This paper provides an adaptive design based on event-based control with unknown control directions. Nussbaum-type functions are used to address the problems associated with nonlinear systems' unknown control dire...
详细信息
ISBN:
(纸本)9798350321050
This paper provides an adaptive design based on event-based control with unknown control directions. Nussbaum-type functions are used to address the problems associated with nonlinear systems' unknown control directions. An event-triggered mechanism, which is based on the measurement error defined by the control signal, is presented to reduce the number of controller updates and save transmission resources. The dynamic surface control is incorporated into the controller architecture to prevent the complexity explosion. Finally, the effectiveness of the designed strategy is assessed using the simulation.
This paper investigates a resilient output formation-containment tracking (FCT) problem for heterogeneous multi-agent systems (MASs) under unknown dynamics and uncertainties. A learning-based control framework using o...
详细信息
This paper investigates a resilient output formation-containment tracking (FCT) problem for heterogeneous multi-agent systems (MASs) under unknown dynamics and uncertainties. A learning-based control framework using online dynamic data is proposed with three hierarchical phases. First, fully distributed observers for agents with various types of objectives are presented under a directed graph. The estimations of tracking reference and time-varying formation are coordinated in terms of both dynamics and states. Second, dynamic data filters based on the internal model principle and partial observations are introduced to reconstruct the MASs information and formulate a virtual tracking system, where the reinforcement learning (RL) technique is applied. Based on two proposed off-policy schemes, the RL algorithm is adapted to a hybrid form under the dynamic data. An ideal tracking controller is uniformly learned and essential dynamics are extracted from the same data. Third, the integrated resilient output FCT controller is further derived using previous learning results. The adaptive neural networks and compensation functions are utilized in a data-driven manner to address unknown faults and uncertainties. The integration of filtering, estimation, and learning broadens a more general control framework than existing results. Finally, validations are demonstrated by numerical simulations.
In this paper, we propose a data-driven adaptive tuning (DDAT) method of traditional incremental integral control laws for unknown nonlinear and non-affine systems. Firstly, the original nonlinear system is transforme...
详细信息
ISBN:
(纸本)9798350321050
In this paper, we propose a data-driven adaptive tuning (DDAT) method of traditional incremental integral control laws for unknown nonlinear and non-affine systems. Firstly, the original nonlinear system is transformed into its equivalent linear model based on the input and output (I/O) data using the compact form dynamic linearization (CFDL) method. Then the learning gain of the integral control law is adjusted by designing and optimizing an objective function. This method only needs the I/O data instead of relying on the model information of the controlled plant. The convergency analysis is also given. Finally, a numerical example and a heat exchanger example are adopted to illustrate the effectiveness of the proposed method.
This article investigates the control problem for a sort of repetitive discrete-time nonlinear systems subject to random packet dropouts and limited communication bandwidth. In order to compensate the impacts from the...
详细信息
This article investigates the control problem for a sort of repetitive discrete-time nonlinear systems subject to random packet dropouts and limited communication bandwidth. In order to compensate the impacts from the constraints on bandwidth, this work designs a communication protocol by designing a two-description coding scheme in combination with the scalar uniform quantization technique. The proposed protocol makes use of two independent channels to transmit data separately, thereby improving the channel utilization efficiency and reducing the probability of packet dropout. Then, with the proposed protocol and the iterative dynamic linearization approach, an adaptive iterative learningcontroller associated with a parameter estimation strategy is provided for the nonlinear system under investigation. The control law is data-driven, which therefore does not require knowledge of the model. Subsequently, the sufficient condition is derived under which the tracking error is forced to convergent. Finally, with the purpose to show the correctness of our theoretical results, we carry out two numerical simulations to test the effectiveness of the proposed control strategy.
This paper introduces a data-driven iterative learningcontrol (ILC) algorithm based on the forgetting factor within the frequency domain. Different from the conventional time-domain design of the forgetting factor al...
详细信息
On the wave of recent advances in data-driven predictive control, we present an explicit controller that can be constructed from a batch of input/output data only. The proposed explicit law is built upon a regularized...
详细信息
On the wave of recent advances in data-driven predictive control, we present an explicit controller that can be constructed from a batch of input/output data only. The proposed explicit law is built upon a regularized data-drivencontrol problem, so as to guarantee the uniqueness of the explicit predictive controller and to endow it with noise handling capabilities. The effectiveness of the retrieved explicit law and the repercussions of regularization on noise handling is analyzed on a benchmark simulation example.
In this paper, the predictive control problem of two-dimensional iterative learning model based on just-in-time learning (JITL) model is studied for batch processes. A new error compensation strategy is proposed based...
详细信息
ISBN:
(纸本)9798350321050
In this paper, the predictive control problem of two-dimensional iterative learning model based on just-in-time learning (JITL) model is studied for batch processes. A new error compensation strategy is proposed based on two-dimensional JITL model by using MPC-ILC integrated control method. Batch axis and time axis are integrated into a comprehensive objective function, and the JITL model is used to solve the problem of large computation of comprehensive objective function. The proposed control algorithm is applied to a typical batch reactor, and the results show that the proposed control strategy has good control performance.
This work investigates the problem of random successive data dropout at the output side of stochastic linear systems and presents a novel successive updating scheme (SUS) based on iterative learningcontrol (ILC) to a...
详细信息
ISBN:
(纸本)9798350321050
This work investigates the problem of random successive data dropout at the output side of stochastic linear systems and presents a novel successive updating scheme (SUS) based on iterative learningcontrol (ILC) to avoid control failures due to data loss. In particular, the successively lost output data in the latest iteration is compensated via predictive information estimated successfully with the same time instant label in the previous iteration by the multi-step predictive model. Mathematical induction is used to demonstrate the convergence of the proposed ILC scheme. Lastly, a simulation example is provided to back up the theoretical analysis.
Optimal control design methods for multiple time-scale systems are a hot research topic in recent years. In this paper, a comprehensive overview of the design methods for optimal control of multiple time-scale systems...
详细信息
ISBN:
(纸本)9798350321050
Optimal control design methods for multiple time-scale systems are a hot research topic in recent years. In this paper, a comprehensive overview of the design methods for optimal control of multiple time-scale systems is presented. Firstly, the mathematical model of the optimal control problem of multiple time-scale systems is given, and the key difficulties of the related research are analysed. Secondly, the design methods for optimal control of multiple time-scale systems based on the model and reinforcement learning (RL) methods are given respectively. Thirdly, the performance analysis and practical application of the multi-time scale system are analyzed. Finally, the current problems in solving the optimization of multiple time-scale systems are analysed, and the research directions of optimal control of multiple time-scale systems are prospected.
data-drivencontrol, which embraces artificial intelligence, machine learning, and experience-based inferencing architectures, has gained significant interest for its ability to provide robust optimization in model-fr...
详细信息
ISBN:
(纸本)9798350355376;9798350355369
data-drivencontrol, which embraces artificial intelligence, machine learning, and experience-based inferencing architectures, has gained significant interest for its ability to provide robust optimization in model-free, nonlinear, and time-varying paradigms. Traditional systems, such as the haptic paddle, used to communicate system dynamics principles in undergraduate curricula, have yet to be adapted to the memory and processing requirements of data-drivencontrol. In this work, we present a modular, open-source 3D printable friction-driven haptic paddle design, building on the designs proposed by the community, using commercial components and simple microelectronic packaging, to enable robust data-drivencontrol for integration in undergraduate education. We make use of the RP2040 microcontroller, a small light-weight logic platform capable of fast online computation and robust memory storage for onboard data-drivencontrol. To validate our design, we first develop an experimental model of the physical dynamics that shows that our 3D printed friction drive is comparable with friction driven paddles and capstan-cable driven paddles. Further, we demonstrate the utility of our design in explicating data-drivencontrol by presenting the development of basic machine learning and reinforcement learning architectures for online, model-free robust control in the presence of time-variable plant dynamics in a trajectory tracking task that is well suited for implementation in undergraduate and introductory graduate system dynamics and controls curricula.
暂无评论