Pub Date : 2017-08-01DOI: 10.1109/COASE.2017.8256323
S. Nagavalli, N. Chakraborty, K. Sycara
Robotic swarms are multi-robot systems whose global behaviors emerge from local interactions between individual robots. Each robot obeys a local control law that can be activated depending on an operator's choice of global swarm behavior. Real missions occur in uncontrolled environments with dynamically arising objectives and require combinations of behaviors. Given a library of swarm behaviors, a supervisory operator commanding the swarm must choose a sequence of behaviors to execute and their execution durations in order to accomplish a particular task during a mission composed of many tasks. In this paper, we address the following problem: given a library of swarm behaviors, the swarm initial state, the final goal and an unordered set of intermediate goals the operator wants to achieve, the objective is to identify a behavior schedule comprised of selected behaviors and associated time intervals of application of the behavior so that the total time to reach the final goal is minimized. Our contributions are as follows: (a) formalization of the problem of behavior scheduling to achieve multiple unordered goals with a robotic swarm, (b) an algorithm that produces a behavior schedule to achieve all intermediate goals and the final goal in minimum time such that the behavior durations are locally optimal and given which the goal sequence has bounded suboptimality and (c) application of this algorithm to configuration control of robot swarms.
{"title":"On time-optimal behavior scheduling of robotic swarms for achieving multiple goals","authors":"S. Nagavalli, N. Chakraborty, K. Sycara","doi":"10.1109/COASE.2017.8256323","DOIUrl":"https://doi.org/10.1109/COASE.2017.8256323","url":null,"abstract":"Robotic swarms are multi-robot systems whose global behaviors emerge from local interactions between individual robots. Each robot obeys a local control law that can be activated depending on an operator's choice of global swarm behavior. Real missions occur in uncontrolled environments with dynamically arising objectives and require combinations of behaviors. Given a library of swarm behaviors, a supervisory operator commanding the swarm must choose a sequence of behaviors to execute and their execution durations in order to accomplish a particular task during a mission composed of many tasks. In this paper, we address the following problem: given a library of swarm behaviors, the swarm initial state, the final goal and an unordered set of intermediate goals the operator wants to achieve, the objective is to identify a behavior schedule comprised of selected behaviors and associated time intervals of application of the behavior so that the total time to reach the final goal is minimized. Our contributions are as follows: (a) formalization of the problem of behavior scheduling to achieve multiple unordered goals with a robotic swarm, (b) an algorithm that produces a behavior schedule to achieve all intermediate goals and the final goal in minimum time such that the behavior durations are locally optimal and given which the goal sequence has bounded suboptimality and (c) application of this algorithm to configuration control of robot swarms.","PeriodicalId":445441,"journal":{"name":"2017 13th IEEE Conference on Automation Science and Engineering (CASE)","volume":"35 8","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133489329","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 : 2017-08-01DOI: 10.1109/COASE.2017.8256202
N. Lu, Xiaodong Ren, Jinbo Song, Yidan Wu
Fall detection is an important problem in the field of public health care, which is especially crucial for instant medical service delivery to the injured elderly due to falls. Ambient camera based fall detection has been a recognized non-intrusive and publicly acceptable method, where video data is employed to discriminate fall event from daily activities. Fall detection with videos usually requires a large dataset to extract features and train the classifier. However, it is hard to collect free-living environment fall data and instead simulated falls by young people have been collected to construct the training dataset, which is controlled intentional behavior and restricted to limited quantity of samples. In addition, the existing video based fall detection methods need segment the subject first, which is inclined to be influenced by image noise, illumination variation and occlusion. To address these problems, a three dimensional convolutional neural network (3D CNN) based method for fall detection is developed which only uses kinetic data to train an automatic feature extractor. Besides the spatial feature in 2D image, the motion information from the video could also be encoded by the three dimensional convolutions over the frames. A LSTM based spatial visual attention scheme is then incorporated, which could enable the network to focus on the key regions. Sports dataset Sports-1M with no fall examples is employed to train the 3D CNN and the visual attention model is trained on the small Multiple Cameras Fall Dataset. Then the visual attention based 3D CNN is employed to extract the features from the videos with fall event and implement fall detection. Experiments have shown the superior performance of the proposed scheme on fall dataset with high detection accuracy of 100%.
{"title":"Visual guided deep learning scheme for fall detection","authors":"N. Lu, Xiaodong Ren, Jinbo Song, Yidan Wu","doi":"10.1109/COASE.2017.8256202","DOIUrl":"https://doi.org/10.1109/COASE.2017.8256202","url":null,"abstract":"Fall detection is an important problem in the field of public health care, which is especially crucial for instant medical service delivery to the injured elderly due to falls. Ambient camera based fall detection has been a recognized non-intrusive and publicly acceptable method, where video data is employed to discriminate fall event from daily activities. Fall detection with videos usually requires a large dataset to extract features and train the classifier. However, it is hard to collect free-living environment fall data and instead simulated falls by young people have been collected to construct the training dataset, which is controlled intentional behavior and restricted to limited quantity of samples. In addition, the existing video based fall detection methods need segment the subject first, which is inclined to be influenced by image noise, illumination variation and occlusion. To address these problems, a three dimensional convolutional neural network (3D CNN) based method for fall detection is developed which only uses kinetic data to train an automatic feature extractor. Besides the spatial feature in 2D image, the motion information from the video could also be encoded by the three dimensional convolutions over the frames. A LSTM based spatial visual attention scheme is then incorporated, which could enable the network to focus on the key regions. Sports dataset Sports-1M with no fall examples is employed to train the 3D CNN and the visual attention model is trained on the small Multiple Cameras Fall Dataset. Then the visual attention based 3D CNN is employed to extract the features from the videos with fall event and implement fall detection. Experiments have shown the superior performance of the proposed scheme on fall dataset with high detection accuracy of 100%.","PeriodicalId":445441,"journal":{"name":"2017 13th IEEE Conference on Automation Science and Engineering (CASE)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132751011","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 : 2017-08-01DOI: 10.1109/COASE.2017.8256237
Timm Ziarnetzky, L. Mönch, Thomas Ponsignon, H. Ehm
Process improvement activities by engineering lots are crucial to stay competitive in the semiconductor market. Production and engineering lots compete for the same equipment. We discuss two production planning formulations for a simplified semiconductor supply chain. The first formulation assumes reduced available capacity for production, while the second one directly incorporates engineering activities. Additional capacity is considered in this formulation because of learning effects that represent process improvements. The integrated formulation outperforms the conventional formulation in a rolling horizon setting with respect to profit.
{"title":"Rolling horizon planning with engineering activities in semiconductor supply chains","authors":"Timm Ziarnetzky, L. Mönch, Thomas Ponsignon, H. Ehm","doi":"10.1109/COASE.2017.8256237","DOIUrl":"https://doi.org/10.1109/COASE.2017.8256237","url":null,"abstract":"Process improvement activities by engineering lots are crucial to stay competitive in the semiconductor market. Production and engineering lots compete for the same equipment. We discuss two production planning formulations for a simplified semiconductor supply chain. The first formulation assumes reduced available capacity for production, while the second one directly incorporates engineering activities. Additional capacity is considered in this formulation because of learning effects that represent process improvements. The integrated formulation outperforms the conventional formulation in a rolling horizon setting with respect to profit.","PeriodicalId":445441,"journal":{"name":"2017 13th IEEE Conference on Automation Science and Engineering (CASE)","volume":"56 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131910538","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 : 2017-08-01DOI: 10.1109/COASE.2017.8256198
Tadayuki Tone, Kenji Suzuki
This paper describes the vibration control of a ferrofluid-based robotic sheet for liquid manipulation. The robotic sheet consists of a ferrofluid wrapped with thin polytetrafluoroethylene sheets. The robotic sheet is able to move the object using the shape deformation properties of the ferrofluid owing to the influence of a magnetic field. On the basis of the previous findings regarding the flexibility of the robotic sheet, we hypothesized that we could manipulate objects whose shape changes dynamically, such as fluids. In this paper, we examine the mixing and transport of a liquid using a robotic sheet. The liquid was mixed by applying a pulsed magnetic field from an electromagnet to the robotic sheet to generate vibration. Moreover, liquid transport was realized by using the gradient of the robotic sheet surface.
{"title":"A ferrofluid-based robotic sheet for liquid manipulation by using vibration control","authors":"Tadayuki Tone, Kenji Suzuki","doi":"10.1109/COASE.2017.8256198","DOIUrl":"https://doi.org/10.1109/COASE.2017.8256198","url":null,"abstract":"This paper describes the vibration control of a ferrofluid-based robotic sheet for liquid manipulation. The robotic sheet consists of a ferrofluid wrapped with thin polytetrafluoroethylene sheets. The robotic sheet is able to move the object using the shape deformation properties of the ferrofluid owing to the influence of a magnetic field. On the basis of the previous findings regarding the flexibility of the robotic sheet, we hypothesized that we could manipulate objects whose shape changes dynamically, such as fluids. In this paper, we examine the mixing and transport of a liquid using a robotic sheet. The liquid was mixed by applying a pulsed magnetic field from an electromagnet to the robotic sheet to generate vibration. Moreover, liquid transport was realized by using the gradient of the robotic sheet surface.","PeriodicalId":445441,"journal":{"name":"2017 13th IEEE Conference on Automation Science and Engineering (CASE)","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115707094","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 : 2017-08-01DOI: 10.1109/COASE.2017.8256243
Hui Xiao
The objective of this research to develop an efficient ranking and selection (R&S) procedure for selecting the best system when its transient mean value is used as the performance measure. In this research, the true underlying mean of each system is not constant but is a function of a discrete index such as observation number or discretely sampled time. This problem is motivated by using simulation to compare the multiple configurations in order to select the configuration with best performance after certain amount of time. For example, selecting the best prototype whose performance is measured by its reliability after a certain amount of time in new product development. Another motivating example can be found in a queuing system that is initially empty and idle. Suppose that we are interested in finding the best configuration of this queuing system such that the waiting time of the 20th customer can be minimized, before the steady state of the system is reached. In this example, the underlying mean of the system is a function of observation number while the discrete index is time in the new product development example.
{"title":"Selecting the best system using transient means with sequential sampling constraints","authors":"Hui Xiao","doi":"10.1109/COASE.2017.8256243","DOIUrl":"https://doi.org/10.1109/COASE.2017.8256243","url":null,"abstract":"The objective of this research to develop an efficient ranking and selection (R&S) procedure for selecting the best system when its transient mean value is used as the performance measure. In this research, the true underlying mean of each system is not constant but is a function of a discrete index such as observation number or discretely sampled time. This problem is motivated by using simulation to compare the multiple configurations in order to select the configuration with best performance after certain amount of time. For example, selecting the best prototype whose performance is measured by its reliability after a certain amount of time in new product development. Another motivating example can be found in a queuing system that is initially empty and idle. Suppose that we are interested in finding the best configuration of this queuing system such that the waiting time of the 20th customer can be minimized, before the steady state of the system is reached. In this example, the underlying mean of the system is a function of observation number while the discrete index is time in the new product development example.","PeriodicalId":445441,"journal":{"name":"2017 13th IEEE Conference on Automation Science and Engineering (CASE)","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124276284","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 : 2017-08-01DOI: 10.1109/COASE.2017.8256183
M. Zou, B. Vogel‐Heuser
Inconsistency management is important and challenging in automated production system design (APSD), because consistency among interdisciplinary models from different stakeholders cannot always be maintained in different design phases. Hence, (semi-)automated techniques assisting domain-specific engineers to resolve arising inconsistencies are essential to increase system reliability and design efficiency. However, few efforts have been made on inconsistency resolution in APSD. This study proposes a systematic method to develop an inconsistency resolution approach from problem definition to approach formulation. The problem addressed is firstly formalized, in order to reach a clear and common understanding across disciplines. Then, the solution space is framed with the feature diagram, which draws on mostly primary features in resolving inconsistencies. The feature diagram also allows approaches in different domains to be compared. As a result, a conceptual approach based on the feature diagram and oriented by the formalized problem is put forward. It is capable to plan resolutions semi-automatically for inter-model inconsistencies in APSD.
{"title":"Feature-based systematic approach development for inconsistency resolution in automated production system design","authors":"M. Zou, B. Vogel‐Heuser","doi":"10.1109/COASE.2017.8256183","DOIUrl":"https://doi.org/10.1109/COASE.2017.8256183","url":null,"abstract":"Inconsistency management is important and challenging in automated production system design (APSD), because consistency among interdisciplinary models from different stakeholders cannot always be maintained in different design phases. Hence, (semi-)automated techniques assisting domain-specific engineers to resolve arising inconsistencies are essential to increase system reliability and design efficiency. However, few efforts have been made on inconsistency resolution in APSD. This study proposes a systematic method to develop an inconsistency resolution approach from problem definition to approach formulation. The problem addressed is firstly formalized, in order to reach a clear and common understanding across disciplines. Then, the solution space is framed with the feature diagram, which draws on mostly primary features in resolving inconsistencies. The feature diagram also allows approaches in different domains to be compared. As a result, a conceptual approach based on the feature diagram and oriented by the formalized problem is put forward. It is capable to plan resolutions semi-automatically for inter-model inconsistencies in APSD.","PeriodicalId":445441,"journal":{"name":"2017 13th IEEE Conference on Automation Science and Engineering (CASE)","volume":"38 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124570592","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 : 2017-08-01DOI: 10.1109/COASE.2017.8256240
Jens Rocholl, L. Mönch
In this paper, we discuss a multiple orders per job (MOJ) scheduling problem. Front opening unified pods (FOUPs) transfer wafers in 300-mm wafer fabs. Several orders can be grouped in one FOUP which can be considered as a job. A lot processing environment is assumed. All the orders have an unrestrictively late common due date. The earliness and tardiness of the orders is minimized. Since the problem is NP-hard, we propose two hybridized genetic algorithms. Computational experiments on randomly generated problem instances are carried out that demonstrate that the genetic algorithms perform well.
{"title":"Genetic algorithms for a single-machine multiple orders per job scheduling problem with a common due date","authors":"Jens Rocholl, L. Mönch","doi":"10.1109/COASE.2017.8256240","DOIUrl":"https://doi.org/10.1109/COASE.2017.8256240","url":null,"abstract":"In this paper, we discuss a multiple orders per job (MOJ) scheduling problem. Front opening unified pods (FOUPs) transfer wafers in 300-mm wafer fabs. Several orders can be grouped in one FOUP which can be considered as a job. A lot processing environment is assumed. All the orders have an unrestrictively late common due date. The earliness and tardiness of the orders is minimized. Since the problem is NP-hard, we propose two hybridized genetic algorithms. Computational experiments on randomly generated problem instances are carried out that demonstrate that the genetic algorithms perform well.","PeriodicalId":445441,"journal":{"name":"2017 13th IEEE Conference on Automation Science and Engineering (CASE)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114766114","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 : 2017-08-01DOI: 10.1109/COASE.2017.8256181
Yaping Zhao, Xiaoyun Xu, Haidong Li
This study considers an assemble-to-order system with periodic-reviews and a general bill of materials. The objective is to maximize the long-run average expected profit through policies including product price and component replenishment and allocation. An upper bound on the objective for all feasible policies is established through a two-stage stochastic programme (SP) which inspires the construction of a myopic policy. Particularly, the first-stage SP solution specifies the price and replenishment decisions, while the allocation decision resembles the second-stage SP recourse solution. A lower bound on the optimal policy is also provided. Numerical experiment results demonstrate that both the upper bound and the proposed policy are effective. This study brings new perspectives to supply chain management of assemble-to-order systems and suggests ways to improve profitability.
{"title":"Maximizing the long-run average expected profit of a periodic-review assemble-to-order system","authors":"Yaping Zhao, Xiaoyun Xu, Haidong Li","doi":"10.1109/COASE.2017.8256181","DOIUrl":"https://doi.org/10.1109/COASE.2017.8256181","url":null,"abstract":"This study considers an assemble-to-order system with periodic-reviews and a general bill of materials. The objective is to maximize the long-run average expected profit through policies including product price and component replenishment and allocation. An upper bound on the objective for all feasible policies is established through a two-stage stochastic programme (SP) which inspires the construction of a myopic policy. Particularly, the first-stage SP solution specifies the price and replenishment decisions, while the allocation decision resembles the second-stage SP recourse solution. A lower bound on the optimal policy is also provided. Numerical experiment results demonstrate that both the upper bound and the proposed policy are effective. This study brings new perspectives to supply chain management of assemble-to-order systems and suggests ways to improve profitability.","PeriodicalId":445441,"journal":{"name":"2017 13th IEEE Conference on Automation Science and Engineering (CASE)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114826830","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 : 2017-08-01DOI: 10.1109/COASE.2017.8256205
G. Fortino, Claudio Savaglio, Mengchu Zhou
Services are expected to represent the real drivers for all Internet of Things (IoT) application scenarios, including Industrial IoT (IIoT). Indeed, rather than focusing only on products, enterprises are now investing on product-service hybrids to improve operational efficiency, boost productivity and, most importantly, build new markets, hence diversifying their revenues. In this direction, IoT service modeling is a crucial activity requiring a multi-facet in-depth examination. In this paper, we first discuss the importance of IoT services by eliciting related benefits and challenges. Then, we tackle service modeling from both a developer-oriented (comprising highlevel descriptive metamodels and ontologies) and an enterprise-oriented (comprising operational models subject to formal verification and execution, such as extensions of the conventional Business Process Models, Petri Nets and WorkFlows) perspective, providing a conceptual mapping between these two complementary approaches. The proposed new paradigm of “Opportunistic IoT Service” extends the available models by explicitly considering essential features for service provision, i.e., dynamicity, context-awareness, co-location and transience.
{"title":"Toward opportunistic services for the industrial Internet of Things","authors":"G. Fortino, Claudio Savaglio, Mengchu Zhou","doi":"10.1109/COASE.2017.8256205","DOIUrl":"https://doi.org/10.1109/COASE.2017.8256205","url":null,"abstract":"Services are expected to represent the real drivers for all Internet of Things (IoT) application scenarios, including Industrial IoT (IIoT). Indeed, rather than focusing only on products, enterprises are now investing on product-service hybrids to improve operational efficiency, boost productivity and, most importantly, build new markets, hence diversifying their revenues. In this direction, IoT service modeling is a crucial activity requiring a multi-facet in-depth examination. In this paper, we first discuss the importance of IoT services by eliciting related benefits and challenges. Then, we tackle service modeling from both a developer-oriented (comprising highlevel descriptive metamodels and ontologies) and an enterprise-oriented (comprising operational models subject to formal verification and execution, such as extensions of the conventional Business Process Models, Petri Nets and WorkFlows) perspective, providing a conceptual mapping between these two complementary approaches. The proposed new paradigm of “Opportunistic IoT Service” extends the available models by explicitly considering essential features for service provision, i.e., dynamicity, context-awareness, co-location and transience.","PeriodicalId":445441,"journal":{"name":"2017 13th IEEE Conference on Automation Science and Engineering (CASE)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114886947","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}
Nowadays, there are more attentions on cost control and yield enhancement in the semiconductor industry. Many manufacturers have the ability to collect the physical data called Status Variables Identification (SVID) by sensors embedded in the advanced machines during the manufacturing process. To maintain the competitive advantages, process monitoring and quick response to yield problem are pivotal in detecting the cause of the faults with the help of the sensor data. To state the physical nature of certain SVID, we usually transform SVID into Fault Detection and Classification parameters (FDC parameters) using statistical indicators. The data containing FDC parameters is called FDC data. This study aims to develop a multivariate analysis model to find out the crucial factors which may lead to process excursion among a large amount of FDC data. We proposed a 2-phase multivariate analysis framework: (1) the Least Absolute Shrinkage and Selection Operator (LASSO) is applied for key operation screening. (2) And Random Forest (RF) is used to rank the FDC parameters based on the key operations. Based on the results, domain engineers can quickly take actions responding to low yield problems.
{"title":"Big data analytic for multivariate fault detection and classification in semiconductor manufacturing","authors":"Ying-Jen Chen, Bo-Cheng Wang, Jei-Zheng Wu, Yi-Chia Wu, Chen-Fu Chien","doi":"10.1109/COASE.2017.8256190","DOIUrl":"https://doi.org/10.1109/COASE.2017.8256190","url":null,"abstract":"Nowadays, there are more attentions on cost control and yield enhancement in the semiconductor industry. Many manufacturers have the ability to collect the physical data called Status Variables Identification (SVID) by sensors embedded in the advanced machines during the manufacturing process. To maintain the competitive advantages, process monitoring and quick response to yield problem are pivotal in detecting the cause of the faults with the help of the sensor data. To state the physical nature of certain SVID, we usually transform SVID into Fault Detection and Classification parameters (FDC parameters) using statistical indicators. The data containing FDC parameters is called FDC data. This study aims to develop a multivariate analysis model to find out the crucial factors which may lead to process excursion among a large amount of FDC data. We proposed a 2-phase multivariate analysis framework: (1) the Least Absolute Shrinkage and Selection Operator (LASSO) is applied for key operation screening. (2) And Random Forest (RF) is used to rank the FDC parameters based on the key operations. Based on the results, domain engineers can quickly take actions responding to low yield problems.","PeriodicalId":445441,"journal":{"name":"2017 13th IEEE Conference on Automation Science and Engineering (CASE)","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2017-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114961064","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：