A5100451327- Robotics and Sensor-Based Localization
- Coal Properties and Utilization
- Advanced Neural Network Applications
- Hydrocarbon exploration and reservoir analysis
- Tactile and Sensory Interactions
- Indoor and Outdoor Localization Technologies
- Gaze Tracking and Assistive Technology
- Rock Mechanics and Modeling
- Multimodal Machine Learning Applications
- Hydraulic Fracturing and Reservoir Analysis
- Domain Adaptation and Few-Shot Learning
- Drilling and Well Engineering
- Human Pose and Action Recognition
- 3D Surveying and Cultural Heritage
- Traffic Prediction and Management Techniques
- Infrastructure Maintenance and Monitoring
- Context-Aware Activity Recognition Systems
- Evacuation and Crowd Dynamics
- Video Surveillance and Tracking Methods
- Robotics and Automated Systems
- Smart Parking Systems Research
- Enhanced Oil Recovery Techniques
- Reservoir Engineering and Simulation Methods
- Geotechnical and Geomechanical Engineering
- Geoscience and Mining Technology
Clemson University
2019-2025
Henan University of Engineering
2018-2025
King Abdullah University of Science and Technology
2024-2025
Beijing Aerospace Flight Control Center
2023-2025
Xuzhou University of Technology
2024-2025
Pacific Northwest National Laboratory
2023
Southwest Petroleum University
2023
Shenyang Center for Disease Control and Prevention
2023
Henan Agricultural University
2022
Chinese Academy of Sciences
2021-2022
This paper presents a new holistic vision-based mobile assistive navigation system to help blind and visually impaired people with indoor independent travel. The detects dynamic obstacles adjusts path planning in real-time improve safety. First, we develop an map editor parse geometric information from architectural models generate semantic consisting of global 2D traversable grid layer context-aware layers. By leveraging the visual positioning service (VPS) within Google Tango device,...
This paper provides a framework for context-aware navigation services vision impaired people. Integrating advanced intelligence into requires knowledge of the semantic properties objects around user's environment. interaction is required to enhance communication about and places improve travel decisions. Our intelligent system human-in-the-loop cyber-physical that interprets ubiquitous entities by interacting with physical world cyber domain, viz., 1) visual cues distance sensing material as...
The concrete aging problem has gained more attention in recent years as bridges and tunnels the United States lack proper maintenance. Though Federal Highway Administration requires these public structures to be inspected regularly, on-site manual inspection by human operators is time-consuming labor-intensive. Conventional approaches for inspection, using RGB imagebased thresholding methods, are not able determine metric information well accurate location assessed defects conditions. To...
Pore is the main adsorption and desorption space of coalbed methane (CBM). size configuration connectivity affect adsorption/desorption hysteresis effect. Using tectonically deformed coal (TDC) original structure medium- high-rank as research objects, through N 2 /CO experiment to analyze pore distribution different scales. We investigate control mechanism heterogeneous evolution in key scales against characteristics during metamorphism deformation by combining CH 4 isothermal under 30°C...
Abstract Human‐made concrete structures require cutting‐edge inspection tools to ensure the quality of construction meet applicable building codes and maintain sustainability aging infrastructure. This paper introduces a wall‐climbing robot for metric that can reach difficult‐to‐access locations with close‐up view visual data collection real‐time flaws detection localization. The is able detect surface (i.e., cracks spalls) produce defect‐highlighted 3D model extracted location clues...
Current metrics for video captioning are mostly based on the text-level comparison between reference and candidate captions. However, they have some insuperable drawbacks, e.g., cannot handle videos without references, may result in biased evaluation due to one-to-many nature of video-to-text neglect visual relevance. From human evaluator's viewpoint, a high-quality caption should be consistent with provided video, but not necessarily similar literal or semantics. Inspired by evaluation, we...
We present a novel collaborative mapping and autonomous parking system for semi-structured multi-story garages, based on cooperative 3-D LiDAR point cloud registration Bayesian probabilistic updating. First, an inertial-enhanced (IE) generalized iterative closest (G-ICP) approach is presented to perform high accuracy odometry, which loosely coupled with inertial measurement unit using multi-state extended Kalman filter fusion. Second, the IE G-ICP utilized reconstruct model each vehicle,...
Coalbed methane wells on the ground in coal mining areas are used to extract a large amount of gas from working face, but activities can cause strata movement, resulting casing deformation. Numerical simulation was study shear stress loading under thick alluvial conditions. The found that will have peak points at interface between aquifer, layer and bedrock, mudstone with intensity 49.9–77.1 MPa. point aquifer appear 4 m below floor as face is mined. junction bedrock increases linearly...
In an aerospace aircraft servo system, the modeling and analysis of permanent magnet synchronous motor (PMSM) faces great challenges due to its nonlinear, multi-variable strong coupling characteristics. To enhance effectiveness, this study introduces a Simulink-based approach. The mechanical model control drive PMSM were constructed by modular modeling, simulation was established on Simulink. accuracy stability establishment verified signal injection under typical working conditions....
Research on the dynamic mechanical properties of concrete under sulfate attack is basis for reasonable design and performance evaluation anti-explosion impact resistance in structures erosion environment. In this study, specimens subjected to compression were measured using split Hopkinson pressure bar (SHPB) test system. The basic (stress-strain curve, compressive strength, elastic modulus, peak strain) subject high strain rate obtained, variation laws macroscopic failure characteristics...
<div class="section abstract"><div class="htmlview paragraph">Towards the goal of real-time navigation autonomous robots, Iterative Closest Point (ICP) based LiDAR odometry methods are a favorable class Simultaneous Localization and Mapping (SLAM) algorithms for their robustness under any light conditions. However, even with recent methods, traditional SLAM challenges persist, where drifts adversarial conditions such as featureless or dynamic environments, well high motion...
In this paper, we exploit the concrete surface flaw inspection through fusion of visual positioning and semantic segmentation approach. The fused result is represented by a 3D metric map with spatial area, width, depth information, which shows advantage over general in image space without info. We also relieve human labor an automatic labeling system composed three hybrid parts: to enable pose association, crack/spalling using deep neural network (pixel level), random field filter for...
This paper proposes a novel assistive navigation system based on simultaneous localization and mapping (SLAM) semantic path planning to help visually impaired users navigate in indoor environments. The integrates multiple wearable sensors feedback devices including RGB-D sensor an inertial measurement unit (IMU) the waist, head mounted camera, microphone earplug/speaker. We develop visual odometry algorithm data estimate user's position orientation, refine orientation error using IMU. employ...
The diffusion coefficient is a crucial parameter to manifest the diffusivity of methane in coal, which obtained by solving mathematical and physical equations that characterize process. model established combining pore structure parameters time-dependent characteristics Euclidean space cannot reflect natural heterogeneity coal. There fundamental difference between fractal space. Based on classic homogeneous spherical Fick coal particles, full-aperture comprehensive dimension decay are...
This paper presents a novel metric inspection robot system using deep neural network to detect and measure surface flaws (i.e., crack spalling) on concrete structures performed by wall-climbing robot. The consists of four modules: robotics data collection module obtain RGB-D images IMU measurement, visual-inertial SLAM generate pose coupled key-frames with depth information, InspectionNet classify each pixel into three classes (back-ground, spalling), 3D registration map fusion register the...
This paper proposes a real-time multi-scale semantic segmentation network (MsNet). MsNet is combination of our novel fusion with matching attention model (MFMA) as the decoding and searched by asymptotic neural architecture search (ANAS) or MobileNetV3 encoding network. The MFMA not only extracts low level spatial features from inputs but also decodes contextual extracted ANAS. Specifically, considering advantages disadvantages addition concatenation fusion, we design (MF) that balances...
Accurate localization on the autonomous driving cars is essential for autonomy and safety, especially complex urban streets search-and-rescue subterranean environments where high-accurate GPS not available. However current odometry estimation may introduce drifting problems in long-term navigation without robust global localization. The main challenges involve scene divergence under interference of dynamic effective perception observation object layout variance from different viewpoints. To...
This paper presents SmartCane - the CCNY Smart Cane system, a robotic white cane and mobile device navigation software for visually impaired people. The system includes Google Tango devices that utilizes simultaneous localization mapping (SLAM) to plan path guide user waypoints within indoor environments. A control panel is mounted on standard enables users communicate with additionally used provide instructions via haptic feedback. Based motion-tracking capabilities of Tango, able generate...