Pub Date : 2021-09-07DOI: 10.1109/ETFA45728.2021.9613360
Ivan Pisa, A. Morell, J. Vicario, R. Vilanova
The incursion of the Industry 4.0 paradigm and the Artificial Neural Networks (ANNs) is changing the way as the industrial systems are conceived and controlled. Now, it is more common to talk about data-driven methods either supporting conventional industrial control strategies, or acting as the control itself. Thus, one can find that in the last years it is more common to find control systems which are purely based on data leaving aside the highly complex mathematical models. However, data-driven models and ANNs have to be correctly trained in order to offer a good performance and therefore, be contemplated as the core part of a control strategy. This can become a time-demanding and tedious process. For that reason, Transfer Learning (TL) techniques can be adopted to ease the conception, design and training processes of the data-based and ANNs methods, since the efforts have to be mainly focused on training a unique net which will be then transferred into the other scenarios. In that sense, we present here a TL approach to design and implement the whole control of a Wastewater Treatment Plant (WWTP). First, the control of the quickest dynamics under control is performed by means of a Long Short-Term Memory cell (LSTM) based Proportional Integral (PI) controller (LSTM-based PI). Once the LSTM is trained and tested, its knowledge will be transferred into the remaining WWTP control loops. In that way, an ease and reduction in the time involved in the design and training of the control as well as in its complexity is achieved. Results have shown a twofold achievement: (i) the LSTM-based PI achieves an improvement of the control performance with respect to a conventional PI controller around a 93.56% and a 99.07% in terms of the Integrated Absolute (IAE) and Integrated Squared (ISE) errors between the desired measurement and the obtained one, respectively, and (ii) the LSTM-based PI controller achieves an average improvement in the IAE and ISE around a 9.55% and 15.25%, respectively, when it is transferred into a different WWTP control loop.
{"title":"Transfer Learning Approach for the Design of Basic Control Loops in Wastewater Treatment Plants","authors":"Ivan Pisa, A. Morell, J. Vicario, R. Vilanova","doi":"10.1109/ETFA45728.2021.9613360","DOIUrl":"https://doi.org/10.1109/ETFA45728.2021.9613360","url":null,"abstract":"The incursion of the Industry 4.0 paradigm and the Artificial Neural Networks (ANNs) is changing the way as the industrial systems are conceived and controlled. Now, it is more common to talk about data-driven methods either supporting conventional industrial control strategies, or acting as the control itself. Thus, one can find that in the last years it is more common to find control systems which are purely based on data leaving aside the highly complex mathematical models. However, data-driven models and ANNs have to be correctly trained in order to offer a good performance and therefore, be contemplated as the core part of a control strategy. This can become a time-demanding and tedious process. For that reason, Transfer Learning (TL) techniques can be adopted to ease the conception, design and training processes of the data-based and ANNs methods, since the efforts have to be mainly focused on training a unique net which will be then transferred into the other scenarios. In that sense, we present here a TL approach to design and implement the whole control of a Wastewater Treatment Plant (WWTP). First, the control of the quickest dynamics under control is performed by means of a Long Short-Term Memory cell (LSTM) based Proportional Integral (PI) controller (LSTM-based PI). Once the LSTM is trained and tested, its knowledge will be transferred into the remaining WWTP control loops. In that way, an ease and reduction in the time involved in the design and training of the control as well as in its complexity is achieved. Results have shown a twofold achievement: (i) the LSTM-based PI achieves an improvement of the control performance with respect to a conventional PI controller around a 93.56% and a 99.07% in terms of the Integrated Absolute (IAE) and Integrated Squared (ISE) errors between the desired measurement and the obtained one, respectively, and (ii) the LSTM-based PI controller achieves an average improvement in the IAE and ISE around a 9.55% and 15.25%, respectively, when it is transferred into a different WWTP control loop.","PeriodicalId":312498,"journal":{"name":"2021 26th IEEE International Conference on Emerging Technologies and Factory Automation (ETFA )","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127567789","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 : 2021-09-07DOI: 10.1109/ETFA45728.2021.9613573
M. Lewke, S. Nielsen, A. List, F. Gärtner, T. Klassen, A. Fay
In recent years, cold spraying (CS) has emerged as a promising technology for repair applications, particularly for oxidation-sensitive materials. In order to obtain an optimum repair result that fulfills the highest requirements regarding material properties, simple geometric shape restoration is not sufficient. Any additive manufacturing process results in particular features in microstructure, possible defects and respective - potentially even anisotropic - mechanical properties. To systematically tailor these microstructures and properties to the specific component and geometry requires complex routines. This work proposes the design of a knowledge-based cold spray repair system that facilitates a complete individual repair procedure for aircraft components. This system includes the elements of (i) reverse engineering to analyze, classify and generate digital data of the damaged component, (ii) pre-processing to obtain the ideal conditions for the CS process, (iii) toolpath planning to optimize robotics for the CS process, (iv) on-line monitoring to ensure process quality, (v) post-processing and (vi) performance testing of the material properties to meet the challenging requirements of the aerospace industry. By using an industrial robot and computer-aided planning of the trajectories, components are to be repaired under cold spray and geometrical conditions for ideal material deposition. The goal is to obtain repaired components that fulfill the required property profile equally well as respective new parts.
{"title":"Knowledge-based Optimization of Cold Spray for Aircraft Component Repair","authors":"M. Lewke, S. Nielsen, A. List, F. Gärtner, T. Klassen, A. Fay","doi":"10.1109/ETFA45728.2021.9613573","DOIUrl":"https://doi.org/10.1109/ETFA45728.2021.9613573","url":null,"abstract":"In recent years, cold spraying (CS) has emerged as a promising technology for repair applications, particularly for oxidation-sensitive materials. In order to obtain an optimum repair result that fulfills the highest requirements regarding material properties, simple geometric shape restoration is not sufficient. Any additive manufacturing process results in particular features in microstructure, possible defects and respective - potentially even anisotropic - mechanical properties. To systematically tailor these microstructures and properties to the specific component and geometry requires complex routines. This work proposes the design of a knowledge-based cold spray repair system that facilitates a complete individual repair procedure for aircraft components. This system includes the elements of (i) reverse engineering to analyze, classify and generate digital data of the damaged component, (ii) pre-processing to obtain the ideal conditions for the CS process, (iii) toolpath planning to optimize robotics for the CS process, (iv) on-line monitoring to ensure process quality, (v) post-processing and (vi) performance testing of the material properties to meet the challenging requirements of the aerospace industry. By using an industrial robot and computer-aided planning of the trajectories, components are to be repaired under cold spray and geometrical conditions for ideal material deposition. The goal is to obtain repaired components that fulfill the required property profile equally well as respective new parts.","PeriodicalId":312498,"journal":{"name":"2021 26th IEEE International Conference on Emerging Technologies and Factory Automation (ETFA )","volume":"174 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132809219","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 : 2021-09-07DOI: 10.1109/ETFA45728.2021.9613505
Giorgio Nicola, S. Ghidoni
In this paper we tackle motion planning in industrial human-robot cooperative scenarios modeled as a reinforcement learning problem solved in a simulated environment. The agent learns the most effective policy to reach the designated target position while avoiding collisions with a human, performing a pick and place task in the robot workspace, and with fixed obstacles. The policy acts as a feedback motion planner (or reactive motion planner), therefore at each time-step it senses the surrounding environment and computes the action to be performed. In this work a novel formulation of the action that guarantees the trajectory derivatives continuity is proposed to create smooth trajectories that are necessary for maximizing the human trust in the robot. The action is defined as the sub-trajectory the agent must follow for the duration of a time-step, therefore the complete trajectory is the concatenation of all the trajectories computed at each time-step. The proposed method does not require to infer the action the human is currently performing and/or foresee the space occupied by the human. Indeed, during the training phase in a simulated environment the agent experience how the human behaves in the specific scenario, therefore it learns the policy that best adapts to the human actions and movements. The proposed method is finally applied in a scenario of human-robot cooperative pick and place.
{"title":"Deep Reinforcement Learning for Motion Planning in Human Robot cooperative Scenarios","authors":"Giorgio Nicola, S. Ghidoni","doi":"10.1109/ETFA45728.2021.9613505","DOIUrl":"https://doi.org/10.1109/ETFA45728.2021.9613505","url":null,"abstract":"In this paper we tackle motion planning in industrial human-robot cooperative scenarios modeled as a reinforcement learning problem solved in a simulated environment. The agent learns the most effective policy to reach the designated target position while avoiding collisions with a human, performing a pick and place task in the robot workspace, and with fixed obstacles. The policy acts as a feedback motion planner (or reactive motion planner), therefore at each time-step it senses the surrounding environment and computes the action to be performed. In this work a novel formulation of the action that guarantees the trajectory derivatives continuity is proposed to create smooth trajectories that are necessary for maximizing the human trust in the robot. The action is defined as the sub-trajectory the agent must follow for the duration of a time-step, therefore the complete trajectory is the concatenation of all the trajectories computed at each time-step. The proposed method does not require to infer the action the human is currently performing and/or foresee the space occupied by the human. Indeed, during the training phase in a simulated environment the agent experience how the human behaves in the specific scenario, therefore it learns the policy that best adapts to the human actions and movements. The proposed method is finally applied in a scenario of human-robot cooperative pick and place.","PeriodicalId":312498,"journal":{"name":"2021 26th IEEE International Conference on Emerging Technologies and Factory Automation (ETFA )","volume":"72 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133977599","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 : 2021-09-07DOI: 10.1109/ETFA45728.2021.9613293
H. Fadhlillah, Bianca Wiesmayr, Michael Oberlehner, Rick Rabiser, Alois Zoitl
Modern production plants are complex Cyber-Physical Production Systems with an ever-increasing share of software controlling and automating their operation. The customization of these systems to the needs of their customers and their frequent evolution over a typically long life-cycle result in a plethora of variants to be managed. However, reuse still is mainly done in an opportunistic way, relying on copy-paste-modify strategies. More systematic, strategic reuse would help to reduce costs and time to market, but requires approaches that can be applied to domain-specific languages for developing distributed control software. In this paper, we propose an approach to manage variability in IEC 61499-based systems using delta-oriented variability modeling. We discuss open challenges and outline a research agenda for variability management in IEC 61499.
{"title":"Towards Delta-Oriented Variability Modeling for IEC 61499","authors":"H. Fadhlillah, Bianca Wiesmayr, Michael Oberlehner, Rick Rabiser, Alois Zoitl","doi":"10.1109/ETFA45728.2021.9613293","DOIUrl":"https://doi.org/10.1109/ETFA45728.2021.9613293","url":null,"abstract":"Modern production plants are complex Cyber-Physical Production Systems with an ever-increasing share of software controlling and automating their operation. The customization of these systems to the needs of their customers and their frequent evolution over a typically long life-cycle result in a plethora of variants to be managed. However, reuse still is mainly done in an opportunistic way, relying on copy-paste-modify strategies. More systematic, strategic reuse would help to reduce costs and time to market, but requires approaches that can be applied to domain-specific languages for developing distributed control software. In this paper, we propose an approach to manage variability in IEC 61499-based systems using delta-oriented variability modeling. We discuss open challenges and outline a research agenda for variability management in IEC 61499.","PeriodicalId":312498,"journal":{"name":"2021 26th IEEE International Conference on Emerging Technologies and Factory Automation (ETFA )","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125441023","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 : 2021-09-07DOI: 10.1109/ETFA45728.2021.9613193
Sebastian Schmied, Daniel Grossmann, Selvine G. Mathias, Ralph Klaus Müller
Information models are defined as a comprehensive semantic description of data within a production system. These systems underlay a constant change. Therefore, the corresponding information models are also subject to continuous evolution. This paper shows approaches for the versioning compliant design of information models and their implementation as well as support strategies to assist such changes.
{"title":"A concept towards the evolution and versioning of aggregated information models","authors":"Sebastian Schmied, Daniel Grossmann, Selvine G. Mathias, Ralph Klaus Müller","doi":"10.1109/ETFA45728.2021.9613193","DOIUrl":"https://doi.org/10.1109/ETFA45728.2021.9613193","url":null,"abstract":"Information models are defined as a comprehensive semantic description of data within a production system. These systems underlay a constant change. Therefore, the corresponding information models are also subject to continuous evolution. This paper shows approaches for the versioning compliant design of information models and their implementation as well as support strategies to assist such changes.","PeriodicalId":312498,"journal":{"name":"2021 26th IEEE International Conference on Emerging Technologies and Factory Automation (ETFA )","volume":"123 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132608324","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 : 2021-09-07DOI: 10.1109/ETFA45728.2021.9613449
A. Bonci, Renat Kermenov, S. Longhi, Giacomo Nabissi
The field of Permanent Magnet Synchronous Motors (PMSMs) diagnosis is of research interest because widely used both in the Industrial environment and in electric vehicles. Amongst various Fault Detection (FD) techniques, the Motor Current Signature Analysis (MCSA) received lots of attention because some defecting frequencies may be monitored through the motor currents in case of steady-state functioning. This latter assumption is not always fulfilled, such e.g. in robotic systems driven by PMSMs, where constant speed assumption is unrealistic in most of the cases. Furthermore, MCSA in not suitable for systems working under non-stationary conditions without using advanced processing techniques. This work investigates the use of load torque information for motor diagnostic purposes under not constant speed assumption. Simulations and experimental results are presented regarding the use of the proposed Motor Torque Analysis (MTA) to overcome these limits.
{"title":"Motor Torque Analysis for diagnosis in PMSMs under non-stationary conditions","authors":"A. Bonci, Renat Kermenov, S. Longhi, Giacomo Nabissi","doi":"10.1109/ETFA45728.2021.9613449","DOIUrl":"https://doi.org/10.1109/ETFA45728.2021.9613449","url":null,"abstract":"The field of Permanent Magnet Synchronous Motors (PMSMs) diagnosis is of research interest because widely used both in the Industrial environment and in electric vehicles. Amongst various Fault Detection (FD) techniques, the Motor Current Signature Analysis (MCSA) received lots of attention because some defecting frequencies may be monitored through the motor currents in case of steady-state functioning. This latter assumption is not always fulfilled, such e.g. in robotic systems driven by PMSMs, where constant speed assumption is unrealistic in most of the cases. Furthermore, MCSA in not suitable for systems working under non-stationary conditions without using advanced processing techniques. This work investigates the use of load torque information for motor diagnostic purposes under not constant speed assumption. Simulations and experimental results are presented regarding the use of the proposed Motor Torque Analysis (MTA) to overcome these limits.","PeriodicalId":312498,"journal":{"name":"2021 26th IEEE International Conference on Emerging Technologies and Factory Automation (ETFA )","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132724851","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 : 2021-09-07DOI: 10.1109/ETFA45728.2021.9613563
Roman Matvejev, Yar Muhammad, Naveed Muhammad
Flow sensing has been investigated in the context of underwater and aerial robotics in the past decade. It has not been explored for applications in autonomous ground robotics. In this work-in-progress paper, we investigate the use of air-flow sensing for the applications in autonomous driving. More precisely, we investigate the use of air-flow sensing for vehicle length estimation by conducting computational-fluid-dynamics (CFD) simulations.
{"title":"Air-flow sensing for vehicle length estimation in autonomous driving applications","authors":"Roman Matvejev, Yar Muhammad, Naveed Muhammad","doi":"10.1109/ETFA45728.2021.9613563","DOIUrl":"https://doi.org/10.1109/ETFA45728.2021.9613563","url":null,"abstract":"Flow sensing has been investigated in the context of underwater and aerial robotics in the past decade. It has not been explored for applications in autonomous ground robotics. In this work-in-progress paper, we investigate the use of air-flow sensing for the applications in autonomous driving. More precisely, we investigate the use of air-flow sensing for vehicle length estimation by conducting computational-fluid-dynamics (CFD) simulations.","PeriodicalId":312498,"journal":{"name":"2021 26th IEEE International Conference on Emerging Technologies and Factory Automation (ETFA )","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117156666","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 : 2021-09-07DOI: 10.1109/ETFA45728.2021.9613218
Ioana Hustiu, C. Mahulea, M. Kloetzer
We consider the problem of obtaining independent trajectories for robots from a team, such that their movement satisfies a global co-safe Linear Temporal Logic (LTL) mission over some regions of interest from the environment. For this, the environment is abstracted into a discrete event system using an underlying partition and an available method is used for decomposing the LTL formula into more parts that can be independently satisfied by a robot. Then, we translate these parts into a conjunction of Boolean formulas and use another approach for planning a team based on Boolean specifications and Petri net models. The proposed combination among the two methods yields independent robot trajectories that are optimal with respect to the number of traversed cells from the partition. The advantages are also illustrated through simulation examples.
{"title":"Optimal task allocation for distributed co-safe LTL specifications","authors":"Ioana Hustiu, C. Mahulea, M. Kloetzer","doi":"10.1109/ETFA45728.2021.9613218","DOIUrl":"https://doi.org/10.1109/ETFA45728.2021.9613218","url":null,"abstract":"We consider the problem of obtaining independent trajectories for robots from a team, such that their movement satisfies a global co-safe Linear Temporal Logic (LTL) mission over some regions of interest from the environment. For this, the environment is abstracted into a discrete event system using an underlying partition and an available method is used for decomposing the LTL formula into more parts that can be independently satisfied by a robot. Then, we translate these parts into a conjunction of Boolean formulas and use another approach for planning a team based on Boolean specifications and Petri net models. The proposed combination among the two methods yields independent robot trajectories that are optimal with respect to the number of traversed cells from the partition. The advantages are also illustrated through simulation examples.","PeriodicalId":312498,"journal":{"name":"2021 26th IEEE International Conference on Emerging Technologies and Factory Automation (ETFA )","volume":"106 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117216511","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 : 2021-09-07DOI: 10.1109/ETFA45728.2021.9613647
Julian Rahm, Daniel Henselmann, L. Urbas
In the process industry, ongoing digitization is leading to a considerable information exchange over the entire life cycle of a process plant. Data models are usually created in the initial planning phases and these change continuously with further progress including the planning phase itself, plant adaptations after commissioning, and maintenance. The various models from the disciplines involved have a considerable semantic overlap, creating dependencies. Changes in one model inevitably lead to changes in another. This paper presents an approach for networking and synchronization between data models based on these overlaps. The resulting synchronization network provides models which are as free of inconsistencies as possible across different disciplines at any point in time. In addition, first approaches for collaborative conflict resolution are presented. Approaches from model-driven software development and web technologies are adapted to the application domain of the process industry. Typical information models created during the planning phase of a process engineering plant serve as an example.
{"title":"Synchronization network of data models in the process industry","authors":"Julian Rahm, Daniel Henselmann, L. Urbas","doi":"10.1109/ETFA45728.2021.9613647","DOIUrl":"https://doi.org/10.1109/ETFA45728.2021.9613647","url":null,"abstract":"In the process industry, ongoing digitization is leading to a considerable information exchange over the entire life cycle of a process plant. Data models are usually created in the initial planning phases and these change continuously with further progress including the planning phase itself, plant adaptations after commissioning, and maintenance. The various models from the disciplines involved have a considerable semantic overlap, creating dependencies. Changes in one model inevitably lead to changes in another. This paper presents an approach for networking and synchronization between data models based on these overlaps. The resulting synchronization network provides models which are as free of inconsistencies as possible across different disciplines at any point in time. In addition, first approaches for collaborative conflict resolution are presented. Approaches from model-driven software development and web technologies are adapted to the application domain of the process industry. Typical information models created during the planning phase of a process engineering plant serve as an example.","PeriodicalId":312498,"journal":{"name":"2021 26th IEEE International Conference on Emerging Technologies and Factory Automation (ETFA )","volume":"205 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116504518","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 : 2021-09-07DOI: 10.1109/ETFA45728.2021.9613641
Jakob Danielsson, T. Seceleanu, Marcus Jägemar, M. Behnam, Mikael Sjödin
In this paper, we present a last-level cache partitioning controller for multi-core systems. Our objective is to control the Quality of Service (QoS) of applications in multi-core systems by monitoring run-time performance and continuously re-sizing cache partition sizes according to the applications' needs. We discuss two different use-cases; one that promotes application fairness and another one that prioritizes applications according to the system engineers' desired execution behavior. We display the performance drawbacks of maintaining a fair schedule for all system tasks and its performance implications for system applications. We, therefore, implement a second control algorithm that enforces cache partition assignments according to user-defined priorities rather than system fairness. Our experiments reveal that it is possible, with non-instrusive (0.3-0.7% CPU utilization) cache controlling measures, to increase performance according to setpoints and maintain the QoS for specific applications in an over-saturated system.
{"title":"Automatic Quality of Service Control in Multi-core Systems using Cache Partitioning","authors":"Jakob Danielsson, T. Seceleanu, Marcus Jägemar, M. Behnam, Mikael Sjödin","doi":"10.1109/ETFA45728.2021.9613641","DOIUrl":"https://doi.org/10.1109/ETFA45728.2021.9613641","url":null,"abstract":"In this paper, we present a last-level cache partitioning controller for multi-core systems. Our objective is to control the Quality of Service (QoS) of applications in multi-core systems by monitoring run-time performance and continuously re-sizing cache partition sizes according to the applications' needs. We discuss two different use-cases; one that promotes application fairness and another one that prioritizes applications according to the system engineers' desired execution behavior. We display the performance drawbacks of maintaining a fair schedule for all system tasks and its performance implications for system applications. We, therefore, implement a second control algorithm that enforces cache partition assignments according to user-defined priorities rather than system fairness. Our experiments reveal that it is possible, with non-instrusive (0.3-0.7% CPU utilization) cache controlling measures, to increase performance according to setpoints and maintain the QoS for specific applications in an over-saturated system.","PeriodicalId":312498,"journal":{"name":"2021 26th IEEE International Conference on Emerging Technologies and Factory Automation (ETFA )","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-09-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116920918","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：