7 results about "Random systems" patented technology

The invention discloses a multi-unmanned aerial vehicle cluster stochastic prediction control method and device based on anti-saturation and packet loss compensation, and the method comprises the steps: building a closed-loop control model of a stochastic system based on a pilot-following multi-unmanned aerial vehicle cluster architecture; through a Markov chain modeling data transmission packet loss process, a forgetting factor is introduced to design a packet loss compensation strategy after random data is randomly lost so as to correct a closed-loop control model of a random system; aiming at the corrected closed-loop control model of the following stochastic system, constructing an anti-actuator saturation control law through convex representation of a saturation function, and further constructing a saturation-limited following stochastic augmentation prediction model; solving a predictive control problem on line based on a random augmentation predictive model; and the control performance index is controlled through rolling optimization, the gain of the feedback controller is updated, and anti-saturation random prediction control following with random following in a packet loss environment is realized. According to the method, the coupling influence of data transmission packet loss and actuator saturation can be effectively solved, and the system performance weakness caused by single problem processing is avoided.

The invention discloses a structural vibration detection method based on an inequality constraint hybrid least square algorithm in the technical field of structural vibration detection. The method comprises the following steps: S1, system model construction: establishing a linear discrete stochastic system state space model; s2, designing an algorithm: proposing a hybrid least square algorithm with input inequality constraints; s3, solving a system model: applying a proposed inequality constraint hybrid least square algorithm to a discrete system to realize simultaneous optimal estimation of a system state and input; and S4, performing comparison simulation: performing simulation by taking a two-layer shear structure as an example, comparing the state and input estimation of the shear structure under the conditions of no constraint and existence of inequality constraint, and verifying the effectiveness of the method. According to the invention, the hybrid least square algorithm of the input inequality constraint is provided, and the filtering effect is optimized by using part of input information, so that the accuracy of structural vibration detection is improved.

The invention provides a linear stochastic system Pareto optimal control method and system and a readable storage medium. The method comprises the following steps: S1, selecting a linear random system, and carrying out simplification and parameterization processing on the linear random system; s2, applying energy stability control input to the linear random system, and recording original data of the system; s3, repeated items are removed by using a '- representation 'method, and a strategy evaluation implementation algorithm is given in combination with a least square method; s4, according to a strategy evaluation implementation algorithm, designing a Pareto optimal control online solving algorithm based on a reinforcement learning method and a ''-expression'' method; and S5, obtaining Pareto optimal control by using the Pareto optimal control online solving algorithm obtained in the step S4. According to the method, a '- representation 'method is applied to the exponential multi-objective reinforcement learning problem, repeated items in the operation process are eliminated, the calculation process is simplified, and the efficiency of Pareto optimal control selection is effectively improved.

The invention belongs to the technical field of random system control and sewage treatment, and discloses a T-S fuzzy Markov jump system control method based on an activated sludge system. According to the method, a nonlinear kinetic model is constructed from key state variables of the system, and the complex biological reaction and substance transfer process in sewage treatment is accurately described. By introducing the Markov jump theory, parameter random jump caused by inlet water quality sudden change and the like is effectively described, and the adaptability and accuracy of the model under the dynamic working condition are enhanced. For possible faults of an actuator (such as a dilution valve and an aerator), a fault model is established and a corresponding fault-tolerant control strategy is designed, so that the reliability and robustness of the system are improved. On the basis of the Lyapunov stability theory, a linear matrix inequality condition which enables the system to be randomly stable under the fault and disturbance and meets the H-infinity performance index is deduced, the effectiveness of the linear matrix inequality condition is verified through strict mathematical derivation, and a theoretical basis is provided for implementation of a controller and stable operation of the system.

The invention provides a sensor network distributed filtering method and system based on adaptive event triggering under replay attack and topology switching, and relates to the technical field of signal processing. The invention aims to solve the problem that the estimation performance is reduced when the existing distributed filtering method has the problems of communication resource limitation, topology random switching, replay attack and the like in a sensor network. The method is characterized by comprising the following steps: establishing a state space model and a sensor output model of a discrete nonlinear time-varying stochastic system, and adopting a self-adaptive event triggering mechanism to enable the measurement of a distributed filter to be available; a replay attack compensation mechanism is established, and the influence caused by the replay attack on the sensor network is processed in time; the developed adaptive event-triggered distributed filter design algorithm is in a recursive form and is very suitable for online application. The method can effectively cope with multiple challenges such as bandwidth limitation, topology change and network security threats, has good applicability and practicability, and is more suitable for practical engineering application.

A controller for optimizing a local control policy of a system for trajectory-centric reinforcement learning is provided. The controller includes performing the following steps: learning a stochastic predictive model of the system using a set of data collected during trial-and-error experiments performed using an initial random control policy; estimating associated mean predictions and uncertainty; determining a local set of deviations of the system from a nominal system state using the learned stochastic system model at a current time step using a control input; determining a system state with a worst-case deviation; determining a gradient of a robustness constraint; using a nonlinear programming to provide and solve a robust policy optimization problem to obtain a system trajectory and simultaneously stabilize the local policy; updating control data according to the solved optimization problem; and outputting the updated control data via an interface.