本期导读

主题

信息物理系统、多标签不平衡、广泛学习系统、数据科学、电力系统、优化控制、图像增强、迭代预测改进、分布式在线优化、联合学习、故障检测滤波器、非线性系统、深度强化学习...

全球科研机构

美国Columbia University、Monmouth University、New Jersey Institute of Technology；加拿大Concordia University；上海交通大学、复旦大学、同济大学、哈尔滨工业大学、华南理工大学、武汉大学、北方工业大学、重庆邮电大学、西南大学、东南大学；鹏城实验室...

 

M. Taheri, K. Khorasani, and  N. Meskin,  “On zero dynamics and controllable cyber-attacks in cyber-physical systems and dynamic coding schemes as their countermeasures,” IEEE/CAA J. Autom. Sinica, vol. 11, no. 11, pp. 2191–2203, Nov. 2024. doi: 10.1109/JAS.2024.124692 

> Vulnerability of CPS to zero dynamics and controllable cyber-attacks is studied.

> Cyber-attacks are derived in terms of nonzero Markov parameters of the CPS and the entries of the observability matrix.

> Number of actuators that need to be compromised for zero dynamics and controllable cyber-attacks is studied.

Y. Lin, Z. Yu, K. Yang, Z. Fan, and  C. L. P. Chen,  “Boosting adaptive weighted broad learning system for multi-label learning,” IEEE/CAA J. Autom. Sinica, vol. 11, no. 11, pp. 2204–2219, Nov. 2024. doi: 10.1109/JAS.2024.124557 

> Aiming at the serious multi-label imbalance problem, this paper innovatively proposes a MLW-BLS.

> Proposes the MLAW-BLS to adaptively adjust corresponding label weights and values of MLW-BLS to construct an efficient imbalanced classifier set.

> Extensive comparative experiments are conducted on 30 datasets with 4 metrics to evaluate the effectiveness of MLAW-BLS compared with 7 mainstream algorithms.

J. Chen, K. Liu, X. Luo, Y. Yuan, K. Sedraoui, Y. Al-Turki, and  M. C. Zhou,  “A state-migration particle swarm optimizer for adaptive latent factor analysis of high-dimensional and incomplete data,” IEEE/CAA J. Autom. Sinica, vol. 11, no. 11, pp. 2220–2235, Nov. 2024. doi: 10.1109/JAS.2024.124575 

> An SPSO algorithm injects particles’ historical position and velocity into the evolution process, enhancing its search ability.

> SPSO’s theoretical convergence is rigorously proved via the analyses of the stochastic convergence conditions on the particles’ position expectations.

> An SPSO-incorporated LFA model implements efficient hyper-parameter adaptation without accuracy loss.

K. Nosrati, J. Belikov, A. Tepljakov, and  E. Petlenkov,  “Revisiting the LQR problem of singular systems,” IEEE/CAA J. Autom. Sinica, vol. 11, no. 11, pp. 2236–2252, Nov. 2024. doi: 10.1109/JAS.2024.124665 

> Examine the conditions for the existence of the LQR algorithm for discrete singular systems.

> Derive LQR algorithm via dynamic programming and penalized LSs over a finite horizon.

> Link the problem to a system using Hamiltonian diagonalization for steady-state analysis.

K. Jiang, R. Wang, Y. Xiao, J. Jiang, X. Xu, and  T. Lu,  “Image enhancement via associated perturbation removal and texture reconstruction learning,” IEEE/CAA J. Autom. Sinica, vol. 11, no. 11, pp. 2253–2269, Nov. 2024. doi: 10.1109/JAS.2024.124521 

> Investigates the image enhancement tasks from a fresh perspective that involves the joint representation of perturbation removal, texture reconstruction and their association.

> Develops a PerTEM to associate degradation simulation and texture restoration, facilitating the learning capability while maintaining the model compactness.

> Experiments on various mainstream image enhancement tasks, such as image deraining, image dehazing and low-light image enhancement have demonstrated that PerTeRNet delivers competitive performance compared to the state-of-the-art method.

Z. Yin, J. Pu, Y. Zhou, and  X. Xue,  “Two-stage approach for targeted knowledge transfer in self-knowledge distillation,” IEEE/CAA J. Autom. Sinica, vol. 11, no. 11, pp. 2270–2283, Nov. 2024. doi: 10.1109/JAS.2024.124629 

> Propose a novel two-stage self-knowledge distillation approach for selective dark knowledge transfer.

> Generate class medoids from logit vectors to represent typical samples per class.

> Distill under-trained data using past predictions on half batch size.

Z. Zhao, Z. Yang, L. Jiang, J. Yang, and  Q. Ge,  “Privacy preserving distributed bandit residual feedback online optimization over time-varying unbalanced graphs,” IEEE/CAA J. Autom. Sinica, vol. 11, no. 11, pp. 2284–2297, Nov. 2024. doi: 10.1109/JAS.2024.124656 

> Differential privacy in distributed online optimization with precise noise control.

> Derives optimal prediction residual feedback boundedness, reducing estimation variance.

> Distributed algorithm with privacy and one-point feedback, handling unbalanced comms.

J. Zhang, B. Du, S. Zhang, and S. Ding, “A double sensitive fault detection filter for positive Markovian jump systems with a hybrid event-triggered mechanism,” IEEE/CAA J. Autom. Sinica, vol. 11, no. 11, pp. 2298–2315, Nov. 2024. doi: 10.1109/JAS.2024.124677 

> A non-monotonic adaptive triggering law is established for PMJSs.

> Asynchronous filters with double sensitivity are proposed for PMJSs.

> A simple analysis and design approach is presented by combining stochastic co-positive Lyapunov function and linear programming.

Z. Song and  P. Li,  “General Lyapunov stability and its application to time-varying convex optimization,” IEEE/CAA J. Autom. Sinica, vol. 11, no. 11, pp. 2316–2326, Nov. 2024. doi: 10.1109/JAS.2024.124374 

> The general Lyapunov stability criteria of nonlinear systems are proposed.

> A less conservative upper bound of settling-time function is provided.

> A fixed-time stable approach is raised for resolving TV convex optimization problem.

M. Yang, G. Liu, Z. Zhou, and  J. Wang,  “Probabilistic automata-based method for enhancing performance of deep reinforcement learning systems,” IEEE/CAA J. Autom. Sinica, vol. 11, no. 11, pp. 2327–2339, Nov. 2024. doi: 10.1109/JAS.2024.124818 

> Develop a novel framework that utilizes probabilistic automata to enhance DRL models.

> Implement reverse breadth-first search to identify and correct key weaknesses in DRL models. Improve the robustness of DRL models through targeted, minimal modifications based on identified vulnerabilities.

> Experiments in different environments verify the effectiveness of the framework in optimizing DRL for real-world industrial applications.

B. Yang, C. Tang, Y. Liu, G. Wen, and G. Chen, “A linear programming-based reinforcement learning mechanism for incomplete-information games,” IEEE/CAA J. Autom. Sinica, vol. 11, no. 11, pp. 2340–2342, Nov. 2024. doi: 10.1109/JAS.2024.124464 

J. Wang, W. Li, and X. Luo, “A distributed adaptive second-order latent factor analysis model,” IEEE/CAA J. Autom. Sinica, vol. 11, no. 11, pp. 2343–2345, Nov. 2024. doi: 10.1109/JAS.2024.124371 

Y. Liu, X. Wu, Y. Bo, J. Wang, and L. Ma, “A transfer learning framework for deep multi-agent reinforcement learning,” IEEE/CAA J. Autom. Sinica, vol. 11, no. 11, pp. 2346–2348, Nov. 2024. doi: 10.1109/JAS.2023.124173 

C.-C. Wang, Y.-L. Wang, and L. Jia, “Multi-USV formation collision avoidance via deep reinforcement learning and COLREGs,” IEEE/CAA J. Autom. Sinica, vol. 11, no. 11, pp. 2349–2351, Nov. 2024. doi: 10.1109/JAS.2023.123846 

Z.-H. Pang, Q. Cao, H. Guo, and Z. Dong, “Prediction-based state estimation and compensation control for networked systems with communication constraints and DoS attacks,” IEEE/CAA J. Autom. Sinica, vol. 11, no. 11, pp. 2352–2354, Nov. 2024. doi: 10.1109/JAS.2024.124605