A5101897449- Robotics and Sensor-Based Localization
- Advanced Image and Video Retrieval Techniques
- 3D Shape Modeling and Analysis
- Human Pose and Action Recognition
- 3D Surveying and Cultural Heritage
- Smart Grid Security and Resilience
- Advanced Vision and Imaging
- Smart Grid and Power Systems
- HVDC Systems and Fault Protection
- Robotic Locomotion and Control
- Multimodal Machine Learning Applications
- Smart Grid Energy Management
- Electric Vehicles and Infrastructure
- Medical Image Segmentation Techniques
- Image Processing and 3D Reconstruction
- Optical measurement and interference techniques
- Gait Recognition and Analysis
- Innovative Microfluidic and Catalytic Techniques Innovation
- Advanced Battery Technologies Research
- Structural Health Monitoring Techniques
- Power Systems Fault Detection
- Electric Power Systems and Control
- Geotechnical Engineering and Underground Structures
- Surgical Simulation and Training
- Machine Fault Diagnosis Techniques
University of Birmingham
2021-2024
Arup Group (United Kingdom)
2024
University of Michigan
2009-2024
Chinese University of Hong Kong
2023-2024
Shanghai Artificial Intelligence Laboratory
2023-2024
Nordion (Canada)
2023
Beijing Academy of Artificial Intelligence
2023
National Grid (United Kingdom)
2015-2021
Zhuhai Institute of Advanced Technology
2021
Storagenergy Technologies (United States)
2020
This paper develops a Bayesian continuous 3D semantic occupancy map from noisy point clouds by generalizing the kernel inference model for building maps, binary problem, to multi-class problem. The proposed method provides unified probabilistic both and probabilities nicely reverts original mapping framework when only one occupied class exists in obtained measurements. spatial relaxes independent grid assumption brings smoothness continuity inference, enabling exploit local correlations...
The rapid development of electric vehicles (EVs) has benefited from the fact that more and countries or regions have begun to attach importance clean energy environmental protection. This paper focuses on optimization EV charging, which cannot be ignored in EVs. increase penetration EVs will generate new electrical loads during charging process, bring challenges local power systems. Moreover, uncoordinated may peak-to-valley difference load, aggravate harmonic distortions, affect auxiliary...
Understanding 3D scenes from multi-view inputs has been proven to alleviate the view discrepancy issue in visual grounding. However, existing methods normally neglect cues embedded text modality and fail weigh relative importance of different views. In this paper, we propose ViewRefer, a framework for grounding exploring how grasp knowledge both modalities. For branch, ViewRefer leverages diverse linguistic large-scale language models, e.g., GPT, expand single multiple geometry-consistent...
The modular multilevel converter (MMC) provides promising development for high-voltage direct current (HVDC) applications, including multiterminal HVDC (MTDC) and renewable energy integration. This paper, considering an offshore wind farm (OWF) integrated MMC MTDC system, investigates its start-up process with three main developments: 1) it further develops the mathematical model of active networks proposes a hierarchical control scheme; 2) terminal which connects OWF, reduced dc voltage...
Serious floor heave in gob-side entry retaining (GER) with fully-mechanized gangue backfilling mining affects the transportation and ventilation safety of mine. A theoretical mechanical model for gob-backfilled GER was established. The effects properties strata, granular compaction area (BFA), vertical support roadside body (RSB), stress concentration solid coal on were studied. results show that increases increase seam buried depth, decreases rock elastic modulus. development depth plastic...
A comprehensive process of the control and protection against a DC fault in voltage source converter (VSC) based high-voltage direct current (HVDC) system typically includes detection, isolation recovery. Regarding an offshore wind farm (OWF) integrated modular multilevel (MMC) multi-terminal HVDC (MTDC) with two paradigms, i.e. master-slave droop under faults, this paper presents isolation, including faulted line section, detailed sequence, which would be useful for practical engineering....
1Department of Civil and Environmental Engineering, University Delaware, Newark, DE 19716, USA 2Department Electrical Computer 3Departments Mathematics Statistics Molecular Physiology Biophysics, The Vermont, Burlington, VT 05405, 4Civil Engineering Specialty, Division Colorado School Mines, Golden, CO 80401,
This work reports a novel Bundle Adjustment (BA) formulation using Reproducing Kernel Hilbert Space (RKHS) representation called RKHS-BA. The proposed is correspondence-free, enables the BA to use RGB-D/LiDAR and semantic labels in optimization directly, provides generalization for photometric loss function commonly used direct methods. RKHS-BA can incorporate appearance within continuous spatial-semantic functional that does not require via image pyramids. We demonstrate its applications...
This paper reports on a novel nonparametric rigid point cloud registration framework, Semantic Continuous Visual Odometry (CVO), that jointly integrates geometric and semantic measurements such as color or labels into the alignment process does not require explicit data association. The clouds are represented functions in reproducible kernel Hilbert space. problem is formulated maximizing inner product between two functions, essentially sum of weighted kernels, each which exploits local...
The introduction of IEC61850 offers many advantages including interoperability and flexibility, helps to reduce downtime that occurs as a result upgrades may be applied within substation, due technology changes. One application is at the process level, where Merging Unit samples output current voltage transformers (CTs/VTs), sends standardized digitized required protection control devices. design development (MU) test system will discussed. A number performance tests carried out on MU.
This article reports on recent progress in robot perception and control methods developed by taking the symmetry of problem into account. Inspired existing mathematical tools for studying structures geometric spaces, sensor registration, state estimator, provide indispensable insights formulations generalization robotics algorithms to challenging unknown environments. When combined with computational learning hard-to-measure quantities, symmetry-preserving unleash tremendous performance. The...
The essence of most shape registration algorithms is to find correspondences between two point clouds and then solve for a rigid body transformation that aligns the geometry. main drawback are obtained by placing sensor at different views; consequently, matched points likely do not correspond same physical in real environment. In other words, cloud discrete representation Alternatively, measurement can be seen as samples from geometry, function learned continuous using regression techniques...
This work develops a learning-based contact estimator for legged robots that bypasses the need physical sensors and takes multi-modal proprioceptive sensory data as input. Unlike vision-based state estimators, estimators are agnostic to perceptually degraded situations such dark or foggy scenes. While some equipped with dedicated detect necessary estimation, do not have sensors, addition of is non-trivial without redesigning hardware. The trained network can estimate events on different...
In this paper, we discuss possible solutions to overcome the critical issues for GaN-on-Si power device popularization including cost competitiveness Si MOSFETs, system level reliability verification, and electromagnetic interference (EMI) mitigation at high switching frequency without compromising loss. Both an advanced epitaxy technology a comprehensive platform of 200 mm electron mobility transistors (HEMTs) mass production are presented. A novel strain engineering is reported realize...
Interoperability testing and analysis tools provide a means for achieving assuring the integrity of multivendor intelligent electronic devices (IEDs) data exchanges. However, are very time consuming error prone, these problems worsen when substation becomes large complex during engineering process, commission, replacement, maintenance, extension. To address this challenge, paper presents virtual digital test system (VDSTS) with interoperability assessing identifying challenges...
IEC 61850 conformance alone does not necessarily guarantee interoperability of devices from different manufacturers. Achieving full in a substation protection and control system involving vendors will increase the confidence power utilities to deploy digital bays or substations on site. This study presents development virtual site test platform assess engineering process performance for fully with multi-vendor bay solutions. A data monitoring tool has been developed visualise flows networks,...
Copper wire and palladium coated copper wires used for interconnections, widely in mass production today. Many forecasts are showing an accelerating trend that the market is moving more into business due to cost. Though technology maturity has increased over these past decades, there still a few limitations, intrinsic material itself, which leads possible bonding process issues. hardness current limitation Front End technologies with delicate pad metal [1]; it can be very difficult define...
In recent years, multimodal large language models (MLLMs) have shown remarkable capabilities in tasks like visual question answering and common sense reasoning, while perception made significant strides tasks, such as detection segmentation. However, MLLMs mainly focus on high-level image-text interpretations struggle with fine-grained understanding, vision usually suffer from open-world distribution shifts due to their limited model capacity. To overcome these challenges, we propose the...
This paper introduces a robust unsupervised SE(3) point cloud registration method that operates without requiring correspondences. The frames clouds as functions in reproducing kernel Hilbert space (RKHS), leveraging SE(3)-equivariant features for direct feature registration. A novel RKHS distance metric is proposed, offering reliable performance amidst noise, outliers, and asymmetrical data. An training approach introduced to effectively handle limited ground truth data, facilitating...