34results about How to "Good nonlinear approximation capability" patented technology

The invention discloses an electric automobile power battery SOC (State of Charge) detection system. The characteristics lie in that the detection system comprises a battery parameter acquisition platform and a battery SOC estimation system, the battery parameter acquisition platform is responsible for real-time parameter acquisition for the voltage, current and temperature of an automobile power battery pack, and the battery SOC estimation system can accurately estimate a battery SOC value through the acquired parameters; and the battery SOC is a non-linear, delayed, multivariable coupling and complex highly demanding real-time system. The detection system effectively solves a problem that the traditional automobile battery SOC estimation method is difficult to achieve an ideal effect.

The invention elates to the technical field of industrial process control, and discloses a method and system applicable to complementing missing data of product quality indicators in the complex industrial process based on selective double-layer ensemble learning. The method comprises the steps of firstly extracting different dimensions of variables of the sampled data to generate multiple sampling sets which serve as training sets of sub-models; then modeling each sub-model by respectively adopting a vector machine method, a BP neural network method and a partial least squares method; and finally putting forward a complementing effect evaluation indicator to perform evaluation on the complementing effect of each sub-model, and selecting a plurality of sub-models with the best complementing effect to perform selective ensemble. The method makes full use of all variables of the training samples, has a good data complementing effect, and facilitates enterprises to obtain the actual operation condition of the production process according to the analysis so as to perform targeted production operation optimization.

The invention relates to a trend prediction method based on double-hidden-layer quantum circuit cyclic unit neural network, comprising the following steps: constructing a permutation entropy set of original operating data; inputting the permutation entropy set into DHL-QCRUNN for training and prediction, and obtaining a predicted permutation entropy set ;Construct the permutation entropy error set of the predicted value and the actual value at each time point; input the permutation entropy error set into DHL-QCRUNN training and prediction, and obtain the predicted normalized permutation entropy error set; denormalize the processing to obtain the final prediction result. The present invention proposes a new type of quantum neural network—a double-hidden layer quantum circuit recurrent unit neural network. The present invention uses the LM algorithm to update the network parameters of DHL-QCRUNN to improve the convergence performance of the neural network, which is comparable to other artificial intelligence methods. Compared with DHL-QCRUNN, DHL-QCRUNN has better nonlinear approximation ability, generalization characteristics and faster convergence speed. The present invention is used to predict the running trend of the monitored object, and achieves higher prediction accuracy, prediction stability and calculation efficiency.

The invention relates to a dual-motor servo system control method based on a segmented neural network friction model, and belongs to the technical field of electromechanical control. Firstly, the double-motor driven servo system with friction is analyzed, and the mathematical model of the double-motor driven servo system with friction is established according to the mechanism modeling method, according to the structure and physical laws of the motor. Then the friction term fi in the model is analyzed, and the friction model of friction nonlinear fi is established by using the segmental neural network. The obtained segmental neural network friction model uses the terminal sliding mode control algorithm to obtain the motor speed synchronous control law, and completes the synchronous tracking control of the dual-motor servo system according to the control law. The method of the invention can eliminate the influence of friction on the dual-motor system, enable the dual-motor system to have better transient performance, effectively improve the tracking response speed of the dual-motor servo system, and ensure the rapid synchronization of the dual-motor system.

The invention discloses an intelligent forecasting device for cucumber greenhouse output based on a recursive neural network. Networked cucumber greenhouse parameter detection platform and greenhouse cucumber output intelligent prediction system; the purpose of the invention is to provide a cucumber greenhouse output intelligent prediction device based on recursive neural network, and the intelligent early warning system detects the environmental parameters, soil parameters and parameters of cucumber greenhouse in real time. Cucumber greenhouse production information, so as to improve the production management of cucumber greenhouse and improve economic benefits.

The invention discloses a location-free detection method used when an SR motor travels to a region with phase electrical inductance changeless with angle variation, which comprises the following steps of: 1, determining the comprehensive performance parameter of an SR motor to be detected and selecting a conventional location-free detection method suitable for the SR motor to be detected; 2, establishing an actual mathematical model of the SR motor to be detected; and 3, carrying out zoning angular location detection on the SR motor to be detected by utilizing a processor: firstly demarcating the region with the phase electrical inductance changeless with the angle variation, and applying the original location-free detection method for location estimation before a conduction phase of the motor enters the region; and when the conduction phase of the motor enters the region with the phase electrical inductance changeless with rotor variation, resetting a constant speed detection region range for realizing the detection of the angle calculated by speed. The invention has simple steps, convenient implementation, strong practicality, high detection precision and small error, and can effectively solve the defects and insufficiency of low detection precision, huge error and the like existing in traditional method used when the SR motor travels to the region with the phase electrical inductance changeless with angle variation.

The invention relates to a CAN bus-based pig house environmental temperature intelligent monitoring system. The system is characterized in that the intelligent monitoring system is composed of a CAN bus-based pig house environment parameter acquisition and intelligent prediction platform, a pig house environment multi-point temperature fusion model and a pig house environment intelligent prediction model. With the CAN bus-based pig house environmental temperature intelligent monitoring system of the present invention adopted, an existing pig house monitoring system cannot intelligently monitor and predict the temperature of the environment of a pig house according to the non-linearity and large lag of the environmental temperature change of the pig house, the large area of the pig house, complicated temperature change and other characteristics, as a result, control and adjustment of the environmental temperature of the pig house are seriously affected, while, with the CAN bus-based pig house environmental temperature intelligent monitoring system adopted, the above problem can be solved.

The present invention discloses a degradation prediction method based on quantum attention cyclic encoding and decoding neural network (QAREDNN). QAREDNN is used in it, and a quantum attention mechanism is introduced to reconstruct the encoder and decoder at the same time so that QAREDNN can fully mine and Emphasize important information while suppressing the interference of redundant information to obtain better nonlinear approximation capabilities; use quantum neurons to construct activity values ​​and weights, and replace encoders and decoders with quantum threshold recurrent units (QGRU) replaced by quantum rotation matrices The traditional recurrent unit in the device can improve the generalization ability and response speed of QAREDNN; in the training process of QAREDNN, the LM algorithm is introduced to realize the rapid update of the rotation angle and attention parameters of the quantum rotation matrix. Due to the advantages of QAREDNN in terms of nonlinear approximation ability, generalization ability, response and training speed, etc., the degradation prediction method based on quantum attention loop encoding and decoding neural network can obtain higher prediction accuracy and computational efficiency.

The present invention discloses a method for improving interpolation position accuracy of a cutting bed based on an improved BP neural network. Parameters of a current node are input into a trained BP neural network interpolation model to obtain predicted values of the node parameters output by the BP neural network interpolation model. The predicted values of the node parameters output by the BP neural network interpolation model are substituted into a B-spline curve to obtain a predicted interpolation node. The deviation between the predicted interpolation node and the actual interpolation node and the corresponding feedback compensation output are calculated. The feedback compensation output is compared with an original machining trajectory curve; the node parameters for the next interpolation are optimized and predicted with the Newton path searching method; the predicted node parameters are substituted into the fitted B-spline curve to obtain the actual interpolation node; and the interpolation operation is performed. The method of the invention has the advantages of good nonlinear approximation ability, high interpolation precision and fast convergence speed.