Article
Automation & Control Systems
Ying Zhang, Guohui Tian, Xuyang Shao, Shaopeng Liu, Mengyang Zhang, Peng Duan
Summary: In this article, an effective solution for efficient robotic object search is presented by leveraging metric-topological map. By constructing a novel metric-topological map and introducing a rapid adjustment method for the robot's viewing angle, the search efficiency is improved, achieving humanlike search behavior.
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2022)
Article
Computer Science, Artificial Intelligence
Ozan Catal, Tim Verbelen, Toon Van de Maele, Bart Dhoedt, Adam Safron
Summary: This paper explores an active inference navigation approach based on a hierarchical generative model, demonstrating consistency with hippocampal function models and implementation on real-world robots. Experimental results show that robots equipped with this model can generate consistent maps and infer correct navigation behavior when a goal location is provided to the system.
Article
Robotics
Jianxian Cai, Fenfen Yan, Yan Shi, Mengying Zhang, Lili Guo
Summary: In this paper, a hierarchical cognitive navigation model (HCNM) is proposed to enhance the self-learning and self-adaptive ability of mobile robots in unknown and complex environments. The HCNM model divides the path planning task into different levels of sub-tasks and solves each sub-task in a smaller state subspace to reduce the dimensions of the state space. Experimental results demonstrate that the HCNM model exhibits strong adaptability and faster convergence time in unknown environments.
Article
Automation & Control Systems
Yuanyang Zhu, Zhi Wang, Chunlin Chen, Daoyi Dong
Summary: This article introduces a rule-based reinforcement learning (RuRL) algorithm for efficient navigation. By employing a wall-following rule to generate a closed-loop trajectory, a reduction rule to shrink the trajectory, and the Pledge rule to guide the exploration strategy, RuRL achieves improved navigation performance in real robot navigation experiments.
IEEE-ASME TRANSACTIONS ON MECHATRONICS
(2022)
Article
Robotics
Fangwen Yu, Yujie Wu, Songchen Ma, Mingkun Xu, Hongyi Li, Huanyu Qu, Chenhang Song, Taoyi Wang, Rong Zhao, Luping Shi
Summary: The research report introduces a brain-inspired general place recognition system called NeuroGPR, which enables robots to recognize places in natural environments by mimicking the neural mechanism of multimodal sensing, encoding, and computing. The system utilizes a multimodal hybrid neural network to encode and integrate cues from different sensors, and a multiscale liquid state machine to process and fuse the information. Experimental results show that NeuroGPR performs well in various environmental conditions.
Article
Engineering, Multidisciplinary
Yong Hun Kim, Hak Ju Kim, Joo Han Lee, San Hee Kang, Eung Ju Kim, Jin Woo Song
Summary: This paper presents a new magnetic map-matching algorithm based on sequential batch fusion, which allows for real-time computation and more stable indoor localization of mobile robots. The algorithm addresses the limitations of conventional magnetic map-matching localization, such as the inability to conduct real-time navigation using a batch process and the occurrence of navigation failures or significant localization errors due to magnetic field distortion. Through weight convergence and magnetic field similarity, the proposed algorithm achieves measurement updates and demonstrates superior performance compared to conventional methods, with a position error of only 1 m in a path where the conventional method fails. As a result, it offers a cost-effective alternative to other sensor-based indoor navigation systems.
Article
Chemistry, Analytical
Fan Wang, Chaofan Zhang, Wen Zhang, Cuiyun Fang, Yingwei Xia, Yong Liu, Hao Dong
Summary: This paper proposes a novel object-level topological visual navigation method that improves the reliability and efficiency of visual navigation through the construction of a topological semantic map and object guidance. Experimental results demonstrate the feasibility and superiority of this method.
Article
Computer Science, Information Systems
Kirill Muravyev, Konstantin Yakovlev
Summary: This study evaluates five open-source topological mapping methods and compares them using novel metrics. The results show that each method has its own advantages and drawbacks, and none of them builds a graph suitable for navigation out of the box.
Article
Engineering, Civil
Zhe Liu, Yu Zhai, Jiaming Li, Guangming Wang, Yanzi Miao, Hesheng Wang
Summary: This paper addresses the problem of mobile robot autonomous navigation in large-scale environments with crowded dynamic objects. Deep reinforcement learning and graph neural networks are used to improve the performance in crowded environments. Simulation and physical experiments validate the effectiveness and applicability of the proposed method in various environments.
IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Dmitrii Dobriborsci, Roman Zashchitin, Mikhail Kakanov, Wolfgang Aumer, Pavel Osinenko
Summary: The application of reinforcement learning in mobile robotics faces challenges in real-world physical environments. This paper presents a local navigation approach for driving a robot without relying on an explicit map, using only laser scan measurements to detect obstacles. An algorithm called stacked Q-learning is proposed, which combines standard reinforcement learning techniques with model-based predictive agents. The results show that the stacked Q-learning algorithm outperforms classical model predictive control in terms of accumulated cost of parking the robot while avoiding obstacles.
JOURNAL OF INTELLIGENT MANUFACTURING
(2023)
Article
Computer Science, Interdisciplinary Applications
Lu Chang, Liang Shan, Weilong Zhang, Yuewei Dai
Summary: This paper presents a hierarchical framework for multi-robot navigation and formation in unknown environments. It includes a learning network, distributed optimization, and velocity adjustment methods. Through these modules, robots can navigate in unknown environments and maintain an optimal formation.
ROBOTICS AND COMPUTER-INTEGRATED MANUFACTURING
(2023)
Article
Engineering, Marine
Zixiao Zhu, Lichuan Zhang, Lu Liu, Dongwei Wu, Shuchang Bai, Ranzhen Ren, Wenlong Geng
Summary: This study proposes a master-slave multi-AUV collaborative navigation method based on hierarchical reinforcement learning. The collaborative navigation system is modeled as a discrete semi-Markov process, and trajectory planning is performed using hierarchical reinforcement learning combined with polar Kalman filter to reduce the positioning error of slave AUVs and enable collaborative navigation.
JOURNAL OF MARINE SCIENCE AND ENGINEERING
(2023)
Article
Engineering, Chemical
Min-Fan Ricky Lee, Sharfiden Hassen Yusuf
Summary: This paper proposes an end-to-end approach that uses deep reinforcement learning for autonomous mobile robot navigation in an unknown environment. The mobile robot can learn collision avoidance and navigation capabilities using deep Q-network and double deep Q-network agents. The simulation and real-world experiments show that the robot can autonomously navigate and reach the target object location without colliding with obstacles.
Article
Computer Science, Information Systems
Muhammad Husnain Haider, Zhonglai Wang, Abdullah Aman Khan, Hub Ali, Hao Zheng, Shaban Usman, Rajesh Kumar, M. Usman Maqbool Bhutta, Pengpeng Zhi
Summary: This paper proposes an adaptive neuro-fuzzy inference system (ANFIS) and global positioning system (GPS) for collision-free navigation of mobile robots. The proposed method automates robot navigation and obstacle avoidance, and controls global path planning and steering using GPS and sensor data fusion. This approach outperforms various state-of-the-art navigation and obstacle avoidance methods in finding an optimal path in unknown environments.
JOURNAL OF KING SAUD UNIVERSITY-COMPUTER AND INFORMATION SCIENCES
(2022)
Review
Computer Science, Information Systems
Kai Zhu, Tao Zhang
Summary: This paper discusses the application of Deep Reinforcement Learning (DRL) in mobile robot navigation, compares the relationships and differences between four typical application scenarios, describes the development of DRL-based navigation, and addresses the challenges and possible solutions facing DRL-based navigation.
TSINGHUA SCIENCE AND TECHNOLOGY
(2021)
Article
Computer Science, Software Engineering
Xinyi Yu, Hanxiong Li, Haidong Yang
Summary: This paper proposes an algorithm for low-light image enhancement that addresses the issues of noise and color distortion by integrating denoising and color restoration. Experimental results demonstrate that this method greatly improves the quality of low-light images.
Article
Computer Science, Artificial Intelligence
Qian-Qian Li, Zi-Peng Wang, Tingwen Huang, Huai-Ning Wu, Han-Xiong Li, Junfei Qiao
Summary: This article addresses fault-tolerant stochastic sampled-data fuzzy control for nonlinear delayed parabolic PDE systems under spatially point measurements. A T-S fuzzy PDE model is used to accurately describe the system. A fault-tolerant SD fuzzy controller with stochastic sampling is designed considering possible actuator failure, and a novel time-dependent Lyapunov functional is constructed to obtain sufficient conditions for the mean square exponential stability of the closed-loop system based on linear matrix inequalities. The effectiveness of the designed approach is illustrated through three examples.
IEEE TRANSACTIONS ON FUZZY SYSTEMS
(2023)
Article
Automation & Control Systems
Peng Wei, Han-Xiong Li
Summary: In this article, a spatiotemporal entropy method is proposed to detect and locate thermal abnormalities of Li-ion battery (LIB) packs. The spatial entropy and temporal entropy are constructed from different scales based on Karhunen-Loeve (KL) decomposition, and then integrated into the comprehensive spatiotemporal entropy. The kernel density estimation is used to derive the detection threshold, and the entropy contribution function is designed for abnormality localization. Experimental results demonstrate the effectiveness of the proposed method in timely detecting and precisely locating abnormal cells at the early stage.
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2023)
Article
Engineering, Electrical & Electronic
Lei Lei, Han-Xiong Li, Hai-Dong Yang
Summary: This article proposes a multiscale convolution-based detection methodology for classifying defects in bare printed circuit boards (PCBs) under uncertainty. A novel window-based loss function is designed to tackle inter-class imbalance and uncertainty. A multiscale convolution network is constructed to process defects with intra-class variance, and large scale extraction features are fused on the small scale to guide the extraction process. Experimental studies demonstrate that the proposed methodology achieves satisfactory detection performance and visual interpretability compared to baseline methods.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
Article
Engineering, Electrical & Electronic
Xinyi Yu, Han-Xiong Li, Haidong Yang
Summary: Surface defect detection of printed circuit boards (PCBs) is a critical stage in ensuring product quality. Existing defect detection methods using deep learning models are limited by image uncertainty and label uncertainty. This paper proposes a novel collaborative learning classification model that addresses these difficulties. Results show that the proposed model achieves excellent performance on various quantitative metrics.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
Article
Automation & Control Systems
Yun Feng, Yaonan Wang, Qin Wan, Xiaogang Zhang, Bing-Chuan Wang, Han-Xiong Li
Summary: So far, fault detection for distributed parameter systems (DPSs) has been mostly model-based and heavily reliant on prior known model information, limiting their usability in industrial applications. In this article, a brand-new framework is proposed for online systems modeling and fault detection of unknown high-dimensional DPSs. The framework includes an interaction between the two parts, where the systems modeling error is transformed into residual signals for fault detection and the online modeling switches to offline mode based on fault-detection results. The effectiveness of the proposed method is validated through experiments on sensor fault diagnosis for the thermal process of a 2-D battery cell.
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2023)
Article
Automation & Control Systems
Peng Wei, Han-Xiong Li
Summary: In this article, a spatio-temporal inference system is proposed to detect and locate thermal abnormalities of battery systems. The system includes three modules: spatio-temporal processing module, abnormality inference module, and spatial inference module. By analyzing the distributed temperatures on the battery system, the monitoring statistic is developed in the spatio-temporal processing module. The abnormality inference module detects abnormalities based on the derived statistic index. The spatial Bayes model is designed to estimate the abnormality location. The experiments demonstrate that the proposed system can detect and locate internal short circuit faults before they result in thermal runaways.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Automation & Control Systems
Yu Zhou, Hua Deng, Han-Xiong Li
Summary: This work develops a Galerkin-spectral model for the 3-D thermal diffusion in pouch cells, which is essential for onboard temperature monitoring of lithium-ion batteries in automobile applications. The model is obtained by space decomposition and simplification of the original physical model, and a low-order representation is extracted using spectral expansion. Experimental and simulation studies demonstrate the satisfactory reduced-order performance and practical validity of the proposed model.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2023)
Article
Automation & Control Systems
Zhijia Zhao, Yiming Liu, Tao Zou, Keum-Shik Hong, Han-Xiong Li
Summary: In this study, a novel adaptive fault-tolerant control strategy is proposed to address the vibration issues in marine risers, ensuring system stability and performance.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Article
Automation & Control Systems
Yiwei Chen, Yu Pan, Daoyi Dong
Summary: This article proposed a neural network model with an entanglement embedding module for quantum language models. The model explicitly captures the nonclassical correlations within word sequences and achieves superior performance on question answering tasks.
IEEE TRANSACTIONS ON CYBERNETICS
(2023)
Article
Automation & Control Systems
Xian-Bing Meng, C. L. Philip Chen, Han-Xiong Li
Summary: In industrial applications, the modeling and online applications of distributed parameter systems (DPSs) are difficult due to their infinite dimension, spatiotemporal coupled dynamics, nonlinearity, and model uncertainties. To address these issues, an online spatiotemporal modeling method is proposed based on confidence-aware multiscale learning. The proposed method integrates evolutionary learning-based spatial basis function, efficient broad learning system for temporal dynamics, and Gaussian process regression for spatiotemporal-scale learning to enable online confidence-aware prediction for DPSs. Experiments based on the curing process in snap curing oven demonstrate the effectiveness of the proposed method.
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2023)
Article
Automation & Control Systems
Zhijia Zhao, Yiming Liu, Ge Ma, Keum-Shik Hong, Han-Xiong Li
Summary: In this article, a new adaptive fuzzy fault-tolerant control (FTC) is proposed for a three-dimensional riser-vessel system with unknown backlash nonlinearity. A model for the smooth inverse dynamics of the backlash is introduced and the control input is divided into an expected input and a compensation error. Fuzzy adaptive technology is employed to achieve compensation considering the imprecision of system modeling and unknown external disturbances. The simulation results demonstrate the effectiveness of the derived scheme.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2023)
Article
Automation & Control Systems
Zhijia Zhao, Yiming Liu, Sentao Cai, Zhifu Li, Yiwen Wang, Keum-Shik Hong, Han-Xiong Li
Summary: This article proposes an adaptive control method for a flexible manipulator to deal with distributed disturbances, unknown dead zones, and input quantization. The unknown dead zone and input quantization are formulated and represented based on essential transformations. An adaptive robust quantized control with online updating laws is developed to ensure robustness, angle position, and reduce vibration. The Lyapunov theoretical analysis is employed to ensure bounded stability. Numerical simulations and experiments are conducted to verify the feasibility and superiority of the proposed method using a Quanser platform.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Yun Feng, Yaonan Wang, Yang Mo, Yiming Jiang, Zhijie Liu, Wei He, Han-Xiong Li
Summary: Fault detection for distributed parameter systems (DPSs) generally requires the complete model information to be known. However, it is common that accurate first-principles physical models are difficult to obtain for many industrial applications. Therefore, the applicability of traditional model-based methods is limited. In this study, an adaptive neural network (AdNN) is constructed to estimate the state variable and the unknown nonlinearity for a class of partially known nonlinear DPSs. Experimental results validate the effectiveness of the proposed method.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Engineering, Electrical & Electronic
Yaxin Wang, Han-Xiong Li, Shengli Xie
Summary: In this article, a spatial model predictive control (MPC) approach is proposed for a nonlinear distributed parameter system (DPS). A data-driven modeling method is utilized to predict the system performance, and a dual adaptation approach is developed to capture the most recent dynamics. Theoretical analysis guarantees stability, and simulations/experiments demonstrate the effectiveness of the proposed method.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
