Industrial time series data are usually time-varying due to multiple factors such as environmental and human disturbances. As traditional time series predicting methods are often based on offline training ignoring the...
详细信息
ISBN:
(纸本)9798350321050
Industrial time series data are usually time-varying due to multiple factors such as environmental and human disturbances. As traditional time series predicting methods are often based on offline training ignoring the changes in working conditions, the prediction results may be inaccurate. In this paper, a time series prediction model based on incremental DBSCAN and KNN with self-learning scheme is proposed to address the problem of time-varying working conditions. The proposed model uses the incremental DBSCAN to automatically identify and expand working conditions with adjusting the number of clusters automatically, and then employs the KNN model to make predictions under different working conditions. Compared with the existing methods, the proposed method is more stable and improves the prediction accuracies of the model under different working conditions.
With the development of drone technology, quadrotor transportation has become an important application direction. Most of the control designs for a quadrotor with a cable-suspended payload (QCSP) are aimed at fixed ca...
详细信息
ISBN:
(纸本)9798350321050
With the development of drone technology, quadrotor transportation has become an important application direction. Most of the control designs for a quadrotor with a cable-suspended payload (QCSP) are aimed at fixed cable lengths, but there are few controller designs for varying cable length QCSP with broad application prospects. In this paper, a controller for varying cable length QCSP is designed using parameter adaptive multi-layer sliding mode structure to control the payload swing, quadrotor position tracking and single-layer sliding mode structure to control the cable length of the suspension system. The nonlinear coupling problem in the QCSP is solved by a simple method, avoiding tedious design reasoning and parameter adjustment. Simulation experiments were conducted and compared with traditional PD controllers, proving the effectiveness of this method in suppressing load swing angles of varying cable length QCSP.
Maintaining indoor thermal comfort with heating, ventilation, and air conditioning (HVAC) systems requires a significant amount of energy, with chillers accounting for over 50% of the total. Therefore, a technical pat...
详细信息
ISBN:
(纸本)9798350321050
Maintaining indoor thermal comfort with heating, ventilation, and air conditioning (HVAC) systems requires a significant amount of energy, with chillers accounting for over 50% of the total. Therefore, a technical path based on chiller load prediction is essential to achieve low energy operation of air conditioning systems in buildings. Many machine learning methods have been widely used for load prediction, but for new buildings, there is still a lack of data to support model training for HVAC chiller load prediction. To address the above load prediction problems, this paper proposes the CEEMDAN-BiLSTM-Attention method for improving chiller load prediction accuracy. The final experimental findings are contrasted with the traditional prediction models like CNN, LSTM and hybrid models (CNN-LSTM, CNN-BiLSTM) and so on. Meanwhile, as for the case of small data samples, this paper builds two migration models CNN-LSTM-TL and CNN-BiLSTM-TL for the comparison. The results demonstrate that the CEEMDAN-BiLSTM-Attention model outperforms the other models in predicting chiller load for both buildings with sufficient data and buildings with small sample data.
This paper proposes a novel method to improve synchronization precision and reduce the influence of stochastic noises in lithography motion stages. Specifically, the multiple-channel adaption-learning-function synchro...
详细信息
ISBN:
(纸本)9798350321050
This paper proposes a novel method to improve synchronization precision and reduce the influence of stochastic noises in lithography motion stages. Specifically, the multiple-channel adaption-learning-function synchronization iterative learningcontrol (MASILC) method we give uses a multiple-channel approach based on A-type of ILC, with adaptive parameters to address the divergence situation caused by stochastic noises in different channels. We verify the superiority of the MASILC method and analyze its stability and convergence performance in theory. Furthermore, we introduce an effective adaptive method inspired by the stochastic acceleration optimization method. Simulation comparisons with analogous approaches demonstrate the effectiveness and superiority of the proposed method in various circumstances.
Aiming at the longitudinal speed control of autonomous buses, an improved Adaptive Feedforward Feedback Iterative learningcontrol (AFF-ILC) algorithm was proposed. The controller structure of this method adopts the P...
详细信息
ISBN:
(纸本)9798350321050
Aiming at the longitudinal speed control of autonomous buses, an improved Adaptive Feedforward Feedback Iterative learningcontrol (AFF-ILC) algorithm was proposed. The controller structure of this method adopts the PD-ILC control structure. At the same time, by introducing the time domain integral operator and the iterative domain differential operator, combined with the adaptively adjusted PD parameters, the feedforward and feedback learning ability of the AFF-ILC control algorithm can be preserved simultaneously. In addition, considering the bus overspeed protection and other factors, the controller design process also considers the saturation constraints of control input and controller parameters. The advantage of the proposed method is that it combines the characteristics of repeated bus operation with the characteristics of iterative learningcontrol algorithm, and the controller design process does not require accurate modeling of the system, only the input and output data can be used for controller design. A series of simulation results verify the effectiveness of the proposed method.
In the context of continuous advancements in modern industry and the escalating demands for the safety and reliability of controlsystems, timely fault diagnosis of controlsystems becomes paramount. driven by this im...
详细信息
In this paper, a novel discrete-time reduced-order extended state observer (ESO) sideslip observer-based mode-free adaptive control (ELOS-MFAC) scheme is developed for underactuated unmanned vehicles. The main contrib...
详细信息
ISBN:
(纸本)9798350321050
In this paper, a novel discrete-time reduced-order extended state observer (ESO) sideslip observer-based mode-free adaptive control (ELOS-MFAC) scheme is developed for underactuated unmanned vehicles. The main contributions are as follows: 1) the time-varying sideslip angle is exactly estimated by a reduced-order ESO, thus achieving a high-precision estimation of the sideslip angle and laying the groundwork for sideslip angle compensation;2) an ESO-based Line of Sight (ELOS) guidance law is proposed to enhance the generalis ability of the LOS guidance law in the case of unknown side slip angles;3) with estimated surge speed and heading guidance, MFAC technology is adopted in the design of speed controllers. The simulation study conclusively demonstrates the efficacy and superiority of the proposed ELOS-MFAC framework.
作者：
Fu, ShengdongYe, LingjianZhejiang Univ
Sch Control Engn Hangzhou 310027 Peoples R China NingboTech Univ
Sch Informat Sci & Engn Ningbo 315100 Peoples R China Huzhou Univ
Huzhou Key Lab Intelligent Sensing & Optimal Cont Sch Engn Huzhou 313000 Peoples R China
The implementation of a well-designed control structure is crucial for efficient and optimal functionality of solid oxide fuel cells (SOFCs). In this work, we propose a self-optimizing control (SOC) solution for a dir...
详细信息
ISBN:
(纸本)9798350321050
The implementation of a well-designed control structure is crucial for efficient and optimal functionality of solid oxide fuel cells (SOFCs). In this work, we propose a self-optimizing control (SOC) solution for a direct internal reforming SOFC, with the objective of maximizing the power generation profit after deducting carbon tax. Based on the SOC methodology, we identify optimal controlled variables (CVs) for the SOFC, which are maintained at constant setpoints to optimize the efficiency, regardless of changes in interference and uncertain parameters. Studies show that the stack temperature is the active constraint which is constrained at the upper bound. Besides, we configure the linear combination of outlet hydrogen fraction, carbon dioxide fraction and voltage as an additional CV to achieve acceptable efficiency loss. The effectiveness of the novel SOC scheme is verified through both steady state and dynamic simulations.
In this study, we focus on the adaptive learningcontrol of high-order uncertain nonlinear multiagent systems (MASs) under replay attacks. The neural-network-based adaptive learning scheme is proposed to approximate t...
详细信息
The state-of-the-art path-tracking control approaches for existing vehicle systems mostly rely on the accurate system dynamics and an initial stabilizing control policy assumption. To overcome those challenges, this p...
详细信息
The state-of-the-art path-tracking control approaches for existing vehicle systems mostly rely on the accurate system dynamics and an initial stabilizing control policy assumption. To overcome those challenges, this paper presents an adaptive learning-based path-tracking control algorithm designed specifically for a completely unknown vehicle system without an initial stabilizing control strategy assumption. Firstly, a new variable is introduced to construct a new matrix thereby affording greater flexibility in selecting controller gains. Subsequently, leveraging this new matrix, a new policy iteration algorithm and an imitation-based policy iteration algorithm are concurrently proposed to achieve model-free learning path-tracking control in an optimal manner. In addition, an advanced data-driven switching policy iteration learning algorithm is developed to inherit the advantages of existing mainstream learning algorithms. When compared to several existing learning algorithms, the proposed algorithm not only eliminates the need for an initial stabilizing policy assumption but also exhibits faster convergence and reduced computational complexity. Finally, numerical simulations and comparisons are conducted to demonstrate the validity of the theoretical analysis. ieee
暂无评论