Basis Function Neural Network For Prediction Research Articles

IoT-based multi-channel information integration method for wireless sensor networks

The speedy progress of the Internet of Things has connected a huge network of smart devices in series and also intensified the pressure of multi-channel information transmission in wireless sensor networks. To solve the transmission of large amount of information, a multi-channel information integration method based on Radial Basis Function neural network, Particle Swarm Optimization algorithm and Dempster-Shafer evidence theory is studied. The method designs an improved counter propagation Radial Basis Function neural network prediction algorithm to correct the abnormal data based on the normal and abnormal values detected by the Grubbs criterion. And then it uses the Dempster-Shafer evidence inference method for information integration on the basis of the complete monitored values after correction obtained using the above method. The lab result indicated that mean absolute error value of the improved counter propagation Radial Basis Function model for anomaly data prediction findings was 4.399, mean square error was 26.820, and root mean squared error value was 5.180. The average time delay of information integration of this research method was 0.24 μs. Compared with other methods, this method has the shortest integration time delay and the highest integration accuracy. The study provides a new method in the multi-channel information integration for wireless sensor networks in the context of internet of things.

Temperature Control Algorithm of Underground Air Conditioning in Station

Aiming at the air-conditioning temperature control of underground station, the air-conditioning temperature controller of underground station is designed based on proportional-integral-differential neural network, the bat algorithm is used to select the initial weights, the radial basis function neural network prediction model is used to optimize the design of the predictor, and the designed control system is simulated on the MATLAB simulation platform. The simulation results show that the dynamic characteristics and steady-state characteristics of the PID controller optimized by bat algorithm are better than those of the PID controller tuned by conventional methods. After optimization by radial basis function neural network prediction model, the rapidity and stability of controller are effectively improved.

Evaluation and prediction on total factor productivity of Chinese petroleum companies via three-stage DEA model and time series neural network model

In this paper, the panel data of 50 listed Chinese petroleum companies from 2009 to 2018 were used. A three-stage Data Envelopment Analysis (DEA) model was used to eliminate environmental factors and random factors, Pearson Epsilon-based Measure (PEBM) distance function and the Malmquist index were combined to evaluate the change of total factor productivity, then Support Vector Regression (SVR) prediction model and Radial Basis Function (RBF) neural network prediction model were compared and it was found that the latter had a small error of prediction on Total Factor Productivity (TFP), hence the radial basis function neural network prediction model was used to predict the total factor productivity of listed petroleum companies in the next two years. The study showed that the average annual growth rate of total factor productivity of the listed petroleum companies was 9.05 %, and its Efficiency Change (EC) index and Scale Efficiency Change (SEC) index were the main driving force for the growth of total factor productivity, while its Technical Change (TC) index and magnitude technical change (MATC) index had an inhibitory effect on the total factor productivity. And the predicted results showed that total factor productivity of the petroleum companies will continue to grow in the next two years. In addition, the research methods provided in this paper can be used not only to evaluate and predict total factor productivity of petroleum field, but also to evaluate and predict total factor productivity of other fields.

Optimal sheet metal fixture locating layout by combining radial basis function neural network and bat algorithm

Considering that sheet metal part has the properties of thin wall, low rigidity, easy to deform, and difficult to locate, this article proposes a new approach to optimizing sheet metal fixture locating layout by combining radial basis function neural network and bat algorithm. First, taking fixture locating layout as design variables based on the “ N-2-1” locating principle, this article generates limited training and testing sample sets by Latin hypercube sampling and finite element analysis. Second, the radial basis function neural network prediction model with the description of the nonlinear mapping relationship between the fixture locating layout and the corresponding sheet metal deformation is constructed through learning from the training sample sets. Third, bat algorithm is applied to search the optimal layout of the “ N” fixture locators for the minimum sheet metal deformation. Finally, two case studies are presented to demonstrate the optimization procedure and the effectiveness of the proposed method.

Application of Radial Basis Function Neural Network on the Prediction of Urban Built-Up Area

The study on the prediction of urban built-up area is the basic issue in urban planning. This paper takes the prediction of urban built-up area of Hefei city as an example, building a factor system that affects built-up area from the economic, social and environmental dimensions. Then, the paper establishes a quantitative prediction model based on the Radial Basis Function neural network. As a comparison, the paper also uses the Back Propagation neural network to predict. The results show that the Radial Basis Function neural network prediction has a higher accuracy and the prediction result is more reasonable and reliable.

An adaptive navigation method for a small unmanned aerial rotorcraft under complex environment

Focusing on low navigation performance of small unmanned aerial rotorcraft under complex environment, a composite navigation method combined with adaptive Kalman filtering and radial basis function neural network prediction method is proposed to improve navigation performance during GPS outages. When the GPS signal is available, an adaptive Kalman filter based on covariance scaling is introduced to deal with the process noise in real time. Meanwhile, a radial basis function neural network is trained on line to construct the projection among input (output of the inertial measurement unit, attitude and GPS losing time) and output (position error and velocity error). During GPS outages, the radial basis function neural network can provide high performance error estimation for position and velocity to improve state information. Finally, a land vehicle test and a flight test have confirmed that the proposed method can improve the navigation performance largely under complex environment.

Performance of Error Back Propagation vis-á-vis Radial Basis Function Neural Network: Part I: Prediction of Properties for Design Engineering of Woven Suiting Fabrics

In this paper, an artificial neural network approach is used to predict the low-stress mechanical, dimensional, and tensile properties of woven suiting fabrics. Radial basis function neural network prediction is studied along with error back propagation based neural networks. Radial basis function neural networks are found to have better predictability and are faster to train and easier to design than back propagation neural networks. The prediction of fabric properties by the neural networks due to changes in fabric constructional parameters is in good agreement with generally accepted trends for fibre, yarn and fabric structure property relationships.