multi-point electric tactile device, tactile signal acquisition and noise reduction processing are completed in the digital media interaction scene. Then, an object packaging box tree model in the virtual environment is constructed and collision detection is performed to optimise tactile feature rendering. Finally, based on the rendering optimisation effect, the feedback force is calculated to complete the interaction control design. The experimental results show that the proposed method has good collision detection performance, high fluency, and can achieve interactive control. The control accuracy is higher than 0.94, and the average response time can reach 0.89 seconds, indicating good application effect.
  Keywords: haptic feedback mechanism; digital media; user interaction; intelligent control.
  DOI: 10.1504/IJSCC.2025.10073558
   
  
• Distributionally robust optimisation-based configuration and scheduling of offshore wind-PV-thermal coupled system considering thermal power dynamics  
  by Zizhan Wu, Lei Pan, Panpan Xia, Weijie Tang, Cedric Caruana 
  Abstract: To address reduced flexibility and security in systems with high offshore wind and PV penetration, this paper develops a wind-PV-thermal coupled system that leverages flexible reserves within thermal units to enhance adaptability under renewable generation uncertainty. Unlike conventional methods reliant on external storage or grid support, it utilises flexible reserves within thermal units to manage uncertainty and enhance adaptability. A distributionally robust optimisation-based scheduling framework is employed to optimise capacity configuration and pre-intraday dispatch, balancing net Profit and system robustness through a multi-objective model. Additionally, an intraday rolling optimisation model is introduced to improve real-time performance and mitigate thermal power fluctuations. Solved via the MOPSO-GUROBI algorithm, simulations show a wind-PV to thermal ratio of 1.45:1 and a 40% thermal deep peak-shaving rate. Under extreme conditions, wind-PV utilisation improve by 5% compared to systems without internal reserves. The results demonstrate enhanced flexibility, stability, and economic efficiency of the proposed system.
  Keywords: offshore wind and PV power; distributionally robust optimisation; DRO; thermal power reserve; capacity configuration; scheduling optimisation.
  DOI: 10.1504/IJSCC.2025.10073559
   
  
• Modelling and control of an intelligent communication system for IMDB sentiment classification  
  by Huafen Yang 
  Abstract: Text data, regarded as a valuable source of human knowledge, plays a vital role in modern research, particularly in sentiment analysis, which determines the polarity of textual content. This study employs natural language processing (NLP) techniques to evaluate several machine learning (ML) models for sentiment classification on the IMDB movie review dataset. The models include logistic regression (LR), CatBoost, XGBoost, support vector classifier (SVC), K-nearest neighbours (KNN), and a neural network (NN) with dense layers. Performance was assessed using accuracy, F-measure, AUC, recall, and precision, with one-hot encoding and three n-gram configurations (unigram, bigram, and trigram). The NN consistently achieved the highest accuracy (0.8842, 0.89584, 0.89512), outperforming all other models. LR also performed well, with accuracies of 86.6%, 86.4%, and 86.3%. The studys novelty lies in demonstrating the NN models superior accuracy and precision, offering a reliable approach for sentiment classification in the IMDB dataset.
  Keywords: sentiment analysis; natural language processing; NLP; IMDB movie reviews; neural networks; one-hot encoding; machine learning classification; N-gram analysis; logistic regression; LR.
  DOI: 10.1504/IJSCC.2025.10073706
   
  
• Cooperative path following control of AUVs based on second-order predefined-time sliding mode control  
  by Xin Chen, Yuehui Ji, Yu Song, Junjie Liu 
  Abstract: Aiming at the collaborative path tracking control problem of underactuated autonomous underwater vehicles (AUVs) in complex marine environments, this paper proposes a sliding mode control strategy with predefined time convergence. First, in the kinematics part, a motion control law based on the line-of-sight (LOS) guidance principle is designed to achieve the collaborative formation of multi-AUV path following. Second, a predefined-time extended state observer (PTESO) is developed, which treats model uncertainties and external disturbances as a lumped disturbance. Its estimation error can rapidly converge to zero within a predefined time, significantly enhancing the systems anti-disturbance capability. Finally, a second-order predefined-time non-singular terminal sliding mode controller (SPTNTSMC) is proposed. A switching strategy is adopted to eliminate the singularity problem of traditional terminal sliding mode, a second-order integral sliding surface is introduced to eliminate steady-state errors, and the predefined-time sliding mode characteristic is utilised to ensure that both the speed errors of AUVs and the sliding mode surface converge to zero within specified times, combining rapidity and robustness. Simulation results show that compared with traditional sliding mode control methods and disturbance observers, the proposed approach exhibits significant improvements in anti-disturbance performance, convergence speed, and stability.
  Keywords: autonomous underwater vehicles; AUVs; coordinated path following; line-of-sight guidance; predefined time sliding mode; predefined time extended state observer.
  DOI: 10.1504/IJSCC.2025.10073740
   
  
• Dynamic load aggregation and optimisation strategy for high penetration distributed power distribution network with multi-source matching  
  by Lizhu Song, Detai Pan, Chaona Yin, Gaobao Li, Cong Li 
  Abstract: The study of dynamic load aggregation and optimisation strategies for distribution networks is of great significance for ensuring dynamic balance of distribution networks. Therefore, a new dynamic load aggregation and optimisation strategy for high penetration distributed power distribution network with multi-source matching is proposed. Using variational autoencoder for multi-source matching of high permeability distributed power distribution network operation data, obtaining high-quality data, and determining the objective function to achieve dynamic load aggregation of the distribution network. Build a dynamic load optimisation model for high penetration distributed power distribution networks, and use an improved grey wolf optimisation algorithm to solve the model and obtain the optimal load optimisation strategy. The experimental results show that the maximum accuracy of the proposed method for dynamic load aggregation in the distribution network is 97.21%, the maximum renewable energy consumption rate is 92.37%, and the total operating cost of the system is 760,500 yuan.
  Keywords: multi-source matching; high penetration; distributed power; distribution network; dynamic load aggregation; optimisation strategy; improved grey wolf optimisation algorithm.
  DOI: 10.1504/IJSCC.2025.10073824
   
  
• Adaptive combination optimisation of switch states on distribution network feeder line based on combination algorithm  
  by Chaona Yin, Detai Pan, Guanliang Chen, Bin Li, Yongnian Huang 
  Abstract: In order to reduce the average duration of power outages and network loss rate in the distribution network, a combination algorithm based adaptive combination optimisation method for switch states on distribution network feeder line is proposed. Analyse the switch state combination strategy without safety constraints and derive the extreme conditions for network loss benefits. After introducing line capacity and voltage constraints, the remaining carrying capacity and voltage threshold were quantified. By combining the global search capability of genetic algorithm with the local refinement adjustment of taboo search, the combination of feeder line switch states in the distribution network was dynamically optimised, achieving the goal of minimising active power loss while satisfying radial constraints and operating conditions. The experimental results show that the average power outage duration of the proposed method in 10 tests is 2.89 to 4.78 minutes per household, with a network loss rate of 3% to 6%.
  Keywords: distribution network feeder line; switch status; adaptive combinatorial optimisation; genetic algorithm; Tabu search.
  DOI: 10.1504/IJSCC.2025.10074022
   
  
• Enhanced thermal demand prediction in buildings using optimised ANFIS models  
  by Huanhuan Wu 
  Abstract: Heating systems consume much energy, mainly from fossil fuels, and harm the environment. Precise thermal demand forecasting enables flexible production, reducing distribution cost, heat loss, and return temperature. For efficient energy use, exploring buildings behaviour and requirements for heating systems is critical, enhancing effectiveness in combined heat and power (CHP) installations. Utilising machine learning (ML) algorithms in forecasting thermal demand in terms of factors such as relative compactness (RCE), orientation (OR), overall height (OVH), and surface area (SA) has been prevalent. In this investigation, ANFIS is employed in forecasting thermal capacity, optimised with algorithms such as Ali Baba and the Forty Thieves (AFT), and the gold rush optimiser (GRO). The conclusion shows that ANGR outperforms ANFIS, with an RMSE of 0.655, approximately half of the RMSE for ANFIS. ANGR proves to have high effectiveness in training with an ideal value for R2 at 0.996.
  Keywords: ANFIS; thermal capacity; anticipation; Ali Baba and the Forty Thieves; AFT golden rush optimiser; GRO.
  DOI: 10.1504/IJSCC.2025.10074113
   
  
• A control-oriented security management approach for power grid topology using the AMIBCF model in networked smart systems  
  by Jingyu Xu 
  Abstract: The increasing complexity and vulnerability of modern power grids necessitate robust security frameworks to protect critical topology data. This research proposes a comprehensive security management system based on the adaptive multi-layer integrated block chain-based cybersecurity framework (AMIBCF) model. The framework integrates access control, anomaly detection, blockchain auditing, cryptographic protection, and fault resilience mechanisms to safeguard grid topology data. An improved AES encryption algorithm and a user authority management scheme were incorporated to enhance data confidentiality and access governance. Simulation results demonstrate significant performance improvements, including a 14% increase in topology anomaly detection accuracy, a 33% reduction in incident response time, and an 83% decrease in annual data breach incidents. The systems effectiveness was validated through machine learning-driven anomaly detection, study presents a scalable and secure architecture for managing the cyber-physical data of power grids in real-time, ensuring resilience against faults, unauthorised access, and cyber threats.
  Keywords: AMIBCF; AI neural networks; ANN; generalised state estimate; GSE; high-voltage; HV; AES; open-source mapping; OSM.
  DOI: 10.1504/IJSCC.2025.10074114
   
  
• Model predictive control of Lipschitz switched dynamics with unstabilisable modes via L2-gain approach  
  by Cuicui Yuan, Cong Li, Qianqian Xu 
  Abstract: This study investigates a coordinated framework combining predictive control with switched rules for switched Lipschitz systems. Unlike conventional methods, the proposed approach relaxes the assumption that all modes are stabilisable and ensures L2 gain performance in the presence of external disturbances. To maintain stability, a multiple Lyapunov function strategy is adopted. Two types of cost functions are introduced: one with an infinite predictive horizon for stabilisable modes, and another with a fixed horizon for unstable modes. The control problem is formulated as an optimisation task using linear matrix inequalities (LMIs), facilitating the computation of control gains. Due to the complexity of constraints arising from stability and cost function requirements, a sequential solution method is employed. This enables selective constraint application at each stage, effectively regulating energy variations across modes and ensuring asymptotic stability. The proposed controller is validated through its application to a chemical process system.
  Keywords: Lipschitz switched systems; model predictive control; MPC; unstabilisable modes; L2 gain performance; multiple Lyapunov functions; persistent dwell time; PDT; linear matrix inequalities; LMI; nonlinear system stabilisation.
  DOI: 10.1504/IJSCC.2025.10074217
   
  
• Virtual reality video news reporting enhancement system based on deep generative adversarial network  
  by Yaolong Wang 
  Abstract: Virtual reality technology is becoming an important medium for news reporting. To enhance the quality of virtual reality video news, this study proposes an image enhancement system based on deep generative adversarial networks. The method incorporates text encoders, discriminators and image enhancement modules to improve image clarity and diversity. After training on the LSUN dataset, the model showed rapid convergence, low loss and strong generalisation ability. In image restoration tests, it achieved the best initial score of 7.4 and the Frechet initial distance of 18.4. It also obtained the lowest mean square error value, the highest peak signal-to-noise ratio and the highest structural similarity value, such as 0.032, 36.5 and 0.92 in natural disaster conditions. The system demonstrated stable response time and good adaptability to different sample sizes. The results confirm outstanding image quality enhancement and provide practical value for future virtual reality video news production.
  Keywords: virtual reality; VR; text encoder; generate adversarial networks; video news reporting; image enhancement.
  DOI: 10.1504/IJSCC.2026.10075636
   
  
• Fault diagnosis of underwater thruster based on multi-algorithm fusion D-S evidence theory  
  by Jiadong Hu, Yao Huang, Zhenzhong Chu 
  Abstract: A multi-algorithm fusion D-S evidence theory method is proposed for the fault diagnosis of underwater thrusters. The classification results of multiple algorithms are used as independent sources of credibility to reduce the potential bias of a single algorithm. To address the issue of insufficient sensitivity in quantifying the degree of evidence conflicts, a differentiated processing method based on Hellinger distance and cosine similarity is proposed. Weights are assigned based on the reliability of the evidence source to reduce the impact of low-quality evidence. A method for allocating conflicting evidence weights is proposed by clarifying the reliability of each group of evidence through information entropy. This ensures the rationality and objectivity of allocation. Finally, the fault diagnosis test verification of the thruster in the pool environment is carried out.
  Keywords: underwater thruster; Hellinger distance; cosine similarity; D-S evidence theory; information entropy.
  DOI: 10.1504/IJSCC.2026.10075698
   
  
• Multi-objective optimisation operation strategy for virtual power plants considering renewable energy consumption rate  
  by Yong Tang, Linlin Yang, Zuoliang Li, Hao Yin, Hao Zhuang 
  Abstract: To address the issues of traditional methods, such as low accuracy in predicting renewable energy consumption rate, low return on investment, and long response time, proposes a multi-objective optimisation operation strategy for virtual power plants considering renewable energy consumption rate. Firstly, photovoltaic and wind power generation models are constructed, and renewable energy consumption rates are accurately calculated. Based on these, a multi-objective optimisation function is established, which comprehensively considers operating costs, power generation, and renewable energy consumption rates. Subsequently, a linearly decreasing parameter setting method is employed to enhance the particle swarm optimisation algorithm. The improved algorithm is then applied to solve the multi-objective optimisation problem, thereby effectively improving its global optimisation capability and solution efficiency. Experimental results demonstrate that the proposed method achieves an average accuracy of 96.41% in predicting renewable energy consumption rate, an average return on investment of 42.37%, a response time ranging from 0.21s to 0.62s.
  Keywords: renewable energy consumption rate; virtual power plant; multi-objective optimisation operation; optimisation function; improved particle swarm optimisation algorithm.
  DOI: 10.1504/IJSCC.2026.10076921
   
  
• Prediction method for unbalanced line loss rate in distribution network considering the uncertainty of green energy output  
  by Yijun Wang 
  Abstract: The research on predicting unbalanced line loss rate can provide scientific basis for optimising the operation and management of distribution networks. Therefore, a prediction method for unbalanced line loss rate in distribution network considering the uncertainty of green energy output is proposed. On the basis of modelling the uncertainty of green energy output, the K-Means++algorithm is used to collect operational data of the distribution network. The collected data is processed by filling in missing data and removing outliers. The processed data is input into the stacking ensemble learning model, which is trained and learned by the base learner and meta learner to obtain relevant prediction results. The experimental results show that the maximum accuracy of the proposed method for modelling the uncertainty of green energy output is 97.12%, the maximum prediction accuracy is 98.33%, and the minimum prediction time is 0.69 s.
  Keywords: green energy output; uncertainty; distribution network; unbalanced; line loss rate; prediction; stacking ensemble learning model.
  DOI: 10.1504/IJSCC.2026.10076922
   
  
• Multispectral image registration and classification incorporating deep learning  
  by Runze Song, Jiazhan Li, Man Zhang, Wenyi Sheng 
  Abstract: To address the issues of low efficiency in multispectral image registration and poor image classification accuracy, this paper proposes an improved method: building upon the LF_NET-BFMatcher-RANSAC, it incorporates spherical cosine distance matching and the PROSAC algorithm to enhance registration accuracy and robustness. Combined with the EfficientNetV2 model, it employs regularisation, integrated attention mechanisms, and loss function optimisation to improve the cleanliness of corn moisture grade classification. Furthermore, a T-SNE visualisation module and evaluation metrics enable quantitative analysis of classification grades. Experiments demonstrate that the proposed method achieves up to 4.22% improvement in SSIM values for registration across the 555 nm, 660 nm, 720 nm and 940 nm bands. For corn moisture classification, our method achieves an average precision of 0.9263, recall of 0.9193, and accuracy of 0.9185 - representing a maximum improvement of 5.67% over the next-best method, validating its effectiveness and practicality.
  Keywords: deep learning; multispectral; image registration; classification.
  DOI: 10.1504/IJSCC.2026.10076923
   
  
• Influencing factors of network media reporting on public perception and adoption of renewable energy technologies  
  by Yuying Lin, Liang Zhang 
  Abstract: This study investigates the impact of online media coverage on public perception and adoption of renewable energy technologies. Through the analysis of extensive online media data, the research reveals that high-frequency reporting significantly increases public attention, with a ten-report increase per month boosting the public interest index by 15%. Content quality also plays a crucial role, as authoritative and in-depth reports enhance public trust by 40%. Sentiment analysis indicates that positive reports foster a favourable public image of renewable energy, increasing acceptance by 20%, while negative reports lead to a 10% decline in public perception. Furthermore, interactive dissemination via social media accelerates information spread, with each social media share increasing public participation by 25%. These findings demonstrate that optimising media reporting strategies can effectively enhance public adoption of renewable energy technologies.
  Keywords: renewable energy policies; media reporting framework; public awareness; content analysis.
  DOI: 10.1504/IJSCC.2026.10076959
   
  
• Advanced control facilitates copper smelting enterprises in achieving dual carbon targets through flue gas treatment  
  by Changliang Guan 
  Abstract: This paper employs advanced control technologies such as model predictive control (MPC), virtual online analyser (VOA), and multimodal control to achieve automatic edge regulation of various valves in the acid-making conversion process. The system effectively handles measurable disturbances such as SO2 content and flue gas flow at the inlet, minimising fluctuations in key controlled variables and control variables. While satisfying operational constraints, the system maximises or minimises control over performance indicators, achieving energy-saving and efficiency-enhancing effects. The system enhances the stability of the acid-making conversion process control, aids in the control of fine acid-making processes, reduces labour intensity, improves response speed to abnormal conditions, and diminishes control level disparities due to reliance on manual operational experience. This contributes to standardised and regulated management, assisting copper smelting enterprises in realising the dual carbon goals of flue gas treatment.
  Keywords: model predictive control; MPC; advanced control; edge regulation; acid-making conversion; dual carbon.
  DOI: 10.1504/IJSCC.2026.10077107
   
  
• Design of multi-sensor data deep fusion algorithm based on fuzzy sets  
  by Xinfa Dong, Qing Cong, Huijuan Zhao 
  Abstract: In response to rotating machinerys low fault diagnosing efficiency and poor sensor data fusion quality, a multi-sensor data deep fusion means using fuzzy sets is proposed. A fault diagnosing means is developed. Multi-sensor data fusion is achieved by utilising the fuzzy support degree of fuzzy sets and dynamic bending distance. Fault diagnosis is achieved by combining complete ensemble empirical mode decomposition with adaptive noise and global average pooling wide kernel multi-scale convolutional neural network. In the experimental verification, the proposed multi-sensor data deep fusion algorithms accuracy in the rotating machinery fault dataset increased by an average of 2.67% compared to other methods. This new fault diagnosing means had an average increase in accuracy of 4.17% and an average reduction in diagnostic time cost of 78.56% when being compared. In bearing fault diagnosis, the diagnosis time cost of the proposed method is reduced by 76.35% and 78.48% respectively compared with other methods, and the accuracy rate is as high as 99.11%. Therefore, multi-sensor data deep fusion based on fuzzy sets can significantly improve the input data quality and promote the fault diagnosis efficiency. The proposed fault diagnosis method has positive application value in diagnosing rotating machinerys fault.
  Keywords: fuzzy set; FS; fault diagnosis; fuzzy support degree; DTW; multi-sensor data fusion.
  DOI: 10.1504/IJSCC.2026.10077249
   
  
• Hierarchical reinforcement learning-based UAV path planning and energy management  
  by Haitao Li, Chao Che, Yuexin Liu, Yuan Li, Li Sun 
  Abstract: Energy management has become a key factor in enhancing its range and economy. However, irrational power allocation strategies seriously challenge UAV energy management. This study proposes an optimisation framework based on hierarchical reinforcement learning (HRL), which decomposes the UAV path planning and power allocation problem. Among them, the first is responsible for generating the optimal path and calculating the demanded power, while the energy management module optimises the overall energy consumption through dynamic power limitation to minimise fuel cell loss and total system energy consumption. To enhance HRLs adaptability, this study introduces an environment-aware layer and priori strategies to improve the decision-making efficiency of the path planning module significantly. The experimental results show that the proposed algorithm reduces the overall cost by 17.68% and 11.56% compared with the single reinforcement learning and single layer reinforcement learning algorithms, respectively, which verifies its superior performance in energy efficiency optimisation.
  Keywords: hierarchical reinforcement learning; HRL; path planning; energy management; hybrid power systems.
  DOI: 10.1504/IJSCC.2026.10077405
   
  
• Multi-UAV path planning methodology for dual-task power inspection scenarios based on genetic algorithm  
  by Haitao Li, Chao Che, Chenyang Duan, Xin Zhang, Li Sun 
  Abstract: With the continuous expansion of power systems, employing unmanned aerial vehicles (UAVs) for power inspection has become a practical necessity. In this paper, two critical inspection tasks in single-depot, multi-UAV scenarios are investigated: target point inspection and transmission line inspection. For the target point inspection problem, a mathematical model incorporating maximum flight distance constraints is established, and an improved genetic algorithm that integrates the 2-opt operator and an adaptive mutation rate mechanism is proposed. Experimental results demonstrate that the improved algorithm achieves superior solution quality and convergence stability. For the transmission line inspection problem, an optimisation model is developed to minimise the total flight distance. An adaptive genetic algorithm (AGA) is proposed to solve this model. Comparative experiments on maps of different scales verify that AGA significantly reduces the total flight distance and optimises the number of dispatched UAVs compared to the classical adaptive genetic algorithm.
  Keywords: unmanned aerial vehicles; UAVs; target point inspection; transmission line inspection; genetic algorithm; GA.
  DOI: 10.1504/IJSCC.2026.10077406
   
  
• Circularly polarised half mode substrate integrated waveguide antenna in X band  
  by Rakhi Kumari, Shweta Srivastava 
  Abstract: In this paper, half mode substrate integrated waveguide (HMSIW)-based antenna with the leakage property has been proposed and analysed for obtaining circular polarisation with axial ratio close to 1 (or 0 dB) in the operating frequency. HMSIW leaky wave antenna is obtained from the square substrate integrated waveguide (SIW) resonator by dividing it along the magnetic wall and preserving its electric field property. When HMSIW is designed from FMSIW by feeding it one side, its resonating modes are reduced and hence, making it more selective for designing the selective antennas. A circularly polarised antenna with negligible side lobes are created by combining the properties of SIW with leaky wave antenna property.
  Keywords: circular polarisation; half mode substrate integrated waveguide; HMSIW; leaky wave; periodic slots; single feed.
  DOI: 10.1504/IJSCC.2026.10077686
   
  
• Automatic sensing method of operation fault in new energy distribution network based on edge computing  
  by Chuanpu Zhu, Cong Li, Fangyuan Fu, Qingyu Wu, Ziyi Zhu 
  Abstract: In order to shorten the data transmission delay and improve the accuracy of automatic fault sensing, a new energy distribution network operation fault detection method based on edge computing is proposed. Firstly, the data transmission path is optimised by queue model and game theory strategy through the internet of things data multi-channel cooperative transmission (cooperative stream grouping, CSG) algorithm. Secondly, the revolving door algorithm is used to dynamically adjust the data acquisition interval, balancing data integrity and processing efficiency. Finally, by combining binary particle swarm optimisation algorithm and D-S evidence theory, the fault data is transformed into probabilistic representations and multi-source evidence is fused to achieve accurate fault perception. The experimental results show that the data transmission delay of the proposed method is as low as 2.2 s, the information loss rate is as low as 1.01%, and the fault perception accuracy is as high as 99.34%.
  Keywords: edge computing; new energy distribution network; fault detection; automatic sensing.
  DOI: 10.1504/IJSCC.2026.10077772
   
  
• A multi-criteria decision-making framework for integrating green and low-carbon technologies into power grid projects: balancing carbon reduction and cost efficiency  
  by Peng Song, Yinlu Zhang, Zhaolong Liu, Weiqi Xiong, Daxiao Hou 
  Abstract: Power grid projects must incorporate green and low-carbon technologies in line with the global push for carbon neutrality. Reducing carbon dioxide emissions while managing costs requires a structured scientific decision-making approach. This research develops a comprehensive multivariate decision framework based on multi-criteria decision-making (MCDM) to support power grid applications of low-carbon technologies. Inspired by zero-carbon park planning, the framework integrates environmental, technological, economic, and sustainability dimensions through stakeholder modelling, fuzzy evaluation, and phased implementation. It employs the analytic hierarchy process (AHP) and Criteria Importance through Intercriteria Correlation (CRITIC) to determine dimension weights, optimised using game theory for unbiased prioritisation. A fuzzy comprehensive evaluation technique is used to analyse technological integration from planning to operational phases. The results highlight the critical role of planning and operations, emphasising economic sustainability and technical stability in reducing emissions. The framework provides strategic resource management guidance to stakeholders, facilitating optimal technology adoption and supporting energy infrastructure development for national and global power sector decarbonisation.
  Keywords: green technologies; low-carbon technologies; power grid integration; multi-criteria decision-making; carbon reduction; cost efficiency; analytic hierarchy process; AHP.
  DOI: 10.1504/IJSCC.2026.10078086
   
  
• High proportion photovoltaic power grid line loss prediction method based on transient adaptive sparrow search algorithm  
  by Ying Zhao, Xingxing Feng, Jun Dong, Yuxin Lu, Yun Zhao, Xipeng Liu 
  Abstract: To enhance the prediction accuracy of line loss power and line loss rate in photovoltaic grids, a high-penetration PV grid line loss prediction method based on transient adaptive sparrow search algorithm is proposed. Feature importance is quantified using gradient boosting decision trees, with redundant variables eliminated through Spearman rank correlation coefficients. Anomaly detection is performed by applying DBSCAN clustering algorithm to isolate noise data via density-based clustering. The temporal modelling capability of Mogrifier LSTM is integrated with the optimisation characteristics of transient adaptive sparrow search algorithm, feeding processed data into the optimised model to generate line loss predictions. Experimental results demonstrate that under PV fluctuation conditions, the proposed method maintains line loss power prediction errors within 0.1 x 104kW-h and controls line loss rate prediction deviations below 1%.
  Keywords: transient adaptive sparrow search algorithm; TASSA; high proportion photovoltaic power grid; Line loss prediction; gradient boosting decision tree; GBDT; mogrifier LSTM; Mg LSTM.
  DOI: 10.1504/IJSCC.2026.10078174
   
  
• Virtual structure formation control of multi-robot systems with high-order sliding mode differentiators  
  by Jie Ren, Chenxi Wang, Mingzheng Feng 
  Abstract: This study addresses the cooperative formation control issue of multiple mobile robots and proposes a dual-loop sliding mode control algorithm based on a higher-order differentiator within a virtual structure framework. The virtual structure approach enhances the stability of the mobile team in maintaining the desired formation and motion. A dual closed-loop control strategy for position and attitude tracking improves trajectory tracking efficiency. The higherorder sliding mode observer effectively suppresses the chattering phenomenon associated with sliding mode control. Simulation results demonstrate that the proposed method offers stable formation control, higher tracking accuracy, and more effective chattering suppression.
  Keywords: robots; formation control; higher-order sliding mode differentiator; virtual structure.
  DOI: 10.1504/IJSCC.2026.10079187
   
  
• Finite time sliding mode control of uncalibrated visual servoing  
  by Jie Ren, Chenyang Ran, Mingzheng Feng 
  Abstract: A SMC algorithm based on a finite time extended state observer (FTESO) is proposed to address the multi-source uncertainty of uncalibrated visual servoing in robotic arms. Uncertainty, coupling, external disturbances are considered as total disturbances. A FTESO is designed to observe the total disturbance, which enhances the systems anti-interference ability. The outer loop adopts SMC law to design state feedback. This suppresses the observation errors of the observer and ensures the stability of the system. It does not rely on specific tasks and system configurations, providing a unified design framework for solving uncalibrated visual servoing problems. Through numerical simulations of three different disturbance cases, it was compared with PI based on LESO and LADRC to verify its good dynamic performance and disturbance rejection performance.
  Keywords: visual servoing; sliding mode control; finite time extended state observer; FTESO; robotic arm.
  DOI: 10.1504/IJSCC.2026.10079188
   
  
• Globally stable dual-loop trajectory tracking control for wheeled mobile robots  
  by Cuiping Pu, Chenxi Wang, Mingqin Fu 
  Abstract: To address the trajectory-tracking problem of wheeled mobile robots, this paper proposes a globally stable dual-loop sliding-mode control scheme. The outer loop generates a position control law using a hyperbolic tangent function. The inner loop designs an angular-velocity control law based on a convergence law to ensure fast heading regulation. To provide the inner loop with the derivative of the outer-loop intermediate command, a fast tracking differentiator is developed to estimate this derivative online. This improves transient performance and alleviates initial oscillations and chattering. The global stability of the resulting closed-loop system is rigorously analysed using Lyapunov theory, and numerical simulations are carried out to validate the effectiveness of the proposed approach.
  Keywords: global stabilisation; trajectory tracking; wheeled mobile robots; dual-loop control; fast tracking differentiator.
  DOI: 10.1504/IJSCC.2026.10079289
   
  
• Accurate perception method for intelligent robot operating environment based on multi-source heterogeneous data  
  by Chunli Wang, Yanyan Liu 
  Abstract: In order to overcome the problems of low data coverage, low perception accuracy, and long time in traditional perception methods, an accurate perception method for intelligent robot operating environment based on multi-source heterogeneous data is proposed. Firstly, a data acquisition system based on RGBD camera, IMU, LiDAR and sonar sensor is constructed; Secondly, a CNN-SA fusion mechanism was designed to achieve adaptive feature fusion of multi-source heterogeneous data through self-attention weighting; finally, the extracted features are input into the designed transformer perception model, which utilises a multi head self-attention mechanism to deeply mine long-range dependencies in the fused features, and combines a multi-layer perception module to achieve high-precision environmental perception. The experimental results show that the proposed method has a data coverage rate of up to 98.86%, a stable perception accuracy of around 95.36%, and a perception time controlled within 1.21 s.
  Keywords: multi-source heterogeneous data; intelligent robot; operating environment; accurate perception; transformer.
  DOI: 10.1504/IJSCC.2026.10079585
   
  
• MEMS-based miniature spectrometer: design optimisation and spectral reconstruction for environmental sensing systems  
  by Shaopeng Chen 
  Abstract: The research provides the design and construction of MEMS focused miniature spectrometer to research advanced environmental sensing instruments. Growing concerns about air quality, water purity, and pollution control have increased the demand for compact, efficient sensing devices capable of reliable real-time spectral analysis. The proposed spectrometer has its foundation on the micro-electro-mechanical systems (MEMS) to allow miniaturisation and simultaneously possess a high sensitivity and resolution. It is designed with the objective of utilising MEMS scanning grating, photodetector arrays and built in optical elements to provide the desired selection and measurement of wavelength on the reduced foot print. The basic approach employed is the spectral signal processing with calibration algorithms that give fine-tuning of the proper detection through the correction of the undesirable noise, temperature difference and non-linearity of the equipment. The system uses a low-energy architecture with embedded control algorithms that enable adaptive environmental monitoring.
  Keywords: MEMS spectrometer; environmental sensing; spectral signal processing; calibration algorithms; miniature optical systems.
  DOI: 10.1504/IJSCC.2026.10079965
   
  
• A compact F-slot planer inverted f antenna with enhanced gain for IoT application  
  by Kumar Vaibhav Srivastava, Rajan Mishra, Rajeev Kumar Chauhan, Praveen Kumar Rao 
  Abstract: This research paper proposes an F-slot planar inverted-F antenna (PIFA) for wireless communication applications. The proposed design addresses the narrow bandwidth limitation inherent to PIFAs. An F-shaped slot is incorporated on the antenna patch to overcome this constraint, enhancing bandwidth performance. The design was developed through a systematic, step-by-step evolution process. The F-slot antenna, with dimensions of 40 mm x 50 mm, is highly compact, enabling seamless integration with small-scale devices such as Zigbee modules, Arduino boards, and Node-MCU (node microcontroller unit) modules. The proposed antenna is suitable for various IoT applications, including smart homes, smart cities, IoT-enabled offices, and small-area network systems. Operating at 2.4 GHz, the antenna is well-suited for transmission via Zigbee, Wi-Fi, and Bluetooth protocols. It provides the impedance matching of -29.17 dB, an efficiency of 73%, and a peak gain of 2.00 dB, outperforming other antennas in this frequency range.
  Keywords: PIFA; enhanced bandwidth; efficiency; zigbee; node-MCU; cloud.
  DOI: 10.1504/IJSCC.2026.10079988
   
  
• Fairness-aware vehicle routing optimisation under intelligent connected environments  
  by Ke Zhang, Yuelong Su 
  Abstract: Traffic congestion is currently a critical issue affecting the well-being of urban residents. With the rapid development of connected and automated vehicles (CAVs), path guidance for connected vehicles can be achieved through vehicle-to-vehicle information exchange and cloud-based control. To address the fairness imbalance in vehicle-road-cloud coordinated management within intelligent connected environments, this study proposes a bilevel programming model, which establishes a bidirectional coupling mechanism between upper-level CAV path guidance optimisation and lower-level human-driven vehicles (HDVs) user equilibrium behaviour, achieving improvement in mixed traffic flow through feedback regulation. Experimental results demonstrate that at the 30% CAV penetration rate, compared to a single CAV optimisation strategy, this model significantly reduces the average travel time gap between connected and human-driven vehicles. Through bilevel iterative optimisation constraints, our model safeguards road usage rights for both conventional and connected vehicles, substantially enhances traffic operational efficiency, and provides a fair solution for mixed traffic management.
  Keywords: intelligently connected vehicles; traffic guidance; bi-level optimisation; traffic equilibrium.
  DOI: 10.1504/IJSCC.2026.10079989
   
  
• Modelling and optimisation of oilfield production command systems based on multi-agent reinforcement learning  
  by Lijun Wang, Jinling Zhao, Ting Wang, Qiang Yin, Jun Wang 
  Abstract: The growing complexity of contemporary oilfield production, notably in response to unpredictable reservoir behaviour, trade-offs with multiple goals, and real-time operational requirements, demands intelligent and adaptive decision-making systems. The conventional centralised control methods cannot cope with large dynamic and heterogeneous environments. In this paper, a Multi-agent reinforcement learning-based production command system (MARL-PCS) is suggested, where the autonomous agents in the system are oilfield parts, which learn adaptive control policies. The hierarchical cooperative actor-critic optimisation (HCACO) mechanism is presented to synchronise the interactions between the agents and to balance decentralised autonomy with centralised goals and to avoid making suboptimal decisions. Also an active Bayesian reward shaping strategy is added to manage uncertainty, through dynamically modifying learning signals based on the changing conditions of operation. The suggested framework increases the efficiency of production, optimisation of energy, and reliability of equipment. MARL-PCS provides a scalable and robust solution, filling the control-optimisation gap, and leading to the next-generation intelligent oilfield management systems.
  Keywords: multi-agent reinforcement learning; oilfield production command systems; hierarchical cooperative actor-critic optimisation; Bayesian reward shaping; intelligent industrial decision-making.
  DOI: 10.1504/IJSCC.2026.10080358
   
  
• Output power fluctuation control method for power storage system based on adaptive wavelet packet decomposition  
  by Wenxuan Liu 
  Abstract: In order to improve the stability of the output power of the power storage system and shorten the control time, a power storage system output power fluctuation control method based on adaptive wavelet packet decomposition is proposed. Firstly, a mathematical model of SOC is constructed for the battery of the power storage system, introducing SOC constraints and charging and discharging power boundary limitations. Secondly, a three-level wavelet packet decomposition tree architecture is constructed, which utilises adaptive wavelet packet basis function selection and signal reconstruction. Finally, an adaptive PID controller integrating fuzzy logic and neural networks is designed to achieve coordinated suppression and precise control of output power fluctuations in the energy storage system through dynamic parameter adjustment. The experimental results show that the proposed method has a smoother output power curve and significantly shorter control response time, with a maximum control response time of 20.5 ms.
  Keywords: adaptive wavelet packet decomposition; electric energy storage system; output power; wave control.
  DOI: 10.1504/IJSCC.2025.10073002
   
  
• Accurate prediction of localisation error in support of reliable wireless sensor network deployment  
  by Donghe Xi 
  Abstract: Accurate localisation in wireless sensor networks (WSNs) is critical for applications such as environmental monitoring, target tracking, and asset management. The average localisation error (ALE) is a key metric reflecting localisation performance and system reliability. This study proposes a novel machine learning (ML)-based framework to predict ALE in WSNs, comparing the performance of histogram gradient boosting regression (HGBR), decision tree regression (DTR), and their hybrid variants integrated with the red-tailed hawk (RTH) respectively. Among these, the DTRT model achieved the highest training phase coefficient of determination (R² = 0.990), outperforming HGRT (R² = 0.975), HGBR (R² = 0.969), and DTR (R² = 0.956). These findings demonstrate the superior predictive accuracy of the DTRT hybrid model for ALE estimation. The study offers valuable insights into ML model performance for ALE prediction, aiding the development of more precise and robust localisation strategies in WSNs.
  Keywords: wireless sensor networks; WSNs; average localisation error; ALE; machine learning; ML; decision tree regression; DTR; histogram-based gradient boosting regression; HGBR; red-tailed hawk optimisation; RTHO; localisation error prediction; hybrid machine learning models.
  DOI: 10.1504/IJSCC.2025.10073381
   
  
• Enhancing autonomous wheelchair mobility: a holistic method for dynamic route planning  
  by Kavitha N. Pillai, Chinthala Ramesh, Kala Sasikumaran, Nalesh Sivanandan 
  Abstract: By combining computer vision and machine learning approaches, this research seeks to improve autonomous wheelchairs and tackles the crucial problem of real-time obstacle detection and avoidance through multi-sensor fusion based autonomous navigation. The proposed sensor fusion based dynamic navigation control algorithm suggests use of machine learning models for processing RGB and depth images to detect and estimate the range of obstacles. The proposed dynamic navigation control algorithm employs Dijkstra's algorithm, renowned for its optimality, to efficiently compute the fastest and most effective path, saving time and energy for the autonomous system. Simulation results demonstrate that the proposed dynamic navigation control algorithm's ability to successfully adjust the wheelchair's planned navigation path across a given grid environment in response to real-time obstacles by identifying and thereby avoiding obstacles. The proposed framework lays the foundation for future advancements in assistive mobility solutions, emphasising adaptability, robustness, and computational efficiency.
  Keywords: artificial intelligence; machine learning; depth images; computer vision; obstacle identification; path planning.
  DOI: 10.1504/IJSCC.2026.10079528
   
  
• Cyber-physical systems for personalised fitness: control, communication, and security in wearable technologies  
  by Honglei Zou, Liu Li 
  Abstract: The integration of wearable fitness devices and mobile health applications has advanced personalised fitness planning and real-time health monitoring in physical education. Despite these benefits, challenges remain in data security, user engagement, and adaptation to diverse educational contexts. This study evaluates the effectiveness, scalability, and security of a personalised digital fitness platform implemented across high schools, universities, and community fitness centres. Over a 12-week period, 142 participants aged 15-40 used wearable trackers to monitor engagement, fitness outcomes, data security perceptions, and satisfaction through structured questionnaires. Data were collected at weeks 0, 6, and 12, and analysed using paired t-tests, Chi-square tests, and descriptive statistics. Results reveal significant correlations, such as between educational environment and engagement, and between perceived data security and fitness improvement. Participants, particularly students, reported high engagement, notable health gains, and satisfaction. These findings underscore the platform's adaptability and potential for broader application in technology-enhanced physical education.
  Keywords: personalised fitness; wearable technology; mobile health applications; digital health platforms; physical education; user engagement; security.
  DOI: 10.1504/IJSCC.2025.10073707
   
  
• Distributed fault correlation perception and self-healing technology for 220 kV substation  
  by Jie Luo, Meng Ye, Jing Li, Hongwei Zhao, Huiyu Shang 
  Abstract: To shorten fault isolation time and improve power restoration success rate, a new distributed fault correlation perception and self-healing technology for 220 kV substations was studied. First, a technology architecture centred on multi-agent systems (MAS) was constructed to coordinate collaborative work between agents and management agents, achieving both local rapid response and global information sharing. Second, multi-source data were collected to construct a fault diagnosis system; through methods like data preprocessing and fuzzy correlation analysis, accurate correlation perception of distributed substation faults was achieved. Finally, a fault self-healing scheme was designed by constructing a recovery model, setting constraints, and developing a fault self-healing recovery process. Experimental results show the proposed method stabilises fault isolation time below 80 ms with power supply restoration success rates reaching 98.7%-99.5%.
  Keywords: 220 kV substation; distributed faults; association perception; self-healing technology.
  DOI: 10.1504/IJSCC.2025.10073248
   
  
• Modelling and forecasting of the Hang Seng index using a hybrid intelligent system framework  
  by Lijia Wang 
  Abstract: Predicting financial outcomes is inherently challenging due to the complex interactions among market dynamics, investor psychology, and economic indicators. Stock price forecasting is particularly difficult given the volatile, non-stationary, and nonlinear nature of time series data. While conventional methods can enhance forecasting accuracy, they often involve high computational complexity, which may increase the risk of prediction error. This study proposes a hybrid approach that integrates the radial basis function (RBF) algorithm with various metaheuristic optimisers, including the moth-flame optimisation algorithm, grasshopper optimisation algorithm, and genetic algorithm. The performance of these hybrid models is evaluated using Hang Seng index data from 2015 to 2023. Results demonstrate that the hybrid models significantly improve forecasting accuracy. Notably, the RBF combined with the grasshopper optimisation algorithm outperforms others, achieving R², MSE, and RMSE values of 0.9949, 28,346, and 168.36, respectively.
  Keywords: stock price; financial markets; radial base function; grasshopper optimisation algorithm; GOA; Hang Seng index; HSI.
  DOI: 10.1504/IJSCC.2025.10074255