Pub Date : 2022-06-16DOI: 10.1109/ICAT54566.2022.9811189
Pamela Njemcevic, Enio Kaljic, A. Maric
In this paper, the error performance of coherent systems in presence of imperfect carrier phase estimation is investigated for signals propagating over the two-ray with diffuse power (TWDP) fading channels, in case when synchronization is performed using pilot carrier located out of the signal’s band-width. In that sense, closed-form approximate average binary error probability (ABEP) expressions are derived for binary and quadrature phase shift keying (BPSK and QPSK) modulated signals, with the carrier extracted using phase-locked loop (PLL) and phase noise approximated by Tikhonov probability density function (PDF). Derived expressions are calculated for various combinations of channel and phase loop parameters, enabling us to observe their effects on overall system performance. The accu-racy of derived expressions is verified through their comparison with the exact ABEPs obtained by numerical integration of the appropriate expressions.
{"title":"The Effects of Pilot-based Carrier Phase Estimation on Performance of Coherently Detected Signals Propagating in TWDP Channels","authors":"Pamela Njemcevic, Enio Kaljic, A. Maric","doi":"10.1109/ICAT54566.2022.9811189","DOIUrl":"https://doi.org/10.1109/ICAT54566.2022.9811189","url":null,"abstract":"In this paper, the error performance of coherent systems in presence of imperfect carrier phase estimation is investigated for signals propagating over the two-ray with diffuse power (TWDP) fading channels, in case when synchronization is performed using pilot carrier located out of the signal’s band-width. In that sense, closed-form approximate average binary error probability (ABEP) expressions are derived for binary and quadrature phase shift keying (BPSK and QPSK) modulated signals, with the carrier extracted using phase-locked loop (PLL) and phase noise approximated by Tikhonov probability density function (PDF). Derived expressions are calculated for various combinations of channel and phase loop parameters, enabling us to observe their effects on overall system performance. The accu-racy of derived expressions is verified through their comparison with the exact ABEPs obtained by numerical integration of the appropriate expressions.","PeriodicalId":414786,"journal":{"name":"2022 XXVIII International Conference on Information, Communication and Automation Technologies (ICAT)","volume":"122 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129076468","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 : 2022-06-16DOI: 10.1109/ICAT54566.2022.9811187
Ritu Walia
This paper examines the implementation of machine learning (ML) capabilities in a test automation suite, specifically for automation of graphical user interface (GUI) testing on an electronic design automation (EDA) tool within an integrated circuit (IC) physical design, verification, and implementation flow. We present a case study using existing tests to extract information and propose an ML implementation framework that consists of three modules, which can be adopted as a systematic pattern for test development. Our study focusses on implementation of the third module in this framework. We use the learnings from iterative testing patterns on a set of EDA tools provided by the Calibre RealTime interfaces from Siemens Digital Industries Software. The goal is to reduce human effort in selection and implementation of test cases and reallocate those resources to integral parts of the testing process like, approving and acting. We first establish metrics and variables, utilize VGG16 architecture for image classification and perform training on test data, and achieve an ML model based on accuracy and precision. Using this result, we present ML implementation as part of the script development process and analyze its impact. Based on our results, we conclude the third module of a framework for inclusion of ML in a regression testing suite for GUI test automation.
本文研究了测试自动化套件中机器学习(ML)功能的实现，特别是在集成电路(IC)物理设计、验证和实现流程中的电子设计自动化(EDA)工具上的图形用户界面(GUI)测试的自动化。我们提出了一个使用现有测试提取信息的案例研究，并提出了一个由三个模块组成的机器学习实现框架，该框架可以作为测试开发的系统模式。我们的研究重点是该框架中第三个模块的实现。我们在Siemens Digital Industries Software的Calibre RealTime接口提供的一组EDA工具上使用从迭代测试模式中学到的知识。目标是减少人类在选择和实现测试用例方面的工作，并将这些资源重新分配到测试过程的组成部分，如批准和执行。我们首先建立度量和变量，利用VGG16架构对图像进行分类，并对测试数据进行训练，实现基于准确度和精度的机器学习模型。使用这个结果，我们将ML实现作为脚本开发过程的一部分，并分析其影响。根据我们的结果，我们总结了框架的第三个模块，用于将ML包含在GUI测试自动化的回归测试套件中。
{"title":"Application of Machine Learning for GUI Test Automation","authors":"Ritu Walia","doi":"10.1109/ICAT54566.2022.9811187","DOIUrl":"https://doi.org/10.1109/ICAT54566.2022.9811187","url":null,"abstract":"This paper examines the implementation of machine learning (ML) capabilities in a test automation suite, specifically for automation of graphical user interface (GUI) testing on an electronic design automation (EDA) tool within an integrated circuit (IC) physical design, verification, and implementation flow. We present a case study using existing tests to extract information and propose an ML implementation framework that consists of three modules, which can be adopted as a systematic pattern for test development. Our study focusses on implementation of the third module in this framework. We use the learnings from iterative testing patterns on a set of EDA tools provided by the Calibre RealTime interfaces from Siemens Digital Industries Software. The goal is to reduce human effort in selection and implementation of test cases and reallocate those resources to integral parts of the testing process like, approving and acting. We first establish metrics and variables, utilize VGG16 architecture for image classification and perform training on test data, and achieve an ML model based on accuracy and precision. Using this result, we present ML implementation as part of the script development process and analyze its impact. Based on our results, we conclude the third module of a framework for inclusion of ML in a regression testing suite for GUI test automation.","PeriodicalId":414786,"journal":{"name":"2022 XXVIII International Conference on Information, Communication and Automation Technologies (ICAT)","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123972836","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 : 2022-06-16DOI: 10.1109/ICAT54566.2022.9811147
Wael Alsabbagh, P. Langendörfer
In this paper, we take the attack approach introduced in our previous work [8] one more step in the direction of exploiting PLCs offline, and extend our experiments to cover the latest and most secured Siemens PLCs line i.e. S7-1500 CPUs. The attack scenario conducted in this work aims at confusing the behavior of the target system when malicious attackers are not connected neither to the victim system nor to its control network at the very moment of the attack. The new approach presented in this paper is comprised of two stages. First, an attacker patches the PLC with a specific interrupt block, Time-of-Day, once he manages successfully to access/compromise an exposed PLC. Then he triggers the block at a later time the attacker wishes when he is completely offline i.e., disconnected to the control network. For a real-world implementation, we tested our approach on a Fischertechnik system using an S7-1500 CPU that supports the newest version of the S7CommPlus protocol i.e. S7CommPlus v3. Our experimental results showed that we could infect the target PLC successfully and conceal our malicious interrupt block in the PLC memory until the very moment we already determined. This makes our attack stealthy as the engineering station can not detect that the PLC got infected. Finally, we presented security and mitigation methods to prevent such a threat.
{"title":"No Need to be Online to Attack - Exploiting S7-1500 PLCs by Time-Of-Day Block","authors":"Wael Alsabbagh, P. Langendörfer","doi":"10.1109/ICAT54566.2022.9811147","DOIUrl":"https://doi.org/10.1109/ICAT54566.2022.9811147","url":null,"abstract":"In this paper, we take the attack approach introduced in our previous work [8] one more step in the direction of exploiting PLCs offline, and extend our experiments to cover the latest and most secured Siemens PLCs line i.e. S7-1500 CPUs. The attack scenario conducted in this work aims at confusing the behavior of the target system when malicious attackers are not connected neither to the victim system nor to its control network at the very moment of the attack. The new approach presented in this paper is comprised of two stages. First, an attacker patches the PLC with a specific interrupt block, Time-of-Day, once he manages successfully to access/compromise an exposed PLC. Then he triggers the block at a later time the attacker wishes when he is completely offline i.e., disconnected to the control network. For a real-world implementation, we tested our approach on a Fischertechnik system using an S7-1500 CPU that supports the newest version of the S7CommPlus protocol i.e. S7CommPlus v3. Our experimental results showed that we could infect the target PLC successfully and conceal our malicious interrupt block in the PLC memory until the very moment we already determined. This makes our attack stealthy as the engineering station can not detect that the PLC got infected. Finally, we presented security and mitigation methods to prevent such a threat.","PeriodicalId":414786,"journal":{"name":"2022 XXVIII International Conference on Information, Communication and Automation Technologies (ICAT)","volume":"89 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117281227","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 : 2022-06-16DOI: 10.1109/ICAT54566.2022.9811233
I. Džafić
Graphs could be used to illustrate a wide range of practical challenges. The word network is usually used to denote a graph in which the elements are associated with the vertices and edges, emphasizing its relevance to power systems. This paper focuses on two common graph theory applications in Advanced Distribution Management Systems (ADMS): topology tracing and fast gain matrix computing. Topology tracing is a critical component of any ADMS. Its primary function is to generate a branch-node model by traversing branches and closed switches. The gain matrix is built during each iteration of the weighted least squares (WLS) state estimation method, which utilizes the normal equations technique. The gain matrix is sparse with a nonzero structure that remains unchanged throughout iterations. This study describes a method for predicting the nonzero structure of the gain matrix directly from the network graph and measurement locations. The suggested method for computing the gain matrix is at least seven times faster than the MATLAB built-in implementation, making it suitable for constructing efficient real-time power system state estimation software for ADMS.
{"title":"Graph Theory as an Engine for Real-Time Advanced Distribution Management System Enhancements","authors":"I. Džafić","doi":"10.1109/ICAT54566.2022.9811233","DOIUrl":"https://doi.org/10.1109/ICAT54566.2022.9811233","url":null,"abstract":"Graphs could be used to illustrate a wide range of practical challenges. The word network is usually used to denote a graph in which the elements are associated with the vertices and edges, emphasizing its relevance to power systems. This paper focuses on two common graph theory applications in Advanced Distribution Management Systems (ADMS): topology tracing and fast gain matrix computing. Topology tracing is a critical component of any ADMS. Its primary function is to generate a branch-node model by traversing branches and closed switches. The gain matrix is built during each iteration of the weighted least squares (WLS) state estimation method, which utilizes the normal equations technique. The gain matrix is sparse with a nonzero structure that remains unchanged throughout iterations. This study describes a method for predicting the nonzero structure of the gain matrix directly from the network graph and measurement locations. The suggested method for computing the gain matrix is at least seven times faster than the MATLAB built-in implementation, making it suitable for constructing efficient real-time power system state estimation software for ADMS.","PeriodicalId":414786,"journal":{"name":"2022 XXVIII International Conference on Information, Communication and Automation Technologies (ICAT)","volume":"123 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116077709","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 : 2022-06-16DOI: 10.1109/ICAT54566.2022.9811126
Amela Drobo, L. S. Becirovic, L. G. Pokvic, Lucija Dzambo, E. Becic, A. Badnjević, Majda Dogic, Alisa Smajovic
Hepatitis C is an inflammatory condition of the liver caused by the hepatitis C virus. Diagnosis of the disease itself is difficult because the incubation period is long, often the disease is initially without some characteristic symptoms, but also due to a lack of laboratory methods. Artificial intelligence is increasingly being used nowadays to make it easier and faster to assess the illness. As hepatitis C is a rising healthcare burden it is of utmost importance to construct effective and reliable screening methods. As AI has already proven useful for diagnosis of a variety of conditions based on clinical parameters, this study focuses on the application of artificial neural network (ANN) for hepatitis C diagnosis. In this study, a database of 1000 respondents divided into two groups was used to develop the ANN: healthy (n = 200) and sick (n = 800). Monitoring parameters were: albumin, alkaline phosphatase, alanine aminotransferase, aspartate aminotransferase, bilirubin, acetylcholinesterase and anti-HCV antibodies. The overall accuracy of the developed ANN was 97,78%, which indicates that the potential of artificial intelligence in diagnosing hepatitis C is enormous, and in the future, attention should be paid to the development of new systems with as much data as possible.
{"title":"Application of artificial neural networks in diagnosis of Hepatitis C","authors":"Amela Drobo, L. S. Becirovic, L. G. Pokvic, Lucija Dzambo, E. Becic, A. Badnjević, Majda Dogic, Alisa Smajovic","doi":"10.1109/ICAT54566.2022.9811126","DOIUrl":"https://doi.org/10.1109/ICAT54566.2022.9811126","url":null,"abstract":"Hepatitis C is an inflammatory condition of the liver caused by the hepatitis C virus. Diagnosis of the disease itself is difficult because the incubation period is long, often the disease is initially without some characteristic symptoms, but also due to a lack of laboratory methods. Artificial intelligence is increasingly being used nowadays to make it easier and faster to assess the illness. As hepatitis C is a rising healthcare burden it is of utmost importance to construct effective and reliable screening methods. As AI has already proven useful for diagnosis of a variety of conditions based on clinical parameters, this study focuses on the application of artificial neural network (ANN) for hepatitis C diagnosis. In this study, a database of 1000 respondents divided into two groups was used to develop the ANN: healthy (n = 200) and sick (n = 800). Monitoring parameters were: albumin, alkaline phosphatase, alanine aminotransferase, aspartate aminotransferase, bilirubin, acetylcholinesterase and anti-HCV antibodies. The overall accuracy of the developed ANN was 97,78%, which indicates that the potential of artificial intelligence in diagnosing hepatitis C is enormous, and in the future, attention should be paid to the development of new systems with as much data as possible.","PeriodicalId":414786,"journal":{"name":"2022 XXVIII International Conference on Information, Communication and Automation Technologies (ICAT)","volume":"76 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124770034","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 : 2022-06-16DOI: 10.1109/ICAT54566.2022.9811159
Ehlimana Krupalija, D. Donko, H. Supic
Social media is an important source of real-world data for sentiment analysis. Hate speech detection models can be trained on data from Twitter and then utilized for content filtering and removal of posts which contain hate speech. This work proposes a new algorithm for calculating user hate speech index based on user post history. Three available datasets were merged for the purpose of acquiring Twitter posts which contained hate speech. Text preprocessing and tokenization was performed, as well as outlier removal and class balancing. The proposed algorithm was used for determining hate speech index of users who posted tweets from the dataset. The preprocessed dataset was used for training and testing multiple machine learning models: k-means clustering without and with principal component analysis, naïve Bayes, decision tree and random forest. Four different feature subsets of the dataset were used for model training and testing. Anomaly detection, data transformation and parameter tuning were used in an attempt to improve classification accuracy. The highest F1 measure was achieved by training the model using a combination of user hate speech index and other user features. The results show that the usage of user hate speech index, with or without other user features, improves the accuracy of hate speech detection.
{"title":"Usage of user hate speech index for improving hate speech detection in Twitter posts","authors":"Ehlimana Krupalija, D. Donko, H. Supic","doi":"10.1109/ICAT54566.2022.9811159","DOIUrl":"https://doi.org/10.1109/ICAT54566.2022.9811159","url":null,"abstract":"Social media is an important source of real-world data for sentiment analysis. Hate speech detection models can be trained on data from Twitter and then utilized for content filtering and removal of posts which contain hate speech. This work proposes a new algorithm for calculating user hate speech index based on user post history. Three available datasets were merged for the purpose of acquiring Twitter posts which contained hate speech. Text preprocessing and tokenization was performed, as well as outlier removal and class balancing. The proposed algorithm was used for determining hate speech index of users who posted tweets from the dataset. The preprocessed dataset was used for training and testing multiple machine learning models: k-means clustering without and with principal component analysis, naïve Bayes, decision tree and random forest. Four different feature subsets of the dataset were used for model training and testing. Anomaly detection, data transformation and parameter tuning were used in an attempt to improve classification accuracy. The highest F1 measure was achieved by training the model using a combination of user hate speech index and other user features. The results show that the usage of user hate speech index, with or without other user features, improves the accuracy of hate speech detection.","PeriodicalId":414786,"journal":{"name":"2022 XXVIII International Conference on Information, Communication and Automation Technologies (ICAT)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130568933","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 : 2022-06-16DOI: 10.1109/ICAT54566.2022.9811180
Shaban Guma, A. Sezgin, N. Bajçinca
We present an adaptive mixed-criticality based algorithm for weight-based task scheduling and communication resource allocation in the context of a Cyber-Physical System (CPS). The weight-based algorithm is motivated by the continuous computation of the task’s criticality and updates the weight of the CPS subsystem to be used in the task scheduler and the cost function of the optimal resource allocation problem. To demonstrate the algorithm performance, we consider a set of robots driving on a grid and performing a set of tasks with a different mixed-criticality profile, controlled and connected via a wireless channel with limited communication resources.
{"title":"Mixed-criticality communication scheme for networked mobile robots","authors":"Shaban Guma, A. Sezgin, N. Bajçinca","doi":"10.1109/ICAT54566.2022.9811180","DOIUrl":"https://doi.org/10.1109/ICAT54566.2022.9811180","url":null,"abstract":"We present an adaptive mixed-criticality based algorithm for weight-based task scheduling and communication resource allocation in the context of a Cyber-Physical System (CPS). The weight-based algorithm is motivated by the continuous computation of the task’s criticality and updates the weight of the CPS subsystem to be used in the task scheduler and the cost function of the optimal resource allocation problem. To demonstrate the algorithm performance, we consider a set of robots driving on a grid and performing a set of tasks with a different mixed-criticality profile, controlled and connected via a wireless channel with limited communication resources.","PeriodicalId":414786,"journal":{"name":"2022 XXVIII International Conference on Information, Communication and Automation Technologies (ICAT)","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134019768","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 : 2022-06-16DOI: 10.1109/ICAT54566.2022.9811177
Vedin Klovo, H. Lačević, I. Džafić
Instantaneous frequency measurement is a critical component of power system control and automation. Recently, electric power distribution networks with a high proportion of renewable energy have been subjected to unprecedented complexity, necessitating more complicated automation solutions. The major reasons for frequency changes include the usage of dispersed generation, the connection of non-linear loads, and the occurrence of some unforeseen system problems. This paper presents two DFT-based power system frequency measuring algorithms. It considers frequency variations from the system’s fundamental frequency, as well as the noise generated by analog to digital converters (ADC). The IEEE Phasor Measurement Unit (PMU) latest Standard specification (IEC/IEEE 60255-118-1:2018) is used to examine these two methodologies. The methodologies are evaluated using test signals that are required to provide PMU quality evaluation and classification while accounting for process noise, ADC conversion noise, and dynamically changing input voltage and current signals. The tradeoff between DFT simplicity in implementation and needed complexity of power systems is put to the test by abrupt variations in frequency and amplitude of the fundamental component.
{"title":"Real-Time Estimation of Instantaneous Power System Fundamental Frequency","authors":"Vedin Klovo, H. Lačević, I. Džafić","doi":"10.1109/ICAT54566.2022.9811177","DOIUrl":"https://doi.org/10.1109/ICAT54566.2022.9811177","url":null,"abstract":"Instantaneous frequency measurement is a critical component of power system control and automation. Recently, electric power distribution networks with a high proportion of renewable energy have been subjected to unprecedented complexity, necessitating more complicated automation solutions. The major reasons for frequency changes include the usage of dispersed generation, the connection of non-linear loads, and the occurrence of some unforeseen system problems. This paper presents two DFT-based power system frequency measuring algorithms. It considers frequency variations from the system’s fundamental frequency, as well as the noise generated by analog to digital converters (ADC). The IEEE Phasor Measurement Unit (PMU) latest Standard specification (IEC/IEEE 60255-118-1:2018) is used to examine these two methodologies. The methodologies are evaluated using test signals that are required to provide PMU quality evaluation and classification while accounting for process noise, ADC conversion noise, and dynamically changing input voltage and current signals. The tradeoff between DFT simplicity in implementation and needed complexity of power systems is put to the test by abrupt variations in frequency and amplitude of the fundamental component.","PeriodicalId":414786,"journal":{"name":"2022 XXVIII International Conference on Information, Communication and Automation Technologies (ICAT)","volume":"484 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132334536","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 : 2022-06-16DOI: 10.1109/ICAT54566.2022.9811165
J. Velagić, Vedin Klovo, H. Lačević
This paper addresses the use of deep learning techniques in 3D point cloud labeling of environment representations for the task of a semantic visual localization of mobile robots. In contrast to standard problems resolved with Convolutional Neural Networks (CNNs), the paper deals with applying CNNs to segment point clouds that are, unlike images, unordered and unstructured. The used point clouds contain laser measurements of 3D positions (x,y,z) as well as captured RGB camera images from the scanned scene to colorize the point cloud (RGB values). The main focus of the paper is on implementation and evaluation of a hand-crafted convolution layer and the ConvPoint CNN architecture that introduces continuous convolutions for point cloud processing. The solution was implemented in the Python programming language using the PyTorch deep learning framework.
{"title":"Semantic Visual Segmentation of a Mobile Robot Environment Using Deep Learning Model","authors":"J. Velagić, Vedin Klovo, H. Lačević","doi":"10.1109/ICAT54566.2022.9811165","DOIUrl":"https://doi.org/10.1109/ICAT54566.2022.9811165","url":null,"abstract":"This paper addresses the use of deep learning techniques in 3D point cloud labeling of environment representations for the task of a semantic visual localization of mobile robots. In contrast to standard problems resolved with Convolutional Neural Networks (CNNs), the paper deals with applying CNNs to segment point clouds that are, unlike images, unordered and unstructured. The used point clouds contain laser measurements of 3D positions (x,y,z) as well as captured RGB camera images from the scanned scene to colorize the point cloud (RGB values). The main focus of the paper is on implementation and evaluation of a hand-crafted convolution layer and the ConvPoint CNN architecture that introduces continuous convolutions for point cloud processing. The solution was implemented in the Python programming language using the PyTorch deep learning framework.","PeriodicalId":414786,"journal":{"name":"2022 XXVIII International Conference on Information, Communication and Automation Technologies (ICAT)","volume":"248 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125709707","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 : 2022-06-16DOI: 10.1109/ICAT54566.2022.9811215
Adis Panjevic, T. Uzunović, Baris Can Ustundag
Ambient conditions, especially temperature and humidity, have a huge impact on the performance of an air quality sensor. In this paper, four correction models were built to compensate the impact of ambient conditions. Linear regression and machine learning algorithms were used for building the models. Correction models were trained by using three types of measurement data. Raw measurement data was used in the first case. Secondly, measurement data was corrected and a significant improvement was shown. Lastly, measurements of various ambient conditions were used as well. Using corrected and extended measurement data brought a great improvement in accuracy of the models. A neural network correction model proved to be the most efficient in all cases. Compensating the impact of ambient conditions on the performance of an air quality sensor by using correction models was efficient and this method could be used in the air quality monitoring applications. This is of particular importance for usage of low-cost sensors in the air quality monitoring.
{"title":"Development of Correction Models for Three-Electrode NO2 Electrochemical Sensor","authors":"Adis Panjevic, T. Uzunović, Baris Can Ustundag","doi":"10.1109/ICAT54566.2022.9811215","DOIUrl":"https://doi.org/10.1109/ICAT54566.2022.9811215","url":null,"abstract":"Ambient conditions, especially temperature and humidity, have a huge impact on the performance of an air quality sensor. In this paper, four correction models were built to compensate the impact of ambient conditions. Linear regression and machine learning algorithms were used for building the models. Correction models were trained by using three types of measurement data. Raw measurement data was used in the first case. Secondly, measurement data was corrected and a significant improvement was shown. Lastly, measurements of various ambient conditions were used as well. Using corrected and extended measurement data brought a great improvement in accuracy of the models. A neural network correction model proved to be the most efficient in all cases. Compensating the impact of ambient conditions on the performance of an air quality sensor by using correction models was efficient and this method could be used in the air quality monitoring applications. This is of particular importance for usage of low-cost sensors in the air quality monitoring.","PeriodicalId":414786,"journal":{"name":"2022 XXVIII International Conference on Information, Communication and Automation Technologies (ICAT)","volume":"7 1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129827866","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：