Article
Robotics
Zhepei Wang, Xin Zhou, Chao Xu, Fei Gao
Summary: In this article, an optimization-based framework for multicopter trajectory planning under geometrical configuration constraints and user-defined dynamic constraints is presented. The framework utilizes a novel trajectory representation based on optimality conditions for control effort minimization. Linear-complexity operations are designed to achieve spatial-temporal deformation for different planning requirements. The framework supports the elimination of geometrical constraints through smooth maps, and is able to handle a variety of state-input constraints through decoupling and backward differentiation. The framework transforms a generally constrained multicopter planning problem into an unconstrained optimization, resulting in reliable and efficient solutions.
IEEE TRANSACTIONS ON ROBOTICS
(2022)
Article
Computer Science, Interdisciplinary Applications
Xin Yang, Yan Ran, Genbao Zhang, Hongwei Wang, Zongyi Mu, Shengguang Zhi
Summary: This study introduces a hybrid approach framework driven by digital twin technology for predicting performance degradation of transmission systems. By combining data-driven and model-driven methods, real-time data is utilized to update state estimations in order to enhance prediction accuracy.
ROBOTICS AND COMPUTER-INTEGRATED MANUFACTURING
(2022)
Article
Computer Science, Artificial Intelligence
Zhe Chen, Xiao-Jun Wu, Josef Kittler
Summary: This paper proposes a Fisher regularized e-dragging framework for image classification, which improves the intraclass compactness and interclass separability of relaxed labels. The Fisher criterion and e-dragging technique are integrated into a unified model, achieving superior performance compared to other classification methods.
IEEE TRANSACTIONS ON COGNITIVE AND DEVELOPMENTAL SYSTEMS
(2023)
Article
Computer Science, Interdisciplinary Applications
Yuxin Li, Wenbin Gu, Minghai Yuan, Yaming Tang
Summary: This paper addresses the dynamic flexible job shop scheduling problem with insufficient transportation resources using deep reinforcement learning. A hybrid deep Q network is developed for this problem, showing superiority and generality compared with current optimization-based approaches through comprehensive experiments.
ROBOTICS AND COMPUTER-INTEGRATED MANUFACTURING
(2022)
Review
Automation & Control Systems
Joshua Pinskier, David Howard
Summary: The emerging field of soft robotics presents a new paradigm for robot design where precision through rigidity is replaced by cognition through compliance. Soft robots, like octopuses, have flexible bodies that can interact with fragile objects and navigate unstructured environments. Currently, the lack of efficient modeling tools means soft robots are primarily designed by hand, but autonomous design methodologies are urgently required for high-performing robots.
ADVANCED INTELLIGENT SYSTEMS
(2022)
Article
Robotics
Bin Fang, Fuchun Sun, Linyuan Wu, Fukang Liu, Xiangxiang Wang, Haiming Huang, Wenbing Huang, Huaping Liu, Li Wen
Summary: This study proposes a soft gripper with multiple grasping modes, utilizing modular soft fingers to achieve various grasping methods, and validating their effectiveness through finite element simulations. The soft gripper features multimode grasping capability, employing different control modes to realize four anthropomorphic grasping modes.
Article
Computer Science, Interdisciplinary Applications
Yuzhang Wei, Qingsong Xu
Summary: This research presents a novel constant-force mechanism end-effector design for robotic polishing to address issues of force overshoot and poor force accuracy. The end-effector regulates contact force passively through robotic control, ensuring consistency in workpiece surface quality.
ROBOTICS AND COMPUTER-INTEGRATED MANUFACTURING
(2022)
Article
Engineering, Multidisciplinary
Jie Xing, Hanli Zhao, Huiling Chen, Ruoxi Deng, Lei Xiao
Summary: In this work, an improved Whale Optimization Algorithm (QGBWOA) is proposed to address the problems of falling into local optimum and slow convergence. Quasi-opposition-based learning and Gaussian barebone mechanism are introduced to enhance the searching ability and diversity of WOA. Experimental results on benchmark datasets demonstrate the significantly improved convergence accuracy and speed of QGBWOA. Furthermore, applications in feature selection and multi-threshold image segmentation validate its capability in solving complex real-world problems.
JOURNAL OF BIONIC ENGINEERING
(2023)
Article
Computer Science, Artificial Intelligence
Ke Wang, Sai Ma, Junlan Chen, Fan Ren, Jianbo Lu
Summary: This article provides an in-depth analysis of recent advances in deep learning-based Visual Odometry (Deep VO) and evaluates its performance improvements in depth estimation, feature extraction and matching, and pose estimation. It also summarizes the application areas of Deep VO in mobile robots, medical robots, augmented and virtual reality. Through literature analysis and comparison, it highlights several open issues and suggests future research directions in this field.
IEEE TRANSACTIONS ON COGNITIVE AND DEVELOPMENTAL SYSTEMS
(2022)
Article
Robotics
Dylan S. Shah, Joran W. Booth, Robert L. Baines, Kun Wang, Massimo Vespignani, Kostas Bekris, Rebecca Kramer-Bottiglio
Summary: Advancements in tensegrity robotics, inspired by biological principles, allow robots to change shape by adjusting internal tension; various design and simulation techniques enable a wide range of locomotion modes; emerging challenges include automated design, state sensing, and kinodynamic motion planning.
Article
Robotics
Mengdi Han, Xiaogang Guo, Xuexian Chen, Cunman Liang, Hangbo Zhao, Qihui Zhang, Wubin Bai, Fan Zhang, Heming Wei, Changsheng Wu, Qinghong Cui, Shenglian Yao, Bohan Sun, Yiyuan Yang, Quansan Yang, Yuhang Ma, Zhaoguo Xue, Jean Won Kwak, Tianqi Jin, Qing Tu, Enming Song, Ziao Tian, Yongfeng Mei, Daining Fang, Haixia Zhang, Yonggang Huang, Yihui Zhang, John A. Rogers
Summary: The research introduces a manufacturing and actuation approach for creating small robots with complex 3D geometries and heterogeneous material construction. Mechanical buckling and shape memory alloy, together with elastic resilience of a shell, enable reversible deformations. The robots can perform bending, twisting, expansion, crawling, walking, turning, and jumping through global heating and laser-induced local thermal actuation.
Article
Computer Science, Artificial Intelligence
Xin Ning, Shaohui Xu, Fangzhe Nan, Qingliang Zeng, Chen Wang, Weiwei Cai, Weijun Li, Yizhang Jiang
Summary: This article proposes a new face editing approach called IricGAN, which can effectively control the attribute intensity while retaining the face's identity and semantic information. It uses a learnable hierarchical feature combination (HFC) function and an attribute regression module (ARM) to achieve gradual modification of face attributes.
IEEE TRANSACTIONS ON COGNITIVE AND DEVELOPMENTAL SYSTEMS
(2023)
Article
Automation & Control Systems
Vishal Patel, Austin Chesmore, Christopher M. Legner, Santosh Pandey
Summary: The workplace has a significant impact on workers' safety, health, and productivity. Technological advancements in wearables are being used to monitor and manage occupational hazards, with a focus on safety, productivity, and health aspects. These tools provide real-time visibility and decision support for frontline workers, improving efficiency and safety compliance.
ADVANCED INTELLIGENT SYSTEMS
(2022)
Article
Robotics
Chunge Bai, Tao Xiao, Yajie Chen, Haoqian Wang, Fang Zhang, Xiang Gao
Summary: This letter presents an incremental voxel-based lidar-inertial odometry (LIO) method for fast-tracking spinning and solid-state lidar scans. By using iVox as the point cloud spatial data structure, the method achieves high tracking speed without the need for complicated tree-based structures or strict k-nearest neighbor queries.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2022)
Article
Automation & Control Systems
Zhihua Chen, Jiehao Li, Shoukun Wang, Junzheng Wang, Liling Ma
Summary: This article proposes a hierarchical control framework with behavior rules for legged stable walking of hexapod wheel-legged robots in unstructured terrain. The framework includes a flexible gait planner and a gait feedback regulator, which can adapt to different terrains by adjusting the foot-end trajectory according to terrain feedback information.
ROBOTICS AND AUTONOMOUS SYSTEMS
(2022)
Article
Robotics
Zheng Huai, Guoquan Huang
Summary: In this paper, a novel robocentric formulation of the visual-inertial navigation system (VINS) is proposed, which achieves consistent motion tracking in challenging environments using only a monocular camera and a six-axis inertial measurement unit (IMU). By reformulating the VINS with respect to a moving local frame, the proposed algorithm obtains higher accuracy relative motion estimates for updating global pose and improves robustness and performance.
INTERNATIONAL JOURNAL OF ROBOTICS RESEARCH
(2022)
Article
Robotics
Erina Baynojir Joyee, Adam Szmelter, David Eddington, Yayue Pan
Summary: This study presents a 3D-printed soft robot with varied material compositions and hierarchical surface structures. The robot exhibits superior locomotion capabilities and versatile functionalities, making it suitable for tasks in harsh environments.
Article
Robotics
K. Park, H. Yuk, M. Yang, J. Cho, H. Lee, J. Kim
Summary: This paper presents a biomimetic robotic skin based on hydrogel-elastomer hybrids and tomographic imaging. The developed skin has softness and resilience similar to human skin, enabling safe, intuitive, and contact-rich human-robot interaction. By measuring resistance changes and touch vibrations through electrodes and microphones, the data is converted into multimodal tactile information.
Article
Computer Science, Interdisciplinary Applications
Liang Guo, Fu Yan, Tian Li, Tao Yang, Yuqian Lu
Summary: The traditional construction of process knowledge base is non-automated and time-consuming, which requires manual work and may lead to ambiguity in knowledge representation. This paper introduces an automatic construction framework based on knowledge graph (KG), which involves steps like classification, annotation, extraction, and representation to improve the efficiency and quality of the process knowledge base.
ROBOTICS AND COMPUTER-INTEGRATED MANUFACTURING
(2022)
Article
Robotics
Chuhua Wang, Yuchen Wang, Mingze Xu, David J. Crandall
Summary: In this paper, a recurrent network named SGNet is proposed for trajectory prediction by estimating and utilizing the goals of observed agents at multiple time scales. The model achieves state-of-the-art results on various datasets, demonstrating its effectiveness in improving accuracy.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2022)
