20505 research outputs found
Sort by
Dataset "Results of centreline extraction based on maximal disks"
No full textMaps of road layouts play an essential role in autonomous driving, and it is often advantageous to represent them in a compact form, using a sparse set of surveyed points of the lane boundaries. While lane centrelines are valuable references in the prediction and planning of trajectories, most centreline extraction methods only achieve satisfactory accuracy with high computational cost and limited performance in sparsely described scenarios. This paper explores the problem of centreline extraction based on a sparse set of border points, evaluating the performance of different approaches on both a self-created and a public dataset, and proposing a novel method to extract the lane centreline by searching and linking the internal maximal circles along the lane. Compared with other centreline extraction methods producing similar numbers of centre points, the proposed approach is significantly more accurate: in our experiments, based on a self-created dataset of road layouts, it achieves a max deviation below 0.15 m and an overall RMSE less than 0.01 m, against the respective values of 1.7 m and 0.35 m for a popular approach based on Voronoi tessellation, and 1 m and 0.25 m for an alternative approach based on distance transform
Emerging decision-making for transportation safety: collaborative agent performance analysis
Get PDFThis article belongs to the Special Issue Emerging Transportation Safety and Operations: Practical Perspectives, 2nd EditionThis paper addresses the challenge of improving decision-making capabilities and safety in autonomous vehicles (AVs) using Agent-Based Modelling (ABM). The study evaluates ABM’s effect on Advanced Driver Assistance Systems (ADASs) in challenging driving situations, like lane merging, by incorporating it into a simulation framework designed for autonomous vehicles. Identifying emergent behaviours that enhance safety and efficiency, verifying the efficacy of ABM in AV decision-making, and investigating the function of hardware acceleration to enable practical application in ADASs are some of the major achievements. According to the simulation results, ABM can greatly improve AV performance, providing a practical and scalable means of enhancing safety in future transportation systems.Vehicle
A set-based design space exploration framework for hybrid-electric aicraft design
No full textEngineering design is characterised by uncertainty caused by a lack of experience and
information. The traditional approach focuses on iterating and refining an initial conceptual
design, which often is similar to the final one. Although this method serves well in
the case of evolutionary design, it is unsuitable for innovation. In fact, without a suitable
initial starting point, many rework iterations may be required to correct early inadequate
design decisions. In addition, it may be challenging to map the requirements directly onto
the design space.
This dissertation aims at developing a methodology to address this problem. The
developed framework starts from the hypothesis, and the knowledge to carry out the mapping
of requirements onto the input parameters is embedded in the simulation model, and
hence no additional rules are required. Instead, a probabilistic surrogate model based on
Gaussian processes is used in conjunction with Bayesian statistics to find and eliminate
unfeasible areas of the design space. This selection criterion is used in a set-based design
approach to explore pockets of the entire continuous design space. Finally, sets with a
sufficient likelihood of satisfying the requirements are searched with a local multidisciplinary
optimisation algorithm to recover the individual design points.
This process reduced the computational cost of the design space exploration by 80%
without sacrificing the number of alternative solutions. Thanks to the large amount of
data obtained, it was possible to produce new knowledge on hybrid-electric aircraft design.
Specifically, it was found that linear segments are sufficient for defining energy
management strategies, and the reduction of NOx emissions and fuel consumption are associated
with climb and cruise, respectively. Furthermore, when studying regional aircraft
operating missions, it was found that partial recharge is necessary to maintain the design
performance. However, this could reduce the duration of the battery. The battery ageing
rate correlates with the EMS’s demand for electrical energy. Finally, it was found that
the battery’s energy density is a determinant of the pack’s durability and the feasibility of
HE aircraft. The rate of improvement in emissions and fuel consumption is non-linear,
suggesting that investing in considerable technological improvements has better returns.
Indeed, the required technological level will not be available until the 2040s without an
exponential increment of the cell energy density.PhD in Aerospac
Opportunities and challenges for monitoring terrestrial biodiversity in the robotics age
No full textWith biodiversity loss escalating globally, a step change is needed in our capacity to accurately monitor species populations across ecosystems. Robotic and autonomous systems (RAS) offer technological solutions that may substantially advance terrestrial biodiversity monitoring, but this potential is yet to be considered systematically. We used a modified Delphi technique to synthesize knowledge from 98 biodiversity experts and 31 RAS experts, who identified the major methodological barriers that currently hinder monitoring, and explored the opportunities and challenges that RAS offer in overcoming these barriers. Biodiversity experts identified four barrier categories: site access, species and individual identification, data handling and storage, and power and network availability. Robotics experts highlighted technologies that could overcome these barriers and identified the developments needed to facilitate RAS-based autonomous biodiversity monitoring. Some existing RAS could be optimized relatively easily to survey species but would require development to be suitable for monitoring of more ‘difficult’ taxa and robust enough to work under uncontrolled conditions within ecosystems. Other nascent technologies (for instance, new sensors and biodegradable robots) need accelerated research. Overall, it was felt that RAS could lead to major progress in monitoring of terrestrial biodiversity by supplementing rather than supplanting existing methods. Transdisciplinarity needs to be fostered between biodiversity and RAS experts so that future ideas and technologies can be codeveloped effectively.This work was funded by the EPSRC UK-RAS Network. D.J.I. is funded by a UK Research and Innovation Future Leaders Fellowship (grant ref: MR/W006316/1), and Z.G.D. and J.C.F. were supported by Research England’s ‘Expanding Excellence in England’ fund.Nature Ecology & Evolutio
Object detection and classification with limited training data
Get PDFSince the rise of deep learning around a decade ago, the field of object detection and classification using a convolutional neural network (CNN) and its variants has grown exponentially...Applied Science
Shock wave overpressure history mapping using high-resolution distributed acoustic sensing
Get PDFThis study explored the use of high-resolution distributed acoustic sensing (HR-DAS) for measuring blast wave overpressures, addressing the limitations of conventional pressure transducers. Shock tube experiments were conducted to evaluate HR-DAS performance in capturing side-on blast overpressures, comparing its strain measurements with reference piezoelectric pressure transducers. The study examined the effects of sensing fibre orientation and mounting methods on sensor sensitivity. Results showed that HR-DAS strain histories aligned well with conventional pressure measurements. The findings demonstrate the feasibility of HR-DAS for blast wave sensing, highlighting its potential for further development and broader applications.This project was funded by the Royal Society (RGS\R1\231115).29th International Conference on Optical Fiber Sensor
Calculating the heat of explosion for non-ideal explosives
Get PDFNon-ideal explosives are typically highly porous, low-density heterogeneous materials. They differ from ideal explosives in that fuel and oxidiser are separated at the molecular level. As such, non-ideal explosives demonstrate lower detonation pressures and velocities than ideal explosives, with more extended detonation reaction zones. These longer detonation reaction zones produce higher heat of explosion due to secondary and tertiary combustion reactions occurring behind the detonation front. Non-ideal explosives are increasingly used as homemade explosives (HME), and it is often challenging to determine likely performance without detailed chemical analysis, or an understanding of loading density, critical diameter, and stoichiometry. Detailed analysis is not an option in security-related settings where preservation of life and property is often challenged by time. Using Hess’ law of enthalpy, and empirical and quantitative thermochemical methods to determine the products of detonation, this paper introduces a simple method to determine the heat of explosion of any non-ideal explosive composition during three specific stages: anoxic detonation; anaerobic combustion; and aerobic combustion. The values calculated compare favourably to heats of explosion for non-ideal explosives published in open-source literature, suggesting that the approach can be developed into a user-friendly application for first responders.The International Explosives Conference (IEC-2024)Next Generation Energetics: 2nd International Explosives Conferenc
Fuelling hydrogen futures? A trust-based model of social acceptance
No full textPublic trust plays a fundamental role in shaping national energy policies in democratic countries, as exemplified by nuclear phase-out in Germany following the Fukushima accident. While trust dynamics have been explored in different contexts of the energy transition, few studies have attempted to quantify the influence of public trust in shaping social acceptance and adoption potential. Moreover, the interaction between public trust and perceived community benefits remains underexplored in the literature, despite the relevance of each factor to facilitating social acceptance and technology uptake. In response, this quantitative analysis closes a parallel research gap by examining the antecedents of public trust and perceived community benefits in the context of deploying hydrogen heating and cooking appliances across parts of the UK housing stock. Drawing on results from a nationally representative online survey (N = 1845), the study advances insights on the consumer perspective of transitioning to ‘hydrogen homes’, which emerged as a topical and controversial aspect of UK energy policy in recent years. Partial least squares structural equation modelling and necessary condition analysis are undertaken to assess the predictive capabilities of a trust-based model, which incorporates aspects of institutional, organisational, interpersonal, epistemic, and social trust. Regarding sufficiency-based logic, social trust is the most influential predictor of public trust, whereas trust in product and service quality corresponds to the most important necessary condition for enabling public trust. Nevertheless, trust in the government, energy sector, and entities involved in research & development are needed to facilitate and strengthen public trust. Overall, this study enriches scholarly understanding of how public trust may shape prospects for trialling novel low-carbon technologies, highlights the need for segment-specific consumer engagement, and advances scholarly understanding of the innovation-decision process in the context of net-zero pathways. As policymakers approach critical decisions on the portfolio of technologies needed to support residential decarbonisation, public trust will prove fundamental to fuelling hydrogen-based energy futures.Engineering and Physical Sciences Research Council|EP/T518104/1This research was supported by the UK Research and Innovation Engineering and Physical Sciences Research Council (UKRI-EPSRC) Grant EP/T518104/1, and sponsored by Cadent Gas Ltd.Sustainable Energy & Fuel
Optimisation of deep learning techniques on remaining useful life prediction of complex engineering systems
Get PDFAddepalli, Pavan - Associate SupervisorPredictive maintenance based on performance degradation is a crucial way to reduce the operation and maintenance costs and potential failures in modern complex engineering systems. Reliable remaining useful life (RUL) prediction is the main criterion for predictive maintenance decisions. Data-driven techniques, especially artificial intelligence (AI) such as deep learning (DL) techniques, have attracted more and more attention in the manufacturing sector with the development of Big Data and Internet-of-Things. Different DL techniques recently have been used for RUL prediction and achieved great success. However, in many cases, the RUL prediction results vary greatly due to the measurement noise and selection of model parameters. This PhD research aims to develop a novel framework to optimise the performance of deep learning methods in the context of predicting the RUL of complex engineering systems. The project consists of four stages including literature review, optimisation investigations, construction of the new prognostic framework and validation. In the first stage, state-of-the-art DL-based approaches for RUL prediction are reviewed. Then, the optimisation investigations on different RNN models, feature engineering, model parameters and RUL target functions are carried out trying to improve the RUL prediction accuracy. After that, the prognostic framework is built based on the result of the optimisation investigation. Meanwhile, a novel three-stage feature selection method and a multi-scale RUL prediction approach are proposed in this stage. In addition, to accommodate the multiple operating conditions of complex engineering systems, this thesis presents an operation-based normalisation method to address the different degradation patterns from data. In the last stage, the C-MAPSS dataset is adopted to validate the proposed framework and the prediction performance is compared with the state-of-the-art RUL prediction approaches. A significant improvement can be observed in the RUL prediction performance using the proposed framework on most of the subsets of the C-MAPSS dataset. Therefore, a reliable and flexible RUL prediction strategy can be made based on this DL-based prognostics framework for complex engineering systemsPhD in Manufacturin
Addressing viral risk in drinking water: evaluating the impact of treatment processes in drinking water treatment
Get PDFJarvis, Peter - Associate Supervisor
Singh, Suniti - Associate SupervisorDrinking water treatment is designed to improve water quality, including the
removal of potential contaminants like viruses. Pilot-scale studies can provide
valuable insights into the effectiveness of full-scale treatment processes in
managing viral risks. While traditional water treatment processes are effective in
removing many contaminants, the specific removal of viruses remains a
significant concern. The World Health Organization (WHO) recommends a 2-log
reduction of enteric viruses from source to tap water. However, the effectiveness
of different treatment processes in achieving this target, especially under varying
conditions, requires further investigation. This study demonstrated that optimized
coagulation, followed by subsequent treatment processes, can effectively remove
Φ-X174 from water, even at high initial concentrations. The results suggest that
conventional drinking water treatment methods can effectively manage the risk
of viral contamination, aligning with the WHO's recommended log removal target.
This study optimised coagulation and conducted challenge trials using a near-
real scale drinking water treatment rig to evaluate the effectiveness of individual
treatment processes and log reduction of Φ-x 174. At a final concentration of 1010
plaque-forming units of Φ-X 174 the somatic coliphage was not detected in the
final water. At a final concentration of 1014 plaque-forming units of Φ-X 174,
coliphage was detected with no coagulant dose but was removed to below the
detection limit for both optimal (10 mg-Fe3+ /L) and sub-optimal (2 mg-Fe3+ /L)
coagulation conditions.MSc by Research in Wate