International Journal of Control Automation and Systems最新文献

Estimation and analysis of energy consumption for machine tool is the basis of energy efficiency improvement. To improve the accuracy of ELM algorithm in CNC machine tool energy consumption prediction, a prediction method based on an improved particle swarm optimization (CAPSO) algorithm and an extreme learning machine (ELM) is proposed. The contribution of the algorithm includes the following three aspects. First, sobol sequence is used to initialize the PSO population to make distribution of initial population more even in solution space. Second, the center wanders and boundary neighborhood updates strategy are used to improve the population quality and convergence rate of PSO. Then, to avoid the optimal local solution, the adaptive inertia weight is introduced to achieve the stochastic perturbation of the population. The performance of the algorithm is tested by ten benchmark function, indicating that the CAPSO ensures the search accuracy and improves the algorithm’s convergence rate. Finally, the CAPSO algorithm is used to optimize the weights and thresholds of an ELM, and the CAPSO-ELM cutting energy consumption prediction model is established. Case analysis and comparative experiments show that the stability, prediction accuracy and generalization ability of CAPSO-ELM model are better than those of other models.

This paper develops a hybrid systems approach to address the self-triggered leader-following consensus problem of nonlinear multi-agent systems. First, a distributed dynamic output feedback consensus control protocol is proposed, and a novel hybrid dynamic self-triggering mechanism (HDSTM), which can provide a pre-designable minimum triggering interval (MITI) for the adjacent events, is developed to reduce the usage of communication resources. Then, by means of internal variables, a hybrid model, including both flow and jump dynamics, is constructed to describe the closed-loop dynamics. Based on this hybrid model, Lyapunov-based consensus analysis and HDSTM design results are derived. Compared to the existing related works, the main superiority of the proposed approach is that the inter-event communication intervals can always be computed in advance and no less than the given MITI. Finally, a numerical example is provided to show the effectiveness.

This paper presents some graph-theoretic conditions for a democratic system controlled by a social network to converge to a regressive or progressive system over time. The democratic system is modeled as a finite state automaton, and a social network of agents is modeled as a directed graph. Agents are controllers making decisions to enable or disable events such that their objectives are to be met. Based on the individual decisions of agents, the final decision is made by the majority rule. Specifically, the conditions obtained imply two strategies for the groups of regressive or progressive agents to achieve their objectives: one is to prevent informed agents in other groups from influencing uninformed agents, and the other is to make at least one uninformed agent in every cycle follow the decisions of the group.

This paper presents a nonsingular sliding mode fault-tolerant control method with fixed-time convergence for a class of robot manipulators with uncertainties, external disturbances, and actuator failures. We estimate self-friction and external disturbances by designing a disturbance observer. Furthermore, based on the disturbance observer, we propose a sliding mode control method for the considered uncertain robot manipulator. Finally, the effectiveness of the proposed method is demonstrated by a numerical example.

In this article, an adaptive continuous sliding mode control (SMC) scheme is presented for the trajectory tracking problem of mechanical systems with parameter uncertainties, external disturbances and actuator faults. The hyperbolic tangent function is widely used to replace the signum function in SMC to ensure that the robust term is continuous and to reduce chattering. Since such an approach is difficult for SMC schemes with adaptive gain to induce system stability through Lyapunov functions, we reconstruct the hyperbolic tangent function by taking both the adaptive control gain and sliding variables as inputs. The designed gain dynamics do not require a priori upper bound on lumped uncertainties, including parameter uncertainties, external disturbances and actuator faults, and ensure no overestimated gains. Besides, an adaptive dual-layer super-twisting (ADLST) observer is adopted to accurately estimate unmeasurable velocities, which achieves the synthesis of an adaptive sliding mode observer and the continuous SMC method with adaptive gain. It is proven through the Lyapunov function that all closed-loop signals are ultimately bounded. Comparative simulations are conducted on a two-link rigid manipulator to demonstrate the effectiveness of the adopted observer and the proposed scheme.

This paper aims at the study of the general decay synchronization of state and spatial diffusion coupled delayed reaction-diffusion memristive neural networks (CDRDMNNs). At first, we present the general decay synchronization concept for the considered network based on ψ-type stability, then we go deeply into the establishment of general decay synchronization criterion of state CDRDMNNs by excogitating a proper controller and utilizing an opportune Lyapunov functional, thereby gain a sufficient condition for attaining general decay synchronization of state CDRDMNNs. Subsequently, we devote ourselves to investigating the general decay synchronization of spatial diffusion CDRDMNNs and derive an adequate condition for realizing the general decay synchronization of this type of network. Next, some sufficient conditions which ensure robust general decay synchronization of state CDRDMNNs and spatial diffusion CDRDMNNs with uncertain parameters are concluded by utilizing various types of inequality techniques. Ultimately, the efficacy of the acquired general decay synchronization results are certified via numerical simulations of two examples.

Bilateral teleoperation with force feedback allows the operators to apply their skills to accomplish challenging tasks safely. Most teleoperation bilateral systems are designed for single interaction scenarios and low-frequency force feedback, which limits their overall performance in complex interaction tasks. Furthermore, the use of passive controllers to ensure system stability can lead to further reductions in force transparency. This paper addresses the hybrid force feedback problem in complex interaction tasks with multiple stages, aiming at enhancing the practicality and robustness of teleoperation systems for complex interaction tasks, as well as reducing the force distortion caused by passive controllers. Firstly, an event-triggered hybrid force feedback architecture is proposed. Within this architecture, we introduce a two-channel fully transparent method with an explicit force controller (FT2-EFC), to enable model-free force tracking during both free motion and vibration contact stages. Besides, an adaptive impedance matching (AIM) algorithm is proposed to improve the physical interaction characteristics in the contact transient stage. Secondly, we present the operator passivity reference dual boundary energy tank (OPRDB-ET) method, which not only ensures the delay stability of the force architecture but also minimizes force distortion resulting from passive damping injection. Finally, experiments demonstrated that the proposed methods ensure the accurate tracking ability of hybrid forces in all stages of complicated interaction tasks and the slight force distortion under communication delay.

Prediction of a driver’s braking intention enables the advanced driver assistance system (ADAS) to intervene in the braking system as early as possible, which may shorten braking distance and improve driving safety. This paper proposes a novel deep learning model called LSTM-CNN-Attention that combines a long short-term memory (LSTM) neural network, convolutional neural network (CNN), and Attention mechanism for extracting spatiotemporal features of multi-sensor data to improve prediction accuracy. The proposed model inherits both temporal and spatial feature extraction abilities from LSTM and CNN. The LSTM-CNN-Attention model has a parallel architecture, which enhances the feature extraction ability of the model for multi-sensor time series data and improves the prediction accuracy of the driver’s braking intention before the braking action. Furthermore, a driving simulator is set up to sample driving data for training and evaluating the proposed method. According to the results of the experiment, the model obtains up to 3.16% higher accuracy than the baseline models such as LSTM, CNN, and bidirectional LTSM (Bi-LSTM). Additionally, the influence of sliding window size and prediction horizon on the performance of the method is investigated. A method of tuning hyperparameters using the genetic algorithm is presented. The results demonstrate that the prediction accuracy increases by about 2% after being optimized by GA.

This paper investigates the exponential synchronization of stochastic time-delayed memristor-based neural networks(MBNNs) with using pinning impulsive control. Different from the traditional impulsive control schemes, a hybrid pinning impulsive control scheme is presented, and some sufficient conditions for exponential synchronization of system are established. Moreover, on the basis of the obtained results, the problem of delayed impulsive stabilization of stochastic time-delayed MBNN is studied. At last, an example is provided to demonstrate the validity of the obtained results.

This paper investigates the adaptive fuzzy fault-tolerant control (FTC) problem for a class of nonlinear underactuated wheeled mobile robot (UWMR) systems with intermittent actuator faults. Considering the UWMR system contains unknown nonlinear dynamics, the fuzzy logic systems (FLSs) are utilized to approximate the unknown nonlinear dynamics. Based on the fault-tolerant control theory and introducing the prescribed performance functions into the backstepping control process, a fuzzy prescribed performance adaptive FTC scheme is developed. It is proved that the proposed fuzzy adaptive prescribed performance FTC approach can guarantee that all the signals of the controlled UWMR system are bounded, and the tracking errors are kept within the prescribed boundaries even in the presence of the intermittent actuator faults. Finally, the computer simulation and comparison results are given to validate the effectiveness of the proposed FTC approach.