Pub Date : 2021-11-25DOI: 10.1109/anzcc53563.2021.9628387
W. Pooja, N. Snehal, K. Sonam, S. Wagh, Navdeep M. Singh
Machine learning is widely being used in medical field for disease diagnostics and research.The area of machine learning is mainly classified into 3 parts: supervised, unsupervised and reinforcement learning.Supervised machine learning (ML) algorithms are used in this paper for modeling and showing the impact of increased testing on the number of daily confirmed cases of COVID-19. The algorithms used to carry out this study are decision tree regression and random forest regression. Machine learning for modeling has proven to be significant for forecasting and hence decision making over the future course of actions. In this paper, Gaussian process regression has been used for modeling as well as forecasting the daily confirmed cases in South Korea. The results obtained show that if the number of tests conducted is increased to the population of South Korea, approximately equal to 51, 286, 183, the peak in the daily cases is obtained earlier and hence the overall number of daily cases is less compared to current cases.
{"title":"Effect of increased number of COVID-19 tests using supervised machine learning models","authors":"W. Pooja, N. Snehal, K. Sonam, S. Wagh, Navdeep M. Singh","doi":"10.1109/anzcc53563.2021.9628387","DOIUrl":"https://doi.org/10.1109/anzcc53563.2021.9628387","url":null,"abstract":"Machine learning is widely being used in medical field for disease diagnostics and research.The area of machine learning is mainly classified into 3 parts: supervised, unsupervised and reinforcement learning.Supervised machine learning (ML) algorithms are used in this paper for modeling and showing the impact of increased testing on the number of daily confirmed cases of COVID-19. The algorithms used to carry out this study are decision tree regression and random forest regression. Machine learning for modeling has proven to be significant for forecasting and hence decision making over the future course of actions. In this paper, Gaussian process regression has been used for modeling as well as forecasting the daily confirmed cases in South Korea. The results obtained show that if the number of tests conducted is increased to the population of South Korea, approximately equal to 51, 286, 183, the peak in the daily cases is obtained earlier and hence the overall number of daily cases is less compared to current cases.","PeriodicalId":246687,"journal":{"name":"2021 Australian & New Zealand Control Conference (ANZCC)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115261313","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 : 2021-11-25DOI: 10.1109/anzcc53563.2021.9628389
Z. Salehi, Yijun Chen, E. Ratnam, I. Petersen, Guodong Shi
In this paper, we study multi-agent systems with distributed resource allocation at individual agents. The agents make local resource allocation decisions including, in some cases, trading decisions — incurring income or expenditure subject to the resource price and system-level resource availability. The agents seek to maximize their individual payoffs, which accrue from both resource allocation income and expenditure. We define a social shaping problem for the system and show that the optimal price is always below a prescribed socially resilient price threshold. By exploring optimality conditions for each agent, we express resource allocation decisions in terms of piece-wise linear functions with respect to the price for unit resource. We further establish a tight range for the coefficients of the linear-quadratic utilities, under which optimal pricing is proven to be always socially resilient.
{"title":"Social Shaping of Linear Quadratic Multi-Agent Systems","authors":"Z. Salehi, Yijun Chen, E. Ratnam, I. Petersen, Guodong Shi","doi":"10.1109/anzcc53563.2021.9628389","DOIUrl":"https://doi.org/10.1109/anzcc53563.2021.9628389","url":null,"abstract":"In this paper, we study multi-agent systems with distributed resource allocation at individual agents. The agents make local resource allocation decisions including, in some cases, trading decisions — incurring income or expenditure subject to the resource price and system-level resource availability. The agents seek to maximize their individual payoffs, which accrue from both resource allocation income and expenditure. We define a social shaping problem for the system and show that the optimal price is always below a prescribed socially resilient price threshold. By exploring optimality conditions for each agent, we express resource allocation decisions in terms of piece-wise linear functions with respect to the price for unit resource. We further establish a tight range for the coefficients of the linear-quadratic utilities, under which optimal pricing is proven to be always socially resilient.","PeriodicalId":246687,"journal":{"name":"2021 Australian & New Zealand Control Conference (ANZCC)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115603453","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 : 2021-11-25DOI: 10.1109/anzcc53563.2021.9628378
T. Veerendar, Deepak Kumar, V. Sreeram
In this paper, a novel meta-heuristic Colliding bodies optimization-based proportional-integral-derivative controller with derivative filter is presented to solve a single-area power system's load frequency control issues. The controller parameters are determined by employing the integral of time multiplied absolute error as the fitness function. A single-area non-reheat thermal power system is considered for establishing the efficacy of the proposed method. The robustness of the proposed controller is also ascertained by inserting perturbation in system parameters. It is observed from the simulation results that the proposed method provides improved dynamic performance over the existing methods.
{"title":"Colliding Bodies Optimization-based PID Controller for Load Frequency Control of single area power system","authors":"T. Veerendar, Deepak Kumar, V. Sreeram","doi":"10.1109/anzcc53563.2021.9628378","DOIUrl":"https://doi.org/10.1109/anzcc53563.2021.9628378","url":null,"abstract":"In this paper, a novel meta-heuristic Colliding bodies optimization-based proportional-integral-derivative controller with derivative filter is presented to solve a single-area power system's load frequency control issues. The controller parameters are determined by employing the integral of time multiplied absolute error as the fitness function. A single-area non-reheat thermal power system is considered for establishing the efficacy of the proposed method. The robustness of the proposed controller is also ascertained by inserting perturbation in system parameters. It is observed from the simulation results that the proposed method provides improved dynamic performance over the existing methods.","PeriodicalId":246687,"journal":{"name":"2021 Australian & New Zealand Control Conference (ANZCC)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125004154","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 : 2021-11-25DOI: 10.1109/anzcc53563.2021.9628203
Lai Wei, R. McCloy, Jie Bao
Motivated by the trend of flexible manufacturing in the process control industry and the uncertain nature of chemical process models, this article aims to achieve offset-free tracking for a family of uncertain nonlinear systems. The proposed control approach employs two main modules: a neural network embedded contraction-based controller to ensure convergence to time-varying references; and an online identification module coupled with a reference generator to provide convergency of the modelled parameters to that of the physical system. The first step in the proposed approach is to provide a guaranteed contraction condition for nonlinear systems, subject to time-varying parametric uncertainty, that are driven by neural network embedded controllers and modelled parameter estimates. The second step is to provide unknown system parameter identification online. By ensuring that uncertain parameter estimates converge to the corresponding physical values, offset-free tracking can be achieved. An illustrative example is included to demonstrate the overall approach.
{"title":"A Neural Network-based Contraction Control with Online Parameter Identification for Uncertain Nonlinear Systems","authors":"Lai Wei, R. McCloy, Jie Bao","doi":"10.1109/anzcc53563.2021.9628203","DOIUrl":"https://doi.org/10.1109/anzcc53563.2021.9628203","url":null,"abstract":"Motivated by the trend of flexible manufacturing in the process control industry and the uncertain nature of chemical process models, this article aims to achieve offset-free tracking for a family of uncertain nonlinear systems. The proposed control approach employs two main modules: a neural network embedded contraction-based controller to ensure convergence to time-varying references; and an online identification module coupled with a reference generator to provide convergency of the modelled parameters to that of the physical system. The first step in the proposed approach is to provide a guaranteed contraction condition for nonlinear systems, subject to time-varying parametric uncertainty, that are driven by neural network embedded controllers and modelled parameter estimates. The second step is to provide unknown system parameter identification online. By ensuring that uncertain parameter estimates converge to the corresponding physical values, offset-free tracking can be achieved. An illustrative example is included to demonstrate the overall approach.","PeriodicalId":246687,"journal":{"name":"2021 Australian & New Zealand Control Conference (ANZCC)","volume":"89 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127173200","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 : 2021-11-25DOI: 10.1109/anzcc53563.2021.9628299
R. M. D. Silva, D. Eckhard
From a practical point of view, adjusting the controller without having to identify the process model has many advantages, for example, when the process is simple but changes a lot during the operation. In this case, there are many direct data-driven methods in the literature which may be employed to adjust a monovariable controller aiming at reference tracking. However, when the control objective is disturbance rejection or regulation, the designer is left with too few choices. The aim of this paper is to provide one new option and show how it can be applied to those control objectives.
{"title":"Data-driven Correlation Approach Applied to Load Disturbance Rejection in a Thermal Process","authors":"R. M. D. Silva, D. Eckhard","doi":"10.1109/anzcc53563.2021.9628299","DOIUrl":"https://doi.org/10.1109/anzcc53563.2021.9628299","url":null,"abstract":"From a practical point of view, adjusting the controller without having to identify the process model has many advantages, for example, when the process is simple but changes a lot during the operation. In this case, there are many direct data-driven methods in the literature which may be employed to adjust a monovariable controller aiming at reference tracking. However, when the control objective is disturbance rejection or regulation, the designer is left with too few choices. The aim of this paper is to provide one new option and show how it can be applied to those control objectives.","PeriodicalId":246687,"journal":{"name":"2021 Australian & New Zealand Control Conference (ANZCC)","volume":"94 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116586423","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 : 2021-11-25DOI: 10.1109/anzcc53563.2021.9628304
E. Boeira, D. Eckhard
This paper addresses the use of the regularization feature on impulse response estimation for systems with colored output noise. Firstly, it is shown that the optimal regularization matrix for this scenario is quite different than the optimal for the white noise case and that there is a direct relationship between the Regularized Weighted Least-Squares with a Bayesian perspective of the identification problem for such case. Also, a new Empirical Bayes method, based on the Bayesian perspective, is introduced to estimate the regularization and noise covariance matrices from data. Finally, a numerical example demonstrates that this new methodology outperforms the traditional Regularized Least-Squares, producing better statistical properties and better results for a model fit measure.
{"title":"Regularized impulse response estimation for systems with colored output noise","authors":"E. Boeira, D. Eckhard","doi":"10.1109/anzcc53563.2021.9628304","DOIUrl":"https://doi.org/10.1109/anzcc53563.2021.9628304","url":null,"abstract":"This paper addresses the use of the regularization feature on impulse response estimation for systems with colored output noise. Firstly, it is shown that the optimal regularization matrix for this scenario is quite different than the optimal for the white noise case and that there is a direct relationship between the Regularized Weighted Least-Squares with a Bayesian perspective of the identification problem for such case. Also, a new Empirical Bayes method, based on the Bayesian perspective, is introduced to estimate the regularization and noise covariance matrices from data. Finally, a numerical example demonstrates that this new methodology outperforms the traditional Regularized Least-Squares, producing better statistical properties and better results for a model fit measure.","PeriodicalId":246687,"journal":{"name":"2021 Australian & New Zealand Control Conference (ANZCC)","volume":"35 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129511931","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}
This paper focuses on the research of the dashboard camera for improving the storage space and object recognition. The experiments showed that the CS method of ISTA-Net (Iterative Shrinkage Thresholding Algorithm with Network) can reduce the storage space by at least 60% and without obviously sacrificing the image quality. Furthermore, the recognition method by YOLOv4 can overcome the variety of environments, which can reach the recognition ratio of over 80% in a small 480x480 pixels. The recognition function can help to quickly catch the key features (ex: car, traffic signal, pedestrian, etc.) in the storage data of the dashboard camera.
{"title":"Integrated Compressed Sensing and YOLOv4 for Application in Image-storage and Object-recognition of Dashboard Camera","authors":"Jim-Wei Wu, Cheng-Chia Wu, Wen-Shan Cen, Shao-An Chao, Jui-Tse Weng","doi":"10.1109/anzcc53563.2021.9628221","DOIUrl":"https://doi.org/10.1109/anzcc53563.2021.9628221","url":null,"abstract":"This paper focuses on the research of the dashboard camera for improving the storage space and object recognition. The experiments showed that the CS method of ISTA-Net (Iterative Shrinkage Thresholding Algorithm with Network) can reduce the storage space by at least 60% and without obviously sacrificing the image quality. Furthermore, the recognition method by YOLOv4 can overcome the variety of environments, which can reach the recognition ratio of over 80% in a small 480x480 pixels. The recognition function can help to quickly catch the key features (ex: car, traffic signal, pedestrian, etc.) in the storage data of the dashboard camera.","PeriodicalId":246687,"journal":{"name":"2021 Australian & New Zealand Control Conference (ANZCC)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122745849","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 : 2021-11-25DOI: 10.1109/anzcc53563.2021.9628293
Tony Dang, Frederik Debrouwere, E. Hostens
Typically, Model Predictive Control (MPC) for highly dynamic systems poses challenges to the computation power needed to optimize the control in real-time. In this paper, we present an explainable methodology to approximate MPCs with low input penalization as a closed form expression, using learning by demonstration. Classical approaches, e.g. using neural networks, result in over-complicated controllers and require huge datasets. In this paper, the prior knowledge on the typical bang-bang behavior of low-input penalized MPC will be exploited to approximate the MPC-law by only sparsely sampling the state space. This is achieved by identifying the switching surface of the sampled MPC-solution using Support Vector Machines (SVMs). The result is a light-weight, interpretable, easy to tune, explicit control law suitable for real-time applications. The methodology is validated in simulation on a benchmark problem from the field of process control (stirred tank reactor), and on a physical set-up of a highly dynamic motion control problem (parallel SCARA). The results, both in simulation and experimentally, show that strong approximation can already be obtained by using very light-weight controllers which, for the SCARA, were able to run on a frequency of at least 2kHz on the experimental setup.
{"title":"Approximating Nonlinear Model Predictive Controllers using Support Vector Machines","authors":"Tony Dang, Frederik Debrouwere, E. Hostens","doi":"10.1109/anzcc53563.2021.9628293","DOIUrl":"https://doi.org/10.1109/anzcc53563.2021.9628293","url":null,"abstract":"Typically, Model Predictive Control (MPC) for highly dynamic systems poses challenges to the computation power needed to optimize the control in real-time. In this paper, we present an explainable methodology to approximate MPCs with low input penalization as a closed form expression, using learning by demonstration. Classical approaches, e.g. using neural networks, result in over-complicated controllers and require huge datasets. In this paper, the prior knowledge on the typical bang-bang behavior of low-input penalized MPC will be exploited to approximate the MPC-law by only sparsely sampling the state space. This is achieved by identifying the switching surface of the sampled MPC-solution using Support Vector Machines (SVMs). The result is a light-weight, interpretable, easy to tune, explicit control law suitable for real-time applications. The methodology is validated in simulation on a benchmark problem from the field of process control (stirred tank reactor), and on a physical set-up of a highly dynamic motion control problem (parallel SCARA). The results, both in simulation and experimentally, show that strong approximation can already be obtained by using very light-weight controllers which, for the SCARA, were able to run on a frequency of at least 2kHz on the experimental setup.","PeriodicalId":246687,"journal":{"name":"2021 Australian & New Zealand Control Conference (ANZCC)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115913149","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 : 2021-11-25DOI: 10.1109/anzcc53563.2021.9628250
Kazuhiko Takahashi, Eri Tano, M. Hashimoto
This study investigates a training method of a quaternion multi–layer neural network based on a gradient– descent method extended to quaternion numbers. The gradient of the cost function is calculated using the generalised ${mathbb{H}}{mathbb{R}}$ calculus to derive the training rule for the network parameters. Computational experiments for identifying and controlling a discrete–time nonlinear plant were conducted to evaluate the proposed method. The results confirmed the feasibility of using the G ${mathbb{H}}{mathbb{R}}$ calculus in the quaternion neural network and showed the capability of using the quaternion neural network for a control system application.
{"title":"Remarks on Quaternion Multi–Layer Neural Network Based on the Generalised HR Calculus","authors":"Kazuhiko Takahashi, Eri Tano, M. Hashimoto","doi":"10.1109/anzcc53563.2021.9628250","DOIUrl":"https://doi.org/10.1109/anzcc53563.2021.9628250","url":null,"abstract":"This study investigates a training method of a quaternion multi–layer neural network based on a gradient– descent method extended to quaternion numbers. The gradient of the cost function is calculated using the generalised ${mathbb{H}}{mathbb{R}}$ calculus to derive the training rule for the network parameters. Computational experiments for identifying and controlling a discrete–time nonlinear plant were conducted to evaluate the proposed method. The results confirmed the feasibility of using the G ${mathbb{H}}{mathbb{R}}$ calculus in the quaternion neural network and showed the capability of using the quaternion neural network for a control system application.","PeriodicalId":246687,"journal":{"name":"2021 Australian & New Zealand Control Conference (ANZCC)","volume":"332 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115975787","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}
This paper investigates the global asymptotic stability of linear time-varying impulsive positive systems (IPSs). Several novel stability criteria of linear time-varying IPSs with different types of impulsive effects are proposed by constructing an indefinite time-varying copositive Lyapunov function. In particular, by using the maximum and average dwell time methods, we discuss the stability of the addressed system with destabilizing impulses and stabilizing impulses, respectively. Moreover, we consider a special case in which the continuous dynamic of the system is not asymptotically stable and the system may contain some destabilizing impulses, and give a slightly stricter stability criterion. Finally, two examples are given to validate the effectiveness of the obtained results.
{"title":"Novel stability conditions of linear time-varying impulsive positive systems based on indefinite Lyapunov functions *","authors":"Niankun Zhang, Peilong Yu, Yuting Kang, Qianqian Zhang","doi":"10.1109/anzcc53563.2021.9628267","DOIUrl":"https://doi.org/10.1109/anzcc53563.2021.9628267","url":null,"abstract":"This paper investigates the global asymptotic stability of linear time-varying impulsive positive systems (IPSs). Several novel stability criteria of linear time-varying IPSs with different types of impulsive effects are proposed by constructing an indefinite time-varying copositive Lyapunov function. In particular, by using the maximum and average dwell time methods, we discuss the stability of the addressed system with destabilizing impulses and stabilizing impulses, respectively. Moreover, we consider a special case in which the continuous dynamic of the system is not asymptotically stable and the system may contain some destabilizing impulses, and give a slightly stricter stability criterion. Finally, two examples are given to validate the effectiveness of the obtained results.","PeriodicalId":246687,"journal":{"name":"2021 Australian & New Zealand Control Conference (ANZCC)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-11-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128072508","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：