19684 research outputs found
Sort by
Vibration characteristics of a compression ignition engine fuelled with different biodiesel-diesel blends
Get PDFBiodiesel has wide application prospects due to its good power performance, fuel economy and emission reduction. Experimental studies have found that the measured engine vibration presents an N-shaped nonlinear trend with the increase of the biodiesel proportion in blends, which cannot be explained solely based on the combustion characteristics of blended fuels. To study the mechanisms for this nonlinear trend of engine vibration, a two-degree-of-freedom nonlinear model of piston–cylinder system was established and verified to analyse the correspondence between in-cylinder combustion behaviour and engine dynamic responses. By correlating simulation results with measured signals, it is found that the root cause of the nonlinear vibration trend is the coupling effect of in-cylinder pressure and piston inertial force. The time integral of piston lateral force in the interval from combustion top dead centre (TDC) to the subsequent piston slap ultimately determines the trend of liner vibrations. These key findings pave the fundamentals for the vibration analysis of engines fuelled with other alternative fuels, which is important for improve engine operation performances including reliability assessment and NVH control.</p
Deep Learning Framework Using Spatial Attention Mechanisms for Adaptable Angle Estimation Across Diverse Array Configurations
No full textRapid advancement of wireless communication systems and the increasing need for accurate, real-time signal processing have driven innovations in direction-of-arrival (DoA) estimation techniques. This paper introduces a novel convolutional neural network (CNN) architecture that combines spatial attention mechanisms with a transfer learning framework to enhance both accuracy and versatility in DoA estimation. The model integrates spatial attention layers to dynamically prioritize signal regions with the highest information value, allowing it to isolate relevant signals and suppress interference in noisy or crowded signal environments. In addition, we utilize a transfer learning framework that enables the model to generalize across various antenna array configurations (i.e., planar, linear, and circular arrays) with minimal additional training. Extensive simulation results benchmark the proposed model against existing state-of-the-art methods for DoA estimation, achieving improved absolute error across diverse conditions. This hybrid approach not only enhances DoA estimation precision, but also significantly reduces retraining requirements when adapting to new array configurations, positioning it as a robust, scalable tool for next-generation wireless communication systems.</p
Deep Reinforcement Learning-Based Resource Allocation for QoE Enhancement in Wireless VR Communications
No full textWireless virtual reality (VR) communication applications have emerged as a transformative technology, offering innovative solutions in various areas of everyday life. However, the successful deployment of these applications faces challenges in ensuring high quality of experience (QoE), especially in environments with limited network resources. This research paper presents a novel approach to address the challenge of enhancing QoE by incorporating deep reinforcement learning (DRL) techniques in the resource allocation process. The proposed model takes into account the quality of service (QoS) parameters of the 5G new radio (NR) network to optimize its operation, ensuring a seamless and immersive VR experience. Specifically, the resource allocation strategy adopts a policy that maximizes the transmission-related QoE value based on the evolving characteristics of the communication channel and user interactions. To evaluate the effectiveness of the proposed approach, extensive simulations and comparative analyses against traditional resource allocation methods are performed. The results demonstrate significant improvements in the transmission-related QoE values and highlight the superiority of the DRL-based resource allocation approach in the dynamic and unpredictable wireless environments.</p
Formal Modeling of Hybrid System Based on Semi-continuous Colored Petri Net:A Case Study of Adaptive Cruise Control System
Get PDFMany Next-Generation consumer electronic devices would be distributed hybrid electronic systems, such as UAVs (Unmanned Aerial Vehicles) and smart electronic cars. The safety and risk control are the key issues for the sustainability of such consumer electronic systems. The modeling of hybrid electronic systems is difficult to be abstracted by traditional Petri Nets. This also makes the reachable marking graph unable to be applied to Petri nets of the hybrid electronic systems. This paper proposes a novel Petri Net to model and analyze the hybrid electronic systems. We name it a Semi-continuous Colored Petri Net (SCPN) that inherits the excellent modeling capabilities and analysis methods of Petri Nets, and can formally depict hybrid quantities. In addition, we propose the construction algorithm for an SCPN reachable marking graph and prove its finiteness. Finally, we model and analyze an Adaptive Cruise Control (ACC) system of smart electronic cars as an example to prove the validity of SCPN. We use the proposed SCPN to model and analyze the running process of an ACC system under the continuous deceleration scenario of the front vehicle. The application study shows that the ACC system has logic flaws under the constant headway strategy when the front vehicle continues to decelerate. Based on this analysis, improvements to the SCPN of the ACC system are made, effectively enhancing its safety and logical correctness
Optimizing Foreign Exchange Trading Performance Through Reinforcement Machine Learning Framework
No full textThe ever-changing financial market of foreign exchange attracts many traders. Traders must make wise decisions to avoid significant losses when buying and selling currencies. This project intends to reduce the chance of suffering from loss by providing a trading strategy. The research on developing a trading strategy specifically for the foreign exchange market is still lacking due to the limitation in selecting the best model to create a trading strategy, which is still a working area. Even with current research on trading strategy, it tends not to work overtime due to unpredictable market trends. Therefore, this paper proposed three models using the algorithms A2C, PPO & DQN to find the best strategy in foreign exchange trading, analyze the impact of individual features on the trading strategy and identify the most influential features to develop the best trading strategy using reinforcement learning and finally evaluate the performance on unseen data using Sharpe Ratio, Sortino Ratio, Omega Ratio, Profit & Loss (%), Maximum Drawdown (%) and Cumulative Score. The experiment result showed that the PPO algorithm performed best on 2 of the currency pairs which is GBP/USD and USD/JPY, with a Sharpe Ratio of 0.23 and 0.70, respectively, and a Profit & Loss of 7.4% and 16.78%, respectively, when tested on unseen data. Meanwhile, when tested on unseen data, the A2C model performed the best on the EUR/USD currency pair with a Sharpe Ratio of 0.16 and a Profit & Loss of 3.34%.</p
Automation and Sustainable urban Transport
No full textThe transition to an automated transport paradigm is a change of unparalleled dimensions. Disrupting a century-long driver-centric mobility status quo will not be an uncomplicated and unimpactful technological triumph but a multi-dimensional mega-shift for the way our cities, societies, livelihoods and our very planet function. autonomous vehicles (AVs) will likely have a critical impact on every sphere of sustainability that we need to proactively try to assess now, that there is still time, to make changes and push towards different directions. Since we are experiencing more and more transport-sensitive climate change and resilience threats we need to use the transformative powers of Artificial Intelligence (AI), in general, and vehicle automation in particular, responsibly so that we craft pathways to more liveable futures. This chapter provides an improved understanding of how transport automation can be linked to the future of our environmental, economic and societal resource preservation. It will also try to predict how automated transport can be more sustainable when we already know about the uneasy alignment of AV-centric technology with genuine sustainability goals because of their possibly contradictory roles and priorities. More specifically, key automated transport issues that represent likely sustainability challenges and opportunities (or both) including motor traffic congestion, local air pollution and greenhouse gas emissions, energy consumption, accident prevention and traffic safety, social inclusion and accessibility, employment market disruption, privacy and cybersecurity are ‘benchmarked’ against the triple bottom line and discussed in detail. The chapter suggests that automation’s pathway for promoting sustainable urban transport futures depends on its ability to be effectively packaged with less hyped but equally transformative functionalities around connectivity, alternative fuelling, public transport and multimodality. Change can be positive only if it can be holistic.</p
An analysis of predictors of conventional and complementary healthcare use in Yorkshire
No full textBackground: Well-recognised trends in types of services used by patients exist within health service utilisation. One increasing area is complementary and alternative medicine (CAM); considered distinct from the use of health professionals working in conventional medicine. Little is known about the contribution of CAM and whether people using CAM with multiple comorbidities make correspondingly less use of conventional health services. Aims: 1) To describe self-reported visits to conventional health professionals and CAM practitioners, and to identify predictors of such visits; 2) To quantify the effect of demographic, health-related and CAM service take-up factors on contact with health services delivered by conventional health professionals. Methods: Data from 70,836 participants in the Yorkshire Health Study, a large-scale population-based cohort study, was analysed descriptively and inferentially to test for associations between variables. Results: 3.5 % of the cohort reported accessing CAM services in the previous three months. Level of contact with conventional health professionals was higher in those accessing practitioner-led CAM services (incidence rate ratio [IRR]=1.28; p < 0.001.) Female gender, older age and increased incidence of mental and physical health conditions were also positively associated with the outcome. Conclusions: Self-reported utilisation of CAM services was low but there were several predictors of recent CAM use based on demographic and health conditions which may be of help in understanding conventional and CAM healthcare utilisation.</p
A contextual framework to standardise the communication of machine learning cyber security characteristics
No full textThe widespread integration of machine learning (ML) across diverse application domains has substantially impacted business and personnel. Notably, ML applications in cybersecurity have gained increased prominence, reflecting a discernible trend towards adoption. However, the decisions surrounding ML adoption are susceptible to external influences, potentially resulting in misinterpreting ML capabilities. The communication used when for incorporating ML into cybersecurity applications lacks standardisation and is influenced by various factors such as personal experience, organisational reputation, and marketing strategies. Furthermore, the application of metrics to assess model performance is characterised by dependence, disarray, and subjectivity, introducing probabilities, uncertainties, and the potential for misinterpretation. The different metrics allow for variability in how capability is communicated, often dependent on the restrictive use case, leading to a lack of certainty in their interpretation. Previous research has highlighted the need for a standardised approach. Building upon our earlier work, this paper aims to authenticate beneficiaries' perception of Machine Learning Cybersecurity (MLCS) capabilities, before consulting with domain experts through a focus group to elucidate a prototype standard for comprehending MLCS capabilities, offering a pivotal roadmap and an initial framework for a comprehensive understanding and effective communication of MLCS capabilities in practical implementations
Access Repository Dashboard
Do you manage Huddersfield Research Portal? Access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard!