A5066085302- 3D Surveying and Cultural Heritage
- IoT and Edge/Fog Computing
- Data Management and Algorithms
- Remote Sensing and LiDAR Applications
- Robotics and Sensor-Based Localization
- Anomaly Detection Techniques and Applications
- Data Stream Mining Techniques
- Energy Efficient Wireless Sensor Networks
- Caching and Content Delivery
- Video Surveillance and Tracking Methods
- Urban and Freight Transport Logistics
- Advanced Adaptive Filtering Techniques
- Time Series Analysis and Forecasting
- 3D Shape Modeling and Analysis
- Data Mining Algorithms and Applications
- Indoor and Outdoor Localization Technologies
- Industrial Vision Systems and Defect Detection
- Digital Transformation in Industry
- Context-Aware Activity Recognition Systems
- Autonomous Vehicle Technology and Safety
- Electricity Theft Detection Techniques
- Geographic Information Systems Studies
- IoT Networks and Protocols
- Distributed Sensor Networks and Detection Algorithms
- Peer-to-Peer Network Technologies
University of Cambridge
2020-2024
European Bioinformatics Institute
2021
Alexandria University
2017-2021
University College London
2018-2019
University of Surrey
2016-2018
The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic will be remembered as one of the defining events 21st century. rapid global outbreak has had significant impacts on human society and is already responsible for millions deaths. Understanding tackling impact virus required a worldwide mobilisation coordination scientific research. COVID-19 Data Portal (https://www.covid19dataportal.org/) was first released part European Platform, April 20th 2020 to facilitate open data...
The Internet of Things (IoT) is revolutionising how energy delivered from producers and used throughout residential households. Optimising the consumption a crucial step toward having greener sustainable production. Such optimisation requires household-centric management system as opposed to one-rule-fits all approach. In this paper, we propose data-driven multi-layer digital twin that aims mirror households’ actual in form household (HDT). When linked production (EDT), HDT empowers model...
Enormous amounts of dynamic observation and measurement data are collected from sensors in Wireless Sensor Networks (WSNs) for the Internet Things (IoT) applications such as environmental monitoring. However, continuous transmission sensed requires high energy consumption. Data between sensor nodes cluster heads (sink nodes) consumes much higher than sensing WSNs. One way reducing consumption is to minimise number transmissions. In this paper, we propose an Adaptive Method Reduction (AM-DR)....
The Internet of Things (IoT) paradigm is revolutionising the world manufacturing into what known as Smart Manufacturing or Industry 4.0. main pillar in smart looks at harnessing IoT data and leveraging machine learning (ML) to automate prediction faults, thus cutting maintenance time cost improving product quality. However, faults real industries are overwhelmingly outweighed by instances good performance (faultless samples); this bias reflected captured devices. Imbalanced limits success ML...
Data owners are creating an ever richer set of information resources online, and these being used for more applications. Spatial data on the Web is becoming ubiquitous voluminous with rapid growth location-based services, spatial technologies, dynamic services published by different organizations. However, heterogeneity peculiarities data, such as use coordinate reference systems, make it difficult users, applications, to discover, interpret in large distributed system that Web. To...
Large volumes of real-world observation and measurement data are collected from sensory devices in the Internet Things (IoT) networks. IoT is often generated highly distributed dynamic environments. Continuous transmission large between sensor head/sink nodes induces a high communication cost for individual nodes. This results significant increase overall energy applications such as environmental monitoring. Decreasing can effectively reduce consumption prolong network lifetime, especially...
This paper presents a data set collected periodically on construction site. The aims to evaluate the performance of simultaneous localization and mapping (SLAM) algorithms used by mobile scanners or autonomous robots. It includes ground-truth scans site using terrestrial laser scanner along with five sequences spatially registered time-synchronized images, lidar scans, inertial coming from our prototypical handheld scanner. We also recover trajectory registering sequential show how use...
The Internet of Things (IoT) concept has attracted a lot attention from the research and innovation community for number years already. One key drivers this hype towards IoT is its applicability to plethora different application domains. However, infrastructures enabling experimental assessment solutions are scarce. Being able test assess behavior performance any piece technology (i.e., protocol, algorithm, application, service, etc.) under real-world circumstances utmost importance increase...
Autoencoders are unsupervised models which have been used for detecting anomalies in multi-sensor environments. A typical use includes training a predictive model with data from sensors operating under normal conditions and using the to detect anomalies. Anomalies can come either real changes environment (real drift) or faulty sensory devices (virtual drift); however, of distinguish between different has not yet considered. To this end, we first propose development Bayesian quantify...
The current Web and data indexing search mechanisms are mainly tailored to process text-based limited in addressing the intrinsic characteristics of distributed, large-scale dynamic Internet Things (IoT) networks. IoT demands novel solutions for create an ecosystem system; however, often numerical, multi-modal heterogeneous. We propose a distributed adaptable mechanism that allows discovery real-world Comparing state-of-the-art approaches, our model does not require any prior knowledge about...
When dealing with a large number of devices, the existing indexing solutions for discovery Internet Things (IoT) sources often fall short to provide an adequate scalability. This is due high computational complexity and communication overhead that required create maintain indices IoT particularly when their attributes are dynamic. paper presents novel approach distributed paves way design data service search gain access data. The proposed method creates concise references by using Gaussian...
There has been a keen interest in detecting abrupt sequential changes streaming data obtained from sensors wireless sensor networks for Internet of Things applications, such as fire/fault detection, activity recognition, and environmental monitoring. Such applications require (near) online detection instantaneous changes. This paper proposes an online, adaptive filtering-based change (OFCD) algorithm. Our method is based on convex combination two decoupled least mean square windowed filters...
An important field in exploratory sensory data analysis is the segmentation of time-series to identify activities interest. In this work, we analyse performance univariate and multi-sensor Bayesian change detection algorithms segmenting accelerometer data. particular, provide theoretical also evaluation on synthetic real-world The results illustrate advantages using multi-sensory variance dynamic (e.g. data).
Abstract Smart cities harness data and technology to enhance the sustainability efficiency of urban areas communities. Walking as a form active travel is essential in promoting sustainable transport. It thus crucial accurately predict pedestrian crossing intention avoid collisions, especially with advent autonomous advanced driver-assisted vehicles. Current research leverages computer vision machine learning advances near-misses; however, this often requires high computation power yield...
Walking as a form of active travel is essential in promoting sustainable transport. It thus crucial to accurately predict pedestrian crossing intention and avoid collisions, especially with the advent autonomous advanced driver-assisted vehicles. Current research leverages computer vision machine learning advances near-misses; however, this often requires high computation power yield reliable results. In contrast, work proposes low-complexity ensemble-learning approach that employs...