Time delays exist widely in real systems, and time-delayed interactions can result in abundant dynamic behaviors and functions in dynamic networks. Inferring the time delays and interactions is challenging due to syst...Time delays exist widely in real systems, and time-delayed interactions can result in abundant dynamic behaviors and functions in dynamic networks. Inferring the time delays and interactions is challenging due to systematic nonlinearity,noises, a lack of information, and so on. Recently, Shi et al. proposed a random state variable resetting method to detect the interactions in a continuous-time dynamic network. By arbitrarily resetting the state variable of a driving node, the equivalent coupling functions of the driving node to any response node in the network can be reconstructed. In this paper,we introduce this method in time-delayed dynamic networks. To infer actual time delays, the nearest neighbor correlation(NNC) function for a given time delay is defined. The significant increments of NNC originate from the delayed effect.Based on the increments, the time delays can be reconstructed and the reconstruction errors depend on the sampling time interval. After time delays are accurately identified, the equivalent coupling functions can also be reconstructed. The numerical results have fully verified the validity of the theoretical analysis.展开更多
A composite anti-disturbance predictive control strategy employing a Multi-dimensional Taylor Network(MTN)is presented for unmanned systems subject to time-delay and multi-source disturbances.First,the multi-source di...A composite anti-disturbance predictive control strategy employing a Multi-dimensional Taylor Network(MTN)is presented for unmanned systems subject to time-delay and multi-source disturbances.First,the multi-source disturbances are addressed according to their specific characteristics as follows:(A)an MTN data-driven model,which is used for uncertainty description,is designed accompanied with the mechanism model to represent the unmanned systems;(B)an adaptive MTN filter is used to remove the influence of the internal disturbance;(C)an MTN disturbance observer is constructed to estimate and compensate for the influence of the external disturbance;(D)the Extended Kalman Filter(EKF)algorithm is utilized as the learning mechanism for MTNs.Second,to address the time-delay effect,a recursiveτstep-ahead MTN predictive model is designed utilizing recursive technology,aiming to mitigate the impact of time-delay,and the EKF algorithm is employed as its learning mechanism.Then,the MTN predictive control law is designed based on the quadratic performance index.By implementing the proposed composite controller to unmanned systems,simultaneous feedforward compensation and feedback suppression to the multi-source disturbances are conducted.Finally,the convergence of the MTN and the stability of the closed-loop system are established utilizing the Lyapunov theorem.Two exemplary applications of unmanned systems involving unmanned vehicle and rigid spacecraft are presented to validate the effectiveness of the proposed approach.展开更多
Mason Reset(MR),a groundbreaking invention by Clesson E.Mason in 1930 that later became a part of“the universal approach to process control instrumentation”,is revisited in this paper and is shown to consists of thr...Mason Reset(MR),a groundbreaking invention by Clesson E.Mason in 1930 that later became a part of“the universal approach to process control instrumentation”,is revisited in this paper and is shown to consists of three actions:fast(errorcorrection),medium(negative feedback for expanded proportional band)and slow(reset for zero steady-state error).The focus of the paper is on the reset action,generated from a positive feedback loop,and its underlying principles with profound implications to our understanding and practice of automatic control,both basic and advanced.For example,we note that reset control and integral control,contrary to common belief,differ fundamentally in design principle and in practicality.Such difference comes to a head in the event of integrator windup:while reset windup is a problem of actuator saturation,the integrator windup is a runaway situation due to controller instability.In fact,there is no advantage gained in replacing MR with an integrator.In other words,one should not integrate the error directly as in standard PID,since doing so makes the closed-loop system internally unstable.With MR-based control formulated in this paper,there is no such threat of instability and,therefore,no need for any anti-windup mechanisms.Furthermore,the integral control is made scalable in this framework as a tradeoff between the steady-state accuracy and the controller stability.This leads to a novel MR-based control design,scalable in gain and in time to accommodate various process characteristics and design specifications.Simple in construction and transparent in principle,this MR-based control,as a basic framework of design,is readily deployable in scale.展开更多
Domain randomization is a widely adopted technique in deep reinforcement learning(DRL)to improve agent generalization by exposing policies to diverse environmental conditions.This paper investigates the impact of diff...Domain randomization is a widely adopted technique in deep reinforcement learning(DRL)to improve agent generalization by exposing policies to diverse environmental conditions.This paper investigates the impact of different reset strategies,normal,non-randomized,and randomized,on agent performance using the Deep Deterministic Policy Gradient(DDPG)and Twin Delayed DDPG(TD3)algorithms within the CarRacing-v2 environment.Two experimental setups were conducted:an extended training regime with DDPG for 1000 steps per episode across 1000 episodes,and a fast execution setup comparing DDPG and TD3 for 30 episodes with 50 steps per episode under constrained computational resources.A step-based reward scaling mechanism was applied under the randomized reset condition to promote broader state exploration.Experimental results showthat randomized resets significantly enhance learning efficiency and generalization,with DDPG demonstrating superior performance across all reset strategies.In particular,DDPG combined with randomized resets achieves the highest smoothed rewards(reaching approximately 15),best stability,and fastest convergence.These differences are statistically significant,as confirmed by t-tests:DDPG outperforms TD3 under randomized(t=−101.91,p<0.0001),normal(t=−21.59,p<0.0001),and non-randomized(t=−62.46,p<0.0001)reset conditions.The findings underscore the critical role of reset strategy and reward shaping in enhancing the robustness and adaptability of DRL agents in continuous control tasks,particularly in environments where computational efficiency and training stability are crucial.展开更多
This study investigates the nonlinear resonance responses of suspended cables subjected to multi-frequency excitations and time-delayed feedback.Two specific combinations and simultaneous resonances are selected for d...This study investigates the nonlinear resonance responses of suspended cables subjected to multi-frequency excitations and time-delayed feedback.Two specific combinations and simultaneous resonances are selected for detailed examination.Initially,utilizing Hamilton’s variational principle,a nonlinear vibration control model of suspended cables under multi-frequency excitations and longitudinal time-delayed velocity feedback is developed,and the Galerkin method is employed to obtain the discrete model.Subsequently,focusing solely on single-mode discretization,analytical solutions for the two simultaneous resonances are derived using the method of multiple scales.The frequency response equations are derived,and the stability analysis is presented for two simultaneous resonance cases.The results demonstrate that suspended cables exhibit complex nonlinearity under multi-frequency excitations.Multiple solutions under multi-frequency excitation can be distinguished through the frequency–response and the detuning-phase curves.By adjusting the control gain and time delay,the resonance range,response amplitude,and phase of suspended cables can be modified.展开更多
In the industrial roller kiln,the time-delay characteristic in heat transfer causes the temperature field to be affected by both the current and historical temperature states.It presents a poor control performance and...In the industrial roller kiln,the time-delay characteristic in heat transfer causes the temperature field to be affected by both the current and historical temperature states.It presents a poor control performance and brings a significant challenge to the process precise control.Considering high complexity of precise modeling,a data-driven time-delay optimal control method for temperature field of roller kiln is proposed based on a large amount of process data.First,the control challenges and problem description brought by time-delay are demonstrated,where the cost function for the time-delay partial differential equation system is constructed.To obtain the optimal control law,the policy iteration in adaptive dynamic programming is adopted to design the time-delay temperature field controller,and neural network is used for the critic network in policy iteration to approximate the optimal time-delay cost function.The closed-loop system stability is proved by designing the Lyapunov function which contains the time-delay information.Finally,through establishing the time-delay temperature field model for roller kiln,the effectiveness and convergence of the proposed method is verified and proved.展开更多
To ensure the safety of power energy transmission channel and mitigate the harm caused by galloping of iced transmission lines,the axial time-delay velocity feedback strategy is adopted to suppress the galloping.The p...To ensure the safety of power energy transmission channel and mitigate the harm caused by galloping of iced transmission lines,the axial time-delay velocity feedback strategy is adopted to suppress the galloping.The par-tial differential equation of galloping with axial time-delay velocity feedback strategy is established based on the variational principle for Hamiltonian.Then,the partial differential equation of galloping is transformed into or-dinary differential equation based on normalization and the Galerkin method.The primary amplitude-frequency response equation,the first-order steady-state approximate solution,and the harmonic amplitude-frequency re-sponse equation are derived by the multiscale method.The impact of different parameters such as time-delay value,control coefficient,and amplitude of external excitation on the galloping response are analyzed.The am-plitude under the primary resonance exhibits periodicity as time-delay value varies.The amplitude diminishes with increased control coefficient and increases with external excitation.Comprehensive consideration of vari-ous influences of parameters on vibration characteristics is crucial when employing the axial time-delay velocity feedback strategy to suppress galloping.Therefore,to achieve the best vibration suppression effect,it is crucial to adjust the time-delay parameter for modifying the range and amplitude of the resonance zone.The conclusions obtained by this study are expected to advance the refinement of active control techniques for iced transmission lines,and may provide valuable insights for practical engineering applications.展开更多
This paper discusses the design of event-triggered output-feedback controller for a class of nonlinear time-delay systems with multiple uncertainties. In sharp contrast to previous works, the considered systems posses...This paper discusses the design of event-triggered output-feedback controller for a class of nonlinear time-delay systems with multiple uncertainties. In sharp contrast to previous works, the considered systems possess two important characteristics: (i) The uncertain nonlinear terms meet the linearly unmeasurable-states dependent growth with the growth rate being an unknown function of the input and output. (ii) There exist input matching uncertainty and unknown measurement sensitivity. By introducing a single dynamic gain and employing a cleverly devised event-triggering mechanism (ETM), we design a new gain-based event-triggered output-feedback controller, which globally regulates all states of the considered systems and maintains global boundedness of the closed-loop system. Furthermore, the estimation of input matching uncertainty achieves convergence towards its actual value, and Zeno behavior does not happen. Two simulation examples including a practical one show that the proposed approach is effective.展开更多
Focusing on the rehabilitation training of hemiplegia patients,this paper proposes a gait-planning strategy based on a central pattern generator and an adaptive time-delay control scheme that utilizes recursive termin...Focusing on the rehabilitation training of hemiplegia patients,this paper proposes a gait-planning strategy based on a central pattern generator and an adaptive time-delay control scheme that utilizes recursive terminal sliding mode for lower limb rehabilitation exoskeleton robots.The central pattern generator network plans a reference gait trajectory for the affected leg,synchronized with the movement of the healthy leg.The proposed adaptive time-delay control scheme possesses a model-independent property due to the mechanism of time-delay estimation,with adaptive control gains that enhance the resilience against system perturbations and a recursive terminal sliding mode control component to achieve a fast convergence rate.According to the Lyapunov stability criterion,it is proved that the gait trajectory-tracking error is uniformly ultimately bounded.Experiments are conducted on a lower limb exoskeleton experimental platform,and the experimental results demonstrate the effectiveness and superiority of the proposed strategies.展开更多
The temperature-humidity models of wood drying were developed based on Time-delay neural network and the identification structures of Time-delay neural network were given. The controlling model and the schedule model,...The temperature-humidity models of wood drying were developed based on Time-delay neural network and the identification structures of Time-delay neural network were given. The controlling model and the schedule model, which revealed the relation between controlling signal and temperature-humidity and the relation between wood moisture content and temperature-humidity of wood drying, were separately presented. The models were simulated by using the measured data of the experimental drying kiln. The numerical simulation results showed that the modeling method was feasible, and the models were effective.展开更多
A novel behavioral model using three-layer time-delay feed-forward neural networks （TDFFNN）is adopted to model radio frequency （RF）power amplifiers exhibiting memory nonlinearities. In order to extract the paramet...A novel behavioral model using three-layer time-delay feed-forward neural networks （TDFFNN）is adopted to model radio frequency （RF）power amplifiers exhibiting memory nonlinearities. In order to extract the parameters, the back- propagation algorithm is applied to train the proposed neural networks. The proposed model is verified by the typical odd- order-only memory polynomial model in simulation, and the performance is compared with different numbers of taped delay lines（TDLs） and perceptrons of the hidden layer. For validating the TDFFNN model by experiments, a digital test bench is set up to collect input and output data of power amplifiers at a 60 × 10^6 sample/s sampling rate. The 3.75 MHz 16-QAM signal generated in the vector signal generator（VSG） is chosen as the input signal, when measuring the dynamic AM/AM and AM/PM characteristics of power amplifiers. By comparisons and analyses, the presented model provides a good performance in convergence, accuracy and efficiency, which is approved by simulation results and experimental results in the time domain and frequency domain.展开更多
文摘Time delays exist widely in real systems, and time-delayed interactions can result in abundant dynamic behaviors and functions in dynamic networks. Inferring the time delays and interactions is challenging due to systematic nonlinearity,noises, a lack of information, and so on. Recently, Shi et al. proposed a random state variable resetting method to detect the interactions in a continuous-time dynamic network. By arbitrarily resetting the state variable of a driving node, the equivalent coupling functions of the driving node to any response node in the network can be reconstructed. In this paper,we introduce this method in time-delayed dynamic networks. To infer actual time delays, the nearest neighbor correlation(NNC) function for a given time delay is defined. The significant increments of NNC originate from the delayed effect.Based on the increments, the time delays can be reconstructed and the reconstruction errors depend on the sampling time interval. After time delays are accurately identified, the equivalent coupling functions can also be reconstructed. The numerical results have fully verified the validity of the theoretical analysis.
基金co-supported by the National Key R&D Program of China(No.2023YFB4704400)the Zhejiang Provincial Natural Science Foundation of China(No.LQ24F030012)the National Natural Science Foundation of China General Project(No.62373033)。
文摘A composite anti-disturbance predictive control strategy employing a Multi-dimensional Taylor Network(MTN)is presented for unmanned systems subject to time-delay and multi-source disturbances.First,the multi-source disturbances are addressed according to their specific characteristics as follows:(A)an MTN data-driven model,which is used for uncertainty description,is designed accompanied with the mechanism model to represent the unmanned systems;(B)an adaptive MTN filter is used to remove the influence of the internal disturbance;(C)an MTN disturbance observer is constructed to estimate and compensate for the influence of the external disturbance;(D)the Extended Kalman Filter(EKF)algorithm is utilized as the learning mechanism for MTNs.Second,to address the time-delay effect,a recursiveτstep-ahead MTN predictive model is designed utilizing recursive technology,aiming to mitigate the impact of time-delay,and the EKF algorithm is employed as its learning mechanism.Then,the MTN predictive control law is designed based on the quadratic performance index.By implementing the proposed composite controller to unmanned systems,simultaneous feedforward compensation and feedback suppression to the multi-source disturbances are conducted.Finally,the convergence of the MTN and the stability of the closed-loop system are established utilizing the Lyapunov theorem.Two exemplary applications of unmanned systems involving unmanned vehicle and rigid spacecraft are presented to validate the effectiveness of the proposed approach.
文摘Mason Reset(MR),a groundbreaking invention by Clesson E.Mason in 1930 that later became a part of“the universal approach to process control instrumentation”,is revisited in this paper and is shown to consists of three actions:fast(errorcorrection),medium(negative feedback for expanded proportional band)and slow(reset for zero steady-state error).The focus of the paper is on the reset action,generated from a positive feedback loop,and its underlying principles with profound implications to our understanding and practice of automatic control,both basic and advanced.For example,we note that reset control and integral control,contrary to common belief,differ fundamentally in design principle and in practicality.Such difference comes to a head in the event of integrator windup:while reset windup is a problem of actuator saturation,the integrator windup is a runaway situation due to controller instability.In fact,there is no advantage gained in replacing MR with an integrator.In other words,one should not integrate the error directly as in standard PID,since doing so makes the closed-loop system internally unstable.With MR-based control formulated in this paper,there is no such threat of instability and,therefore,no need for any anti-windup mechanisms.Furthermore,the integral control is made scalable in this framework as a tradeoff between the steady-state accuracy and the controller stability.This leads to a novel MR-based control design,scalable in gain and in time to accommodate various process characteristics and design specifications.Simple in construction and transparent in principle,this MR-based control,as a basic framework of design,is readily deployable in scale.
基金supported by the Deputyship for Research&Innovation,Ministry of Education in Saudi Arabia(Project No.MoE-IF-UJ-R2-22-04220773-1).
文摘Domain randomization is a widely adopted technique in deep reinforcement learning(DRL)to improve agent generalization by exposing policies to diverse environmental conditions.This paper investigates the impact of different reset strategies,normal,non-randomized,and randomized,on agent performance using the Deep Deterministic Policy Gradient(DDPG)and Twin Delayed DDPG(TD3)algorithms within the CarRacing-v2 environment.Two experimental setups were conducted:an extended training regime with DDPG for 1000 steps per episode across 1000 episodes,and a fast execution setup comparing DDPG and TD3 for 30 episodes with 50 steps per episode under constrained computational resources.A step-based reward scaling mechanism was applied under the randomized reset condition to promote broader state exploration.Experimental results showthat randomized resets significantly enhance learning efficiency and generalization,with DDPG demonstrating superior performance across all reset strategies.In particular,DDPG combined with randomized resets achieves the highest smoothed rewards(reaching approximately 15),best stability,and fastest convergence.These differences are statistically significant,as confirmed by t-tests:DDPG outperforms TD3 under randomized(t=−101.91,p<0.0001),normal(t=−21.59,p<0.0001),and non-randomized(t=−62.46,p<0.0001)reset conditions.The findings underscore the critical role of reset strategy and reward shaping in enhancing the robustness and adaptability of DRL agents in continuous control tasks,particularly in environments where computational efficiency and training stability are crucial.
基金supported in part by the National Natural Science Foundation of China(Grant No.12432001)Natural Science Foundation of Hunan Province(Grant Nos.2023JJ60527,2023JJ30152,and 2023JJ30259)the Natural Science Foundation of Changsha(KQ2202133).
文摘This study investigates the nonlinear resonance responses of suspended cables subjected to multi-frequency excitations and time-delayed feedback.Two specific combinations and simultaneous resonances are selected for detailed examination.Initially,utilizing Hamilton’s variational principle,a nonlinear vibration control model of suspended cables under multi-frequency excitations and longitudinal time-delayed velocity feedback is developed,and the Galerkin method is employed to obtain the discrete model.Subsequently,focusing solely on single-mode discretization,analytical solutions for the two simultaneous resonances are derived using the method of multiple scales.The frequency response equations are derived,and the stability analysis is presented for two simultaneous resonance cases.The results demonstrate that suspended cables exhibit complex nonlinearity under multi-frequency excitations.Multiple solutions under multi-frequency excitation can be distinguished through the frequency–response and the detuning-phase curves.By adjusting the control gain and time delay,the resonance range,response amplitude,and phase of suspended cables can be modified.
基金supported in part by the Key Program of National Natural Science Foundation of China(62033014)the Application Projects of Integrated Standardization and New Paradigm for Intelligent Manufacturing from the Ministry of Industry and Information Technology of China in 2016the Fundamental Research Funds for the Central Universities of Central South University(2021zzts0700).
文摘In the industrial roller kiln,the time-delay characteristic in heat transfer causes the temperature field to be affected by both the current and historical temperature states.It presents a poor control performance and brings a significant challenge to the process precise control.Considering high complexity of precise modeling,a data-driven time-delay optimal control method for temperature field of roller kiln is proposed based on a large amount of process data.First,the control challenges and problem description brought by time-delay are demonstrated,where the cost function for the time-delay partial differential equation system is constructed.To obtain the optimal control law,the policy iteration in adaptive dynamic programming is adopted to design the time-delay temperature field controller,and neural network is used for the critic network in policy iteration to approximate the optimal time-delay cost function.The closed-loop system stability is proved by designing the Lyapunov function which contains the time-delay information.Finally,through establishing the time-delay temperature field model for roller kiln,the effectiveness and convergence of the proposed method is verified and proved.
基金supported by the National Natural Science Foundation of China(Grant No.51507106)China Postdoctoral Science Foundation(Grant No.2021M702371)。
文摘To ensure the safety of power energy transmission channel and mitigate the harm caused by galloping of iced transmission lines,the axial time-delay velocity feedback strategy is adopted to suppress the galloping.The par-tial differential equation of galloping with axial time-delay velocity feedback strategy is established based on the variational principle for Hamiltonian.Then,the partial differential equation of galloping is transformed into or-dinary differential equation based on normalization and the Galerkin method.The primary amplitude-frequency response equation,the first-order steady-state approximate solution,and the harmonic amplitude-frequency re-sponse equation are derived by the multiscale method.The impact of different parameters such as time-delay value,control coefficient,and amplitude of external excitation on the galloping response are analyzed.The am-plitude under the primary resonance exhibits periodicity as time-delay value varies.The amplitude diminishes with increased control coefficient and increases with external excitation.Comprehensive consideration of vari-ous influences of parameters on vibration characteristics is crucial when employing the axial time-delay velocity feedback strategy to suppress galloping.Therefore,to achieve the best vibration suppression effect,it is crucial to adjust the time-delay parameter for modifying the range and amplitude of the resonance zone.The conclusions obtained by this study are expected to advance the refinement of active control techniques for iced transmission lines,and may provide valuable insights for practical engineering applications.
基金supported by the fund of Beijing Municipal Commission of Education(KM202210017001 and 22019821001)the Natural Science Foundation of Henan Province(222300420253).
文摘This paper discusses the design of event-triggered output-feedback controller for a class of nonlinear time-delay systems with multiple uncertainties. In sharp contrast to previous works, the considered systems possess two important characteristics: (i) The uncertain nonlinear terms meet the linearly unmeasurable-states dependent growth with the growth rate being an unknown function of the input and output. (ii) There exist input matching uncertainty and unknown measurement sensitivity. By introducing a single dynamic gain and employing a cleverly devised event-triggering mechanism (ETM), we design a new gain-based event-triggered output-feedback controller, which globally regulates all states of the considered systems and maintains global boundedness of the closed-loop system. Furthermore, the estimation of input matching uncertainty achieves convergence towards its actual value, and Zeno behavior does not happen. Two simulation examples including a practical one show that the proposed approach is effective.
基金supported by the National Natural Science Foundation of China(62473337,62003305)the Key Research and Development Program of Zhejiang Province(2024C03040,2022C03029)the funding of Leading Innovative and Entrepreneur Team Introduction Program of Zhejiang Province(2023R01006).
文摘Focusing on the rehabilitation training of hemiplegia patients,this paper proposes a gait-planning strategy based on a central pattern generator and an adaptive time-delay control scheme that utilizes recursive terminal sliding mode for lower limb rehabilitation exoskeleton robots.The central pattern generator network plans a reference gait trajectory for the affected leg,synchronized with the movement of the healthy leg.The proposed adaptive time-delay control scheme possesses a model-independent property due to the mechanism of time-delay estimation,with adaptive control gains that enhance the resilience against system perturbations and a recursive terminal sliding mode control component to achieve a fast convergence rate.According to the Lyapunov stability criterion,it is proved that the gait trajectory-tracking error is uniformly ultimately bounded.Experiments are conducted on a lower limb exoskeleton experimental platform,and the experimental results demonstrate the effectiveness and superiority of the proposed strategies.
基金This study was supported by the Key Program of Ministry of Education of China (01066)
文摘The temperature-humidity models of wood drying were developed based on Time-delay neural network and the identification structures of Time-delay neural network were given. The controlling model and the schedule model, which revealed the relation between controlling signal and temperature-humidity and the relation between wood moisture content and temperature-humidity of wood drying, were separately presented. The models were simulated by using the measured data of the experimental drying kiln. The numerical simulation results showed that the modeling method was feasible, and the models were effective.
基金The National Natural Science Foundation of China(No.60621002)the National High Technology Research and Development Pro-gram of China(863 Program)(No.2007AA01Z2B4).
文摘A novel behavioral model using three-layer time-delay feed-forward neural networks （TDFFNN）is adopted to model radio frequency （RF）power amplifiers exhibiting memory nonlinearities. In order to extract the parameters, the back- propagation algorithm is applied to train the proposed neural networks. The proposed model is verified by the typical odd- order-only memory polynomial model in simulation, and the performance is compared with different numbers of taped delay lines（TDLs） and perceptrons of the hidden layer. For validating the TDFFNN model by experiments, a digital test bench is set up to collect input and output data of power amplifiers at a 60 × 10^6 sample/s sampling rate. The 3.75 MHz 16-QAM signal generated in the vector signal generator（VSG） is chosen as the input signal, when measuring the dynamic AM/AM and AM/PM characteristics of power amplifiers. By comparisons and analyses, the presented model provides a good performance in convergence, accuracy and efficiency, which is approved by simulation results and experimental results in the time domain and frequency domain.
