Search results for: scheduling algorithms

Authors: Dangut Maren David, Skaf Zakwan

Abstract:

Adequate monitoring of vehicle component in other to obtain high uptime is the goal of predictive maintenance, the major challenge faced by businesses in industries is the significant cost associated with a delay in service delivery due to system downtime. Most of those businesses are interested in predicting those problems and proactively prevent them in advance before it occurs, which is the core advantage of Prognostic Health Management (PHM) application. The recent emergence of industry 4.0 or industrial internet of things (IIoT) has led to the need for monitoring systems activities and enhancing system-to-system or component-to- component interactions, this has resulted to a large generation of data known as big data. Analysis of big data represents an increasingly important, however, due to complexity inherently in the dataset such as imbalance classification problems, it becomes extremely difficult to build a model with accurate high precision. Data-driven predictive modeling for condition-based maintenance (CBM) has recently drowned research interest with growing attention to both academics and industries. The large data generated from industrial process inherently comes with a different degree of complexity which posed a challenge for analytics. Thus, imbalance classification problem exists perversely in industrial datasets which can affect the performance of learning algorithms yielding to poor classifier accuracy in model development. Misclassification of faults can result in unplanned breakdown leading economic loss. In this paper, an advanced approach for handling imbalance classification problem is proposed and then a prognostic model for predicting aircraft component replacement is developed to predict component replacement in advanced by exploring aircraft historical data, the approached is based on hybrid ensemble-based method which improves the prediction of the minority class during learning, we also investigate the impact of our approach on multiclass imbalance problem. We validate the feasibility and effectiveness in terms of the performance of our approach using real-world aircraft operation and maintenance datasets, which spans over 7 years. Our approach shows better performance compared to other similar approaches. We also validate our approach strength for handling multiclass imbalanced dataset, our results also show good performance compared to other based classifiers.

Keywords: prognostics, data-driven, imbalance classification, deep learning

Procedia PDF Downloads 149

Authors: Laura Micheli, Majd Hijazi, Mahmoud Faytarouni

Abstract:

The damage reported by oil and gas industrial facilities revealed the utmost vulnerability of steel liquid storage tanks to seismic events. The failure of steel storage tanks may yield devastating and long-lasting consequences on built and natural environments, including the release of hazardous substances, uncontrolled fires, and soil contamination with hazardous materials. It is, therefore, fundamental to reliably predict the damage that steel liquid storage tanks will likely experience under future seismic hazard events. The seismic performance of steel liquid storage tanks is usually assessed using vulnerability curves obtained from the numerical simulation of a tank under different hazard scenarios. However, the computational demand of high-fidelity numerical simulation models, such as finite element models, makes the vulnerability assessment of liquid storage tanks time-consuming and often impractical. As a solution, this paper presents a surrogate model-based strategy for predicting seismic-induced damage in steel liquid storage tanks. In the proposed strategy, the surrogate model is leveraged to reduce the computational demand of time-consuming numerical simulations. To create the data set for training the surrogate model, field damage data from past earthquakes reconnaissance surveys and reports are collected. Features representative of steel liquid storage tank characteristics (e.g., diameter, height, liquid level, yielding stress) and seismic excitation parameters (e.g., peak ground acceleration, magnitude) are extracted from the field damage data. The collected data are then utilized to train a surrogate model that maps the relationship between tank characteristics, seismic hazard parameters, and seismic-induced damage via a data-driven surrogate model. Different types of surrogate algorithms, including naïve Bayes, k-nearest neighbors, decision tree, and random forest, are investigated, and results in terms of accuracy are reported. The model that yields the most accurate predictions is employed to predict future damage as a function of tank characteristics and seismic hazard intensity level. Results show that the proposed approach can be used to estimate the extent of damage in steel liquid storage tanks, where the use of data-driven surrogates represents a viable alternative to computationally expensive numerical simulation models.

Keywords: damage prediction , data-driven model, seismic performance, steel liquid storage tanks, surrogate model

Procedia PDF Downloads 122

Authors: M. Aboalazayem, Essam A. Gouda, Ahmed M. Moussa, Amr E. Flifl

Abstract:

Field data collecting is considered one of the most difficult work due to the difficulty of accessing large zones such as large lakes. Also, it is well known that the cost of obtaining field data is very expensive. Remotely monitoring of lake water quality (WQ) provides an economically feasible approach comparing to field data collection. Researchers have shown that lake WQ can be properly monitored via Remote sensing (RS) analyses. Using satellite images as a method of WQ detection provides a realistic technique to measure quality parameters across huge areas. Landsat (LS) data provides full free access to often occurring and repeating satellite photos. This enables researchers to undertake large-scale temporal comparisons of parameters related to lake WQ. Satellite measurements have been extensively utilized to develop algorithms for predicting critical water quality parameters (WQPs). The goal of this paper is to use RS to derive WQ indicators in Aswan High Dam Reservoir (AHDR), which is considered Egypt's primary and strategic reservoir of freshwater. This study focuses on using Landsat8 (L-8) band surface reflectance (SR) observations to predict water-quality characteristics which are limited to Turbidity (TUR), total suspended solids (TSS), and chlorophyll-a (Chl-a). ArcGIS pro is used to retrieve L-8 SR data for the study region. Multiple linear regression analysis was used to derive new correlations between observed optical water-quality indicators in April and L-8 SR which were atmospherically corrected by values of various bands, band ratios, and or combinations. Field measurements taken in the month of May were used to validate WQP obtained from SR data of L-8 Operational Land Imager (OLI) satellite. The findings demonstrate a strong correlation between indicators of WQ and L-8 .For TUR, the best validation correlation with OLI SR bands blue, green, and red, were derived with high values of Coefficient of correlation (R2) and Root Mean Square Error (RMSE) equal 0.96 and 3.1 NTU, respectively. For TSS, Two equations were strongly correlated and verified with band ratios and combinations. A logarithm of the ratio of blue and green SR was determined to be the best performing model with values of R2 and RMSE equal to 0.9861 and 1.84 mg/l, respectively. For Chl-a, eight methods were presented for calculating its value within the study area. A mix of blue, red, shortwave infrared 1(SWR1) and panchromatic SR yielded the greatest validation results with values of R2 and RMSE equal 0.98 and 1.4 mg/l, respectively.

Keywords: remote sensing, landsat 8, nasser lake, water quality

Procedia PDF Downloads 70

Authors: Nareshkumar Harale, B. B. Meshram

Abstract:

The continued exponential growth of successful cyber intrusions against today’s businesses has made it abundantly clear that traditional perimeter security measures are no longer adequate and effective. We evolved the network trust architecture from trust-untrust to Zero-Trust, With Zero Trust, essential security capabilities are deployed in a way that provides policy enforcement and protection for all users, devices, applications, data resources, and the communications traffic between them, regardless of their location. Information exchange over the Internet, in spite of inclusion of advanced security controls, is always under innovative, inventive and prone to cyberattacks. TCP/IP protocol stack, the adapted standard for communication over network, suffers from inherent design vulnerabilities such as communication and session management protocols, routing protocols and security protocols are the major cause of major attacks. With the explosion of cyber security threats, such as viruses, worms, rootkits, malwares, Denial of Service attacks, accomplishing efficient and effective intrusion detection and prevention is become crucial and challenging too. In this paper, we propose a design and analysis model for next generation network intrusion detection and protection system as part of layered security strategy. The proposed system design provides intrusion detection for wide range of attacks with layered architecture and framework. The proposed network intrusion classification framework deals with cyberattacks on standard TCP/IP protocol, routing protocols and security protocols. It thereby forms the basis for detection of attack classes and applies signature based matching for known cyberattacks and data mining based machine learning approaches for unknown cyberattacks. Our proposed implemented software can effectively detect attacks even when malicious connections are hidden within normal events. The unsupervised learning algorithm applied to network audit data trails results in unknown intrusion detection. Association rule mining algorithms generate new rules from collected audit trail data resulting in increased intrusion prevention though integrated firewall systems. Intrusion response mechanisms can be initiated in real-time thereby minimizing the impact of network intrusions. Finally, we have shown that our approach can be validated and how the analysis results can be used for detecting and protection from the new network anomalies.

Keywords: network intrusion detection, network intrusion prevention, association rule mining, system analysis and design

Procedia PDF Downloads 200

Authors: Harris Niavis, Dimitra Politaki

Abstract:

The manufacturing industry was collecting vast amounts of data for monitoring product quality thanks to the advances in the ICT sector and dedicated IoT infrastructure is deployed to track and trace the production line. However, industries have not yet managed to unleash the full potential of these data due to defective data collection methods and untrusted data storage and sharing. Blockchain is gaining increasing ground as a key technology enabler for Industry 4.0 and the smart manufacturing domain, as it enables the secure storage and exchange of data between stakeholders. On the other hand, AI techniques are more and more used to detect anomalies in batch and time-series data that enable the identification of unusual behaviors. The proposed scheme is based on smart contracts to enable automation and transparency in the data exchange, coupled with anomaly detection algorithms to enable reliable data ingestion in the system. Before sensor measurements are fed to the blockchain component and the smart contracts, the anomaly detection mechanism uniquely combines artificial intelligence models to effectively detect unusual values such as outliers and extreme deviations in data coming from them. Specifically, Autoregressive integrated moving average, Long short-term memory (LSTM) and Dense-based autoencoders, as well as Generative adversarial networks (GAN) models, are used to detect both point and collective anomalies. Towards the goal of preserving the privacy of industries' information, the smart contracts employ techniques to ensure that only anonymized pointers to the actual data are stored on the ledger while sensitive information remains off-chain. In the same spirit, blockchain technology guarantees the security of the data storage through strong cryptography as well as the integrity of the data through the decentralization of the network and the execution of the smart contracts by the majority of the blockchain network actors. The blockchain component of the Data Traceability Software is based on the Hyperledger Fabric framework, which lays the ground for the deployment of smart contracts and APIs to expose the functionality to the end-users. The results of this work demonstrate that such a system can increase the quality of the end-products and the trustworthiness of the monitoring process in the smart manufacturing domain. The proposed AI-enabled data traceability software can be employed by industries to accurately trace and verify records about quality through the entire production chain and take advantage of the multitude of monitoring records in their databases.

Keywords: blockchain, data quality, industry4.0, product quality

Procedia PDF Downloads 151

Authors: Dmitry A. Koptelov, Leonid M. Mestetskiy

Abstract:

Voronoi Diagram (VD) of finite set of disjoint simple polygons, called sites, is a partition of plane into loci (for each site at the locus) – regions, consisting of points that are closer to a given site than to all other. Set of polygons is a universal model for many applications in engineering, geoinformatics, design, computer vision, and graphics. VD of polygons construction usually done with a reduction to task of constructing VD of segments, for which there are effective O(n log n) algorithms for n segments. Preprocessing – constructing segments from polygons’ sides, and postprocessing – polygon’s loci construction by merging the loci of the sides of each polygon are also included in reduction. This approach doesn’t take into account two specific properties of the resulting segment sites. Firstly, all this segments are connected in pairs in the vertices of the polygons. Secondly, on the one side of each segment lies the interior of the polygon. The polygon is obviously included in its locus. Using this properties in the algorithm for VD construction is a resource to reduce computations. The article proposes an algorithm for the direct construction of VD of polygonal sites. Algorithm is based on sweepline paradigm, allowing to effectively take into account these properties. The solution is performed based on reduction. Preprocessing is the constructing of set of sites from vertices and edges of polygons. Each site has an orientation such that the interior of the polygon lies to the left of it. Proposed algorithm constructs VD for set of oriented sites with sweepline paradigm. Postprocessing is a selecting of edges of this VD formed by the centers of empty circles touching different polygons. Improving the efficiency of the proposed sweepline algorithm in comparison with the general Fortune algorithm is achieved due to the following fundamental solutions: 1. Algorithm constructs only such VD edges, which are on the outside of polygons. Concept of oriented sites allowed to avoid construction of VD edges located inside the polygons. 2. The list of events in sweepline algorithm has a special property: the majority of events are connected with “medium” polygon vertices, where one incident polygon side lies behind the sweepline and the other in front of it. The proposed algorithm processes such events in constant time and not in logarithmic time, as in the general Fortune algorithm. The proposed algorithm is fully implemented and tested on a large number of examples. The high reliability and efficiency of the algorithm is also confirmed by computational experiments with complex sets of several thousand polygons. It should be noted that, despite the considerable time that has passed since the publication of Fortune's algorithm in 1986, a full-scale implementation of this algorithm for an arbitrary set of segment sites has not been made. The proposed algorithm fills this gap for an important special case - a set of sites formed by polygons.

Keywords: voronoi diagram, sweepline, polygon sites, fortunes' algorithm, segment sites

Procedia PDF Downloads 144

Authors: Julie M. Novak, Simone K. Brennan, Lacey Brim

Abstract:

Undergraduate medical education (UME) curriculum notably addresses micro-level communications (e.g., patient-provider, intercultural, inter-professional), yet frequently under-examines the role and impact of organizational communication, a more macro-level. Organizational communication, however, functions as foundation and through systemic structures of an organization and thereby serves as hidden curriculum and influences learning experiences and outcomes. Yet, little available research exists fully examining how students experience organizational communication while in medical school. Extant literature and best practices provide insufficient guidance for UME programs, in particular. The purpose of this study was to map and examine current organizational communication systems and processes in a UME program. Employing a phenomenology-grounded and participatory approach, this study sought to understand the organizational communication system from medical students' perspective. The research team consisted of a core team and 13 medical student co-investigators. This research employed multiple methods, including focus groups, individual interviews, and two surveys (one reflective of focus group questions, the other requesting students to submit ‘examples’ of communications). To provide context for student responses, nonstudent participants (faculty, administrators, and staff) were sampled, as they too express concerns about communication. Over 400 students across all cohorts and 17 nonstudents participated. Data were iteratively analyzed and checked for triangulation. Findings reveal the complex nature of organizational communication and student-oriented communications. They reveal program-impactful strengths, weaknesses, gaps, and tensions and speak to the role of organizational communication practices influencing both climate and culture. With regard to communications, students receive multiple, simultaneous communications from multiple sources/channels, both formal (e.g., official email) and informal (e.g., social media). Students identified organizational strengths including the desire to improve student voice, and message frequency. They also identified weaknesses related to over-reliance on emails, numerous platforms with inconsistent utilization, incorrect information, insufficient transparency, assessment/input fatigue, tacit expectations, scheduling/deadlines, responsiveness, and mental health confidentiality concerns. Moreover, they noted gaps related to lack of coordination/organization, ambiguous point-persons, student ‘voice-only’, open communication loops, lack of core centralization and consistency, and mental health bridges. Findings also revealed organizational identity and cultural characteristics as impactful on the medical school experience. Cultural characteristics included program size, diversity, urban setting, student organizations, community-engagement, crisis framing, learning for exams, inefficient bureaucracy, and professionalism. Moreover, they identified system structures that do not always leverage cultural strengths or reduce cultural problematics. Based on the results, opportunities for productive change are identified. These include leadership visibly supporting and enacting overall organizational narratives, making greater efforts in consistently ‘closing the loop’, regularly sharing how student input effects change, employing strategies of crisis communication more often, strengthening communication infrastructure, ensuring structures facilitate effective operations and change efforts, and highlighting change efforts in informational communication. Organizational communication and communications are not soft-skills, or of secondary concern within organizations, rather they are foundational in nature and serve to educate/inform all stakeholders. As primary stakeholders, students and their success directly affect the accomplishment of organizational goals. This study demonstrates how inquiries about how students navigate their educational experience extends research-based knowledge and provides actionable knowledge for the improvement of organizational operations in UME.

Keywords: medical education programs, organizational communication, participatory research, qualitative mixed methods

Procedia PDF Downloads 92

Authors: Mohamed Alimoussa, Alice Porebski, Nicolas Vandenbroucke, Rachid Oulad Haj Thami, Sana El Fkihi

Abstract:

Color and texture are highly discriminant visual cues that provide an essential information in many types of images. Color texture representation and classification is therefore one of the most challenging problems in computer vision and image processing applications. Color textures can be represented in different color spaces by using multiple image descriptors which generate a high dimensional set of texture features. In order to reduce the dimensionality of the feature set, feature selection techniques can be used. The goal of feature selection is to find a relevant subset from an original feature space that can improve the accuracy and efficiency of a classification algorithm. Traditionally, feature selection is focused on removing irrelevant features, neglecting the possible redundancy between relevant ones. This is why some feature selection approaches prefer to use feature clustering analysis to aid and guide the search. These techniques can be divided into two categories. i) Feature clustering-based ranking algorithm uses feature clustering as an analysis that comes before feature ranking. Indeed, after dividing the feature set into groups, these approaches perform a feature ranking in order to select the most discriminant feature of each group. ii) Feature clustering-based subset search algorithms can use feature clustering following one of three strategies; as an initial step that comes before the search, binded and combined with the search or as the search alternative and replacement. In this paper, we propose a new feature clustering-based sequential selection approach for the purpose of color texture representation and classification. Our approach is a three step algorithm. First, irrelevant features are removed from the feature set thanks to a class-correlation measure. Then, introducing a new automatic feature clustering algorithm, the feature set is divided into several feature clusters. Finally, a sequential search algorithm, based on a filter model and a separability measure, builds a relevant and non redundant feature subset: at each step, a feature is selected and features of the same cluster are removed and thus not considered thereafter. This allows to significantly speed up the selection process since large number of redundant features are eliminated at each step. The proposed algorithm uses the clustering algorithm binded and combined with the search. Experiments using a combination of two well known texture descriptors, namely Haralick features extracted from Reduced Size Chromatic Co-occurence Matrices (RSCCMs) and features extracted from Local Binary patterns (LBP) image histograms, on five color texture data sets, Outex, NewBarktex, Parquet, Stex and USPtex demonstrate the efficiency of our method compared to seven of the state of the art methods in terms of accuracy and computation time.

Keywords: feature selection, color texture classification, feature clustering, color LBP, chromatic cooccurrence matrix

Procedia PDF Downloads 101

Authors: Farina Riaz, Shahab Abdulla, Srinjoy Ganguly, Hajime Suzuki, Ravinesh C. Deo, Susan Hopkins

Abstract:

Recently, quantum machine learning has received significant attention. For various types of data, including text and images, numerous quantum machine learning (QML) models have been created and are being tested. Images are exceedingly complex data components that demand more processing power. Despite being mature, classical machine learning still has difficulties with big data applications. Furthermore, quantum technology has revolutionized how machine learning is thought of, by employing quantum features to address optimization issues. Since quantum hardware is currently extremely noisy, it is not practicable to run machine learning algorithms on it without risking the production of inaccurate results. To discover the advantages of quantum versus classical approaches, this research has concentrated on colored image data. Deep learning classification models are currently being created on Quantum platforms, but they are still in a very early stage. Black and white benchmark image datasets like MNIST and Fashion MINIST have been used in recent research. MNIST and CIFAR-10 were compared for binary classification, but the comparison showed that MNIST performed more accurately than colored CIFAR-10. This research will evaluate the performance of the QML algorithm on the colored benchmark dataset CIFAR-10 to advance QML's real-time applicability. However, deep learning classification models have not been developed to compare colored images like Quantum Convolutional Neural Network (QCNN) to determine how much it is better to classical. Only a few models, such as quantum variational circuits, take colored images. The methodology adopted in this research is a hybrid approach by using penny lane as a simulator. To process the 10 classes of CIFAR-10, the image data has been translated into grey scale and the 28 × 28-pixel image containing 10,000 test and 50,000 training images were used. The objective of this work is to determine how much the quantum approach can outperform a classical approach for a comprehensive dataset of color images. After pre-processing 50,000 images from a classical computer, the QCNN model adopted a hybrid method and encoded the images into a quantum simulator for feature extraction using quantum gate rotations. The measurements were carried out on the classical computer after the rotations were applied. According to the results, we note that the QCNN approach is ~12% more effective than the traditional classical CNN approaches and it is possible that applying data augmentation may increase the accuracy. This study has demonstrated that quantum machine and deep learning models can be relatively superior to the classical machine learning approaches in terms of their processing speed and accuracy when used to perform classification on colored classes.

Keywords: CIFAR-10, quantum convolutional neural networks, quantum deep learning, quantum machine learning

Procedia PDF Downloads 94

Authors: Ameni Mejri, Moufida Hajjaj, Salem Hasnaoui, Ridha Bouallegue

Abstract:

The technology advances have most lately made the millimeter wave (mmWave) communication possible. Due to the huge amount of spectrum that is available in MmWave frequency bands, this promising candidate is considered as a key technology for the deployment of 5G cellular networks. In order to enhance system capacity and achieve spectral efficiency, very large antenna arrays are employed at mmWave systems by exploiting array gain. However, it has been shown that conventional beamforming strategies are not suitable for mmWave hardware implementation. Therefore, new features are required for mmWave cellular applications. Unlike traditional multiple-input-multiple-output (MIMO) systems for which only digital precoders are essential to accomplish precoding, MIMO technology seems to be different at mmWave because of digital precoding limitations. Moreover, precoding implements a greater number of radio frequency (RF) chains supporting more signal mixers and analog-to-digital converters. As RF chain cost and power consumption is increasing, we need to resort to another alternative. Although the hybrid precoding architecture has been regarded as the best solution based on a combination between a baseband precoder and an RF precoder, we still do not get the optimal design of hybrid precoders. According to the mapping strategies from RF chains to the different antenna elements, there are two main categories of hybrid precoding architecture. Given as a hybrid precoding sub-array architecture, the partially-connected structure reduces hardware complexity by using a less number of phase shifters, whereas it sacrifices some beamforming gain. In this paper, we treat the hybrid precoder design in mmWave MIMO systems as a problem of matrix factorization. Thus, we adopt the alternating minimization principle in order to solve the design problem. Further, we present our proposed algorithm for the partially-connected structure, which is based on the iterative hard thresholding method. Through simulation results, we show that our hybrid precoding algorithm provides significant performance gains over existing algorithms. We also show that the proposed approach reduces significantly the computational complexity. Furthermore, valuable design insights are provided when we use the proposed algorithm to make simulation comparisons between the hybrid precoding partially-connected structure and the fully-connected structure.

Keywords: alternating minimization, hybrid precoding, iterative hard thresholding, low-complexity, millimeter wave communication, partially-connected structure

Procedia PDF Downloads 293

Authors: Lalith Edirisinghe

Abstract:

Due to the variability in global trading patterns, some ports in the world experience a shortage of shipping containers while the rest of the ports have excess container stocks. According to this study, carriers operate and manage their container inventories independently, leading to enormous container repositioning costs. In contrast, the researcher suggests that costs can be minimized if carriers exchange containers among them. In other words, rather than repositioning excess containers, a carrier could offer them to another carrier in the same port that has a shortage and vice versa. However, this is easier said than done because there is huge complexity in global container management as it involves many operational parameters such as multiple types and sizes of containers, the varying transit times of different carriers, etc., and the exchange may take place in various ports globally. Therefore, the exchange should be facilitated through a fully comprehensive automated computer system that could consider all the parameters that impact the possibility of exchange containers. Accordingly, the research used mixed research methods, combining qualitative and quantitative approaches. Data analysis is conducted using SPSS tools, and a prototype is developed as the output of the research. The proposed mathematical solution will proactively scan through the container size, type, and volume of every member carrier in each port and map how the deficit and excess quantities could be shared among them and set off the imbalance of empty container reposition at ports of their interest. The approach includes obtaining and processing container inventory information from multiple parties in real time for assessing container data associated with each party for each port at a given time. Using the container data, container inventories for each party at each port for a defined time are forecasted. A first party having surplus (offeror) and deficit (offeree) of empty containers at a first and a second port at a first time, respectively, is determined. A second party having a deficit and surplus of empty containers at the first time, respectively, is determined. Offering the first and the second party a container exchange opportunity to enable the first party to supply surplus empty containers to the second party at the first port based on the first container characteristics and the second party to supply surplus empty containers to the first party at the second port based on the second container characteristics. After the offeree obtains containers, they will be shipped to a port determined by the exporters. To ensure the sustainability of this method, the system should provide equal benefits to both the offeror and the offeree. Accordingly, the system will consider not only the number of containers exchanged but also the duration the offeree may hold them in its custody. This reduces container repositioning costs by utilizing mathematical modeling, algorithms, big data, machine learning, and artificial intelligence. This method and system may reduce the container repositioning cost by twenty percent.

Keywords: container inventory, benefit of exchange, reposition, imbalance, shipping, carriers, offeree, offeror

Procedia PDF Downloads 34

Authors: Asma Omri, Noureddine Benothman, Sofiane Sayahi, Fethi Tlili, Hichem Besbes

Abstract:

Cycling, a popular mode of transportation, can also be perilous due to cyclists' vulnerability to collisions with vehicles and obstacles. This paper presents an innovative cyclist safety system based on radar technology designed to offer real-time collision risk warnings to cyclists. The system incorporates a low-power radar sensor affixed to the bicycle and connected to a microcontroller. It leverages radar point cloud detections, a clustering algorithm, and a supervised classifier. These algorithms are optimized for efficiency to run on the TI’s AWR 1843 BOOST radar, utilizing a coarse classification approach distinguishing between cars, trucks, two-wheeled vehicles, and other objects. To enhance the performance of clustering techniques, we propose a 2-Level clustering approach. This approach builds on the state-of-the-art Density-based spatial clustering of applications with noise (DBSCAN). The objective is to first cluster objects based on their velocity, then refine the analysis by clustering based on position. The initial level identifies groups of objects with similar velocities and movement patterns. The subsequent level refines the analysis by considering the spatial distribution of these objects. The clusters obtained from the first level serve as input for the second level of clustering. Our proposed technique surpasses the classical DBSCAN algorithm in terms of geometrical metrics, including homogeneity, completeness, and V-score. Relevant cluster features are extracted and utilized to classify objects using an SVM classifier. Potential obstacles are identified based on their velocity and proximity to the cyclist. To optimize the system, we used the View of Delft dataset for hyperparameter selection and SVM classifier training. The system's performance was assessed using our collected dataset of radar point clouds synchronized with a camera on an Nvidia Jetson Nano board. The radar-based cyclist safety system is a practical solution that can be easily installed on any bicycle and connected to smartphones or other devices, offering real-time feedback and navigation assistance to cyclists. We conducted experiments to validate the system's feasibility, achieving an impressive 85% accuracy in the classification task. This system has the potential to significantly reduce the number of accidents involving cyclists and enhance their safety on the road.

Keywords: 2-level clustering, coarse classification, cyclist safety, warning system based on radar technology

Procedia PDF Downloads 46

Authors: Dosun Shin, Assegid Kidane, Pavan Turaga

Abstract:

According to the National Institute of Health, living a sedentary lifestyle leads to a number of health issues, including increased risk of cardiovascular dis-ease, type 2 diabetes, obesity, and certain types of cancers. This project brings together experts in multiple disciplines to bring product design, sensor design, algorithms, and health intervention studies to develop a product and system that helps reduce the amount of time sitting at the workplace. This paper illustrates ongoing improvements to prototypes the research team developed in initial research; including working prototypes with a software application, which were developed and demonstrated for users. Additional modifications were made to improve functionality, aesthetics, and ease of use, which will be discussed in this paper. Extending on the foundations created in the initial phase, our approach sought to further improve the product by conducting additional human factor research, studying deficiencies in competitive products, testing various materials/forms, developing working prototypes, and obtaining feedback from additional potential users. The solution consisted of an aesthetically pleasing seat cover cushion that easily attaches to common office chairs found in most workplaces, ensuring a wide variety of people can use the product. The product discreetly contains sensors that track when the user sits on their chair, sending information to a phone app that triggers reminders for users to stand up and move around after sitting for a set amount of time. This paper also presents the analyzed typical office aesthetics and selected materials, colors, and forms that complimented the working environment. Comfort and ease of use remained a high priority as the design team sought to provide a product and system that integrated into the workplace. As the research team continues to test, improve, and implement this solution for the sedentary workplace, the team seeks to create a viable product that acts as an impetus for a more active workday and lifestyle, further decreasing the proliferation of chronic disease and health issues for sedentary working people. This paper illustrates in detail the processes of engineering, product design, methodology, and testing results.

Keywords: anti-sedentary work behavior, new product development, sensor design, health intervention studies

Procedia PDF Downloads 127

Authors: África Vicario, Fernando J. Álvarez, Felipe Parralejo, Fernando Aranda

Abstract:

The irrigation systems of traditional agriculture, such as gravity-fed irrigation, produce a great waste of water because, generally, there is no control over the amount of water supplied in relation to the water needed. The AIRON Project tries to solve this problem by implementing an IoT-based system to sensor the irrigation plots so that the state of the crops and the amount of water used for irrigation can be known remotely. The IoT system consists of a sensor network that measures the humidity of the soil, the weather conditions (temperature, relative humidity, wind and solar radiation) and the irrigation water flow. The communication between this network and a central gateway is conducted by means of long-range wireless communication that depends on the characteristics of the irrigation plot. The main objective of the AIRON project is to deploy an IoT sensor network in two different plots of the irrigation community of Aranjuez in the Spanish region of Madrid. The first plot is 2 km away from the central gateway, so LoRa has been used as the base communication technology. The problem with this plot is the absence of mains electric power, so devices with energy-saving modes have had to be used to maximize the external batteries' use time. An ESP32 SOC board with a LoRa module is employed in this case to gather data from the sensor network and send them to a gateway consisting of a Raspberry Pi with a LoRa hat. The second plot is located 18 km away from the gateway, a range that hampers the use of LoRa technology. In order to establish reliable communication in this case, the long-term evolution (LTE) standard is used, which makes it possible to reach much greater distances by using the cellular network. As mains electric power is available in this plot, a Raspberry Pi has been used instead of the ESP32 board to collect sensor data. All data received from the two plots are stored on a proprietary server located at the irrigation management company's headquarters. The analysis of these data by means of machine learning algorithms that are currently under development should allow a short-term prediction of the irrigation water demand that would significantly reduce the waste of this increasingly valuable natural resource. The major finding of this work is the real possibility of deploying a remote sensing system for irrigated plots by using Commercial-Off-The-Shelf (COTS) devices, easily scalable and adaptable to design requirements such as the distance to the control center or the availability of mains electrical power at the site.

Keywords: internet of things, irrigation water control, LoRa, LTE, smart farming

Procedia PDF Downloads 44

Authors: Zainab Atia, Hongwei He, Panagiotis Sarantopoulos

Abstract:

Over the past few years, firms have witnessed a massive increase in sophisticated algorithmic deployment, which has become quite pervasive in today’s modern society. With the wide availability of data for retailers, the ability to track consumers using algorithmic pricing has become an integral option in online platforms. As more companies are transforming their businesses and relying more on massive technological advancement, pricing algorithmic systems have brought attention and given rise to its wide adoption, with many accompanying benefits and challenges to be found within its usage. With the overall aim of increasing profits by organizations, algorithmic pricing is becoming a sound option by enabling suppliers to cut costs, allowing better services, improving efficiency and product availability, and enhancing overall consumer experiences. The adoption of algorithms in retail has been pioneered and widely used in literature across varied fields, including marketing, computer science, engineering, economics, and public policy. However, what is more, alarming today is the comprehensive understanding and focus of this technology and its associated ethical influence on consumers’ perceptions and behaviours. Indeed, due to algorithmic ethical concerns, consumers are found to be reluctant in some instances to share their personal data with retailers, which reduces their retention and leads to negative consumer outcomes in some instances. This, in its turn, raises the question of whether firms can still manifest the acceptance of such technologies by consumers while minimizing the ethical transgressions accompanied by their deployment. As recent modest research within the area of marketing and consumer behavior, the current research advances the literature on algorithmic pricing, pricing ethics, consumers’ perceptions, and price fairness literature. With its empirical focus, this paper aims to contribute to the literature by applying the distinction of the two common types of algorithmic pricing, dynamic and personalized, while measuring their relative effect on consumers’ behavioural outcomes. From a managerial perspective, this research offers significant implications that pertain to providing a better human-machine interactive environment (whether online or offline) to improve both businesses’ overall performance and consumers’ wellbeing. Therefore, by allowing more transparent pricing systems, businesses can harness their generated ethical strategies, which fosters consumers’ loyalty and extend their post-purchase behaviour. Thus, by defining the correct balance of pricing and right measures, whether using dynamic or personalized (or both), managers can hence approach consumers more ethically while taking their expectations and responses at a critical stance.

Keywords: algorithmic pricing, dynamic pricing, personalized pricing, price ethicality

Procedia PDF Downloads 59

Authors: Jerome Joshi

Abstract:

The paper presents the modelling of financial markets using the Stochastic Pi Calculus model. The Stochastic Pi Calculus model is mainly used for biological applications; however, the feature of this model promotes its use in financial markets, more prominently in high frequency trading. The trading system can be broadly classified into exchange, market makers or intermediary traders and fundamental traders. The exchange is where the action of the trade is executed, and the two types of traders act as market participants in the exchange. High frequency trading, with its complex networks and numerous market participants (intermediary and fundamental traders) poses a difficulty while modelling. It involves the participants to seek the advantage of complex trading algorithms and high execution speeds to carry out large volumes of trades. To earn profits from each trade, the trader must be at the top of the order book quite frequently by executing or processing multiple trades simultaneously. This would require highly automated systems as well as the right sentiment to outperform other traders. However, always being at the top of the book is also not best for the trader, since it was the reason for the outbreak of the ‘Hot – Potato Effect,’ which in turn demands for a better and more efficient model. The characteristics of the model should be such that it should be flexible and have diverse applications. Therefore, a model which has its application in a similar field characterized by such difficulty should be chosen. It should also be flexible in its simulation so that it can be further extended and adapted for future research as well as be equipped with certain tools so that it can be perfectly used in the field of finance. In this case, the Stochastic Pi Calculus model seems to be an ideal fit for financial applications, owing to its expertise in the field of biology. It is an extension of the original Pi Calculus model and acts as a solution and an alternative to the previously flawed algorithm, provided the application of this model is further extended. This model would focus on solving the problem which led to the ‘Flash Crash’ which is the ‘Hot –Potato Effect.’ The model consists of small sub-systems, which can be integrated to form a large system. It is designed in way such that the behavior of ‘noise traders’ is considered as a random process or noise in the system. While modelling, to get a better understanding of the problem, a broader picture is taken into consideration with the trader, the system, and the market participants. The paper goes on to explain trading in exchanges, types of traders, high frequency trading, ‘Flash Crash,’ ‘Hot-Potato Effect,’ evaluation of orders and time delay in further detail. For the future, there is a need to focus on the calibration of the module so that they would interact perfectly with other modules. This model, with its application extended, would provide a basis for researchers for further research in the field of finance and computing.

Keywords: concurrent computing, high frequency trading, financial markets, stochastic pi calculus

Procedia PDF Downloads 50

Authors: Katharina Eberhardt, Florian Klaus Kaiser, Frank Schultmann

Abstract:

Last-mile distribution is one of the most critical parts of a disaster relief operation. Various uncertainties, such as infrastructure conditions, resource availability, and fluctuating beneficiary demand, render last-mile distribution challenging in disaster relief operations. The need to balance critical performance criteria like response time, meeting demand and cost-effectiveness further complicates the task. The occurrence of disasters cannot be controlled, and the magnitude is often challenging to assess. In summary, these uncertainties create a need for additional flexibility, agility, and preparedness in logistics operations. As a result, strategic planning and efficient network design are critical for an effective and efficient response. Furthermore, the increasing frequency of disasters and the rising cost of logistical operations amplify the need to provide robust and resilient solutions in this area. Therefore, we formulate a scenario-based bi-objective optimization model that integrates pre-positioning, allocation, and distribution of relief supplies extending the general form of a covering location problem. The proposed model aims to minimize underlying logistics costs while maximizing demand coverage. Using a set of disruption scenarios, the model allows decision-makers to identify optimal network solutions to address the risk of disruptions. We provide an empirical case study of the public authorities’ emergency food storage strategy in Germany to illustrate the potential applicability of the model and provide implications for decision-makers in a real-world setting. Also, we conduct a sensitivity analysis focusing on the impact of varying stockpile capacities, single-site outages, and limited transportation capacities on the objective value. The results show that the stockpiling strategy needs to be consistent with the optimal number of depots and inventory based on minimizing costs and maximizing demand satisfaction. The strategy has the potential for optimization, as network coverage is insufficient and relies on very high transportation and personnel capacity levels. As such, the model provides decision support for public authorities to determine an efficient stockpiling strategy and distribution network and provides recommendations for increased resilience. However, certain factors have yet to be considered in this study and should be addressed in future works, such as additional network constraints and heuristic algorithms.

Keywords: humanitarian logistics, bi-objective optimization, pre-positioning, last mile distribution, decision support, disaster relief networks

Procedia PDF Downloads 48

Authors: Yuta Nakano, Kozo Kajiwara, Atsushi Hori, Takeshi Fujita

Abstract:

In association with the wave of electric vehicles (EV), low energy consumption systems have become more and more important. One of the key technologies to realize low energy consumption is a dynamic vision sensor (DVS), or we can call it an event sensor, neuromorphic vision sensor and so on. This sensor has several features, such as high temporal resolution, which can achieve 1 Mframe/s, and a high dynamic range (120 DB). However, the point that can contribute to low energy consumption the most is its sparsity; to be more specific, this sensor only captures the pixels that have intensity change. In other words, there is no signal in the area that does not have any intensity change. That is to say, this sensor is more energy efficient than conventional sensors such as RGB cameras because we can remove redundant data. On the other side of the advantages, it is difficult to handle the data because the data format is completely different from RGB image; for example, acquired signals are asynchronous and sparse, and each signal is composed of x-y coordinate, polarity (two values: +1 or -1) and time stamp, it does not include intensity such as RGB values. Therefore, as we cannot use existing algorithms straightforwardly, we have to design a new processing algorithm to cope with DVS data. In order to solve difficulties caused by data format differences, most of the prior arts make a frame data and feed it to deep learning such as Convolutional Neural Networks (CNN) for object detection and recognition purposes. However, even though we can feed the data, it is still difficult to achieve good performance due to a lack of intensity information. Although polarity is often used as intensity instead of RGB pixel value, it is apparent that polarity information is not rich enough. Considering this context, we proposed to use the timestamp information as a data representation that is fed to deep learning. Concretely, at first, we also make frame data divided by a certain time period, then give intensity value in response to the timestamp in each frame; for example, a high value is given on a recent signal. We expected that this data representation could capture the features, especially of moving objects, because timestamp represents the movement direction and speed. By using this proposal method, we made our own dataset by DVS fixed on a parked car to develop an application for a surveillance system that can detect persons around the car. We think DVS is one of the ideal sensors for surveillance purposes because this sensor can run for a long time with low energy consumption in a NOT dynamic situation. For comparison purposes, we reproduced state of the art method as a benchmark, which makes frames the same as us and feeds polarity information to CNN. Then, we measured the object detection performances of the benchmark and ours on the same dataset. As a result, our method achieved a maximum of 7 points greater than the benchmark in the F1 score.

Keywords: event camera, dynamic vision sensor, deep learning, data representation, object recognition, low energy consumption

Procedia PDF Downloads 65

Authors: Merve Yünlü, Nihat Kandemir, Aylin Ersoy

Abstract:

Temperature sensors are used in home appliances not only to monitor the basic functions of the machine but also to minimize energy consumption and ensure safe operation. In parallel with the development of smart home applications and IoT algorithms, these sensors produce important data such as the frequency of use of the machine, user preferences, and the compilation of critical data in terms of diagnostic processes for fault detection throughout an appliance's operational lifespan. Commercially available thin-film resistive temperature sensors have a well-established manufacturing procedure that allows them to operate over a wide temperature range. However, these sensors are over-designed for white goods applications. The operating temperature range of these sensors is between -70°C and 850°C, while the temperature range requirement in home appliance applications is between 23°C and 500°C. To ensure the operation of commercial sensors in this wide temperature range, usually, a platinum coating of approximately 1-micron thickness is applied to the wafer. However, the use of platinum in coating and the high coating thickness extends the sensor production process time and therefore increases sensor costs. In this study, an attempt was made to develop a low-cost temperature sensor design and production method that meets the technical requirements of white goods applications. For this purpose, a custom design was made, and design parameters (length, width, trim points, and thin film deposition thickness) were optimized by using statistical methods to achieve the desired resistivity value. To develop thin film resistive temperature sensors, one side polished sapphire wafer was used. To enhance adhesion and insulation 100 nm silicon dioxide was coated by inductively coupled plasma chemical vapor deposition technique. The lithography process was performed by a direct laser writer. The lift-off process was performed after the e-beam evaporation of 10 nm titanium and 280 nm platinum layers. Standard four-point probe sheet resistance measurements were done at room temperature. The annealing process was performed. Resistivity measurements were done with a probe station before and after annealing at 600°C by using a rapid thermal processing machine. Temperature dependence between 25-300 °C was also tested. As a result of this study, a temperature sensor has been developed that has a lower coating thickness than commercial sensors but can produce reliable data in the white goods application temperature range. A relatively simplified but optimized production method has also been developed to produce this sensor.

Keywords: thin film resistive sensor, temperature sensor, household appliance, sustainability, energy efficiency

Procedia PDF Downloads 45

Authors: Loris Franchi, Daniele Calvi, Sabrina Corpino

Abstract:

This paper presents a design methodology in which stakeholders are assisted with the exploration of a so-called negotiation space, aiming to the maximization of both group social welfare and single stakeholder’s perceived utility. The outcome results in less design iterations needed for design convergence while obtaining a higher solution effectiveness. During the early stage of a space project, not only the knowledge about the system but also the decision outcomes often are unknown. The scenario is exacerbated by the fact that decisions taken in this stage imply delayed costs associated with them. Hence, it is necessary to have a clear definition of the problem under analysis, especially in the initial definition. This can be obtained thanks to a robust generation and exploration of design alternatives. This process must consider that design usually involves various individuals, who take decisions affecting one another. An effective coordination among these decision-makers is critical. Finding mutual agreement solution will reduce the iterations involved in the design process. To handle this scenario, the paper proposes a design methodology which, aims to speed-up the process of pushing the mission’s concept maturity level. This push up is obtained thanks to a guided negotiation space exploration, which involves autonomously exploration and optimization of trade opportunities among stakeholders via Artificial Intelligence algorithms. The negotiation space is generated via a multidisciplinary collaborative optimization method, infused by game theory and multi-attribute utility theory. In particular, game theory is able to model the negotiation process to reach the equilibria among stakeholder needs. Because of the huge dimension of the negotiation space, a collaborative optimization framework with evolutionary algorithm has been integrated in order to guide the game process to efficiently and rapidly searching for the Pareto equilibria among stakeholders. At last, the concept of utility constituted the mechanism to bridge the language barrier between experts of different backgrounds and differing needs, using the elicited and modeled needs to evaluate a multitude of alternatives. To highlight the benefits of the proposed methodology, the paper presents the design of a CubeSat mission for the observation of lunar radiation environment. The derived solution results able to balance all stakeholders needs and guaranteeing the effectiveness of the selection mission concept thanks to its robustness in valuable changeability. The benefits provided by the proposed design methodology are highlighted, and further development proposed.

Keywords: concurrent engineering, artificial intelligence, negotiation in engineering design, multidisciplinary optimization

Procedia PDF Downloads 107

Authors: Christoph Opperer

Abstract:

Space syntax is a theory and method for analyzing the spatial layout of buildings and urban environments to understand how they can influence patterns of human movement, social interaction, and behavior. While direct visibility is a key factor in space syntax research, important visual information such as light, color, texture, etc., are typically not considered, even though psychological studies have shown a strong correlation to the human perceptual experience within physical space – with light and color, for example, playing a crucial role in shaping the perception of spaciousness. Furthermore, these surface properties are often the visual features that are most salient and responsible for drawing attention to certain elements within the environment. This paper explores the potential of integrating these factors into general space syntax methods and visibility-based analysis of space, particularly for architectural spatial layouts. To this end, we use a combination of geometric (isovist) and topological (visibility graph) approaches together with image-based methods, allowing a comprehensive exploration of the relationship between spatial geometry, visual aesthetics, and human experience. Custom-coded ray-tracing techniques are employed to generate spherical panorama images, encoding three-dimensional spatial data in the form of two-dimensional images. These images are then processed through computer vision algorithms to generate saliency-maps, which serve as a visual representation of areas most likely to attract human attention based on their visual properties. The maps are subsequently used to weight the vertices of isovists and the visibility graph, placing greater emphasis on areas with high saliency. Compared to traditional methods, our weighted visibility analysis introduces an additional layer of information density by assigning different weights or importance levels to various aspects within the field of view. This extends general space syntax measures to provide a more nuanced understanding of visibility patterns that better reflect the dynamics of human attention and perception. Furthermore, by drawing parallels to traditional isovist and VGA analysis, our weighted approach emphasizes a crucial distinction, which has been pointed out by Ervin and Steinitz: the difference between what is possible to see and what is likely to be seen. Therefore, this paper emphasizes the importance of including surface properties in visibility-based analysis to gain deeper insights into how people interact with their surroundings and to establish a stronger connection with human attention and perception.

Keywords: space syntax, visibility analysis, isovist, visibility graph, visual features, human perception, saliency detection, raytracing, spherical images

Procedia PDF Downloads 41

Authors: Megan Burfoot, Ali GhaffarianHoseini, Nicola Naismith, Amirhosein GhaffarianHoseini

Abstract:

Intelligent Acoustic Technology (IAT) is a novel architectural device used in buildings to automatically vary the acoustic conditions of space. IAT is realized by integrating two components: Variable Acoustic Technology (VAT) and an intelligent system. The VAT passively alters the RT by changing the total sound absorption in a room. In doing so, the Reverberation Time (RT) is changed and thus, the sound strength and clarity are altered. The intelligent system detects sound waves in real-time to identify the aural situation, and the RT is adjusted accordingly based on pre-programmed algorithms. IAT - the synthesis of these two components - can dramatically improve acoustic comfort, as the acoustic condition is automatically optimized for any detected aural situation. This paper presents an evaluation of the improvements of acoustic comfort in an existing tertiary classroom located at Auckland University of Technology in New Zealand. This is a pilot case study, the first of its’ kind attempting to quantify the benefits of IAT. Naturally, the potential acoustic improvements from IAT can be actualized by only installing the VAT component of IAT and by manually adjusting it rather than utilizing an intelligent system. Such a simplified methodology is adopted for this case study to understand the potential significance of IAT without adopting a time and cost-intensive strategy. For this study, the VAT is built by overlaying reflective, rotating louvers over sound absorption panels. RT's are measured according to international standards before and after installing VAT in the classroom. The louvers are manually rotated in increments by the experimenter and further RT measurements are recorded. The results are compared with recommended guidelines and reference values from national standards for spaces intended for speech and communication. The results obtained from the measurements are used to quantify the potential improvements in classroom acoustic comfort, where IAT to be used. This evaluation reveals the current existence of poor acoustic conditions in the classroom caused by high RT's. The poor acoustics are also largely attributed to the classrooms’ inability to vary acoustic parameters for changing aural situations. The classroom experiences one static acoustic state, neglecting to recognize the nature of classrooms as flexible, dynamic spaces. Evidently, when using VAT the classroom is prescribed with a wide range of RTs it can achieve. Namely, acoustic requirements for varying teaching approaches are satisfied, and acoustic comfort is improved. By quantifying the benefits of using VAT, it can confidently suggest these same benefits are achieved with IAT. Nevertheless, it is encouraged that future studies continue this line of research toward the eventual development of IAT and its’ acceptance into mainstream architecture.

Keywords: acoustic comfort, classroom acoustics, intelligent acoustics, variable acoustics

Procedia PDF Downloads 165

Authors: M. Bahari Mehrabani, Hua-Peng Chen

Abstract:

Management and maintenance of coastal defence structures during the expected life cycle have become a real challenge for decision makers and engineers. Accurate evaluation of the current condition and future performance of flood defence structures is essential for effective practical maintenance strategies on the basis of available field inspection data. Moreover, as coastal defence structures age, it becomes more challenging to implement maintenance and management plans to avoid structural failure. Therefore, condition inspection data are essential for assessing damage and forecasting deterioration of ageing flood defence structures in order to keep the structures in an acceptable condition. The inspection data for flood defence structures are often collected using discrete visual condition rating schemes. In order to evaluate future condition of the structure, a probabilistic deterioration model needs to be utilised. However, existing deterioration models may not provide a reliable prediction of performance deterioration for a long period due to uncertainties. To tackle the limitation, a time-dependent condition-based model associated with a transition probability needs to be developed on the basis of condition grade scheme for flood defences. This paper presents a probabilistic method for predicting future performance deterioration of coastal flood defence structures based on condition grading inspection data and deterioration curves estimated by expert judgement. In condition-based deterioration modelling, the main task is to estimate transition probability matrices. The deterioration process of the structure related to the transition states is modelled according to Markov chain process, and a reliability-based approach is used to estimate the probability of structural failure. Visual inspection data according to the United Kingdom Condition Assessment Manual are used to obtain the initial condition grade curve of the coastal flood defences. The initial curves then modified in order to develop transition probabilities through non-linear regression based optimisation algorithms. The Monte Carlo simulations are then used to evaluate the future performance of the structure on the basis of the estimated transition probabilities. Finally, a case study is given to demonstrate the applicability of the proposed method under no-maintenance and medium-maintenance scenarios. Results show that the proposed method can provide an effective predictive model for various situations in terms of available condition grading data. The proposed model also provides useful information on time-dependent probability of failure in coastal flood defences.

Keywords: condition grading, flood defense, performance assessment, stochastic deterioration modelling

Procedia PDF Downloads 210

Authors: Szymon Rutkowski, Tomasz Korbak

Abstract:

We present a reinforcement-learning-based method of training neural machine translation models without parallel corpora. The standard encoder-decoder approach to machine translation suffers from two problems we aim to address. First, it needs parallel corpora, which are scarce, especially for low-resource languages. Second, it lacks psychological plausibility of learning procedure: learning a foreign language is about learning to communicate useful information, not merely learning to transduce from one language’s 'encoding' to another. We instead pose the problem of learning to translate as learning a policy in a communication game between two agents: the translator and the classifier. The classifier is trained beforehand on a natural language inference task (determining the entailment relation between a premise and a hypothesis) in the target language. The translator produces a sequence of actions that correspond to generating translations of both the hypothesis and premise, which are then passed to the classifier. The translator is rewarded for classifier’s performance on determining entailment between sentences translated by the translator to disciple’s native language. Translator’s performance thus reflects its ability to communicate useful information to the classifier. In effect, we train a machine translation model without the need for parallel corpora altogether. While similar reinforcement learning formulations for zero-shot translation were proposed before, there is a number of improvements we introduce. While prior research aimed at grounding the translation task in the physical world by evaluating agents on an image captioning task, we found that using a linguistic task is more sample-efficient. Natural language inference (also known as recognizing textual entailment) captures semantic properties of sentence pairs that are poorly correlated with semantic similarity, thus enforcing basic understanding of the role played by compositionality. It has been shown that models trained recognizing textual entailment produce high-quality general-purpose sentence embeddings transferrable to other tasks. We use stanford natural language inference (SNLI) dataset as well as its analogous datasets for French (XNLI) and Polish (CDSCorpus). Textual entailment corpora can be obtained relatively easily for any language, which makes our approach more extensible to low-resource languages than traditional approaches based on parallel corpora. We evaluated a number of reinforcement learning algorithms (including policy gradients and actor-critic) to solve the problem of translator’s policy optimization and found that our attempts yield some promising improvements over previous approaches to reinforcement-learning based zero-shot machine translation.

Keywords: agent-based language learning, low-resource translation, natural language inference, neural machine translation, reinforcement learning

Procedia PDF Downloads 98

Authors: Emmanuel Ogala

Abstract:

Cloud computing has been envisioned as the next-generation architecture of Information Technology (IT) enterprise. In contrast to traditional solutions where IT services are under physical, logical and personnel controls, cloud computing moves the application software and databases to the large data centres, where the management of the data and services may not be fully trustworthy. This is due to the fact that the systems are opened to the whole world and as people tries to have access into the system, many people also are there trying day-in day-out on having unauthorized access into the system. This research contributes to the improvement of cloud computing security for better operation. The work is motivated by two problems: first, the observed easy access to cloud computing resources and complexity of attacks to vital cloud computing data system NIC requires that dynamic security mechanism evolves to stay capable of preventing illegitimate access. Second; lack of good methodology for performance test and evaluation of biometric security algorithms for securing records in cloud computing environment. The aim of this research was to evaluate the performance of an integrated security system (ISS) for securing exams records in cloud computing environment. In this research, we designed and implemented an ISS consisting of three security mechanisms of biometric (fingerprint), auto-PIN and password into one stream of access control and used for securing examination records in Kogi State University, Anyigba. Conclusively, the system we built has been able to overcome guessing abilities of hackers who guesses people password or pin. We are certain about this because the added security system (fingerprint) needs the presence of the user of the software before a login access can be granted. This is based on the placement of his finger on the fingerprint biometrics scanner for capturing and verification purpose for user’s authenticity confirmation. The study adopted the conceptual of quantitative design. Object oriented and design methodology was adopted. In the analysis and design, PHP, HTML5, CSS, Visual Studio Java Script, and web 2.0 technologies were used to implement the model of ISS for cloud computing environment. Note; PHP, HTML5, CSS were used in conjunction with visual Studio front end engine design tools and MySQL + Access 7.0 were used for the backend engine and Java Script was used for object arrangement and also validation of user input for security check. Finally, the performance of the developed framework was evaluated by comparing with two other existing security systems (Auto-PIN and password) within the school and the results showed that the developed approach (fingerprint) allows overcoming the two main weaknesses of the existing systems and will work perfectly well if fully implemented.

Keywords: performance evaluation, fingerprint, auto-pin, password-based, security systems, cloud computing environment

Procedia PDF Downloads 113

Authors: Jamuna Loganath

Abstract:

Education is the key to the development of any individual’s personality. Today’s students will be tomorrow’s citizens of the global society. The education of the student is the edifice on which his/her future will be built. Schools therefore should provide an all-round development of students so as to foster a healthy society. The behaviors and the attitude of the students in school play an essential role for the success of the education process. Frequent reports of misbehaviors such as clowning, harassing classmates, verbal insults are becoming common in schools today. If this issue is left unattended, it may develop a negative attitude and increase the delinquent behavior. So, the need of the hour is to find a solution to this problem. To solve this issue, it is important to monitor the students’ behaviors in school and give necessary feedback and mentor them to develop a positive attitude and help them to become a successful grownup. Nevertheless, measuring students’ behavior and attitude is extremely challenging. None of the present technology has proven to be effective in this measurement process because actions, reactions, interactions, response of the students are rarely used in the course of the data due to complexity. The purpose of this proposal is to recommend an effective supervising system after carrying out a feasibility study by measuring the behavior of the Students. This can be achieved by equipping schools with CCTV cameras. These CCTV cameras installed in various schools of the world capture the facial expressions and interactions of the students inside and outside their classroom. The real time raw videos captured from the CCTV can be uploaded to the cloud with the help of a network. The video feeds get scooped into various nodes in the same rack or on the different racks in the same cluster in Hadoop HDFS. The video feeds are converted into small frames and analyzed using various Pattern recognition algorithms and MapReduce algorithm. Then, the video frames are compared with the bench marking database (good behavior). When misbehavior is detected, an alert message can be sent to the counseling department which helps them in mentoring the students. This will help in improving the effectiveness of the education process. As Video feeds come from multiple geographical areas (schools from different parts of the world), BIG DATA helps in real time analysis as it analyzes computationally to reveal patterns, trends, and associations, especially relating to human behavior and interactions. It also analyzes data that can’t be analyzed by traditional software applications such as RDBMS, OODBMS. It has also proven successful in handling human reactions with ease. Therefore, BIG DATA could certainly play a vital role in handling this issue. Thus, effectiveness of the education process can be enhanced with the help of video analytics using the latest BIG DATA technology.

Keywords: big data, cloud, CCTV, education process

Procedia PDF Downloads 218

Authors: O. B. Iskender, K. V. Ling, V. Dubanchet, L. Simonini

Abstract:

This paper discusses the recent advances and future prospects of spacecraft position and attitude control using Model Predictive Control (MPC). First, the challenges of the space missions are summarized, in particular, taking into account the errors, uncertainties, and constraints imposed by the mission, spacecraft and, onboard processing capabilities. The summary of space mission errors and uncertainties provided in categories; initial condition errors, unmodeled disturbances, sensor, and actuator errors. These previous constraints are classified into two categories: physical and geometric constraints. Last, real-time implementation capability is discussed regarding the required computation time and the impact of sensor and actuator errors based on the Hardware-In-The-Loop (HIL) experiments. The rationales behind the scenarios’ are also presented in the scope of space applications as formation flying, attitude control, rendezvous and docking, rover steering, and precision landing. The objectives of these missions are explained, and the generic constrained MPC problem formulations are summarized. Three key design elements used in MPC design: the prediction model, the constraints formulation and the objective cost function are discussed. The prediction models can be linear time invariant or time varying depending on the geometry of the orbit, whether it is circular or elliptic. The constraints can be given as linear inequalities for input or output constraints, which can be written in the same form. Moreover, the recent convexification techniques for the non-convex geometrical constraints (i.e., plume impingement, Field-of-View (FOV)) are presented in detail. Next, different objectives are provided in a mathematical framework and explained accordingly. Thirdly, because MPC implementation relies on finding in real-time the solution to constrained optimization problems, computational aspects are also examined. In particular, high-speed implementation capabilities and HIL challenges are presented towards representative space avionics. This covers an analysis of future space processors as well as the requirements of sensors and actuators on the HIL experiments outputs. The HIL tests are investigated for kinematic and dynamic tests where robotic arms and floating robots are used respectively. Eventually, the proposed algorithms and experimental setups are introduced and compared with the authors' previous work and future plans. The paper concludes with a conjecture that MPC paradigm is a promising framework at the crossroads of space applications while could be further advanced based on the challenges mentioned throughout the paper and the unaddressed gap.

Keywords: convex optimization, model predictive control, rendezvous and docking, spacecraft autonomy

Procedia PDF Downloads 86

Authors: Varun Agarwal

Abstract:

Introduction: With the advent of personalized medicine, histopathological review of whole slide images (WSIs) for cancer diagnosis presents an exceedingly time-consuming, complex task. Specifically, detecting metastatic regions in WSIs of sentinel lymph node biopsies necessitates a full-scanned, holistic evaluation of the image. Thus, digital pathology, low-level image manipulation algorithms, and machine learning provide significant advancements in improving the efficiency and accuracy of WSI analysis. Using Camelyon16 data, this paper proposes a deep learning pipeline to automate and ameliorate breast cancer metastasis localization and WSI classification. Methodology: The model broadly follows five stages -region of interest detection, WSI partitioning into image tiles, convolutional neural network (CNN) image-segment classifications, probabilistic mapping of tumor localizations, and further processing for whole WSI classification. Transfer learning is applied to the task, with the implementation of Inception-ResNetV2 - an effective CNN classifier that uses residual connections to enhance feature representation, adding convolved outputs in the inception unit to the proceeding input data. Moreover, in order to augment the performance of the transfer learning CNN, a stack of convolutional denoising autoencoders (CDAE) is applied to produce embeddings that enrich image representation. Through a saliency-detection algorithm, visual training segments are generated, which are then processed through a denoising autoencoder -primarily consisting of convolutional, leaky rectified linear unit, and batch normalization layers- and subsequently a contrast-normalization function. A spatial pyramid pooling algorithm extracts the key features from the processed image, creating a viable feature map for the CNN that minimizes spatial resolution and noise. Results and Conclusion: The simplified and effective architecture of the fine-tuned transfer learning Inception-ResNetV2 network enhanced with the CDAE stack yields state of the art performance in WSI classification and tumor localization, achieving AUC scores of 0.947 and 0.753, respectively. The convolutional feature retention and compilation with the residual connections to inception units synergized with the input denoising algorithm enable the pipeline to serve as an effective, efficient tool in the histopathological review of WSIs.

Keywords: breast cancer, convolutional neural networks, metastasis mapping, whole slide images

Procedia PDF Downloads 106

Authors: Stefano Iannello, Massimiliano Materazzi

Abstract:

Fluidized bed reactor technologies are one of the most valuable pathways for thermochemical conversions of biogenic fuels due to their good operating flexibility. Nevertheless, there are still issues related to the mixing and separation of heterogeneous phases during operation with highly volatile feedstocks, including biomass and waste. At high temperatures, the volatile content of the feedstock is released in the form of the so-called endogenous bubbles, which generally exert a “lift” effect on the particle itself by dragging it up to the bed surface. Such phenomenon leads to high release of volatile matter into the freeboard and limited mass and heat transfer with particles of the bed inventory. The aim of this work is to get a better understanding of the behaviour of a single reacting particle in a hot fluidized bed reactor during the devolatilization stage. The analysis has been undertaken at different fluidization regimes and temperatures to closely mirror the operating conditions of waste-to-energy processes. Beechwood and polypropylene particles were used to resemble the biomass and plastic fractions present in waste materials, respectively. The non-invasive X-ray technique was coupled to particle tracking algorithms to characterize the motion of a single feedstock particle during the devolatilization with high resolution. A high-energy X-ray beam passes through the vessel where absorption occurs, depending on the distribution and amount of solids and fluids along the beam path. A high-speed video camera is synchronised to the beam and provides frame-by-frame imaging of the flow patterns of fluids and solids within the fluidized bed up to 72 fps (frames per second). A comprehensive mathematical model has been developed in order to validate the experimental results. Beech wood and polypropylene particles have shown a very different dynamic behaviour during the pyrolysis stage. When the feedstock is fed from the bottom, the plastic material tends to spend more time within the bed than the biomass. This behaviour can be attributed to the presence of the endogenous bubbles, which drag effect is more pronounced during the devolatilization of biomass, resulting in a lower residence time of the particle within the bed. At the typical operating temperatures of thermochemical conversions, the synthetic polymer softens and melts, and the bed particles attach on its outer surface, generating a wet plastic-sand agglomerate. Consequently, this additional layer of sand may hinder the rapid evolution of volatiles in the form of endogenous bubbles, and therefore the establishment of a poor drag effect acting on the feedstock itself. Information about the mixing and segregation of solid feedstock is of prime importance for the design and development of more efficient industrial-scale operations.

Keywords: fluidized bed, pyrolysis, waste feedstock, X-ray

Procedia PDF Downloads 145

Authors: Olena Kavats, Dmitry Khramov, Kateryna Sergieieva, Vladimir Vasyliev, Iurii Kavats

Abstract:

Geoinformation technologies of space agromonitoring are a means of operative decision making support in the tasks of managing the agricultural sector of the economy. Existing technologies use satellite images in the optical range of electromagnetic spectrum. Time series of optical images often contain gaps due to the presence of clouds and haze. A geoinformation technology is created. It allows to fill gaps in time series of optical images (Sentinel-2, Landsat-8, PROBA-V, MODIS) with radar survey data (Sentinel-1) and use information about agrometeorological conditions of the growing season for individual monitoring years. The technology allows to perform crop classification and mapping for spring-summer (winter and spring crops) and autumn-winter (winter crops) periods of vegetation, monitoring the dynamics of crop state seasonal changes, crop yield forecasting. Crop classification is based on supervised classification algorithms, takes into account the peculiarities of crop growth at different vegetation stages (dates of sowing, emergence, active vegetation, and harvesting) and agriculture land state characteristics (row spacing, seedling density, etc.). A catalog of samples of the main agricultural crops (Ukraine) is created and crop spectral signatures are calculated with the preliminary removal of row spacing, cloud cover, and cloud shadows in order to construct time series of crop growth characteristics. The obtained data is used in grain crop growth tracking and in timely detection of growth trends deviations from reference samples of a given crop for a selected date. Statistical models of crop yield forecast are created in the forms of linear and nonlinear interconnections between crop yield indicators and crop state characteristics (temperature, precipitation, vegetation indices, etc.). Predicted values of grain crop yield are evaluated with an accuracy up to 95%. The developed technology was used for agricultural areas monitoring in a number of Great Britain and Ukraine regions using EOS Crop Monitoring Platform (https://crop-monitoring.eos.com). The obtained results allow to conclude that joint use of Sentinel-1 and Sentinel-2 images improve separation of winter crops (rapeseed, wheat, barley) in the early stages of vegetation (October-December). It allows to separate successfully the soybean, corn, and sunflower sowing areas that are quite similar in their spectral characteristics.

Keywords: geoinformation technology, crop classification, crop yield prediction, agricultural monitoring, EOS Crop Monitoring Platform

Procedia PDF Downloads 402