A5063033361- Robotics and Sensor-Based Localization
- Modular Robots and Swarm Intelligence
- Robotic Path Planning Algorithms
- Planetary Science and Exploration
- Advanced Neural Network Applications
- Spectroscopy and Chemometric Analyses
- Advanced Materials and Mechanics
- Video Surveillance and Tracking Methods
- Advanced Image and Video Retrieval Techniques
- Spectroscopy Techniques in Biomedical and Chemical Research
- Soft Robotics and Applications
- Fatigue and fracture mechanics
- Astro and Planetary Science
- Space Satellite Systems and Control
- Microstructure and Mechanical Properties of Steels
- Frequency Control in Power Systems
- Robot Manipulation and Learning
- Image and Object Detection Techniques
- Additive Manufacturing and 3D Printing Technologies
- Advanced Vision and Imaging
- Injection Molding Process and Properties
- Laser-induced spectroscopy and plasma
- Adaptive Control of Nonlinear Systems
- High Temperature Alloys and Creep
- Infrared Target Detection Methodologies
Jilin University
2019-2024
Ministry of Education of the People's Republic of China
2023-2024
Jilin Medical University
2022-2024
Harbin Institute of Technology
1988-2021
Beijing Academy of Artificial Intelligence
2021
École Polytechnique Fédérale de Lausanne
2018
The University of Texas at Arlington
1998-2002
Rubber Research Institute
1999
Modeling the quantitative relationship between target components and measured spectral information is an essential part of laser-induced breakdown spectroscopy (LIBS) analysis. However, many traditional multivariate analysis algorithms must reduce dimension or extract characteristic lines in advance, which may result loss reduced accuracy. Indeed, improving precision interpretability LIBS a critical challenge Mars exploration. To solve this problem, paper proposes end-to-end lightweight...
Semantic segmentation of Mars scenes has a crucial role in rovers science missions. Current CNN-based composition U-Net powerful information extraction capabilities; however, convolutional localization suffers from the limited global context modeling capability. Although Transformer performed well, it still encounters obstacles and retention low-level features. This issue is particularly relevant for Martian rocks with their varying shapes, textures, sizes scenes. In this paper, we propose...
Martian rock segmentation aims to separate pixels from background, which plays a crucial role in downstream tasks, such as traversing and geologic analysis by Mars rovers. The U-Nets have achieved certain results segmentation. However, due the inherent locality of convolution operations, are inadequate modeling global context long-range spatial dependencies. Although emerging Transformers can solve this, they suffer difficulties extracting retaining sufficient low-level local information....
With the increase in number of types spectrometers use, calibration models cannot be shared among different instruments, however, this problem can solved via transfer (CT). In study, a variety modern process analysis technology (PAT) data are taken as research object. After preprocessing spectra using principal component (PCA) and cubic spline interpolation, TrAdaBoost algorithm learning combined with extreme machine (ELM), i.e., TrAdaBoost-ELM, is used to master model slave instruments make...
Martian surface, as a typical unstructured terrain, is extremely challenging for Mars exploration missions. Commonly, rovers require multiple sensors to explore such harsh environment, depth cameras, range finder and other devices. However, the onboard load, power storage of are not sufficient achieve high-level stereoscopic perception, which can be adverse downstream tasks visual navigation scientific exploration. To this end, in paper we propose awareness perception light-weight framework...
Presented in this paper are the design and simulation results of a new AGC logic. The logic is specifically designed to work under NERC's Control Performance Standard Disturbance with improved effectiveness efficiency. old Criteria briefly described compared. considerations constructing environment included. Texas Utilities Electric Company (TU), joint effort University at Arlington, has begun implementation
Reconfigurability in versatile systems of modular robots is achieved by changing the morphology overall structure as well connecting and disconnecting modules. Recurrent connectivity changes can cause misalignment that leads to mechanical failure system. This paper presents a new approach reconfiguration, inspired art origami, eliminates during transformation. Our method consists an energy-optimal reconfiguration planner generates initial 2D assembly pattern actuation sequence units, both...
An accurate, nondestructive, and low-cost measurement system was developed using a portable near-infrared (NIR) spectrometer (DLP NIRscan Nano), Raspberry Pi board, display, lithium battery, self-made three-dimensional printed shell. NIR data were collected through two modes (column Hadamard transform) based on digital light processing. With this equipment, detection models of soluble solid content (SSC) firmness, essential quality indicators the fruit, established via quantitative analysis...
Martian terrain segmentation aims to assign all pixels of an input image with various labels, which provides a firm support for the downstream research on rover traversing and geologic analysis tasks. However, existing studies in this field suffer from limitations two aspects: one is lack large-scale high-quality datasets, other over-reliance purely supervised learning that very data-hungry sensitive domain shifts among different datasets. In article, we overcome these perspective both data...
In this paper, we propose a multi-object tracking framework for videos captured by UAVs, considering motion imperfection in the following two aspects: 1) blurring of objects due to high-speed UAV and objects, deteriorating performance detector; 2) coupling global movement camera with object motion, resulting nonlinearity trajectories adjacent frames further more difficult predict. For blurring, paper proposes hybrid deblurring module that deals blurred while retaining clear frames, trading...
This article presents two kernel-based rock detection methods for a Mars rover. Rock on planetary surfaces is particularly pivotal vehicles regarding navigation and obstacle avoidance. However, the diverse morphologies of Martian rocks, sparsity pixel-wise features, engineering constraints are great challenges to current object methods, resulting in inaccurate delayed location recognition. We therefore propose region-wise framework design algorithms, kernel principle component analysis...
Modular robots hold the promise of changing their shape and even dimension to adapt various tasks environments.To realize this superiority, it is essential find appropriate morphology its corresponding behavior simultaneously ensure optimality reconfiguration.However, achieving co-optimization challenging because robotic configuration motion are interactive coupled with each other, as well optimization processes.To end, we proposed a framework based on hierarchical Deep Reinforcement...