Pub Date : 2019-06-10DOI: 10.1109/MECO.2019.8759994
Jens Rudolf, Daniel Gis, S. Stieber, C. Haubelt, R. Dorsch
In IoT domain energy aware firmware development is critical for applications that run on mobile or battery constrained devices. Virtual system prototypes (VSP) empower developers to assess the application power consumption behavior before actual hardware prototypes become available. The SystemC modeling language has become the widely adopted industry standard for the implementation of such VSPs. In this paper, we present a novel approach for extending SystemC based VSPs with pluggable, pre-compiled power models that can be configured during runtime to generate accurate power profiles for the simulated firmware. The necessary modifications to the VSP are kept minimal. We demonstrate the application of our approach by instrumenting a pre-existing SystemC model for a state-of-the-art MEMS-based inertial sensor with a power model and show that the generated power profile estimation matches closely the energy consumption measured from its hardware prototype. As an additional advantage of our proposed precompiled approach, manufactures can ship their power models to costumers without disclosing implementation IP.
{"title":"SystemC Power Profiling for IoT Device Firmware using Runtime Configurable Models","authors":"Jens Rudolf, Daniel Gis, S. Stieber, C. Haubelt, R. Dorsch","doi":"10.1109/MECO.2019.8759994","DOIUrl":"https://doi.org/10.1109/MECO.2019.8759994","url":null,"abstract":"In IoT domain energy aware firmware development is critical for applications that run on mobile or battery constrained devices. Virtual system prototypes (VSP) empower developers to assess the application power consumption behavior before actual hardware prototypes become available. The SystemC modeling language has become the widely adopted industry standard for the implementation of such VSPs. In this paper, we present a novel approach for extending SystemC based VSPs with pluggable, pre-compiled power models that can be configured during runtime to generate accurate power profiles for the simulated firmware. The necessary modifications to the VSP are kept minimal. We demonstrate the application of our approach by instrumenting a pre-existing SystemC model for a state-of-the-art MEMS-based inertial sensor with a power model and show that the generated power profile estimation matches closely the energy consumption measured from its hardware prototype. As an additional advantage of our proposed precompiled approach, manufactures can ship their power models to costumers without disclosing implementation IP.","PeriodicalId":141324,"journal":{"name":"2019 8th Mediterranean Conference on Embedded Computing (MECO)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129106260","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-06-10DOI: 10.1109/MECO.2019.8760125
P. Babayan, M. Ershov, D. Y. Erokhin
In our research we compare various neural network architectures that are used for object detection and recognition. In this work vehicles and pedestrians are considered objects of interest. Modern artificial neural networks are able to detect and localize objects of known classes. This allows them to be used in various technical vision systems and video analysis systems. In this paper we compare three architectures (YOLO, Faster R-CNN, SSD) by the following criteria: processing speed, mAP, precision and recall.
{"title":"Neural Network-Based Vehicle and Pedestrian Detection for Video Analysis System","authors":"P. Babayan, M. Ershov, D. Y. Erokhin","doi":"10.1109/MECO.2019.8760125","DOIUrl":"https://doi.org/10.1109/MECO.2019.8760125","url":null,"abstract":"In our research we compare various neural network architectures that are used for object detection and recognition. In this work vehicles and pedestrians are considered objects of interest. Modern artificial neural networks are able to detect and localize objects of known classes. This allows them to be used in various technical vision systems and video analysis systems. In this paper we compare three architectures (YOLO, Faster R-CNN, SSD) by the following criteria: processing speed, mAP, precision and recall.","PeriodicalId":141324,"journal":{"name":"2019 8th Mediterranean Conference on Embedded Computing (MECO)","volume":"184 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123180123","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-06-10DOI: 10.1109/MECO.2019.8760162
M. K. Halidu, P. B. Zadeh, A. S. Akbari, R. Behringer
Face recognition has become a field of interest in many applications such as security and entertainments. In surveillance system, the quality of the recoded footage is sometimes insufficient due to the distance and angle of the camera from the scene. This causes the object of interest, e.g. the face of a person in the scene to be of low resolution, which increases the difficulty in recognition process. Image resolution enhancement is a potential solution for enlarging low-resolution images for real time face recognition. An enlarged image is then compared to available database of images to either identify or verify the individuals. However, the optimal performance of face recognition techniques when various image enlargement methods have been applied to them has not been investigated. In this research, the performance of PCA based face recognition method, with the three most well-known image enlargement techniques (Nearest Neighbour, Bilinear, Bicubic) is investigated. First, an input image is down sampled to six different resolutions. The down-sampled image is then enlarged to its original size using the three named image enlargement techniques. The enlarged image is then input to a PCA face recognition system for the recognition process. The simulation results using images from the SCFace database show that PCA based face recognition illustrates superior results when input images enlarged using Nearest Neighbour technique, while the performance of Bicubic and Bilinear techniques is slightly lower than Nearest Neighbour method.
{"title":"PCA in the context of Face Recognition with the Image Enlargement Techniques","authors":"M. K. Halidu, P. B. Zadeh, A. S. Akbari, R. Behringer","doi":"10.1109/MECO.2019.8760162","DOIUrl":"https://doi.org/10.1109/MECO.2019.8760162","url":null,"abstract":"Face recognition has become a field of interest in many applications such as security and entertainments. In surveillance system, the quality of the recoded footage is sometimes insufficient due to the distance and angle of the camera from the scene. This causes the object of interest, e.g. the face of a person in the scene to be of low resolution, which increases the difficulty in recognition process. Image resolution enhancement is a potential solution for enlarging low-resolution images for real time face recognition. An enlarged image is then compared to available database of images to either identify or verify the individuals. However, the optimal performance of face recognition techniques when various image enlargement methods have been applied to them has not been investigated. In this research, the performance of PCA based face recognition method, with the three most well-known image enlargement techniques (Nearest Neighbour, Bilinear, Bicubic) is investigated. First, an input image is down sampled to six different resolutions. The down-sampled image is then enlarged to its original size using the three named image enlargement techniques. The enlarged image is then input to a PCA face recognition system for the recognition process. The simulation results using images from the SCFace database show that PCA based face recognition illustrates superior results when input images enlarged using Nearest Neighbour technique, while the performance of Bicubic and Bilinear techniques is slightly lower than Nearest Neighbour method.","PeriodicalId":141324,"journal":{"name":"2019 8th Mediterranean Conference on Embedded Computing (MECO)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131184825","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-06-10DOI: 10.1109/MECO.2019.8760004
Ilir Keka, B. Çiço
Visualization as a technique is increasingly being used in the data science. The aim of this paper is to analyze the data, in this case the data of Load Profiles and to find the trend detection of these data. In addition, the intention of this work is to find the correlation of the variables in Multiple Linear Regression Model. As tool there is used R Programming Language, which is very suitable for data visualization.
{"title":"Data Visualization as Helping Technique for Data Analysis, Trend Detection and Correlation of Variables Using R Programming Language","authors":"Ilir Keka, B. Çiço","doi":"10.1109/MECO.2019.8760004","DOIUrl":"https://doi.org/10.1109/MECO.2019.8760004","url":null,"abstract":"Visualization as a technique is increasingly being used in the data science. The aim of this paper is to analyze the data, in this case the data of Load Profiles and to find the trend detection of these data. In addition, the intention of this work is to find the correlation of the variables in Multiple Linear Regression Model. As tool there is used R Programming Language, which is very suitable for data visualization.","PeriodicalId":141324,"journal":{"name":"2019 8th Mediterranean Conference on Embedded Computing (MECO)","volume":"132 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132093275","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-06-10DOI: 10.1109/MECO.2019.8760108
Dordije Boskovic, M. Orlandić, Sivert Bakken, T. Johansen
Hyperspectral images obtained by imaging spectrometer contain a vast amount of data which require techniques such as target detection to extract useful information. This article presents an implementation of the target detection method Adaptive Cosine Estimator (ACE) for hyperspectral images. The algorithm is implemented as hardware-software partitioned system on Zynq-7000 development platform. The computationally intensive operations are accelerated on FPGA with the speedup factor of 28.54. The timing analysis presents results for the partitioned system as well as for the software implementation on Zynq processing system used for comparison. The detection performance of the implemented algorithm is tested and verified using publicly available hyperspectral scenes with ground truth data.
{"title":"HW/SW Implementation of Hyperspectral Target Detection Algorithm","authors":"Dordije Boskovic, M. Orlandić, Sivert Bakken, T. Johansen","doi":"10.1109/MECO.2019.8760108","DOIUrl":"https://doi.org/10.1109/MECO.2019.8760108","url":null,"abstract":"Hyperspectral images obtained by imaging spectrometer contain a vast amount of data which require techniques such as target detection to extract useful information. This article presents an implementation of the target detection method Adaptive Cosine Estimator (ACE) for hyperspectral images. The algorithm is implemented as hardware-software partitioned system on Zynq-7000 development platform. The computationally intensive operations are accelerated on FPGA with the speedup factor of 28.54. The timing analysis presents results for the partitioned system as well as for the software implementation on Zynq processing system used for comparison. The detection performance of the implemented algorithm is tested and verified using publicly available hyperspectral scenes with ground truth data.","PeriodicalId":141324,"journal":{"name":"2019 8th Mediterranean Conference on Embedded Computing (MECO)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116568308","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-06-10DOI: 10.1109/MECO.2019.8760180
A. Zabasta, A. Zhiravetska, N. Kunicina, K. Kondratjevs
The modern autonomous beekeeping system developed in this research is the real example of Internet of Things technologies (IoT) in the beekeeping sector. It performs a bee colony maintenance control without interfering with its processes, while optimizing frequency of the apiary inspection. The system helps to analyze data correlation with video, meteo data, mass changes in time as well as interpretation of nest temperature, humidity and linking to local geographic and biological conditions. It allows a beekeeper to request and receive key data indicators and in accordance with the indicators to react on time and provide the best required maintenance of the bee colony. By implementing the autonomous beekeeping, the hives conditions can be tracked remotely, e.g. whether the inside temperature is critical, if the family is missing feed, therefore the critical deviation can be detected and prevented in time.
{"title":"Technical Implementation of IoT Concept for Bee Colony Monitoring","authors":"A. Zabasta, A. Zhiravetska, N. Kunicina, K. Kondratjevs","doi":"10.1109/MECO.2019.8760180","DOIUrl":"https://doi.org/10.1109/MECO.2019.8760180","url":null,"abstract":"The modern autonomous beekeeping system developed in this research is the real example of Internet of Things technologies (IoT) in the beekeeping sector. It performs a bee colony maintenance control without interfering with its processes, while optimizing frequency of the apiary inspection. The system helps to analyze data correlation with video, meteo data, mass changes in time as well as interpretation of nest temperature, humidity and linking to local geographic and biological conditions. It allows a beekeeper to request and receive key data indicators and in accordance with the indicators to react on time and provide the best required maintenance of the bee colony. By implementing the autonomous beekeeping, the hives conditions can be tracked remotely, e.g. whether the inside temperature is critical, if the family is missing feed, therefore the critical deviation can be detected and prevented in time.","PeriodicalId":141324,"journal":{"name":"2019 8th Mediterranean Conference on Embedded Computing (MECO)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126656198","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-06-10DOI: 10.1109/MECO.2019.8760109
Veselin N. Ivanović, Srdjan Jovanovski
Estimation of nonstationary one-dimensional and two-dimensional signals represents very challenging problem that has efficiently been solved by using time-frequency and space/spatial-frequency analysis tools, respectively. However, these solutions provide high quality results only in the cases of linear frequency modulated (FM) signals. To this end, regions of support of the developed solutions correspond to the instantaneous frequency (IF) of the estimated signals, whereas the filtering problem is reduced to the IF estimation. Contrarily, non-linear signals occupy certain ranges of frequencies in a time instant, so that the IF estimation-based solutions cannon produce high quality results in this case. Therefore, in this paper the time-frequency filtering solution suitable for the non-linear FM signal estimation is considered.
{"title":"Real-Time Procedure for Development of an Optimal Time-Frequency Filter Suitable for Non-Linear Highly Nonstationary FM Signals Estimation","authors":"Veselin N. Ivanović, Srdjan Jovanovski","doi":"10.1109/MECO.2019.8760109","DOIUrl":"https://doi.org/10.1109/MECO.2019.8760109","url":null,"abstract":"Estimation of nonstationary one-dimensional and two-dimensional signals represents very challenging problem that has efficiently been solved by using time-frequency and space/spatial-frequency analysis tools, respectively. However, these solutions provide high quality results only in the cases of linear frequency modulated (FM) signals. To this end, regions of support of the developed solutions correspond to the instantaneous frequency (IF) of the estimated signals, whereas the filtering problem is reduced to the IF estimation. Contrarily, non-linear signals occupy certain ranges of frequencies in a time instant, so that the IF estimation-based solutions cannon produce high quality results in this case. Therefore, in this paper the time-frequency filtering solution suitable for the non-linear FM signal estimation is considered.","PeriodicalId":141324,"journal":{"name":"2019 8th Mediterranean Conference on Embedded Computing (MECO)","volume":"54 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122674805","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-06-10DOI: 10.1109/MECO.2019.8760289
M. Engin
One of the most important limitations to be considered in the design phase of the embedded system is the energy consumption. The energy consumption is more efficient, especially in the embedded devices fed by the battery, because the charging time is an obstacle to the use of the device. In addition, energy efficiency in terms of power dissipation and system health is important in other embedded applications such as biomedical, test and measurement, industrial control and robots. Reducing energy consumption during the design phase of the embedded system is generally considered as the task of the hardware. In fact, the software should also undertake the task of improving energy efficiency. In this research, the control algorithm of the embedded system is written in two different methods and their energy consumption was compared.
{"title":"Energy Efficiency of Embedded Controllers","authors":"M. Engin","doi":"10.1109/MECO.2019.8760289","DOIUrl":"https://doi.org/10.1109/MECO.2019.8760289","url":null,"abstract":"One of the most important limitations to be considered in the design phase of the embedded system is the energy consumption. The energy consumption is more efficient, especially in the embedded devices fed by the battery, because the charging time is an obstacle to the use of the device. In addition, energy efficiency in terms of power dissipation and system health is important in other embedded applications such as biomedical, test and measurement, industrial control and robots. Reducing energy consumption during the design phase of the embedded system is generally considered as the task of the hardware. In fact, the software should also undertake the task of improving energy efficiency. In this research, the control algorithm of the embedded system is written in two different methods and their energy consumption was compared.","PeriodicalId":141324,"journal":{"name":"2019 8th Mediterranean Conference on Embedded Computing (MECO)","volume":"64 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124495120","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-06-10DOI: 10.1109/MECO.2019.8760189
Manar N. Shaker, A. Hussien, G. Alkady, H. Amer, I. Adly
In the Automotive Industry, many applications are currently implemented on Field Programmable Gate Arrays (FPGAs). Nowadays, due to the continuous shrinking of transistor dimensions, FPGAs are subjected to Multiple Event Upsets (MEUs) in addition to the well-studied Single Event Upsets (SEUs). Fault tolerance is often used to mitigate this problem. This paper explains why the currently utilized fault-tolerant techniques such as scrubbing will probably produce some erroneous outputs; further more Triple Modular Redundancy may not recover from MEUs. Penta Modular Redundancy can efficiently recover from MEUs as well as SEUs; however, it cannot detect some faulty scenarios. This problem is solved by using the Hexa Modular Redundancy fault tolerant technique. The reliabilities of both Penta and Hexa Modular Redundancy are calculated using Markov models to investigate whether the expected increase in system reliability outweighs the cost of extra added redundancy. Finally, the extra power consumed by the architecture due to the added redundancy is estimated using Xilinx Vivado tools.
{"title":"Mitigating the Effect of Multiple Event Upsets in FPGA-Based Automotive Applications","authors":"Manar N. Shaker, A. Hussien, G. Alkady, H. Amer, I. Adly","doi":"10.1109/MECO.2019.8760189","DOIUrl":"https://doi.org/10.1109/MECO.2019.8760189","url":null,"abstract":"In the Automotive Industry, many applications are currently implemented on Field Programmable Gate Arrays (FPGAs). Nowadays, due to the continuous shrinking of transistor dimensions, FPGAs are subjected to Multiple Event Upsets (MEUs) in addition to the well-studied Single Event Upsets (SEUs). Fault tolerance is often used to mitigate this problem. This paper explains why the currently utilized fault-tolerant techniques such as scrubbing will probably produce some erroneous outputs; further more Triple Modular Redundancy may not recover from MEUs. Penta Modular Redundancy can efficiently recover from MEUs as well as SEUs; however, it cannot detect some faulty scenarios. This problem is solved by using the Hexa Modular Redundancy fault tolerant technique. The reliabilities of both Penta and Hexa Modular Redundancy are calculated using Markov models to investigate whether the expected increase in system reliability outweighs the cost of extra added redundancy. Finally, the extra power consumed by the architecture due to the added redundancy is estimated using Xilinx Vivado tools.","PeriodicalId":141324,"journal":{"name":"2019 8th Mediterranean Conference on Embedded Computing (MECO)","volume":"93 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130922297","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2019-06-10DOI: 10.1109/MECO.2019.8760195
M. Hagara, O. Ondrácek, P. Kubinec, R. Stojanovic
Over the last four decades, hundreds of methods for edge detection in digital images have been published. Each new edge detector was tested on a smaller or larger number of images. Sometimes the researchers have used the images, encoded in lossy JPEG format to test their proposed algorithm. The goal of this paper is to show whether lossy image compression can affect the quality of edge detection. The results presented in this article show that lossy image compression can impair the efficiency of edge detection by up to six percent.
{"title":"Edge Detection in JPEG Grayscale Images","authors":"M. Hagara, O. Ondrácek, P. Kubinec, R. Stojanovic","doi":"10.1109/MECO.2019.8760195","DOIUrl":"https://doi.org/10.1109/MECO.2019.8760195","url":null,"abstract":"Over the last four decades, hundreds of methods for edge detection in digital images have been published. Each new edge detector was tested on a smaller or larger number of images. Sometimes the researchers have used the images, encoded in lossy JPEG format to test their proposed algorithm. The goal of this paper is to show whether lossy image compression can affect the quality of edge detection. The results presented in this article show that lossy image compression can impair the efficiency of edge detection by up to six percent.","PeriodicalId":141324,"journal":{"name":"2019 8th Mediterranean Conference on Embedded Computing (MECO)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114712335","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：