Pub Date : 2020-07-01DOI: 10.23919/CCC50068.2020.9188547
Youtian Guo, Qi Gao, Feng Pan
Autonomous-Driving technology has begun to bring great convenience to daily trip, transportation, and surveying harsh environment. Considering that deep reinforcement learning has requirements for the convergence performance of the training results, and the actual training results sometimes cannot converge steadily or fail to reach the training goals, in this paper, the trained model reuse method was proposed, which can use the trained model generates Q(St, At) and can be used as a part of Deep Reinforcement Learning model, and this model was built based on the value function that could predict the Q value corresponding to the various actions performed in the environment state. In the Pygame platform, a simplified traffic simulation environment was set, it is observed that the Autonomous-Driving vehicle could run smoothly without collision in a fixed-length test simulation environment, and this trained model reuse method could help autonomous vehicle accelerate the learning process, obtain better simulation results during most of the training process, save simulation time and computing resources.
{"title":"Trained Model Reuse of Autonomous-Driving in Pygame with Deep Reinforcement Learning","authors":"Youtian Guo, Qi Gao, Feng Pan","doi":"10.23919/CCC50068.2020.9188547","DOIUrl":"https://doi.org/10.23919/CCC50068.2020.9188547","url":null,"abstract":"Autonomous-Driving technology has begun to bring great convenience to daily trip, transportation, and surveying harsh environment. Considering that deep reinforcement learning has requirements for the convergence performance of the training results, and the actual training results sometimes cannot converge steadily or fail to reach the training goals, in this paper, the trained model reuse method was proposed, which can use the trained model generates Q(St, At) and can be used as a part of Deep Reinforcement Learning model, and this model was built based on the value function that could predict the Q value corresponding to the various actions performed in the environment state. In the Pygame platform, a simplified traffic simulation environment was set, it is observed that the Autonomous-Driving vehicle could run smoothly without collision in a fixed-length test simulation environment, and this trained model reuse method could help autonomous vehicle accelerate the learning process, obtain better simulation results during most of the training process, save simulation time and computing resources.","PeriodicalId":255872,"journal":{"name":"2020 39th Chinese Control Conference (CCC)","volume":"94 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130005376","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 : 2020-07-01DOI: 10.23919/CCC50068.2020.9188420
Tong Liu, Xingyu Li, M. Fu, Zhaoxiang Liang
Regional geomagnetic maps are widely used in geomagnetic navigation and magnetic anomaly detection. However, the complexity of geomagnetic spatial trend changes and the spatial sparseness of the geomagnetic data affect the accuracy of regional geomagnetic map construction. In order to improve the accuracy of regional geomagnetic maps, this paper proposes the Support Vector Machine Residual Kriging method (SVMRKriging). First, Support Vector Machine (SVM) is used to fit the geomagnetic trend changes, then the residual component is interpolated by ordinary Kriging, and finally these two parts are added to construct a regional geomagnetic map. Experiments were performed using geomagnetic grid data and aeromagnetic data. The experiment results show that SVMRKriging method can improve the accuracy of regional geomagnetic maps with geomagnetic trend changes.
{"title":"Regional Geomagnetic Map Construction based on Support Vector Machine Residual Kriging","authors":"Tong Liu, Xingyu Li, M. Fu, Zhaoxiang Liang","doi":"10.23919/CCC50068.2020.9188420","DOIUrl":"https://doi.org/10.23919/CCC50068.2020.9188420","url":null,"abstract":"Regional geomagnetic maps are widely used in geomagnetic navigation and magnetic anomaly detection. However, the complexity of geomagnetic spatial trend changes and the spatial sparseness of the geomagnetic data affect the accuracy of regional geomagnetic map construction. In order to improve the accuracy of regional geomagnetic maps, this paper proposes the Support Vector Machine Residual Kriging method (SVMRKriging). First, Support Vector Machine (SVM) is used to fit the geomagnetic trend changes, then the residual component is interpolated by ordinary Kriging, and finally these two parts are added to construct a regional geomagnetic map. Experiments were performed using geomagnetic grid data and aeromagnetic data. The experiment results show that SVMRKriging method can improve the accuracy of regional geomagnetic maps with geomagnetic trend changes.","PeriodicalId":255872,"journal":{"name":"2020 39th Chinese Control Conference (CCC)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130098846","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 : 2020-07-01DOI: 10.23919/CCC50068.2020.9188463
Tianwei Dai, Z. Ding
The Internet of Things (IoT) extends the electronic connectivity into millions of IoT nodes in our city, which collect, share and fuse information to comprehend the status of the city. In order to achieve the autonomy to make control decisions based on the collected and analyzed information, a promising artificial intelligence method, reinforcement learning (RL), is for smart entities to leverage. In this paper, we propose a distributed learning approach using the deep RL method and consensus theories to solve the coordinated sensing coverage problem in wireless sensor and actuator networks. Also, evaluation works show the proposed algorithm emerges powerful capability, and this approach provides important operational advantages over traditional centralized and distributed approaches.
{"title":"Coordinated Sensing Coverage with Distributed Deep Reinforcement Learning","authors":"Tianwei Dai, Z. Ding","doi":"10.23919/CCC50068.2020.9188463","DOIUrl":"https://doi.org/10.23919/CCC50068.2020.9188463","url":null,"abstract":"The Internet of Things (IoT) extends the electronic connectivity into millions of IoT nodes in our city, which collect, share and fuse information to comprehend the status of the city. In order to achieve the autonomy to make control decisions based on the collected and analyzed information, a promising artificial intelligence method, reinforcement learning (RL), is for smart entities to leverage. In this paper, we propose a distributed learning approach using the deep RL method and consensus theories to solve the coordinated sensing coverage problem in wireless sensor and actuator networks. Also, evaluation works show the proposed algorithm emerges powerful capability, and this approach provides important operational advantages over traditional centralized and distributed approaches.","PeriodicalId":255872,"journal":{"name":"2020 39th Chinese Control Conference (CCC)","volume":"70 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130121445","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 : 2020-07-01DOI: 10.23919/CCC50068.2020.9188705
Lusheng Yao
In this paper, the problem of finite-time stability of linear systems with single delay is considered. A set of conditions equivalent to the definition are derived. These conditions are in the form of continuous multivariate eigenvalue problem or Karush–Kuhn–Tucker conditions. By these conditions, finite-time stability of linear time delay system can be checked numerically. A numerical example is given to illustrate the potentialities of these conditions.
{"title":"Why Finite-Time Stability is So Special: Operator Norm and Multivariate Eigenvalue Problem Behind the Curtain","authors":"Lusheng Yao","doi":"10.23919/CCC50068.2020.9188705","DOIUrl":"https://doi.org/10.23919/CCC50068.2020.9188705","url":null,"abstract":"In this paper, the problem of finite-time stability of linear systems with single delay is considered. A set of conditions equivalent to the definition are derived. These conditions are in the form of continuous multivariate eigenvalue problem or Karush–Kuhn–Tucker conditions. By these conditions, finite-time stability of linear time delay system can be checked numerically. A numerical example is given to illustrate the potentialities of these conditions.","PeriodicalId":255872,"journal":{"name":"2020 39th Chinese Control Conference (CCC)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130194295","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 : 2020-07-01DOI: 10.23919/CCC50068.2020.9188773
Chao Ma, Wei Wu, Yidao Ji, Hang Fu
This paper investigates the formation problem of autonomous underwater vehicles with unreliable switching communication topologies and time-varying transmission delays. More precisely, the switching topologies contains achievable and unachievable sub-topologies, which can better describe the practical underwater communication environment. By performing model transformation and constructing appropriate multiple Lyapunov-function method, sufficient conditions are established based on admissible edge-dependent average dwell time, such that the desired formation configuration can be achieved with transmission delays. Finally, an illustrative example is given at last to verify the effectiveness of the main results.
{"title":"Distributed Formation of Autonomous Underwater Vehicles with Unreliable Switching Topologies and Transmission Delays","authors":"Chao Ma, Wei Wu, Yidao Ji, Hang Fu","doi":"10.23919/CCC50068.2020.9188773","DOIUrl":"https://doi.org/10.23919/CCC50068.2020.9188773","url":null,"abstract":"This paper investigates the formation problem of autonomous underwater vehicles with unreliable switching communication topologies and time-varying transmission delays. More precisely, the switching topologies contains achievable and unachievable sub-topologies, which can better describe the practical underwater communication environment. By performing model transformation and constructing appropriate multiple Lyapunov-function method, sufficient conditions are established based on admissible edge-dependent average dwell time, such that the desired formation configuration can be achieved with transmission delays. Finally, an illustrative example is given at last to verify the effectiveness of the main results.","PeriodicalId":255872,"journal":{"name":"2020 39th Chinese Control Conference (CCC)","volume":"20 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134139241","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 : 2020-07-01DOI: 10.23919/CCC50068.2020.9188923
Zhaochen Lin, Xiyang Liu, Yinbao Niu, Ning Hao, Fenghua He
Inthis paper, a quadrotor’s tracking control problem is investigated in which the feedback loop has a time delay caused by transmission. First, the relative motion model is established. Then, an improved Linear Active Disturbance Rejection Controller(LADRC) algorithm is proposed to deal with the time delay in feedback. Finally, simulation and experimental results show the effectiveness of the proposed algorithm.
{"title":"An Improved LADRC Algorithm for Quadrotors","authors":"Zhaochen Lin, Xiyang Liu, Yinbao Niu, Ning Hao, Fenghua He","doi":"10.23919/CCC50068.2020.9188923","DOIUrl":"https://doi.org/10.23919/CCC50068.2020.9188923","url":null,"abstract":"Inthis paper, a quadrotor’s tracking control problem is investigated in which the feedback loop has a time delay caused by transmission. First, the relative motion model is established. Then, an improved Linear Active Disturbance Rejection Controller(LADRC) algorithm is proposed to deal with the time delay in feedback. Finally, simulation and experimental results show the effectiveness of the proposed algorithm.","PeriodicalId":255872,"journal":{"name":"2020 39th Chinese Control Conference (CCC)","volume":"54 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134192548","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 : 2020-07-01DOI: 10.23919/CCC50068.2020.9189564
Xiaoxing Ren, Dewei Li, Y. Xi, Lulu Pan, Haibin Shao
We consider the distributed optimization problem on signed networks. Each agent has a local function which depends on a subset of the components of the variable and is subject to a local constraint set. A primal-dual algorithm with fixed step size is proposed. The algorithm ensures that the agents' estimates converge to a subset of the components of an optimal solution or its opposite. Note that each component of the variable is allowed to be associated with more than one agents, our algorithm guarantees that those coupled agents achieve bipartite consensus on estimates for the intersection components. Numerical results are provided to demonstrate the theoretical analysis.
{"title":"Primal-dual algorithm for distributed optimization with local domains on signed networks","authors":"Xiaoxing Ren, Dewei Li, Y. Xi, Lulu Pan, Haibin Shao","doi":"10.23919/CCC50068.2020.9189564","DOIUrl":"https://doi.org/10.23919/CCC50068.2020.9189564","url":null,"abstract":"We consider the distributed optimization problem on signed networks. Each agent has a local function which depends on a subset of the components of the variable and is subject to a local constraint set. A primal-dual algorithm with fixed step size is proposed. The algorithm ensures that the agents' estimates converge to a subset of the components of an optimal solution or its opposite. Note that each component of the variable is allowed to be associated with more than one agents, our algorithm guarantees that those coupled agents achieve bipartite consensus on estimates for the intersection components. Numerical results are provided to demonstrate the theoretical analysis.","PeriodicalId":255872,"journal":{"name":"2020 39th Chinese Control Conference (CCC)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134514522","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 : 2020-07-01DOI: 10.23919/CCC50068.2020.9188698
Xiuqin Fang, Han Liu, Guo Xie, Youmin Zhang, Ding Liu
In this paper a method based on the combination of product quantization and pruning to compress deep neural network with large size model and great amount of calculation is proposed. First of all, we use pruning to reduce redundant parameters in deep neural network, and then refine the tune network for fine tuning. Then we use product quantization to quantize the parameters of the neural network to 8 bits, which reduces the storage overhead so that the deep neural network can be deployed in embedded devices. For the classification tasks in the Mnist dataset and Cifar10 dataset, the network models such as LeNet5, AlexNet, ResNet are compressed to 23 to 38 times without losing accuracy as much as possible.
{"title":"Deep Neural Network Compression Method Based on Product Quantization","authors":"Xiuqin Fang, Han Liu, Guo Xie, Youmin Zhang, Ding Liu","doi":"10.23919/CCC50068.2020.9188698","DOIUrl":"https://doi.org/10.23919/CCC50068.2020.9188698","url":null,"abstract":"In this paper a method based on the combination of product quantization and pruning to compress deep neural network with large size model and great amount of calculation is proposed. First of all, we use pruning to reduce redundant parameters in deep neural network, and then refine the tune network for fine tuning. Then we use product quantization to quantize the parameters of the neural network to 8 bits, which reduces the storage overhead so that the deep neural network can be deployed in embedded devices. For the classification tasks in the Mnist dataset and Cifar10 dataset, the network models such as LeNet5, AlexNet, ResNet are compressed to 23 to 38 times without losing accuracy as much as possible.","PeriodicalId":255872,"journal":{"name":"2020 39th Chinese Control Conference (CCC)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134560124","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 : 2020-07-01DOI: 10.23919/CCC50068.2020.9189249
Xu Lubing, Ye Yangfei, Yi Yang
In this paper, a novel DOB identification and control algorithm is developed for hypersonic flight vehicle. The DOBC technique is combined with the DNNs Models so as to identify attitude and velocity in the presence of disturbance. By using an adaptive method to adjust the weight matrices and compensate the unknown parameters, the control input is built with the Nussbaum gain matrix and feedback control gain. The stability proof is provided by using Lyapunov method. Finally, the simulation results shows DOBC technique is combined with DNNs models can obtained satisfactory dynamical identification and anti-disturbance performance.
{"title":"DOB Identification and Anti-Disturbance Control for Hypersonic Flight Vehicle Systems","authors":"Xu Lubing, Ye Yangfei, Yi Yang","doi":"10.23919/CCC50068.2020.9189249","DOIUrl":"https://doi.org/10.23919/CCC50068.2020.9189249","url":null,"abstract":"In this paper, a novel DOB identification and control algorithm is developed for hypersonic flight vehicle. The DOBC technique is combined with the DNNs Models so as to identify attitude and velocity in the presence of disturbance. By using an adaptive method to adjust the weight matrices and compensate the unknown parameters, the control input is built with the Nussbaum gain matrix and feedback control gain. The stability proof is provided by using Lyapunov method. Finally, the simulation results shows DOBC technique is combined with DNNs models can obtained satisfactory dynamical identification and anti-disturbance performance.","PeriodicalId":255872,"journal":{"name":"2020 39th Chinese Control Conference (CCC)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134569830","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 : 2020-07-01DOI: 10.23919/CCC50068.2020.9189568
Jiaxing Wang, Dazhi Wang, Xinghua Wang
The performance of the industrial robot servo system (IRSS) depends on two factors, one is the control algorithm and mechanical processing accuracy during system design, and the other is maintenance during system operation. Based on the strategy of condition-based maintenance, the long-term stable high-performance operation of the industrial robot servo system can be maintained. In order to improve the performance of the servo system through the predictive maintenance of industrial robots, we need to monitor the operating state of the equipment during its operation and use intelligent algorithms to identify the operating state. The fault diagnosis of industrial robots represented by bearing fault diagnosis plays a crucial role in the optimization of IRSS. In the early stages of faults, online and accurate diagnosis can achieve predictive maintenance and improve the performance of IRSS. In this paper, a new multi-sensor information fusion technology is proposed, which uses the signals of multiple sensors as the input of a one-dimensional (1D) convolutional neural network (CNN), and implements a fault classification method through an improved CNN. This method is verified on the public data set of Case Western Reserve University and the IMS bearing database of the University of Cincinnati. Compared with the traditional 1D or 2D CNN and other fault classification methods, the model is simplified and can be used more Less data and simpler calculation complexity achieve higher fault classification accuracy.
{"title":"Fault Diagnosis of Industrial Robots Based on Multi-sensor Information Fusion and 1D Convolutional Neural Network","authors":"Jiaxing Wang, Dazhi Wang, Xinghua Wang","doi":"10.23919/CCC50068.2020.9189568","DOIUrl":"https://doi.org/10.23919/CCC50068.2020.9189568","url":null,"abstract":"The performance of the industrial robot servo system (IRSS) depends on two factors, one is the control algorithm and mechanical processing accuracy during system design, and the other is maintenance during system operation. Based on the strategy of condition-based maintenance, the long-term stable high-performance operation of the industrial robot servo system can be maintained. In order to improve the performance of the servo system through the predictive maintenance of industrial robots, we need to monitor the operating state of the equipment during its operation and use intelligent algorithms to identify the operating state. The fault diagnosis of industrial robots represented by bearing fault diagnosis plays a crucial role in the optimization of IRSS. In the early stages of faults, online and accurate diagnosis can achieve predictive maintenance and improve the performance of IRSS. In this paper, a new multi-sensor information fusion technology is proposed, which uses the signals of multiple sensors as the input of a one-dimensional (1D) convolutional neural network (CNN), and implements a fault classification method through an improved CNN. This method is verified on the public data set of Case Western Reserve University and the IMS bearing database of the University of Cincinnati. Compared with the traditional 1D or 2D CNN and other fault classification methods, the model is simplified and can be used more Less data and simpler calculation complexity achieve higher fault classification accuracy.","PeriodicalId":255872,"journal":{"name":"2020 39th Chinese Control Conference (CCC)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133907993","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：