International Journal of Systems, Control and Communications (5 papers in press)

Regular Issues

• Sliding mode active disturbance rejection control for uncalibrated visual servoing  
  by Cuiping Pu, Yi Wu, Jie Ren, Chenyang Ran 
  Abstract: A sliding mode control algorithm based on a linear extended state observer is proposed to address the multi-source uncertainty of uncalibrated visual servoing in robotic arms. Uncertainty, nonlinearity, coupling, external disturbances, etc. are considered as total disturbances. The linear expansion state observer is designed to observe the total disturbance and compensate for it. This enhances the systems anti-interference ability. The outer loop adopts sliding mode control law as state feedback. It suppresses the observation errors of the observer and ensures the stability of the system. Sliding mode control based on linear extended observer does not depend on specific tasks and system configurations. It provides a unified design framework for solving the problem of uncalibrated visual servoing. Through numerical simulations of three different disturbance cases, it was compared with PI linear extended state observer and linear active disturbance rejection to verify its good dynamic performance and disturbance rejection performance.
  Keywords: active disturbance rejection control; ADRC; sliding mode control; SMC; visual servoing; mechanical arm.
  
  
• Spatiotemporal and region relationship network for micro-expression recognition  
  by Zhiming Zhou, Ai Guan, Qiaoling Han, Qiuyan Zheng, Yue Zhao, Baoqing Zhu 
  Abstract: Micro-expressions are subtle facial expressions that occur when a person fails to suppress his emotional response. A significant issue is the intricate correlation between micro-expression motion and various facial regions, which hinders the extraction of effective feature. This paper proposes a novel spatiotemporal and regional relationship network (STRNet) to address this issue. STRNet consists of two branches and a specialised feature fusion module. Specifically, one branch performs multi-level feature extraction. The other branch focuses on modelling the relationships between different facial regions. The outputs of the two branches are then combined through a feature fusion module; this module enhances the generalisation of the micro-expression features extracted by the network. Experiments on four public datasets (SAMM, CASMEII, SMIC and CASME3) validate STRNets effectiveness. On CASMEII, it achieved UF1 and UAR scores of 0.9792 and 0.9764, respectively, and on CASME3, 0.5848 and 0.5601. STRNet outperformed existing methods, demonstrating superior performance.
  Keywords: micro-expression recognition; affective computing; image classification; deep learning; convolutional neural network.
  
  
• Non-fragile H control for 2D Markov jump systems with partially unknown probabilities and its application in metal rolling process  
  by Peng Cui, Zhenghao Ni, Feng Li 
  Abstract: This article studies the asynchronous H non-fragile control problem of two-dimensional Markov jump systems with imperfect probability information based on the Roesser systems. First, to make the system better simulate actual engineering applications, its transition probability and observation probability are considered to be completely known, partially unknown and completely unknown. Second, the hidden Markov model addresses the asynchronous phenomenon between the controller and the system, given that the system mode might not be precisely achieved in the actual environment. Simultaneously, non-fragile control is used to overcome the influence of controller gain fluctuation during the action process. Furthermore, a suitable Lyapunov function is constructed, and the associated theorem is used to deduce adequate requirements for the closed-loop systems stability. Lastly, the industrial steel rolling process and the Darboux equation example are used to confirm the feasibility of the suggested asynchronous non-fragile controller.
  Keywords: Roesser systems; hidden Markov model; HMM; non-fragile control; partially information; 2D systems.
  
  
• Linear active disturbance rejection control of fighter aircraft based on MADDPG algorithm  
  by Yetong Lin, Yuehui Ji, Yu Song, Junjie Liu 
  Abstract: To address nonlinearity, strong coupling, and disturbances in fighter aircraft attitude control, this paper proposes an intelligent control method based on multi-agent deep deterministic policy gradient (MADDPG) and linear active disturbance rejection control (LADRC). A three-channel LADRC controller estimates and compensates for disturbances, uncertainties, and coupling terms in real time. To overcome the dimensionality curse in multi-parameter optimization, an independent DDPG controller is assigned to each channel, ensuring cooperative control via reward sharing and enhancing dynamic response and disturbance rejection. Additionally, long short-term memory (LSTM) and self-attention mechanisms are integrated into the policy and value networks to improve representation and decision-making. A piecewise combined reward function mitigates sparse rewards. Simulations show that, compared to MADQN-based and traditional LADRC methods, the proposed approach achieves superior control performance and significantly reduces manual parameter tuning efforts.
  Keywords: Fighter aircraft; attitude control; deep reinforcement learning; LSTM; self-attention.
  
  
• Multimodal comparative learning chip defect detection algorithm based on GLIP guidance  
  by Ziyi He, Bingqi Wang, Li Ma, Jingjing Fang 
  Abstract: In the production of semiconductor chips, the existing process technology and the working environment have an impact on the quality of the chip, so defect detection on the chip surface is crucial. However, in real-world environments, it is challenging to collect a sufficiently large and highly representative sample of defects. In this paper, we propose a multimodal comparative learning approach with GLIP for location guidance and Multi scale fusion modules for different multiscale fusions to localise defect locations of different shapes and sizes. In the testing phase, samples from different chip types in the training set were used to demonstrate the good generalisation ability and accuracy of our model and tested on the MVTEC dataset to demonstrate the superiority of our method, where the image-level and pixel level accuracies on our privately owned chip dataset can reach 91.3 and 92.6, and the pixel-level accuracy on the MVTEC is 92.3.
  Keywords: defect detection; visual language model; zero-sample inference; comparative learning; transfer learning.