Pub Date : 2020-08-01DOI: 10.1109/CASE48305.2020.9216876
Zhixuan Jia, Shuangxi Huang, Yushun Fan
With the further development of service Internet, how to effectively manage, control and optimize the entire service Internet system has become a research hotspot. However, the current research on service Internet optimization mostly focuses on service selection and service composition. And few studies have provided the global optimization strategy for service Internet. So, in this paper, first of all, we propose an ecological model of service Internet by using ecological theory and business process characteristics. Then, based on this model, considering the time-varying uncertainty of service demands, a global service Internet optimization policy learning algorithm called GSIOA is proposed, which uses deep reinforcement learning to automatically formulate global management, control, and optimization strategies of service Internet. Finally, through the comparison of simulation results, it can be concluded that the performance of GSIOA is better than other baseline methods. Our algorithm based on deep reinforcement learning can well solve the global optimization problem of service Internet.
{"title":"The Ecological Model and Global optimization Algorithm for Service Internet","authors":"Zhixuan Jia, Shuangxi Huang, Yushun Fan","doi":"10.1109/CASE48305.2020.9216876","DOIUrl":"https://doi.org/10.1109/CASE48305.2020.9216876","url":null,"abstract":"With the further development of service Internet, how to effectively manage, control and optimize the entire service Internet system has become a research hotspot. However, the current research on service Internet optimization mostly focuses on service selection and service composition. And few studies have provided the global optimization strategy for service Internet. So, in this paper, first of all, we propose an ecological model of service Internet by using ecological theory and business process characteristics. Then, based on this model, considering the time-varying uncertainty of service demands, a global service Internet optimization policy learning algorithm called GSIOA is proposed, which uses deep reinforcement learning to automatically formulate global management, control, and optimization strategies of service Internet. Finally, through the comparison of simulation results, it can be concluded that the performance of GSIOA is better than other baseline methods. Our algorithm based on deep reinforcement learning can well solve the global optimization problem of service Internet.","PeriodicalId":212181,"journal":{"name":"2020 IEEE 16th International Conference on Automation Science and Engineering (CASE)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132718924","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 : 2020-08-01DOI: 10.1109/CASE48305.2020.9216878
Ahmad Hajeer, Lei Chen, E. Hu
A new methodology that mixes the already well-established classifications of Wall Climbing Robots (WCR) based on locomotion and attachment is proposed. The need for industrial inspections of processing assets, with vertical flat surfaces, like tanks and pressure vessels that are continuously exposed to fatigue, corrosion and erosion are identified. The types of industrial inspections tasks vary in a wide range from, Visual Inspections (VI) and Non-Destructive Testing (NDT) to cleaning and repair. Mobile robots are used to perform inspection tasks. There are few commercial WCR that are available for industrial applications. The classification method is based on locomotion mechanisms, and attachment techniques rather than classical classification assumed in all previous reviews of WCR. A general design framework is suggested to derive the custom-building design of the inspection WCR. The performance of the WCR is compared against the main features of maximum speed and payload capacity. The main locomotion types are wheeled (W), tracked (T), walking (L) and hybrid (H). The primary attachment methods are magnetic (G), pneumatic (P), mechanical (M), electrostatic (E), chemical (C) and hybrid (H). As another application, the classification is used to identify and differentiate WCR from other types of Climbing Robots (CR) to cover almost all kinds of WCR. In addition, the classification can be employed in the customisation of WCR that are built to perform specialised inspection tasks. The customisation of WCR results from the need for simplicity in their structure, adapting testing equipment and improving operations.
{"title":"Review of Classification for Wall Climbing Robots for Industrial Inspection Applications","authors":"Ahmad Hajeer, Lei Chen, E. Hu","doi":"10.1109/CASE48305.2020.9216878","DOIUrl":"https://doi.org/10.1109/CASE48305.2020.9216878","url":null,"abstract":"A new methodology that mixes the already well-established classifications of Wall Climbing Robots (WCR) based on locomotion and attachment is proposed. The need for industrial inspections of processing assets, with vertical flat surfaces, like tanks and pressure vessels that are continuously exposed to fatigue, corrosion and erosion are identified. The types of industrial inspections tasks vary in a wide range from, Visual Inspections (VI) and Non-Destructive Testing (NDT) to cleaning and repair. Mobile robots are used to perform inspection tasks. There are few commercial WCR that are available for industrial applications. The classification method is based on locomotion mechanisms, and attachment techniques rather than classical classification assumed in all previous reviews of WCR. A general design framework is suggested to derive the custom-building design of the inspection WCR. The performance of the WCR is compared against the main features of maximum speed and payload capacity. The main locomotion types are wheeled (W), tracked (T), walking (L) and hybrid (H). The primary attachment methods are magnetic (G), pneumatic (P), mechanical (M), electrostatic (E), chemical (C) and hybrid (H). As another application, the classification is used to identify and differentiate WCR from other types of Climbing Robots (CR) to cover almost all kinds of WCR. In addition, the classification can be employed in the customisation of WCR that are built to perform specialised inspection tasks. The customisation of WCR results from the need for simplicity in their structure, adapting testing equipment and improving operations.","PeriodicalId":212181,"journal":{"name":"2020 IEEE 16th International Conference on Automation Science and Engineering (CASE)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133191522","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}
Industrial Internet of Things (IIoT) is regarded as a promising transformation technology for manufacturing industry. We expect the information flow through along the process of material flow, gauging and collective devices feed IIoT platform with essential data. AI and big data analytical techniques can then be used to process the data to support a series of decision-making. Users at both managerial and operational levels can take actions accordingly, which significantly improve management efficiency in one organization, enhance supply chain coordination, and attract investment, etc. Powerful as bigdata and AI technology, they cannot operate without data. In centralized, third-party-managed IIoT platform, few manufactures are willing to share their critical data. This study extends IIoT framework with blockchain technology to provide a solution for trustless data sharing. The solution embeds encrypted ledger to avoid credential data tampering, employs m-in-n-out smart contracts to secure data exchange, and utilizes consensus mechanism to improve cyber security. The proposed framework integrates blockchain with IIoT to operate the entire industrial data exchange in a decentralized network, namely Industrial Blockchain of Things (IBoT).
{"title":"Industrial Blockchain of Things: A Solution for Trustless Industrial Data Sharing and Beyond","authors":"Zheng Zhang, Liang Huang, Renzhong Tang, T. Peng, Lihang Guo, Xingwei Xiang","doi":"10.1109/CASE48305.2020.9216817","DOIUrl":"https://doi.org/10.1109/CASE48305.2020.9216817","url":null,"abstract":"Industrial Internet of Things (IIoT) is regarded as a promising transformation technology for manufacturing industry. We expect the information flow through along the process of material flow, gauging and collective devices feed IIoT platform with essential data. AI and big data analytical techniques can then be used to process the data to support a series of decision-making. Users at both managerial and operational levels can take actions accordingly, which significantly improve management efficiency in one organization, enhance supply chain coordination, and attract investment, etc. Powerful as bigdata and AI technology, they cannot operate without data. In centralized, third-party-managed IIoT platform, few manufactures are willing to share their critical data. This study extends IIoT framework with blockchain technology to provide a solution for trustless data sharing. The solution embeds encrypted ledger to avoid credential data tampering, employs m-in-n-out smart contracts to secure data exchange, and utilizes consensus mechanism to improve cyber security. The proposed framework integrates blockchain with IIoT to operate the entire industrial data exchange in a decentralized network, namely Industrial Blockchain of Things (IBoT).","PeriodicalId":212181,"journal":{"name":"2020 IEEE 16th International Conference on Automation Science and Engineering (CASE)","volume":"62 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114392894","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 : 2020-08-01DOI: 10.1109/CASE48305.2020.9217027
Takayuki Murooka, A. I. Károly, Felix von Drigalski, Yoshihisa Ijiri
Motion planning is an essential component of robotic systems. Gradient-based planning has been proposed to produce smooth paths under constraints for manipulation tasks such as picking and placing objects. However, it does not deal well with discontinuities, which occur in many manipulation problems, e.g. when deciding whether to pick an object from the side or from the top. Sampling-based planning is robust against such discontinuities, but often produces paths with unnecessary motions that require heavy post-processing. In this paper, we propose a novel method to solve a complete manipulation task using gradient-based optimization while preserving the advantages of sampling-based planning. By dividing the surface of the object into regions where the grasp pose can be extracted quasi-continuously, we define multiple optimization problems in parallel, which are evaluated independently. As our method generates the motion path and grasp plan for the entire task, constraints that arise from each moment of the task are propagated automatically to the optimization of the entire task, facilitating the setup. We show the effectiveness of the proposed method in simulation and with a real robot.
{"title":"Simultaneous Planning of Grasp and Motion using Sample Regions and Gradient-Based Optimization","authors":"Takayuki Murooka, A. I. Károly, Felix von Drigalski, Yoshihisa Ijiri","doi":"10.1109/CASE48305.2020.9217027","DOIUrl":"https://doi.org/10.1109/CASE48305.2020.9217027","url":null,"abstract":"Motion planning is an essential component of robotic systems. Gradient-based planning has been proposed to produce smooth paths under constraints for manipulation tasks such as picking and placing objects. However, it does not deal well with discontinuities, which occur in many manipulation problems, e.g. when deciding whether to pick an object from the side or from the top. Sampling-based planning is robust against such discontinuities, but often produces paths with unnecessary motions that require heavy post-processing. In this paper, we propose a novel method to solve a complete manipulation task using gradient-based optimization while preserving the advantages of sampling-based planning. By dividing the surface of the object into regions where the grasp pose can be extracted quasi-continuously, we define multiple optimization problems in parallel, which are evaluated independently. As our method generates the motion path and grasp plan for the entire task, constraints that arise from each moment of the task are propagated automatically to the optimization of the entire task, facilitating the setup. We show the effectiveness of the proposed method in simulation and with a real robot.","PeriodicalId":212181,"journal":{"name":"2020 IEEE 16th International Conference on Automation Science and Engineering (CASE)","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114806718","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 : 2020-08-01DOI: 10.1109/CASE48305.2020.9216783
Xuedong Zhu, Weihao Wang, Xinxing Guo, Leyuan Shi
This paper investigates a more general flexible job shop scheduling problem with multiple process plans which is common in the modern manufacturing system. As an extension of the traditional flexible job shop scheduling problem, various realistic flexibility such as processing flexibility, machine flexibility and sequencing flexibility are considered in this problem. Due to the high complexity and the real-time requirement of this problem, a genetic programming-based evolutionary approach is proposed to automatically generate effective dispatching rules for this problem, and an evaluation method is developed to evaluate the generated dispatching rules. Three experiments are conducted to evaluate the performance of the proposed approach for real cases with large-scale test problems. Numerical results show that the proposed approach outperforms the classical dispatching rules and the state-of-theart algorithms, and is able to provide higher-quality solutions with less computational time.
{"title":"A Genetic Programming-Based Evolutionary Approach for Flexible Job Shop Scheduling with Multiple Process Plans","authors":"Xuedong Zhu, Weihao Wang, Xinxing Guo, Leyuan Shi","doi":"10.1109/CASE48305.2020.9216783","DOIUrl":"https://doi.org/10.1109/CASE48305.2020.9216783","url":null,"abstract":"This paper investigates a more general flexible job shop scheduling problem with multiple process plans which is common in the modern manufacturing system. As an extension of the traditional flexible job shop scheduling problem, various realistic flexibility such as processing flexibility, machine flexibility and sequencing flexibility are considered in this problem. Due to the high complexity and the real-time requirement of this problem, a genetic programming-based evolutionary approach is proposed to automatically generate effective dispatching rules for this problem, and an evaluation method is developed to evaluate the generated dispatching rules. Three experiments are conducted to evaluate the performance of the proposed approach for real cases with large-scale test problems. Numerical results show that the proposed approach outperforms the classical dispatching rules and the state-of-theart algorithms, and is able to provide higher-quality solutions with less computational time.","PeriodicalId":212181,"journal":{"name":"2020 IEEE 16th International Conference on Automation Science and Engineering (CASE)","volume":"133 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114668646","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 : 2020-08-01DOI: 10.1109/CASE48305.2020.9217038
Reem Muhammed Al-Gaifi, Veit Müller, N. Elkmann
This paper demonstrates the effectiveness of using high-resolution tactile sensors for a two-finger gripper for reactive grasping under conditions featuring uncertainty, whereas uncertainty will be understood as positioning inaccuracy induced by grasping planners and vision systems. Reactive grasping is defined here as the adoption of gripper pose in accordance to measured tactile sensor feedback. We propose a reactive grasping algorithm that in combination with high-resolution tactile sensors achieves grasping of unknown and novel objects with only one correction move. Extensive tests of the algorithm in different configurations showed a success rate of 91 % compared to 31 % without it.
{"title":"Reactive grasping using high-resolution tactile sensors","authors":"Reem Muhammed Al-Gaifi, Veit Müller, N. Elkmann","doi":"10.1109/CASE48305.2020.9217038","DOIUrl":"https://doi.org/10.1109/CASE48305.2020.9217038","url":null,"abstract":"This paper demonstrates the effectiveness of using high-resolution tactile sensors for a two-finger gripper for reactive grasping under conditions featuring uncertainty, whereas uncertainty will be understood as positioning inaccuracy induced by grasping planners and vision systems. Reactive grasping is defined here as the adoption of gripper pose in accordance to measured tactile sensor feedback. We propose a reactive grasping algorithm that in combination with high-resolution tactile sensors achieves grasping of unknown and novel objects with only one correction move. Extensive tests of the algorithm in different configurations showed a success rate of 91 % compared to 31 % without it.","PeriodicalId":212181,"journal":{"name":"2020 IEEE 16th International Conference on Automation Science and Engineering (CASE)","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121969713","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 : 2020-08-01DOI: 10.1109/CASE48305.2020.9216846
Canrong Zhang, Hao Guan
The container relocation problem is one of important issues in seaport terminals which could bring a significant saving on the operating cost even with a slight improvement due to the huge number of containers processed across the world each year. Given a specific layout and container retrieval priorities, the container relocation problem aims to find the optimal movement sequence to minimize the total number of container relocation operations. In this paper, we propose a new upper bound method called MLUB that incorporates branch pruners. These pruners are derived from some machine learning techniques through using the optimal solution values of many small-scale instances. The tightened upper bounds generated by MLUB are used subsequently in the exact branchand-bound algorithm called IB&B. Moreover, we also provide a tighter lower bound for the problem by additionally considering the interaction between consecutive target containers. Based on the benchmark data published recently in the literature, extensive experiments are conducted to test the performance of the proposed algorithms.
{"title":"A data-driven exact algorithm for the container relocation problem","authors":"Canrong Zhang, Hao Guan","doi":"10.1109/CASE48305.2020.9216846","DOIUrl":"https://doi.org/10.1109/CASE48305.2020.9216846","url":null,"abstract":"The container relocation problem is one of important issues in seaport terminals which could bring a significant saving on the operating cost even with a slight improvement due to the huge number of containers processed across the world each year. Given a specific layout and container retrieval priorities, the container relocation problem aims to find the optimal movement sequence to minimize the total number of container relocation operations. In this paper, we propose a new upper bound method called MLUB that incorporates branch pruners. These pruners are derived from some machine learning techniques through using the optimal solution values of many small-scale instances. The tightened upper bounds generated by MLUB are used subsequently in the exact branchand-bound algorithm called IB&B. Moreover, we also provide a tighter lower bound for the problem by additionally considering the interaction between consecutive target containers. Based on the benchmark data published recently in the literature, extensive experiments are conducted to test the performance of the proposed algorithms.","PeriodicalId":212181,"journal":{"name":"2020 IEEE 16th International Conference on Automation Science and Engineering (CASE)","volume":"16 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124756330","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 : 2020-08-01DOI: 10.1109/CASE48305.2020.9216746
Fredy Raptopoulos, Maria Koskinopoulou, M. Maniadakis
Incorporating robots into industrial settings is not a new concept, but their use in the waste recycling industry is critical. Recently AI-assisted robots are used to support waste sorting and improve the quantity and quality of recovered materials. This article aims to study and apply a new transfer paradigm for recyclable sorting using Delta robots, which is based on replacing the usual Pick-and-Place process with the much faster Pick-and-Toss process. Current robotic sorting systems can sort one item per second, Pick-and-Toss intends to significantly advance this score. We quantitatively and qualitatively assess the tossing approach by comparing it to Pick-and-Place, in terms of accuracy and robustness, both in simulation and on a real waste sorting lab-setup equipped with an ABB-IRB360 Delta robot. Overall, the Pick-and-Toss approach proves to be a powerful mechanism that succeeds faster sorting of waste streams in comparison to the standard Pick-and-Place procedure.
{"title":"Robotic Pick-and-Toss Facilitates Urban Waste Sorting *","authors":"Fredy Raptopoulos, Maria Koskinopoulou, M. Maniadakis","doi":"10.1109/CASE48305.2020.9216746","DOIUrl":"https://doi.org/10.1109/CASE48305.2020.9216746","url":null,"abstract":"Incorporating robots into industrial settings is not a new concept, but their use in the waste recycling industry is critical. Recently AI-assisted robots are used to support waste sorting and improve the quantity and quality of recovered materials. This article aims to study and apply a new transfer paradigm for recyclable sorting using Delta robots, which is based on replacing the usual Pick-and-Place process with the much faster Pick-and-Toss process. Current robotic sorting systems can sort one item per second, Pick-and-Toss intends to significantly advance this score. We quantitatively and qualitatively assess the tossing approach by comparing it to Pick-and-Place, in terms of accuracy and robustness, both in simulation and on a real waste sorting lab-setup equipped with an ABB-IRB360 Delta robot. Overall, the Pick-and-Toss approach proves to be a powerful mechanism that succeeds faster sorting of waste streams in comparison to the standard Pick-and-Place procedure.","PeriodicalId":212181,"journal":{"name":"2020 IEEE 16th International Conference on Automation Science and Engineering (CASE)","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129064087","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 : 2020-08-01DOI: 10.1109/CASE48305.2020.9216786
M. Lieret, Jonas Fertsch, J. Franke
Within the last years unmanned aerial vehicles (UAV) have found their way into several industrial applications and are currently considered to be a highly efficient solution to deliver parcels to end costumers. However, a power shutdown or emergency stop of the propulsion system in the case of a failure is compared to stationary or mobile ground robots no suitable reaction as it leads to a fall of the UAV and hence creates a safety hazard for persons in the surroundings. Thus it must be ensured that the UAV is not only capable of detecting occurring failures but also of remaining airworthy and navigable at any time.Within this work we therefore present a fault detection approach for autonomous multirotor systems using a redundant flight control architecture. The presented fault detection methodology is capable of analysing flight control units (FCU) based on different hard- and software and does not require a precise time synchronization. To deal with different and unsynchronized FCUs we design and implement an inexact voter to continuously compare states and functionalities of the individual FCUs. In the case of a failure or break down of the active FCU the developed architecture has the ability of switching to one of the fallback systems.The approach and algorithms used for failure detection are presented in detail and afterwards the overall system is evaluated using a hexarotor in different flight states and conditions.
{"title":"Fault detection for autonomous multirotors using a redundant flight control architecture","authors":"M. Lieret, Jonas Fertsch, J. Franke","doi":"10.1109/CASE48305.2020.9216786","DOIUrl":"https://doi.org/10.1109/CASE48305.2020.9216786","url":null,"abstract":"Within the last years unmanned aerial vehicles (UAV) have found their way into several industrial applications and are currently considered to be a highly efficient solution to deliver parcels to end costumers. However, a power shutdown or emergency stop of the propulsion system in the case of a failure is compared to stationary or mobile ground robots no suitable reaction as it leads to a fall of the UAV and hence creates a safety hazard for persons in the surroundings. Thus it must be ensured that the UAV is not only capable of detecting occurring failures but also of remaining airworthy and navigable at any time.Within this work we therefore present a fault detection approach for autonomous multirotor systems using a redundant flight control architecture. The presented fault detection methodology is capable of analysing flight control units (FCU) based on different hard- and software and does not require a precise time synchronization. To deal with different and unsynchronized FCUs we design and implement an inexact voter to continuously compare states and functionalities of the individual FCUs. In the case of a failure or break down of the active FCU the developed architecture has the ability of switching to one of the fallback systems.The approach and algorithms used for failure detection are presented in detail and afterwards the overall system is evaluated using a hexarotor in different flight states and conditions.","PeriodicalId":212181,"journal":{"name":"2020 IEEE 16th International Conference on Automation Science and Engineering (CASE)","volume":"T156 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125658902","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 : 2020-08-01DOI: 10.1109/CASE48305.2020.9216981
Yifan Zhao, Gangyan Xu, Yali Du, Meng Fang
Adaptive traffic signal control is widely recognized as an effective solution to improve urban mobility and reduce congestion in metropolises. Recently, reinforcement learning has been adopted for this transportation problem. While centralized reinforcement learning inevitably faces action space explosion, decentralized reinforcement learning allows agents to develop policies based on local observations but suffers from unstable training. In this paper, we present CommNetPF, a multi-agent decentralized reinforcement learning model incorporating communication and neighbourhood policy fingerprints for adaptive traffic signal control. With policy fingerprints in communication, agents learn to produce cooperative policies and the model converges faster. Experiments in scenarios of adaptive traffic signal control show that CommNetPF outperforms several strong baselines in terms of control performance and convergence speed.
{"title":"Learning Multi-Agent Communication with Policy Fingerprints for Adaptive Traffic Signal Control","authors":"Yifan Zhao, Gangyan Xu, Yali Du, Meng Fang","doi":"10.1109/CASE48305.2020.9216981","DOIUrl":"https://doi.org/10.1109/CASE48305.2020.9216981","url":null,"abstract":"Adaptive traffic signal control is widely recognized as an effective solution to improve urban mobility and reduce congestion in metropolises. Recently, reinforcement learning has been adopted for this transportation problem. While centralized reinforcement learning inevitably faces action space explosion, decentralized reinforcement learning allows agents to develop policies based on local observations but suffers from unstable training. In this paper, we present CommNetPF, a multi-agent decentralized reinforcement learning model incorporating communication and neighbourhood policy fingerprints for adaptive traffic signal control. With policy fingerprints in communication, agents learn to produce cooperative policies and the model converges faster. Experiments in scenarios of adaptive traffic signal control show that CommNetPF outperforms several strong baselines in terms of control performance and convergence speed.","PeriodicalId":212181,"journal":{"name":"2020 IEEE 16th International Conference on Automation Science and Engineering (CASE)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130647144","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：