Hybrid neural network

About: Hybrid neural network is a research topic. Over the lifetime, 1305 publications have been published within this topic receiving 18223 citations.

Hybrid neural network and fuzzy logic approaches for rendezvous and capture in space

Abstract: The nonlinear behavior of many practical systems and unavailability of quantitative data regarding the input-output relations makes the analytical modeling of these systems very difficult. On the other hand, approximate reasoning-based controllers which do not require analytical models have demonstrated a number of successful applications such as the subway system in the city of Sendai. These applications have mainly concentrated on emulating the performance of a skilled human operator in the form of linguistic rules. However, the process of learning and tuning the control rules to achieve the desired performance remains a difficult task. Fuzzy Logic Control is based on fuzzy set theory. A fuzzy set is an extension of a crisp set. Crisp sets only allow full membership or no membership at all, whereas fuzzy sets allow partial membership. In other words, an element may partially belong to a set.

Hybrid neural network based fault diagnosis of rotating machinery

Abstract: Vibration fault is the main fault of hydraulic generator set. From the analysis of vibration signal, it provides a wealthy of information for fault diagnosis. This paper presents a hybrid approach of neural network to realize automatic diagnosis. Pulse coupled neural network (PCNN) has very strong capability in the feature extraction, and entropy time signature from a PCNN has the property of insensitive to rotation, scaling and translation, it is used to extract the feature vector of vibration signal. Probability neural network (PNN) has excellent performance in the pattern recognition. Therefore, it is used in the vibration fault classification. Experimental results show the proposed method greatly robust to diagnose the fault, by comparison with another artificial neural network.

Image recognition method, device and equipment based on hybrid neural network model

Abstract: The invention discloses an image recognition method, a device and equipment based on a hybrid neural network model, and a computer readable storage medium. The method comprises the steps of inputtinga to-be-recognized image into a convolutional auto-encoder for preprocessing; extracting image features of the preprocessed to-be-identified image by using a feature extractor constructed based on transfer learning; extracting internal time sequence features of the preprocessed to-be-identified image by using a long short-term memory network model; utilizing a feature fusion door and a feature screening door to fuse and screen the image features and the internal time sequence features to obtain target features of the recognition image; and utilizing a softmax classifier to classify the targetfeatures to obtain a classification result of the to-be-identified image. According to the method, the device, the equipment and the computer readable storage medium provided by the invention, the number of images required for training the neural network model can be greatly reduced, and the accuracy of image recognition is improved.

Data stream reconstruction method and reconstructable data stream processor

Abstract: The invention discloses a data stream reconstruction method and a reconfigurable data stream processor. Particularly, data flow reconstruction for a hybrid artificial neural network is carried out; according to different neural network layers, corresponding function configuration dynamic changes are carried out on resources such as a computing unit, a storage unit and a data flow unit; neural network layers with different functions are realized by multiplexing hardware on a large scale, and for a hybrid neural network structure formed by a plurality of neural network layers, the effects of improving the hardware utilization rate, improving the operation speed, reducing the power consumption and the like are achieved. Particularly, the reusable configuration is confirmed by acquiring the characteristic information of other novel neural network layers, so that a resource reuse basis can be provided for subsequent research and construction of other novel neural network layers and implementation of a hybrid neural network based on the novel neural network layers, and the universality is extremely high.

A hybrid neural network-based behavioral model of microwave active components

Abstract: In this work, the development, implementation and verification of a novel behavioral model for microwave active components is presented The application of neural networks to microwave device modeling has been demonstrated to be very useful In order to model an active component such a packaged amplifier, it is necessary to produce models of the active components as well as the passive elements of the module This is typically a lengthy, specialized process The authors demonstrate a simplified process of developing models for discrete transistors and packaged amplifiers using common device measurements, and integrating neural network techniques within the behavioral modeling process