Search results for: real output

Authors: Jon-Chao Hong, Tsai-Ru Fan, Shih-Min Hsu

Abstract:

Virtual reality (VR) can generate mirror neurons that facilitate learners to transfer virtual skills to a real environment in skill training, and most studies approved the positive effect of applying in many domains. However, rare studies have focused on the experiential values of participants from a gender perspective. To address this issue, the present study used a VR program named kitchen assistant training, focusing on cutting vegetables and invited 400 students to practice for 20 minutes. Useful data from 367 were subjected to statistical analysis. The results indicated that male participants. From the comparison of average, it seems that females perceived higher than males in learning effectiveness. Expectedly, the VR-Cutting vegetables can be used for pre-training of real vegetable cutting.

Keywords: exponential value, facilitate learning, gender difference, virtual reality

Procedia PDF Downloads 94

Authors: Waddah Abdelbagi Talha, Mohammed Abdullah Elmaleeh, John Ojur Dennis

Abstract:

This paper focuses on the simulation and implementation of a resonator micro-sensor for low magnetic field sensing based on a U-shaped cantilever and piezoresistive configuration, which works based on Lorentz force physical phenomena. The resonance frequency is an important parameter that depends upon the highest response and sensitivity through the frequency domain (frequency response) of any vibrated micro-scale of an electromechanical system (MEMS) device. And it is important to determine the direction of the detected magnetic field. The deflection of the cantilever is considered for vibrated mode with different frequencies in the range of (0 Hz to 7000 Hz); for the purpose of observing the frequency response. A simple electronic circuit-based polysilicon piezoresistors in Wheatstone's bridge configuration are used to transduce the response of the cantilever to electrical measurements at various voltages. Microcontroller-based Arduino program and PROTEUS electronic software are used to analyze the output signals from the sensor. The highest output voltage amplitude of about 4.7 mV is spotted at about 3 kHz of the frequency domain, indicating the highest sensitivity, which can be called resonant sensitivity. Based on the resonant frequency value, the mode of vibration is determined (up-down vibration), and based on that, the vector of the magnetic field is also determined.

Keywords: resonant frequency, sensitivity, Wheatstone bridge, UNO-microcontroller

Procedia PDF Downloads 127

Authors: C. J. Rossouw, T. I. van Niekerk

Abstract:

The global manufacturing industry is utilizing the internet and cloud-based services to further explore the anatomy and optimize manufacturing processes in support of the movement into the Fourth Industrial Revolution (4IR). The 4IR from a third world and African perspective is hindered by the fact that many manufacturing systems that were developed in the third industrial revolution are not inherently equipped to utilize the internet and services of the 4IR, hindering the progression of third world manufacturing industries into the 4IR. This research focuses on the development of a non-invasive and cost-effective cyber-physical IoT system that will exploit a machine’s vibration to expose semantic characteristics in the manufacturing process and utilize these results through a real-time cloud-based machine condition monitoring system with the intention to optimize the system. A microcontroller-based IoT sensor was designed to acquire a machine’s mechanical vibration data, process it in real-time, and transmit it to a cloud-based platform via Wi-Fi and the internet. Time-frequency Fourier analysis was applied to the vibration data to form an image representation of the machine’s behaviour. This data was used to train a Convolutional Neural Network (CNN) to learn semantic characteristics in the machine’s behaviour and relate them to a state of operation. The same data was also used to train a Convolutional Autoencoder (CAE) to detect anomalies in the data. Real-time edge-based artificial intelligence was achieved by deploying the CNN and CAE on the sensor to analyse the vibration. A cloud platform was deployed to visualize the vibration data and the results of the CNN and CAE in real-time. The cyber-physical IoT system was deployed on a semi-automated metal granulation machine with a set of trained machine learning models. Using a single sensor, the system was able to accurately visualize three states of the machine’s operation in real-time. The system was also able to detect a variance in the material being granulated. The research demonstrates how non-IoT manufacturing systems can be equipped with edge-based artificial intelligence to establish a remote machine condition monitoring system.

Keywords: IoT, cyber-physical systems, artificial intelligence, manufacturing, vibration analytics, continuous machine condition monitoring

Procedia PDF Downloads 88

Authors: R. Dariani, S. Schmidt, R. Kasper

Abstract:

Based on a non-linear single track model which describes the dynamics of vehicle, an optimal path planning strategy is developed. Real time optimization is used to generate reference control values to allow leading the vehicle alongside a calculated lane which is optimal for different objectives such as energy consumption, run time, safety or comfort characteristics. Strict mathematic formulation of the autonomous driving allows taking decision on undefined situation such as lane change or obstacle avoidance. Based on position of the vehicle, lane situation and obstacle position, the optimization problem is reformulated in real-time to avoid the obstacle and any car crash.

Keywords: autonomous driving, obstacle avoidance, optimal control, path planning

Procedia PDF Downloads 369

Authors: Pedro P. Drumond, Priscilla M. Moura, Marcia M. L. P. Coelho

Abstract:

The purpose of On-site Stormwater Detention (OSD) system is to promote the detention of addition stormwater runoff caused by impervious areas, in order to maintain the peak flow the same as the pre-urbanization condition. In recent decades, these systems have been built in many cities around the world. However, its real efficiency continues to be unknown due to the lack of research, especially with regard to monitoring its real performance. Thus, this study aims to compare the water level monitoring data of an OSD built in Belo Horizonte/Brazil with the results of theoretical methods simulations, usually adopted in OSD design. There were made two theoretical simulations, one using the Rational Method and Modified Puls method and another using the Soil Conservation Service (SCS) method and Modified Puls method. The monitoring data were obtained with a water level sensor, installed inside the reservoir and connected to a data logger. The comparison of OSD performance was made for 48 rainfall events recorded from April/2015 to March/2017. The comparison of maximum water levels in the OSD showed that the results of the simulations with Rational/Puls and SCS/Puls methods were, on average 33% and 73%, respectively, lower than those monitored. The Rational/Puls results were significantly higher than the SCS/Puls results, only in the events with greater frequency. In the events with average recurrence interval of 5, 10 and 200 years, the maximum water heights were similar in both simulations. Also, the results showed that the duration of rainfall events was close to the duration of monitored hydrograph. The rising time and recession time of the hydrographs calculated with the Rational Method represented better the monitored hydrograph than SCS Method. The comparison indicates that the real discharge coefficient value could be higher than 0.61, adopted in Puls simulations. New researches evaluating OSD real performance should be developed. In order to verify the peak flow damping efficiency and the value of the discharge coefficient is necessary to monitor the inflow and outflow of an OSD, in addition to monitor the water level inside it.

Keywords: best management practices, on-site stormwater detention, source control, urban drainage

Procedia PDF Downloads 188

Authors: A. T. Kassem

Abstract:

Fires areconsidered a common hazardous action that any building may face. Most buildings’ structural elements are designed, taking into consideration precautions for fire safety, using deterministic design approaches. Public and highly important buildings are commonly designed considering standard fire rating and, in many cases, contain large compartments with central domes. Real fire scenarios are not commonly brought into action in structural design of buildings because of complexities in both scenarios and analysis tools. This paper presents a modern approach towards analysis of spherical domes in real fire condition via implementation of building information modelling, and adopting a probabilistic approach. BIMhas been implemented to bridge the gap between various software packages enabling them to function interactively to model both real fire and corresponding structural response. Ventilation controlled fires scenarios have been modeled using both “Revit” and “Pyrosim”. Monte Carlo simulation has been adopted to engage the probabilistic analysis approach in dealing with various parameters. Conclusions regarding failure probability and fire endurance, in addition to the effects of various parameters, have been extracted.

Keywords: concrete, spherical domes, ventilation controlled fires, BIM, monte carlo simulation, pyrosim, revit

Procedia PDF Downloads 95

Authors: Assem El-Ansary

Abstract:

This paper illustrates the event-oriented Behavioral Pattern Analysis (BPA) modeling approach using the Healthcare Multi-Agent System. The Event defined in BPA is a real-life conceptual entity that is unrelated to any implementation. The major contributions of this research are: The Behavioral Pattern Analysis (BPA) modeling methodology. The development of an interactive software tool (DECISION), which is based on a combination of the Analytic Hierarchy Process (AHP) and the ELECTRE Multi-Criteria Decision Making (MCDM) methods.

Keywords: analysis, modeling methodology, software modeling, event-oriented, behavioral pattern, use cases, Healthcare Multi-Agent System

Procedia PDF Downloads 550

Authors: Mankour Mohamed, Miloudi Mohamed

Abstract:

A thorough and accurate simulation is necessary for the study of a High Voltage Direct Current (HVDC) link system during various types of disturbances, including internal faults on both converters, either on the rectifier or on the inverter, as well as external faults, such as AC or DC faults on both converter sides inside the DC link party. In this study, we examine how an HVDC inverter responds to three different types of failures, including faults at the inverter valve, system control faults, and single-phase-to-ground AC faults at the sending end of the inverter side. As this phenomenon represents the most frequent problem that may affect inverter valves, particularly those based on thyristor valves (LCC (line-Commutated converter)), it is more precise to explore which circumstance generates and raises the commutation failure on inverter valves. Because of the techniques used to accelerate the simulation, digital real-time simulators are now the most potent tools that provide simulation results. The real-time-lab RT-LAB platform HYPERSIM OP-5600 is used to implement the Simulation in the Loop (SIL) technique, which is used to validate the results. It is demonstrated how to recover from both the internal faults and the AC problem. The simulation findings show how crucial a role the control system plays in fault recovery.

Keywords: hypersim simulator, HVDC systems, mono-polar link, AC faults, misfiring faults

Procedia PDF Downloads 94

Authors: Zhiwen Yang, Jingyuan Li, Zhenkai Xie, Jian Ling, Jiguang Wang, Mengliang Li

Abstract:

To reflect the effects of different emission control strategies, such as retrofitting after-treatment system and replacing with natural gas-fueled vehicles, on particle number (PN), particle mass (PM) and nitrogen oxides (NOx) emissions emitted by urban bus, a portable emission measurement system (PEMS) was employed herein to conduct real-world driving emission measurements on a diesel oxidation catalytic converter (DOC) and catalyzed diesel particulate filter (CDPF) retrofitting China IV diesel bus, a China IV diesel bus, and a China V natural gas bus. The results show that both tested diesel buses possess markedly advantages in NOx emission control when compared to the lean-burn natural gas bus equipped without any NOx after-treatment system. As to PN and PM, only the DOC+CDPF retrofitting diesel bus exhibits enormous benefits on emission control relate to the natural gas bus, especially the normal diesel bus. Meanwhile, the differences in PM and PN emissions between retrofitted and normal diesel buses generally increase with the increase in vehicle-specific power (VSP). Furthermore, the differences in PM emissions, especially those in the higher VSP ranges, are more significant than those in PN. In addition, the maximum peak PN particle size (32 nm) of the retrofitted diesel bus was significantly lower than that of the normal diesel bus (100 nm). These phenomena indicate that the CDPF retrofitting can effectively reduce diesel bus exhaust particle emissions, especially those with large particle sizes.

Keywords: CDPF, diesel, natural gas, real-world emissions

Procedia PDF Downloads 297

Authors: Ramonika Sengupta, Stuti Kachhwaha, Asha Adhiya, K. Satya Raja Sekhar, Rajwinder Kaur

Abstract:

Numerous efforts have been done in the past decade to develop the methods of secure communication that are free from interception and eavesdropping. In fiber optic communication, chaotic optical carrier signals are used for data encryption in secure data transmission. Mach-Zehnder Modulators (MZM) are the key components for generating the chaotic signals to be used as optical carriers. This paper presents the dynamics of a lithium niobate MZM modulator under various biasing conditions. The chaotic fluctuations of the intensity of a laser diode have been generated using the electro-optic MZM modulator operating in a highly nonlinear regime. The modulator is driven in closed loop by its own output at an earlier time. When used as an electro-optic oscillator employing delayed feedback, the MZM displays a wide range of output waveforms of varying complexity. The dynamical behavior of the system ranges from periodic to nonlinear oscillations. The nonlinearity displayed by the system is reproducible and is easily controllable. In this paper, we demonstrate a wide variety of optical signals generated by MZM using easily controllable device parameters in both open and close loop bias conditions.

Keywords: chaotic carrier, fiber optic communication, Mach-Zehnder modulator, secure data transmission

Procedia PDF Downloads 272

Authors: Payam Haghparast, Mikhail V. Sorin, Hakim Nesreddine

Abstract:

The complex oblique shock phenomenon can be simply assumed as a normal shock at the constant area section to simulate a sharp pressure increase and velocity decrease in 1-D thermodynamic models. The assumed normal shock location is one of the greatest sources of error in ejector thermodynamic models. Most researchers consider an arbitrary location without justifying it. Our study compares the effect of normal shock place on ejector dimensions in 1-D models. To this aim, two different ejector experimental test benches, a constant area-mixing ejector (CAM) and a constant pressure-mixing (CPM) are considered, with different known geometries, operating conditions and working fluids (R245fa, R141b). In the first step, in order to evaluate the real value of the efficiencies in the different ejector parts and critical back pressure, a CFD model was built and validated by experimental data for two types of ejectors. These reference data are then used as input to the 1D model to calculate the lengths and the diameters of the ejectors. Afterwards, the design output geometry calculated by the 1D model is compared directly with the corresponding experimental geometry. It was found that there is a good agreement between the ejector dimensions obtained by the 1D model, for both CAM and CPM, with experimental ejector data. Furthermore, it is shown that normal shock place affects only the constant area length as it is proven that the inlet normal shock assumption results in more accurate length. Taking into account previous 1D models, the results suggest the use of the assumed normal shock location at the inlet of the constant area duct to design the supersonic ejectors.

Keywords: 1D model, constant area-mixing, constant pressure-mixing, normal shock location, ejector dimensions

Procedia PDF Downloads 194

Authors: Tien-Li Chang, Huang-Chi Huang, Zhao-Chi Chen, Wun-Yi Chen

Abstract:

The droplet-based microfluidic are used as micro-reactors for chemical and biological assays. Hence, the precise addition of reagents into the droplets is essential for this function in the scope of lab-on-a-chip applications. To obtain the characteristics (size, velocity, pressure, and frequency of production) of droplets, this study describes an integrated on-chip method of real-time signal detection. By controlling and manipulating the fluids, the flow behavior can be obtained in the droplet-based microfluidics. The detection method is used a type of infrared sensor. Through the varieties of droplets in the microfluidic devices, the real-time conditions of velocity and pressure are gained from the sensors. Here the microfluidic devices are fabricated by polydimethylsiloxane (PDMS). To measure the droplets, the signal acquisition of sensor and LabVIEW program control must be established in the microchannel devices. The devices can generate the different size droplets where the flow rate of oil phase is fixed 30 μl/hr and the flow rates of water phase range are from 20 μl/hr to 80 μl/hr. The experimental results demonstrate that the sensors are able to measure the time difference of droplets under the different velocity at the voltage from 0 V to 2 V. Consequently, the droplets are measured the fastest speed of 1.6 mm/s and related flow behaviors that can be helpful to develop and integrate the practical microfluidic applications.

Keywords: microfluidic, droplets, sensors, single detection

Procedia PDF Downloads 493

Authors: Tawfiq Chekifi, Brahim Dennai, Rachid Khelfaoui

Abstract:

Over the last few decades, modeling immiscible fluids such as oil and water have been a classical research topic. Droplet-based microfluidics presents a unique platform for mixing, reaction, separation, dispersion of drops, and numerous other functions. For this purpose, several devices were studied, as well as microfluidic oscillator. The latter was obtained from wall attachment microfluidic amplifiers using a feedback loop from the outputs to the control inputs, nevertheless this device have not well used for microdrops applications. In this paper, we suggest a numerical CFD study of a microfluidic oscillator with two different lengths of feedback loop. In order to produce simultaneous microdrops of gasoil on water, a typical geometry that includes double T-junction is connected to the fluidic oscillator. The generation of microdrops is computed by volume-of-fluid method (VOF). Flow oscillations of microdrops were triggered by the Coanda effect of jet flow. The aim of work is to obtain a high oscillation frequency in output of this passive device, the influence of hydrodynamics and physics parameters on the microdrops frequency in the output of our microsystem is also analyzed, The computational results show that, the length of feedback loop, applied pressure on T-junction and interfacial tension have a significant effect on the dispersion of microdrops and its oscillation frequency. Across the range of low Reynold number, the microdrops generation and its dynamics have been accurately controlled by adjusting applying pressure ratio of two phases.

Keywords: fluidic oscillator, microdrops manipulation, VOF (volume of fluid method), microfluidic oscillator

Procedia PDF Downloads 397

Authors: Pengfei Meng, Shuangcheng Jia, Qian Li

Abstract:

We proposed a real-time high-precision semantic segmentation network based on a multi-resolution feature fusion module, the auxiliary feature extracting module, upsampling module, and atrous spatial pyramid pooling (ASPP) module. We designed a feature fusion structure, which is integrated with sufficient features of different resolutions. We also studied the effect of side-branch structure on the network and made discoveries. Based on the discoveries about the side-branch of the network structure, we used a side-branch auxiliary feature extraction layer in the network to improve the effectiveness of the network. We also designed upsampling module, which has better results than the original upsampling module. In addition, we also re-considered the locations and number of atrous spatial pyramid pooling (ASPP) modules and modified the network structure according to the experimental results to further improve the effectiveness of the network. The network presented in this paper takes the backbone network of Bisenetv2 as a basic network, based on which we constructed a network structure on which we made improvements. We named this network deep multiple-resolution bilateral networks for real-time, referred to as DMBR-Net. After experimental testing, our proposed DMBR-Net network achieved 81.2% mIoU at 119FPS on the Cityscapes validation dataset, 80.7% mIoU at 109FPS on the CamVid test dataset, 29.9% mIoU at 78FPS on the COCOStuff test dataset. Compared with all lightweight real-time semantic segmentation networks, our network achieves the highest accuracy at an appropriate speed.

Keywords: multi-resolution feature fusion, atrous convolutional, bilateral networks, pyramid pooling

Procedia PDF Downloads 150

Authors: Li Yu, Chenlu Jiao, Hongrun Xin, Yan Wang, Kaiyang Wang

Abstract:

In order to study the impact of various factors on the housing price, we propose to build different prediction models based on deep learning to determine the existing data of the real estate in order to more accurately predict the housing price or its changing trend in the future. Considering that the factors which affect the housing price vary widely, the proposed prediction models include two categories. The first one is based on multiple characteristic factors of the real estate. We built Convolution Neural Network (CNN) prediction model and Long Short-Term Memory (LSTM) neural network prediction model based on deep learning, and logical regression model was implemented to make a comparison between these three models. Another prediction model is time series model. Based on deep learning, we proposed an LSTM-1 model purely regard to time series, then implementing and comparing the LSTM model and the Auto-Regressive and Moving Average (ARMA) model. In this paper, comprehensive study of the second-hand housing price in Beijing has been conducted from three aspects: crawling and analyzing, housing price predicting, and the result comparing. Ultimately the best model program was produced, which is of great significance to evaluation and prediction of the housing price in the real estate industry.

Keywords: deep learning, convolutional neural network, LSTM, housing prediction

Procedia PDF Downloads 306

Authors: Nurul Aini Abdul Wahab, Nor Syamim Halidin, Sayidatina Aisah Masnan, Nur Izzati Romli

Abstract:

There are many applications of principle component method for reducing the large set of variables in various fields. Fisher’s Discriminant function is also a popular tool for classification. In this research, the researcher focuses on studying the performance of Principle Component-Fisher’s Discriminant function in helping to classify rice kernels to their defined classes. The data were collected on the smells or odour of the rice kernel using odour-detection sensor, Cyranose. 32 variables were captured by this electronic nose (e-nose). The objective of this research is to measure how well a combination model, between principle component and linear discriminant, to be as a classification model. Principle component method was used to reduce all 32 variables to a smaller and manageable set of components. Then, the reduced components were used to develop the Fisher’s Discriminant function. In this research, there are 4 defined classes of rice kernel which are Aromatic, Brown, Ordinary and Others. Based on the output from principle component method, the 32 variables were reduced to only 2 components. Based on the output of classification table from the discriminant analysis, 40.76% from the total observations were correctly classified into their classes by the PC-Discriminant function. Indirectly, it gives an idea that the classification model developed has committed to more than 50% of misclassifying the observations. As a conclusion, the Fisher’s Discriminant function that was built on a 2-component from PCA (2-PC-D) is not satisfying to classify the rice kernels into its defined classes.

Keywords: classification model, discriminant function, principle component analysis, variable reduction

Procedia PDF Downloads 332

Authors: Niloofar Ashktorab, Negar Ashktorab

Abstract:

Oil exports fund nearly half of Iran’s government expenditures, since many years other countries have been imposed different sanctions against Iran. Sanctions that primarily target Iran’s key energy sector have harmed Iran’s economy. The strategic effects of sanctions might be reduction as Iran adjusts to them economically. In this study, we evaluate the impact of oil price and sanctions against Iran on food commodity prices by using panel data method. Here, we find that the food commodity prices, the oil price and real exchange rate are stationary. The results show positive effect of oil price changes, real exchange rate and sanctions on food commodity prices.

Keywords: oil price, food basket, sanctions, panel data, Iran

Procedia PDF Downloads 356

Authors: M. Iommi, G. Losco

Abstract:

Current building sector is focused on reduction of energy requirements, on renewable energy generation and on regeneration of existing urban areas. These targets need to be solved with a systemic approach, considering several aspects simultaneously such as climate conditions, lighting conditions, solar radiation, PV potential, etc. The solar access analysis is an already known method to analyze the solar potentials, but in current years, simulation tools have provided more effective opportunities to perform this type of analysis, in particular in the early design stage. Nowadays, the study of the solar access is related to the easiness of the use of simulation tools, in rapid and easy way, during the design process. This study presents a comparison of three simulation tools, from the point of view of the user, with the aim to highlight differences in the easy-to-use of these tools. Using a real urban context as case study, three tools; Ecotect, Townscope and Heliodon, are tested, performing models and simulations and examining the capabilities and output results of solar access analysis. The evaluation of the ease-to-use of these tools is based on some detected parameters and features, such as the types of simulation, requirements of input data, types of results, etc. As a result, a framework is provided in which features and capabilities of each tool are shown. This framework shows the differences among these tools about functions, features and capabilities. The aim of this study is to support users and to improve the integration of simulation tools for solar access with the design process.

Keywords: energy building design tools, solar access analysis, solar potential, urban planning

Procedia PDF Downloads 340

Authors: Linyu Wang, Jiahui Ma, Jianhong Xiang, Hanyu Jiang

Abstract:

For the pilot design of the sparse channel estimation model in Multiple Input Multiple Output Orthogonal Frequency Division Multiplexing (MIMO-OFDM) systems, the observation matrix constructed according to the matrix cross-correlation criterion, total correlation criterion and other optimization criteria are not optimal, resulting in inaccurate channel estimation and high bit error rate at the receiver. This paper proposes a pilot design method combining high-power sum and high-power variance criteria, which can more accurately estimate the channel. First, the pilot insertion position is designed according to the high-power variance criterion under the condition of equal power. Then, according to the high power sum criterion, the pilot power allocation is converted into a cone programming problem, and the power allocation is carried out. Finally, the optimal pilot is determined by calculating the weighted sum of the high power sum and the high power variance. Compared with the traditional pilot frequency, under the same conditions, the constructed MIMO-OFDM system uses the optimal pilot frequency for channel estimation, and the communication bit error rate performance obtains a gain of 6~7dB.

Keywords: MIMO-OFDM, pilot optimization, compressed sensing, channel estimation

Procedia PDF Downloads 149

Authors: Emmanuel U. Oleka, Anil Khanal, Gary L. Lebby, Ali R. Osareh

Abstract:

Synchrophasor technology is fast being deployed in electric power grids all over the world and is fast changing the way the grids are managed. This trend is to continue until the entire power grids are fully connected so they can be monitored and controlled in real-time. Much achievement has been made in the synchrophasor technology development and deployment, and much more are yet to be achieved. Real-time power grid control and protection potentials of synchrophasor are yet to be explored. It is of necessity that researchers keep in view the various challenges that still need to be overcome in expanding the frontiers of synchrophasor technology. This paper outlines the major challenges that should be dealt with in order to achieve the goal of total power grid visualization, monitoring and control using synchrophasor technology.

Keywords: electric power grid, grid visualization, phasor measurement unit, synchrophasor

Procedia PDF Downloads 556

Authors: Min Jindong, Wang Mingxia

Abstract:

With the widespread use of artificial intelligence in life, computer vision, especially deep convolutional neural network models, has developed rapidly. With the increase of the complexity of the real visual target detection task and the improvement of the recognition accuracy, the target detection network model is also very large. The huge deep neural network model is not conducive to deployment on edge devices with limited resources, and the timeliness of network model inference is poor. In this paper, knowledge distillation is used to compress the huge and complex deep neural network model, and the knowledge contained in the complex network model is comprehensively transferred to another lightweight network model. Different from traditional knowledge distillation methods, we propose a novel knowledge distillation that incorporates multi-faceted features, called M-KD. In this paper, when training and optimizing the deep neural network model for target detection, the knowledge of the soft target output of the teacher network in knowledge distillation, the relationship between the layers of the teacher network and the feature attention map of the hidden layer of the teacher network are transferred to the student network as all knowledge. in the model. At the same time, we also introduce an intermediate transition layer, that is, an intermediate guidance layer, between the teacher network and the student network to make up for the huge difference between the teacher network and the student network. Finally, this paper adds an exploration module to the traditional knowledge distillation teacher-student network model. The student network model not only inherits the knowledge of the teacher network but also explores some new knowledge and characteristics. Comprehensive experiments in this paper using different distillation parameter configurations across multiple datasets and convolutional neural network models demonstrate that our proposed new network model achieves substantial improvements in speed and accuracy performance.

Keywords: object detection, knowledge distillation, convolutional network, model compression

Procedia PDF Downloads 278

Authors: Likewin Thomas, M. V. Manoj Kumar, B. Annappa

Abstract:

Traditional execution process follows the path of execution drawn by the process analyst without observing the behaviour of resource and other real-time constraints. Identifying process model, predicting the behaviour of resource and recommending the optimal path of execution for a real time process is challenging. The proposed AlfyMiner: αyM iner gives a new dimension in process execution with the novel techniques Process Model Analyser: PMAMiner and Resource behaviour Analyser: RBAMiner for recommending the probable path of execution. PMAMiner discovers next probable activity for currently executing activity in an online process using variant matching technique to identify the set of next probable activity, among which the next probable activity is discovered using decision tree model. RBAMiner identifies the resource suitable for performing the discovered next probable activity and observe the behaviour based on; load and performance using polynomial regression model, and waiting time using queueing theory. Based on the observed behaviour αyM iner recommend the probable path of execution with; next probable activity and the best suitable resource for performing it. Experiments were conducted on process logs of CoSeLoG Project1 and 72% of accuracy is obtained in identifying and recommending next probable activity and the efficiency of resource performance was optimised by 59% by decreasing their load.

Keywords: cross-organization process mining, process behaviour, path of execution, polynomial regression model

Procedia PDF Downloads 334

Authors: Andrea Pietrelli, Vincenzo Ferrara, Bruno Allard, Francois Buret, Irene Bavasso, Nicola Lovecchio, Francesca Costantini, Firas Khaled

Abstract:

This paper aims to show some applications of microbial fuel cells (MFCs), an energy harvesting technique, as clean power source to supply low power device for application like wireless sensor network (WSN) for environmental monitoring. Furthermore, MFC can be used directly as biosensor to analyse parameters like pH and temperature or arranged in form of cluster devices in order to use as small power plant. An MFC is a bioreactor that converts energy stored in chemical bonds of organic matter into electrical energy, through a series of reactions catalysed by microorganisms. We have developed a lab-scale terrestrial microbial fuel cell (TMFC), based on soil that acts as source of bacteria and flow of nutrient and a lab-scale waste water microbial fuel cell (WWMFC), where waste water acts as flow of nutrient and bacteria. We performed large series of tests to exploit the capability as biosensor. The pH value has strong influence on the open circuit voltage (OCV) delivered from TMFCs. We analyzed three condition: test A and B were filled with same soil but changing pH from 6 to 6.63, test C was prepared using a different soil with a pH value of 6.3. Experimental results clearly show how with higher pH value a higher OCV was produced; indeed reactors are influenced by different values of pH which increases the voltage in case of a higher pH value until the best pH value of 7 is achieved. The influence of pH on OCV of lab-scales WWMFC was analyzed at pH value of 6.5, 7, 7.2, 7.5 and 8. WWMFCs are influenced from temperature more than TMFCs. We tested the power performance of WWMFCs considering four imposed values of ambient temperature. Results show how power performance increase proportionally with higher temperature values, doubling the output power from 20° to 40°. The best value of power produced from our lab-scale TMFC was equal to 310 μW using peaty soil, at 1KΩ, corresponding to a current of 0.5 mA. A TMFC can supply proper energy to low power devices of a WSN by means of the design of three stages scheme of an energy management system, which adapts voltage level of TMFC to those required by a WSN node, as 3.3V. Using a commercial DC/DC boost converter, that needs an input voltage of 700 mV, the current source of 0.5 mA, charges a capacitor of 6.8 mF until it will have accumulated an amount of charge equal to 700 mV in a time of 10 s. The output stage includes an output switch that close the circuit after a time of 10s + 1.5ms because the converter can boost the voltage from 0.7V to 3.3V in 1.5 ms. Furthermore, we tested in form of clusters connected in series up to 20 WWMFCs, we have obtained a high voltage value as output, around 10V, but low current value. MFC can be considered a suitable clean energy source to be used to supply low power devices as a WSN node or to be used directly as biosensor.

Keywords: energy harvesting, low power electronics, microbial fuel cell, terrestrial microbial fuel cell, waste-water microbial fuel cell, wireless sensor network

Procedia PDF Downloads 207

Authors: Adelina Vrancianu

Abstract:

The problem of artificial real accession will be analyzed in this study both in terms of old and current Civil Code provisions and in terms of comparative law, European legal and Canadian systems. The current Civil Code from 2009 has brought new changes about the application and solutions regarding artificial real accession. The hypothesis in which a person is making works with his own materials on the real estate belonging to another person is developed and analyzed in detail from national and international point of view in relation with the good faith. The scope of this analysis is to point out what are the changes issued from case-law and which ones are new, inspired from other law systems in regard to the good/bad faith. The new civil code has promoted a definition for this notion. Is this definition a new one inspired from the comparative law or is it inspired from the case-law? Is it explained for every case scenario of accession or is a general notion? The study tries to respond to these questions and to present the new aspects in the area. has reserved a special place for the situation of execution of works with own materials exceeding the border with violation of another’s right of property, where the variety of solutions brings into discussion the case of expropriation for private interest. The new Civil Code is greatly influenced by the Civil Code from Quebec in comparison with the old code of French influence. The civil reform was needed and has brought into attention new solutions inspired from the Canadian system which has mitigated the permanent conflict between the constructor and the immovable owner.

Keywords: accession, good faith, new civil code, comparative law

Procedia PDF Downloads 462

Authors: Shorena Tsiklauri

Abstract:

Out-migration is an important issue for Georgia as well as since independence has loosed due to emigration one fifth of its population. During Soviet time out-migration from USSR was almost impossible and one of the most important instruments in regulating population movement within the Soviet Union was the system of compulsory residential registrations, so-called “propiska”. Since independent here was not any regulation for migration from Georgia. The majorities of Georgian migrants go abroad by tourist visa and then overstay, becoming the irregular labor migrants. The official statistics on migration published for this period was based on the administrative system of population registration, were insignificant in terms of numbers and did not represent the real scope of these migration movements. This paper discusses the data quality and methodology of migration statistics in Georgia and we are going to answer the questions: what is the real reason of increasing immigration flows according to the official numbers since 2000s?

Keywords: data quality, Georgia, methodology, migration

Procedia PDF Downloads 417

Authors: Farzan Farhangian, Behnam Shakibafar, Bobda Cedric, Rene Jr. Landry

Abstract:

Among the air navigation technologies, many of them are capable of increasing aviation sustainability as well as accuracy improvement in Alternative Positioning, Navigation, and Timing (APNT), especially avionics Distance Measuring Equipment (DME), Very high-frequency Omni-directional Range (VOR), etc. The integration of these air navigation solutions could make a robust and efficient accuracy in air mobility, air traffic management and autonomous operations. Designing a proper RF front-end, power amplifier and software-defined transponder could pave the way for reaching an optimized avionics navigation solution. In this article, the possibility of reaching an optimum front-end to be used with single low-cost Software-Defined Radio (SDR) has been investigated in order to reach a software-defined DME architecture. Our software-defined approach uses the firmware possibilities to design a real-time software architecture compatible with a Multi Input Multi Output (MIMO) BladeRF to estimate an accurate time delay between a Transmission (Tx) and the reception (Rx) channels using the synchronous scheduled communication. We could design a novel power amplifier for the transmission channel of the DME to pass the minimum transmission power. This article also investigates designing proper pair pulses based on the DO-178B avionics standard. Various guidelines have been tested, and the possibility of passing the certification process for each standard term has been analyzed. Finally, the performance of the DME was tested in the laboratory environment using an IFR6000, which showed that the proposed architecture reached an accuracy of less than 0.23 Nautical mile (Nmi) with 98% probability.

Keywords: avionics, DME, software defined radio, navigation

Procedia PDF Downloads 79

Authors: Radek Malinský, Ivan Jelínek

Abstract:

The current web has become a modern encyclopedia, where people share their thoughts and ideas on various topics around them. Such kind of encyclopedia is very useful for other people who are looking for answers to their questions. However, with the growing popularity of social networking and blogging and ever expanding network services, there has also been a growing diversity of technologies along with different structure of individual websites. It is, therefore, difficult to directly find a relevant answer for a common Internet user. This paper presents a web application for the real-time end-to-end analysis of selected Internet trends; where the trend can be whatever the people post online. The application integrates fully configurable tools for data collection and analysis using selected webometric algorithms, and for its chronological visualization to user. It can be assumed that the application facilitates the users to evaluate the quality of various products that are mentioned online.

Keywords: Trend, visualizer, web analysis, web 2.0.

Procedia PDF Downloads 264

Authors: Alexander Vandesompele, Joni Dambre

Abstract:

The diversity in dendritic arborization, as first illustrated by Santiago Ramon y Cajal, has always suggested a role for dendrites in the functionality of neurons. In the past decades, thanks to new recording techniques and optical stimulation methods, it has become clear that dendrites are not merely passive electrical components. They are observed to integrate inputs in a non-linear fashion and actively participate in computations. Regardless, in simulations of neural networks dendritic structure and functionality are often overlooked. Especially in a machine learning context, when designing artificial neural networks, point neuron models such as the leaky-integrate-and-fire (LIF) model are dominant. These models mimic the integration of inputs at the neuron soma, and ignore the existence of dendrites. In this work, the LIF point neuron model is extended with a simple form of dendritic computation. This gives the LIF neuron increased capacity to discriminate spatiotemporal input sequences, a dendritic functionality as observed in another study. Simulations of the spiking neurons are performed using the Bindsnet framework. In the common LIF model, incoming synapses are independent. Here, we introduce a dependency between incoming synapses such that the post-synaptic impact of a spike is not only determined by the weight of the synapse, but also by the activity of other synapses. This is a form of short term plasticity where synapses are potentiated or depressed by the preceding activity of neighbouring synapses. This is a straightforward way to prevent inputs from simply summing linearly at the soma. To implement this, each pair of synapses on a neuron is assigned a variable,representing the synaptic relation. This variable determines the magnitude ofthe short term plasticity. These variables can be chosen randomly or, more interestingly, can be learned using a form of Hebbian learning. We use Spike-Time-Dependent-Plasticity (STDP), commonly used to learn synaptic strength magnitudes. If all neurons in a layer receive the same input, they tend to learn the same through STDP. Adding inhibitory connections between the neurons creates a winner-take-all (WTA) network. This causes the different neurons to learn different input sequences. To illustrate the impact of the proposed dendritic mechanism, even without learning, we attach five input neurons to two output neurons. One output neuron isa regular LIF neuron, the other output neuron is a LIF neuron with dendritic relationships. Then, the five input neurons are allowed to fire in a particular order. The membrane potentials are reset and subsequently the five input neurons are fired in the reversed order. As the regular LIF neuron linearly integrates its inputs at the soma, the membrane potential response to both sequences is similar in magnitude. In the other output neuron, due to the dendritic mechanism, the membrane potential response is different for both sequences. Hence, the dendritic mechanism improves the neuron’s capacity for discriminating spa-tiotemporal sequences. Dendritic computations improve LIF neurons even if the relationships between synapses are established randomly. Ideally however, a learning rule is used to improve the dendritic relationships based on input data. It is possible to learn synaptic strength with STDP, to make a neuron more sensitive to its input. Similarly, it is possible to learn dendritic relationships with STDP, to make the neuron more sensitive to spatiotemporal input sequences. Feeding structured data to a WTA network with dendritic computation leads to a significantly higher number of discriminated input patterns. Without the dendritic computation, output neurons are less specific and may, for instance, be activated by a sequence in reverse order.

Keywords: dendritic computation, spiking neural networks, point neuron model

Procedia PDF Downloads 133

Authors: V. Jaime Contreras, S. Iliana Nunez, S. Mario Sanchez

Abstract:

Inside a manufacturing facility, we can find innumerable automatic and manual operations, all of which are relevant to the production process. Some of these processes add more value to the products more than others. Manual operations tend to add value to the product since they can be found in the final assembly area o final operations of the process. In this areas, where a mistake or accident can increase the cost of waste exponentially. To reduce or mitigate these costly mistakes, one approach is to rely on automation to eliminate the operator from the production line - requires a hefty investment and development of specialized machinery. In our approach, the center of the solution is the operator through sufficient and adequate instrumentation, real-time reporting and ergonomics. Efficiency and reduced cycle time can be achieved thorough the integration of Internet-of-Things (IoT) ready technologies into assembly operations to enhance the ergonomics of the workstations. Augmented reality visual aids, RFID triggered personalized workstation dimensions and real-time data transfer and reporting can help achieve these goals. In this case study, a standard work cell will be used for real-life data acquisition and a simulation software to extend the data points beyond the test cycle. Three comparison scenarios will run in the work cell. Each scenario will introduce a dimension of the ergonomics to measure its impact independently. Furthermore, the separate test will determine the limitations of the technology and provide a reference for operating costs and investment required. With the ability, to monitor costs, productivity, cycle time and scrap/waste in real-time the ROI (return on investment) can be determined at the different levels to integration. This case study will help to show that ergonomics in the assembly lines can make significant impact when IoT technologies are introduced. Ergonomics can effectively reduce waste and increase productivity with minimal investment if compared with setting up to custom machine.

Keywords: augmented reality visual aids, ergonomics, real-time data acquisition and reporting, RFID triggered workstation dimensions

Procedia PDF Downloads 214

Authors: Jaeyoung Lee

Abstract:

Autonomous driving systems require high reliability to provide people with a safe and comfortable driving experience. However, despite the development of a number of vehicle sensors, it is difficult to always provide high perceived performance in driving environments that vary from time to season. The image segmentation method using deep learning, which has recently evolved rapidly, provides high recognition performance in various road environments stably. However, since the system controls a vehicle in real time, a highly complex deep learning network cannot be used due to time and memory constraints. Moreover, efficient networks are optimized for GPU environments, which degrade performance in embedded processor environments equipped simple hardware accelerators. In this paper, a semantic segmentation network, matrix multiplication accelerator network (MMANet), optimized for matrix multiplication accelerator (MMA) on Texas instrument digital signal processors (TI DSP) is proposed to improve the recognition performance of autonomous driving system. The proposed method is designed to maximize the number of layers that can be performed in a limited time to provide reliable driving environment information in real time. First, the number of channels in the activation map is fixed to fit the structure of MMA. By increasing the number of parallel branches, the lack of information caused by fixing the number of channels is resolved. Second, an efficient convolution is selected depending on the size of the activation. Since MMA is a fixed, it may be more efficient for normal convolution than depthwise separable convolution depending on memory access overhead. Thus, a convolution type is decided according to output stride to increase network depth. In addition, memory access time is minimized by processing operations only in L3 cache. Lastly, reliable contexts are extracted using the extended atrous spatial pyramid pooling (ASPP). The suggested method gets stable features from an extended path by increasing the kernel size and accessing consecutive data. In addition, it consists of two ASPPs to obtain high quality contexts using the restored shape without global average pooling paths since the layer uses MMA as a simple adder. To verify the proposed method, an experiment is conducted using perfsim, a timing simulator, and the Cityscapes validation sets. The proposed network can process an image with 640 x 480 resolution for 6.67 ms, so six cameras can be used to identify the surroundings of the vehicle as 20 frame per second (FPS). In addition, it achieves 73.1% mean intersection over union (mIoU) which is the highest recognition rate among embedded networks on the Cityscapes validation set.

Keywords: edge network, embedded network, MMA, matrix multiplication accelerator, semantic segmentation network

Procedia PDF Downloads 129