Pub Date : 2022-06-08DOI: 10.23919/ACC53348.2022.9867144
Zhenlei Chen, Qing Guo, Yan Shi, Yao Yan
A distributed interaction control based on impedance frame is proposed in Euler-Lagrangian network system to achieve multi-agents interaction synchronization. Considering the different control objective of task space, a unified impedance dynamics is studied to realize two control modes transformation from position tracking to interaction synchronization by the impedance parameter regulation. The constrained conditions of variable impedance parameters is discussed to design a reasonable distributed controller by Lyapunov technique. Finally, The effectiveness of the proposed controller is validated by simulations on a multi-robots network system with 4 nodes.
{"title":"Distributed Cooperative Control from Position Motion to Interaction Synchronization","authors":"Zhenlei Chen, Qing Guo, Yan Shi, Yao Yan","doi":"10.23919/ACC53348.2022.9867144","DOIUrl":"https://doi.org/10.23919/ACC53348.2022.9867144","url":null,"abstract":"A distributed interaction control based on impedance frame is proposed in Euler-Lagrangian network system to achieve multi-agents interaction synchronization. Considering the different control objective of task space, a unified impedance dynamics is studied to realize two control modes transformation from position tracking to interaction synchronization by the impedance parameter regulation. The constrained conditions of variable impedance parameters is discussed to design a reasonable distributed controller by Lyapunov technique. Finally, The effectiveness of the proposed controller is validated by simulations on a multi-robots network system with 4 nodes.","PeriodicalId":366299,"journal":{"name":"2022 American Control Conference (ACC)","volume":"70 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131033132","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 : 2022-06-08DOI: 10.23919/ACC53348.2022.9867650
D. Landgraf, Andreas Völz, K. Graichen
This paper shows how model predictive control of a nonlinear system with a time-dependent autocorrelated disturbance can be realized in a computationally efficient way. To this end, the disturbance is modeled as a Gaussian process that can be reformulated as a linear state space model driven by white Gaussian noise. This disturbance model is combined with a first-principles physical system model resulting in a latent force model. In this way, the current states and the disturbance can be estimated using the unscented Kalman filter. Moreover, the disturbance model can be used to predict future values of the disturbance, which is necessary for predictive control. In order to further reduce the computational effort of the algorithm, a separate predictor for the linear disturbance subsystem is outlined. The predicted disturbance and the estimated states are used to formulate the optimization problem of a model predictive controller. The proposed approach is evaluated numerically using the example of a two-dimensional overhead crane. It is shown that the algorithm is real-time capable with computation times in the sub-millisecond range.
{"title":"Nonlinear Model Predictive Control with Latent Force Models","authors":"D. Landgraf, Andreas Völz, K. Graichen","doi":"10.23919/ACC53348.2022.9867650","DOIUrl":"https://doi.org/10.23919/ACC53348.2022.9867650","url":null,"abstract":"This paper shows how model predictive control of a nonlinear system with a time-dependent autocorrelated disturbance can be realized in a computationally efficient way. To this end, the disturbance is modeled as a Gaussian process that can be reformulated as a linear state space model driven by white Gaussian noise. This disturbance model is combined with a first-principles physical system model resulting in a latent force model. In this way, the current states and the disturbance can be estimated using the unscented Kalman filter. Moreover, the disturbance model can be used to predict future values of the disturbance, which is necessary for predictive control. In order to further reduce the computational effort of the algorithm, a separate predictor for the linear disturbance subsystem is outlined. The predicted disturbance and the estimated states are used to formulate the optimization problem of a model predictive controller. The proposed approach is evaluated numerically using the example of a two-dimensional overhead crane. It is shown that the algorithm is real-time capable with computation times in the sub-millisecond range.","PeriodicalId":366299,"journal":{"name":"2022 American Control Conference (ACC)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131156683","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 : 2022-06-08DOI: 10.23919/ACC53348.2022.9867244
Yatong Bai, Tanmay Gautam, Yujie Gai, S. Sojoudi
As neural networks become more prevalent in safety-critical systems, ensuring their robustness against adversaries becomes essential. "Adversarial training" is one of the most common methods for training robust networks. Current adversarial training algorithms solve highly non-convex bi-level optimization problems. These algorithms suffer from the lack of convergence guarantees and can exhibit unstable behaviors. A recent work has shown that the standard training formulation of a one-hidden-layer, scalar-output fully-connected neural network with rectified linear unit (ReLU) activations can be reformulated as a finite-dimensional convex program, addressing the aforementioned issues for training non-robust networks. In this paper, we leverage this "convex training" framework to tackle the problem of adversarial training. Unfortunately, the scale of the convex training program proposed in the literature grows exponentially in the data size. We prove that a stochastic approximation procedure that scales linearly yields high-quality solutions. With the complexity roadblock removed, we derive convex optimization models that train robust neural networks. Our convex methods provably produce an upper bound on the global optimum of the adversarial training objective and can be applied to binary classification and regression. We demonstrate in experiments that the proposed method achieves a superior robustness compared with the existing methods.
{"title":"Practical Convex Formulations of One-hidden-layer Neural Network Adversarial Training","authors":"Yatong Bai, Tanmay Gautam, Yujie Gai, S. Sojoudi","doi":"10.23919/ACC53348.2022.9867244","DOIUrl":"https://doi.org/10.23919/ACC53348.2022.9867244","url":null,"abstract":"As neural networks become more prevalent in safety-critical systems, ensuring their robustness against adversaries becomes essential. \"Adversarial training\" is one of the most common methods for training robust networks. Current adversarial training algorithms solve highly non-convex bi-level optimization problems. These algorithms suffer from the lack of convergence guarantees and can exhibit unstable behaviors. A recent work has shown that the standard training formulation of a one-hidden-layer, scalar-output fully-connected neural network with rectified linear unit (ReLU) activations can be reformulated as a finite-dimensional convex program, addressing the aforementioned issues for training non-robust networks. In this paper, we leverage this \"convex training\" framework to tackle the problem of adversarial training. Unfortunately, the scale of the convex training program proposed in the literature grows exponentially in the data size. We prove that a stochastic approximation procedure that scales linearly yields high-quality solutions. With the complexity roadblock removed, we derive convex optimization models that train robust neural networks. Our convex methods provably produce an upper bound on the global optimum of the adversarial training objective and can be applied to binary classification and regression. We demonstrate in experiments that the proposed method achieves a superior robustness compared with the existing methods.","PeriodicalId":366299,"journal":{"name":"2022 American Control Conference (ACC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130736459","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 : 2022-06-08DOI: 10.23919/ACC53348.2022.9867659
Ao Li, Yan Chen, Wen-Chiao Lin, Xinyu Du
Influenced by tire effective radius change, suspension rearrangement, and pitch/roll disturbance due to tire blowout, the vehicle center of gravity (CG) can significantly relocate toward the blown-out tire position. This paper proposes an estimation method of the CG relocation for ground vehicles with tire blowout by utilizing vertical force variations and geometric relationships in tire blowout events. Based on a new recursive least square (RLS) formulation in this paper, the three-dimensional (3D) CG relocation (i.e., the height, the longitudinal and lateral positions) can be estimated simultaneously. Matlab/Simulink and CarSim® co-simulation results for different tire blowout locations validate that the proposed estimation method can effectively and accurately capture the vehicle 3D CG relocation after tire blowout.
{"title":"Estimation of Three-Dimensional Center of Gravity Relocation for Ground Vehicles with Tire Blowout","authors":"Ao Li, Yan Chen, Wen-Chiao Lin, Xinyu Du","doi":"10.23919/ACC53348.2022.9867659","DOIUrl":"https://doi.org/10.23919/ACC53348.2022.9867659","url":null,"abstract":"Influenced by tire effective radius change, suspension rearrangement, and pitch/roll disturbance due to tire blowout, the vehicle center of gravity (CG) can significantly relocate toward the blown-out tire position. This paper proposes an estimation method of the CG relocation for ground vehicles with tire blowout by utilizing vertical force variations and geometric relationships in tire blowout events. Based on a new recursive least square (RLS) formulation in this paper, the three-dimensional (3D) CG relocation (i.e., the height, the longitudinal and lateral positions) can be estimated simultaneously. Matlab/Simulink and CarSim® co-simulation results for different tire blowout locations validate that the proposed estimation method can effectively and accurately capture the vehicle 3D CG relocation after tire blowout.","PeriodicalId":366299,"journal":{"name":"2022 American Control Conference (ACC)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130841560","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 : 2022-06-08DOI: 10.23919/ACC53348.2022.9867414
M. Maghenem, Adnane Saoud, A. Loría
In this and the companion paper [1] we propose a direct-adaptive-control framework for hybrid dynamical systems with unknown parameters. The approach addresses both the tracking-control and the parameter-estimation problems and relies on Lyapunov theory for hybrid systems. In this paper, we extend the main results of [1] to deal with hybrid systems that contain a regressor that is of linear order of growth, thereby relaxing the boundedness restriction imposed in [1]. As in the latter reference, the statements rely on Lyapunov theory for hybrid systems and we establish uniform global asymptotic stability in closed loop. In particular, parameter-estimation convergence is guaranteed when a generic hybrid persistence of excitation condition on the pair of discrete and continuous regressor functions holds. On the other hand, the relaxation of the boundedness assumption relies on a higher-order adaptation law.
{"title":"Adaptive Control/Identification for Hybrid Systems, Part II: with Linear-growth-order Discrete Regressor","authors":"M. Maghenem, Adnane Saoud, A. Loría","doi":"10.23919/ACC53348.2022.9867414","DOIUrl":"https://doi.org/10.23919/ACC53348.2022.9867414","url":null,"abstract":"In this and the companion paper [1] we propose a direct-adaptive-control framework for hybrid dynamical systems with unknown parameters. The approach addresses both the tracking-control and the parameter-estimation problems and relies on Lyapunov theory for hybrid systems. In this paper, we extend the main results of [1] to deal with hybrid systems that contain a regressor that is of linear order of growth, thereby relaxing the boundedness restriction imposed in [1]. As in the latter reference, the statements rely on Lyapunov theory for hybrid systems and we establish uniform global asymptotic stability in closed loop. In particular, parameter-estimation convergence is guaranteed when a generic hybrid persistence of excitation condition on the pair of discrete and continuous regressor functions holds. On the other hand, the relaxation of the boundedness assumption relies on a higher-order adaptation law.","PeriodicalId":366299,"journal":{"name":"2022 American Control Conference (ACC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130957547","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 : 2022-06-08DOI: 10.23919/ACC53348.2022.9867471
A. Aghajan, B. Touri
We consider the averaging-based distributed optimization over random independent but time-varying networks where we generalize the results for independently and identically distributed (i.i.d.) networks to independent networks. In this paper, we show that under some assumptions on average connectivity the averaging-based distributed optimization methods still work for independent random networks when the averaging weight matrices are column-stochastic in-expectation. This result has interesting implications including implications on the robustness of the averaging-based distributed optimization methods to link-failure.
{"title":"On Distributed Optimization Over Random Independent Networks","authors":"A. Aghajan, B. Touri","doi":"10.23919/ACC53348.2022.9867471","DOIUrl":"https://doi.org/10.23919/ACC53348.2022.9867471","url":null,"abstract":"We consider the averaging-based distributed optimization over random independent but time-varying networks where we generalize the results for independently and identically distributed (i.i.d.) networks to independent networks. In this paper, we show that under some assumptions on average connectivity the averaging-based distributed optimization methods still work for independent random networks when the averaging weight matrices are column-stochastic in-expectation. This result has interesting implications including implications on the robustness of the averaging-based distributed optimization methods to link-failure.","PeriodicalId":366299,"journal":{"name":"2022 American Control Conference (ACC)","volume":"55 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133041182","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 : 2022-06-08DOI: 10.23919/ACC53348.2022.9867732
G. Das, Daigo Shishika
In this paper we consider a Target-guarding differential game where the Defender must protect a linearly moving line segment by intercepting the Attacker who tries to reach it. In contrast to common Target-guarding problems, we assume that the Defender is attached to the Target and moves along with it. This assumption affects the Defender’s maximum speed depending on its heading direction. A zero-sum differential game of degree for the Attacker-winning scenario is studied, where the payoff is defined to be the distance between the two agents at the time of reaching the Target. We derive the equilibrium strategies and the Value function by leveraging the solution for the infinite-length Target scenario. The zero-level set of this Value function provides the barrier surface that divides the state space into Defender-winning and Attacker-winning regions. We present simulation results at the end to demonstrate the theoretical results.
{"title":"Guarding a Translating Line with an Attached Defender","authors":"G. Das, Daigo Shishika","doi":"10.23919/ACC53348.2022.9867732","DOIUrl":"https://doi.org/10.23919/ACC53348.2022.9867732","url":null,"abstract":"In this paper we consider a Target-guarding differential game where the Defender must protect a linearly moving line segment by intercepting the Attacker who tries to reach it. In contrast to common Target-guarding problems, we assume that the Defender is attached to the Target and moves along with it. This assumption affects the Defender’s maximum speed depending on its heading direction. A zero-sum differential game of degree for the Attacker-winning scenario is studied, where the payoff is defined to be the distance between the two agents at the time of reaching the Target. We derive the equilibrium strategies and the Value function by leveraging the solution for the infinite-length Target scenario. The zero-level set of this Value function provides the barrier surface that divides the state space into Defender-winning and Attacker-winning regions. We present simulation results at the end to demonstrate the theoretical results.","PeriodicalId":366299,"journal":{"name":"2022 American Control Conference (ACC)","volume":"62 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133386700","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 : 2022-06-08DOI: 10.23919/ACC53348.2022.9867668
Tatsuya Miyano, J. Romberg, M. Egerstedt
We consider the problem of collectively transporting multiple objects by multiple agents. The objective is to find the optimal matching between the objects and agents that minimizes the energy of the overall system. We show that combining a proximal gradient method with continuous relaxation yields a distributed algorithm which converges to a near-optimal solution for the associated optimization problem. Furthermore, by using this solution as an initial solution, a distributed negative-cycle canceling algorithm, which monotonically decreases the matching cost at each step, provides the globally optimal solution for the problem. Numerical simulations demonstrate the performance on practical problems.
{"title":"Distributed Optimal Assignment Algorithm for Collective Foraging *","authors":"Tatsuya Miyano, J. Romberg, M. Egerstedt","doi":"10.23919/ACC53348.2022.9867668","DOIUrl":"https://doi.org/10.23919/ACC53348.2022.9867668","url":null,"abstract":"We consider the problem of collectively transporting multiple objects by multiple agents. The objective is to find the optimal matching between the objects and agents that minimizes the energy of the overall system. We show that combining a proximal gradient method with continuous relaxation yields a distributed algorithm which converges to a near-optimal solution for the associated optimization problem. Furthermore, by using this solution as an initial solution, a distributed negative-cycle canceling algorithm, which monotonically decreases the matching cost at each step, provides the globally optimal solution for the problem. Numerical simulations demonstrate the performance on practical problems.","PeriodicalId":366299,"journal":{"name":"2022 American Control Conference (ACC)","volume":"534 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133498011","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 : 2022-06-08DOI: 10.23919/ACC53348.2022.9867425
Pooja Sharma, N. Satyanarayana
The strong delay independent stability (DIS) for commensurate multiple time delay systems (CMTDSs) is studied in this paper. The necessary and sufficient strong DIS condition is derived for an augmented time delay system (TDS) originated from the original multiple time delay system (MTDS). A linear matrix inequality (LMI) condition is used to deal with the strong DIS, whose solution is derived by means of Kalman-Yakubovich-Popov (KYP) lemma and Kronecker properties. Two numerical examples are given to demonstrate the advantages and applicability of the proposed approach.
{"title":"Strong delay independent stability condition for commensurate time delay systems","authors":"Pooja Sharma, N. Satyanarayana","doi":"10.23919/ACC53348.2022.9867425","DOIUrl":"https://doi.org/10.23919/ACC53348.2022.9867425","url":null,"abstract":"The strong delay independent stability (DIS) for commensurate multiple time delay systems (CMTDSs) is studied in this paper. The necessary and sufficient strong DIS condition is derived for an augmented time delay system (TDS) originated from the original multiple time delay system (MTDS). A linear matrix inequality (LMI) condition is used to deal with the strong DIS, whose solution is derived by means of Kalman-Yakubovich-Popov (KYP) lemma and Kronecker properties. Two numerical examples are given to demonstrate the advantages and applicability of the proposed approach.","PeriodicalId":366299,"journal":{"name":"2022 American Control Conference (ACC)","volume":"66 5-6","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132693017","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 : 2022-06-08DOI: 10.23919/ACC53348.2022.9867465
Jonas Hansson, Alain Govaert, R. Pates, E. Tegling, K. Soltesz
Case isolation, that is, detection and isolation of infected individuals in order to prevent spread, is a strategy to curb infectious disease epidemics. Here, we study the efficiency of a case isolation strategy subject to time delays in terms of its ability to stabilize the epidemic spread in heterogeneous contact networks. For an SIR epidemic model, we characterize the stability boundary analytically and show how it depends on the time delay between infection and isolation as well as the heterogeneity of the inter-individual contact network, quantified by the variance in contact rates. We show that network heterogeneity accounts for a restricting correction factor to previously derived stability results for homogeneous SIR models (with uniform contact rates), which are therefore too optimistic on the relevant time scales. We illustrate the results and the underlying mechanisms through insightful numerical examples.
{"title":"Limitations of time-delayed case isolation in heterogeneous SIR models","authors":"Jonas Hansson, Alain Govaert, R. Pates, E. Tegling, K. Soltesz","doi":"10.23919/ACC53348.2022.9867465","DOIUrl":"https://doi.org/10.23919/ACC53348.2022.9867465","url":null,"abstract":"Case isolation, that is, detection and isolation of infected individuals in order to prevent spread, is a strategy to curb infectious disease epidemics. Here, we study the efficiency of a case isolation strategy subject to time delays in terms of its ability to stabilize the epidemic spread in heterogeneous contact networks. For an SIR epidemic model, we characterize the stability boundary analytically and show how it depends on the time delay between infection and isolation as well as the heterogeneity of the inter-individual contact network, quantified by the variance in contact rates. We show that network heterogeneity accounts for a restricting correction factor to previously derived stability results for homogeneous SIR models (with uniform contact rates), which are therefore too optimistic on the relevant time scales. We illustrate the results and the underlying mechanisms through insightful numerical examples.","PeriodicalId":366299,"journal":{"name":"2022 American Control Conference (ACC)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-06-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132724556","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：