Preset Constraint Control For Lower Triangular Nonlinear Stochastic System

Many actual systems can be modeled or converted into lower triangular nonlinear systems,such as quad-rotor aircraft and chemical reactor systems,and actual systems often have multiple constraints.Then,the preset constraint control of lower triangular nonlinear systems has received widespread attention from domestic and foreign scholars.In practical industrial systems,sensors,controlled objects,and actuators often interact with each other through networks.Because of limited communication network bandwidth and computing resources,it is necessary to reduce the occupation of network resources while ensuring system performance.Meanwhile,owing to physical structure and security limitations,the actuator output is often bounded,making it sometimes difficult to reach the given value of the control signal.In addition,the industrial environment is complex and ever-changing,and often operates for a long time,which is highly susceptible to stochastic disturbances,sensor failures(changes in measurement sensitivity),and actuator failures.Moreover,some state information of industrial systems may be difficult to obtain,and effective observers need to be constructed to facilitate controller design.Therefore,this thesis takes into account factors such as input saturation,limited network resources,difficulty in obtaining the state,unknown measurement sensitivity and actuator failure,and investigates the predefined constraint control problem for the lower-triangular nonlinear stochastic system.The main research contents are as follows:(1)For a class of uncertain nonlinear stochastic systems,the transformed stateconstrained state feedback control problem based on input saturation is considered.Based on the difference between the designed control signal and the actual control signal,a fullorder high-gain compensation system is constructed to eliminate the influence of input saturation characteristics.Meanwhile,an asymmetric tangent barrier Lyapunov function is established for a transformed state,which converts the predefined constraint problem of the transformed state into the boundedness problem of the barrier Lyapunov function.Subsequently,a full-state constraint adaptive tracking control algorithm is designed.This algorithm successfully constrains the transformed state within asymmetric boundaries and ensures the probability stability of the closed-loop system.(2)For a class of interconnected nonlinear stochastic systems,the state feedback hybrid event-triggered control problem under constant constraints is studied.To achieve direct constraints on a state,a state-dependent nonlinear transform function is constructed,and the state constraint control problem is converted into a bounded problem of the nonlinear transform function.Meanwhile,a hybrid event-triggered mechanism is established to balance the relationship between system performance and controller update times.Subsequently,based on dynamic surface control technology and neural network theory,a full-state constrained adaptive tracking control algorithm based on state feedback is designed,which not only effectively reduces controller updates,but also ensures the probability stability of the closed-loop system and achieves direct constraints on the state.(3)For a class of uncertain nonlinear stochastic systems,the finite-time control problem of state feedback dynamic event-triggered under time-varying constraints is studied.With the help of the constraint control theory,a predefined-time performance function and a barrier function are constructed.If the barrier function is bounded,the state remains in a specific region within a predefined time.Meanwhile,by introducing a dynamic variable,a dynamic event-triggered mechanism is established to reduce the number of controller updates.Subsequently,by means of finite-time stability theory,a full-state constraint finitetime tracking control algorithm triggered by dynamic events is constructed.This algorithm achieves the constraint of the state to specified boundaries within a predefined time,and the closed-loop system is probability stable within a finite time.(4)For a class of nonlinear stochastic systems with completely unknown measurement sensitivity and arbitrary relative degree,the dynamic event-triggered control problem of output feedback under preset time constraints is studied.Since only the output is measurable and the sensor is not ideal,a dynamic gain reduced-order K-filter is established to reconstruct the unmeasurable state.Meanwhile,a preset-time performance function and a barrier Lyapunov function are constructed to ensure that output operates within the preset range.Subsequently,a dynamic event-triggered mechanism is proposed,whose trigger threshold is directly related to the designed dynamic variable,reducing the controller update frequency.Finally,based on the Nussbaum function,a dynamic event-triggered prescribed performance control algorithm under completely unknown measurement sensitivity is proposed.When the measurement sensitivity amplitude is unknown and the symbol changes,this algorithm ensures that the closed-loop system is stable in a probabilistic sense and that the output is limited to a given area within a preset time.(5)For a class of complex nonlinear stochastic systems with unknown measurement sensitivity and arbitrary relative degree,the output feedback predefined performance control problem considering actuator replacement time is studied.Based on the output measured based on non-ideal sensors,a low-order K-filter is constructed to estimate the unmeasured state,and dynamic gain is introduced to overcome the difficulties caused by nonlinear functions.Meanwhile,a shift function and a funnel-like constraint function are established,which relaxes the requirement that the initial value of the constraint variable needs to be within the constraint boundaries,providing the possibility to restore the tracking error to the original constraint range.Subsequently,a monitoring function is constructed to monitor actuator faults,and a predefined performance fault-tolerant control algorithm is designed.Taking into account the actuator replacement time,this algorithm can ensure that the tracking error is restored to the original constraint boundaries and maintain the stable operation of the closed-loop system.