Abstract:To solve the problem of scheduling for the minimum completion time of tasks in a flexible manufacturing system,this paper proposes a scheduling algorithm based on timed Petri nets(TdPN)and cost function.Firstly,by analyzing the existing TdPN-based scheduling algorithms,we introduce a novel cost function taking into account the transition firing vectors from the current marking and subsequent markings to the target one.Then,through traversing the cost functions of markings in the partial reachable graph,the firing of next transition is selected.Meanwhile,the employment of backtracking method prevents deadlock markings and those markings that do not meet the system specifications,thereby obtaining logical transition sequences of the TdPN system.By transforming the logical transition sequences into timed ones,the minimum time transition sequences are computed,and then the scheduling scheme for the minimum completion time of the system can be obtained.Finally,the feasibility and effectiveness of the approach presented in this paper are validated through practical instances.

Abstract:Considering a class of nonlinear systems under output error constraints and external disturbances,a filtered backstepping control strategy is proposed by combining with a fixed-time disturbance observer.Firstly,the prescribed performance function is designed to solve the problem of output error constraint.Secondly,a fixed time filter and a fixed-time disturbance observer are designed to respectively tackle the"explosion of complexity"problem and the external disturbances problem,and to ensure that the disturbance estimation error converges to zero within a fixed time.The Lyapunov theory is utilized to analyze the fixed time stability of the closed-loop system,at the same time,the output error meets the constraint requirements in the preset time.Finally,comparative simulation is used to verify the effectiveness of the control strategy.

Abstract:Conventional PI control is difficult to ensure stable operation for the complex ore feeding process.Firstly,this article analyzes the complexity mechanism characteristics of the ore feeding process,and establishes a hybrid dynamic model consisting of a low order linear model and an unknown high-order nonlinear dynamic system.Secondly,an intelligent PI control method is proposed,which is composed of signal compensation method,rule inference control and switching mechanism.This method utilizes a one-step optimal control law to design a signal compensator,which weakens the influence of nonlinear disturbances and transmission delay on the ore feeding process,and combines rule inference to improve the closed-loop control effect.Regarding switching problem of the ore feeding machines,an intelligent switching mechanism has been designed to ensure the reasonable allocation of five ore feeding machines.Finally,the proposed method is applied to complex ore feeding process,the actual application results show that the proposed method can control the ore feeding amount and frequency within the target value range required by the process.

Abstract:In this paper,a consensus problem is studied for a group of linear agents with input constraints and state constraints.To address the problem of frequent communication between agents and regular update of information between decentralized controllers,the dynamic event-triggered control mechanism(DETC)and the dual-mode model predictive control principles(MPC)are integrated to formulate the distributed consensus control protocols,i.e.DETC-MPC.Under this control strategy,the agents can meet the constraints on system states and inputs when they are steered to reach consensus with an exponential convergence rate.Both of the feasibility of local MPC optimization and the stability of resultant closed-loop system are guaranteed.Meanwhile,a minimum inter-event time is ensured between any two consecutive triggering instants and thus no agent exhibits Zeno behavior.To avoid continuous verification of event-triggering condition,a non-periodic and non-persistent method is designed to check triggering conditions.The numerical simulation shows the effectiveness of the proposed DETC-MPC.

Abstract:Model predictive control(MPC)has been widely applied in modern control due to its advantages in handling multivariable and constrained systems.However,the uncertainties and input delays of the systems pose significant chal-lenges to the design and implementation of MPC.This paper proposes an adaptive Tube-MPC algorithm for linear systems with parameter uncertainties and input delays.Firstly,the original system is transformed into an augmented system using the state variable extension to address the delay problem.An adaptive updating law with time-varying update rates is de-signed for parameter estimation,ensuring that the estimation error remains bounded.Secondly,the MPC controller uses ellipsoidal sets to parameterize the state Tube to capture the state trajectories.The offline and online optimization problems are converted into semi-definite programming(SDP)problems,simplifying computational complexity and achieving robust control of the system.Finally,two sets of simulation examples are used to verify the effectiveness of the proposed adaptive Tube-MPC algorithm.

Abstract:Combining the direct torque control and the model predictive control,the model predictive torque control for permanent magnet synchronous motor has the advantages of simple control and high dynamic response.However,there are two problems of voltage vector combination failure and complex duty-cycle calculation when multiple voltage vectors are applied during the control cycle to eliminate steady-state errors.This paper proposes a voltage vector duty cycle optimization strategy for direct model predictive control of permanent magnet synchronous motor,which improves the effectiveness of the operation time of dual voltage vector.Based on the principle of deadbeat control,the relationship between magnetic flux vector and voltage vector is derived.The control target is normalized to the reference voltage vector,and thus solving the problem of tuning weighting factor.Furthermore,the optimal voltage vector combination can be directly obtained by dividing the sub-regions of the sector.Finally,a sum and a ratio constraint conditions are proposed for the action time of voltage vectors based on the principle of vector synthesis.Therefore,the rationality of each voltage vector and the effectiveness of synthesized voltage vector are improved.The experimental results indicate that the proposed method has better steady-state performance and good dynamic performance than the traditional methods.

Abstract:This paper investigates the trajectory tracking problem of quadcopter UAV under external disturbances and proposes an iterative learning control method based on MPC.Firstly,design the MPC controller as the feedback controller of the quadcopter UAV system,and introduce prediction error into the MPC feedback loop for feedback correction.On this basis,the iterative learning controller is designed as the feedforward controller of the quadcopter UAV system.Employing hyperbolic tangent function optimizes the learning gain of the iterative learning controller,which can reduce the number of iterations and improve trajectory tracking accuracy.The composite control of MPC feedback and iterative feedforward can effectively overcome the impact of external disturbances and achieve high-precision trajectory tracking of UAV.Finally,the feasibility and effectiveness of the proposed method are verified through the simulation.By the experimental comparison,the method designed in this paper has higher tracking accuracy and strong robustness.

Abstract:The stochastic uncertainty of wind speed presents a great challenge for achieving stable power control in wind energy conversion system(WECS).Due to the excellence in handling the uncertainties based on probabilistic descriptions,stochastic model predictive control(SMPC)has been widely applied in the power control of WECS.However,the fluctua-tion characteristic of wind speed can lead to frequent changes in the operating points of WECS,and the traditional SMPC can only ensure the feasibility and stability of WECS at one single predesigned operating point.To address the above issue,a stable SMPC strategy for uncertain WECS is proposed in this paper to ensure the feasibility and stability of WECS under changing operating points over the whole operating regions.A Luenberger observer is employed to estimate model-plant mismatch introduced by linearization process,while a tube-based control framework is deployed to cope with stochastic wind speed disturbance.The feasibility of SMPC is ensured by incorporating artificial steady targets as optimization vari-ables,while the stability of WECS is guaranteed by modifying the cost function and extending the terminal constraint.The effectiveness of the proposed strategy is validated through simulations and experiments by fatigue,aerodynamics,structures and turbulence(FAST)under different scenarios.

Abstract:In this paper,a multi-finger hybrid robotic arm control system and its grasping control method are designed.The rigid-flexible hybrid structure enhances both the accuracy of control and the safety of object interaction during grasping tasks.Kinematic modeling of the rigid and flexible components of the robotic arm is performed using screw theory and the piecewise constant curvature method,respectively.An integrated model of the rigid-flexible hybrid robotic arm,based on the Jacobian matrix,is then derived.To mitigate the impact of model inaccuracies on system performance,a novel approach is proposed that combines data-driven techniques with model predictive control.This method replaces portions of the imprecise system model with historical data and constructs the current state and control inputs through a linear combination of this data.The effectiveness of this approach in controlling the posture of the flexible gripper is demonstrated through favorable trajectory tracking results in simulations.Building upon this,an accurate grasping and enveloping grasping method based on the posture of the grasped object is developed,and the validity of the method is confirmed through ArUco cube grasping experiments.

Abstract:Fuel cell stack(FCS)is one of the primary energy sources for the application of new energy vehicles,and the equivalent fuel economy is a focal point of concern for many researchers.This article proposes an adaptive implementation control strategy to achieve optimal equivalent fuel economy by allocating power demand between the fuel cell and the power battery.In order to incorporate the influence of temperature into the control strategy,an FCS model with a thermal model and a power battery model considering temperature are established.Bayesian inference is employed to analyze and predict future power demand.Based on the FCS model,battery model,and predicted power demand,a real-time control strategy is designed,and optimization is conducted within the Pontryagin minimum principle.The proposed control strategy is validated through simulations and hardware-in-the-loop(Hil)experiments on an FCS with power of 40 kW.Comparison with rule-based control strategy and loss minimum strategy indicates that,considering the influence of temperature,the proposed control strategy effectively reduces equivalent fuel consumption by 4％.

Abstract:In the grid-connected system of a half-bridge converter series Y-connection microgrid(HCSY-MG),fluc-tuations in renewable energy power will introduce DC and fundamental frequency fluctuation components in the grid-connected current.To address this specific issue in the HCSY-MG system,a grid-connected current model predictive control strategy based on adaptive compensation of fluctuation components is proposed.Under the condition of stochas-tic fluctuations in renewable energy,the grid-connected current expression of the HCSY-MG system is derived,and its characteristics are analyzed.Based on these characteristics,the grid-connected current expression is utilized as the predic-tive model,and the artemisnin optimization algorithm is improved using chaotic and adaptive mechanisms to enhance the search speed in the rolling optimization process.This approach effectively reduces the fluctuation components in the grid-connected current while improving control performance and accuracy.Finally,simulations and experimental comparisons with existing methods verify the feasibility,effectiveness,and specificity of the proposed control strategy.

Abstract:With consideration of the complexity of the three-dimensional bin packing problem(3D-BPP)in the multi-bin semi-online scenarios,a multi-agent hierarchical reinforcement learning algorithm is proposed to improve packing efficiency and space utilization.The proposed algorithm models the problem by using a multi-agent Markov decision process(MAMDP),including three fully cooperative agents responsible for item selection,bin selection,and placement planning,respectively.A distributional learning method is introduced to enhance the stability and convergence of the algorithm.Experimental results demonstrate that the algorithm exhibits superior packing performance across various envi-ronmental configurations,significantly improving space utilization and the number of packed items.It also shows strong generalization capabilities in multi-bin and multi-item selection scenarios.Compared to traditional heuristic algorithms,the proposed method has clear advantages in dynamic decision-making and adaptive optimization,particularly demonstrating robustness when handling items with unknown size distributions.The innovation lies in the first application of a multi-agent hierarchical reinforcement learning framework to the 3D-BPP,achieving end-to-end optimization of packing decisions and providing a novel solution for complex packing scenarios.

Abstract:This paper presents a novel distributed economic model predictive control(EMPC)strategy with stability guarantees for state-coupled constrained nonlinear systems with communication delays.The optimal economic equilibrium point of the system is computed discretely,and a stabilizing optimal control problem related to this point is constructed.The optimal value function of the stabilizing control problem is then used to formulate implicit contractive constraints for the original distributed EMPC optimization.By employing terminal constraints and the input-to-state stability(ISS)lemma,the recursive feasibility of the distributed EMPC and the ISS property of the closed-loop system with respect to state deviations from the optimal economic equilibrium point are established.Finally,the effectiveness of the approach is demonstrated using a cascade of nonlinear continuous stirred tank reactors.

Abstract:As a key foundation of the entire production process of copper products,ingredient plan is directly related to the continuous stability of the subsequent process and the economic benefits of the enterprise.Considering the unsatisfied resource utilization of ingredient plans provided by existing methods,an ingredient optimization model with the objective of maximizing resource utilization and ingredient consistency is established.Then,an improved multifactorial differential evolution algorithm is proposed.To solve premature convergence and infeasible solutions caused by complex constraints in the traditional multifactorial differential evolution algorithm,the improved algorithm first designs a mutation strategy based on the sine cosine algorithm to balance cross task knowledge transfer,global search,and local development capabilities of the algorithm.Secondly,the Levy optimal perturbation strategy is introduced to improve the diversity of solutions,and prevent the algorithm from getting stuck in local optima.In addition,a repair strategy for infeasible solutions with lower computational cost is proposed to ensure search efficiency.Finally,simulation experiments based on real data from a domestic copper industry and multitask benchmark problems demonstrate that compared with other methods,the model and method proposed in this paper can significantly improve the resource utilization rate and consistency of ingredient plan,providing good support for stable production in the copper industry.And the proposed algorithm exhibits good stability in complex multitask benchmark problems.

Abstract:Aiming at the motion control problem of the manipulator,first,the DH parameter table and kinematic equation corresponding to the physical manipulator are established.Then,a position predictive controller of the manipulator is designed using kinematics.Subsequently,to address the convex optimization problem of the controller model,the terminal cost and terminal constraints are introduced,and the enhanced iterative model predictive control strategy is employed to achieve the motion control of the multi-axis manipulator.The Error-Summing Enhanced Newton algorithm is used to improve the iterative efficiency in the rolling optimization process.In addition,a MESEN algorithm is proposed which further improves the convergence speed of the system and reduces the dependence of the optimization process on the initial system deviation by effectively controlling the step size of the algorithm and the correlation coefficient in the model.Finally,the convergence of the MESEN algorithm is analyzed using Lyapunov's theorem,and the six-axis manipulator is simulated and verified.Experimental results show that the proposed algorithm performs well in the large-scale constrained model.

Abstract:This paper considers a tracking control problem for a type of discrete-time nonlinear fully actuated systems with external disturbances.A fully actuated system(FAS)anti-disturbance predictive control method is proposed to address this problem.Firstly,a high-order disturbance observer is designed by adopting a difference operator and its high-order form to accurately estimate the external disturbances under a less conservatism assumption,which provides a better foundation to construct a disturbance preview.Secondly,an incremental FAS(IFAS)prediction model with the disturbance preview is established by utilizing a Diophantine Equation.Based on this IFAS prediction model,the multistep ahead predictions are derived to minimize an objective function for obtaining an optimal anti-disturbance controller,such that the tracking performance can be maintained.Then,a sufficient condition is presented to analyze the bounded stability and tracking performance of the closed-loop FASs in the further discussion.Finally,the proposed FAS anti-disturbance predictive control provides a solution to spacecraft attitude control for the verification of effectiveness.

Abstract:Predictive control based on fully actuated system(FAS)approaches employs nonlinear input transformations to map the original inputs into the desired linear closed-loop system inputs.This enables the construction of distributed linear predictive models,effectively reducing the complexity of solving the optimization problem.However,when the system owns input saturation,such input transformations introduce highly nonlinear constraint issues to the desired predic-tive model.To address this,this paper proposes a cascaded predictive control method and designs a cascaded predictive controller with a two-layer optimization structure.In the first layer of optimization,the predictive input sequence from the previous instant is used to optimize the linear boundaries of the transformed inputs within the predictive horizon.In the sec-ond layer of optimization,based on the newly defined linear constraints,a series of distributed linear optimization problems with slack factors are solved.This cascaded predictive method effectively avoids the nonlinear constraint issues caused by input transformations,reduces optimization complexity,improves the solvability of nonlinear optimization problems,and ensures the stability of the closed-loop system.Finally,the effectiveness of the proposed algorithm is verified through simulations on fully-actuated spacecraft attitude system and the under-actuated rotational translational actuator system.

Abstract:Considering the requirements for mold switching,job transportation and the impact of processing load on machine energy consumption in green production,the setting time,transportation time,variable machine energy consump-tion and the pre-maintenance mechanism are incorporated into the scheduling problem model.A flexible job-shop green scheduling model is established with objectives of minimizing the completion time,energy consumption and total noise.The Jaya algorithm based on objective balance is proposed to solve it.Considering the influence of the initial population quality on the evolutionary efficiency,a hybrid initialization strategy is designed,which enhances the objective balance of the initial population and accelerates the convergence speed of the algorithm.For all non-optimal individuals,a diversified Jaya optimization operator is put forward to increase population diversity and global search capability.To prevent the algo-rithm from getting stuck in a local optimum and enhance the distribution of the solutions,three local search strategies based on objective balance are designed.The effectiveness of the proposed strategies is verified through ablation experiments.The experimental comparison results with four algorithms have demonstrated the superiority of the proposed algorithm.

Abstract:To address the trajectory planning challenges of polishing robots under complex working conditions involving joint constraints,frictional disturbances,and external end-effector forces,a dual-layer predictive control method integrated with force feedback is proposed to achieve dynamic trajectory planning and enhance system robustness.The hierarchical framework consists of two layers:the upper layer generates globally optimal trajectories using model predictive control(MPC)under ideal conditions,while the lower layer dynamically adjusts the control output through fuzzy-compensated pre-dictive control.Simultaneously,force sensor feedback is integrated to construct a dynamic constraint optimization model,which adjusts pressure thresholds based on task requirements to prevent overshoot and ensure safe interaction.Simula-tion results demonstrate that the proposed method achieves stable and smooth trajectory tracking in complex environments through the synergistic mechanism of hierarchical optimization and real-time compensation,with significant improvements in both precision and energy efficiency.

Abstract:The model uncertainty in the model predictive control(MPC)method mainly arises from factors such as external disturbances,input noise,and time-varying parameters,all of which may lead to prediction errors and thus reduce control performance.This paper proposes a novel stochastic model predictive control(SMPC)algorithm for a class of discrete time-varying systems with additive disturbances.The algorithm can effectively predict the system's future behavior and optimize the control input under specific constraints to achieve predetermined performance goals.To implement the control method proposed in this paper,the Lyapunov stability theory is employed to ensure the closed-loop stability of the system,convex combination techniques are used to address the time-varying nature of the system parameters,and convex optimization methods are utilized to calculate the control gain of the nominal system.Additionally,the inverse cumulative distribution function is used to transform probabilistic constraints into deterministic constraints.Finally,the paper verifies the proposed method through numerical simulation experiments.The simulation results show that the SMPC strategy can better adapt to the randomly changing environment,maintain high control accuracy when facing random disturbances,and has relatively low conservatism.