In order to realize the prediction of air quality of pollution sources, this paper uses the method of wavelet neural network. In this paper, the pollutant concentration values of pollution sources were predicted, so as to realize the monitoring and early warning of pollution source emission. In this...

A short penetrating distance is a drawback of traditional ultra wideband radar systems. A pulse compression technique has been used in UWB radar systems to increase the penetrating distance. However, this leads to an increase in the blind zone. The paper proposes a new method to address this issue which...

This paper proposes a algorithm for 3D near-field source localization, which uses the cross array and fourth-order cumulant. We construct two high-dimensional matrices by using these six fourth-order cumulant matrices, according to the subspace theory, we can extract the rotation factors, which contain...

Early warning radar is able to effectively detect the target utilizing Doppler Effects between the target and radar. The detection effectiveness of the airborne early warning radar is affected by its detection range as well as its detection probability. In this paper, a method for simulating the detection...

Abstract.P300 Event Related Potential (ERP) has drawn much interest in different application of psychological background. An endogenous ERP component, P300, has been widely inspected in detecting deception. Detecting P300 from EEG signals is hard to find because P300 is overloaded with noise and other...

An experimental method and reconstruction arithmetic of temperature field are proposed. The fast Fourier transform followed by a digital low-pass filter and the inverse Fourier transform were applied to interference fringes for removing speckle noise, so that airflow field phase information is in high-contrast...

Abstract.In this paper, we use the method of permutation entropy symbolization to symbolize the original data. It mainly uses the size relation of the adjacent data to symbolize the original data, preserves the details of the original data, and overcomes the problem that global static symbolization method...

Image recognition has the characteristics of large sample size, high complexity and redundant information, which has become a hot and difficult topic in the present study. To solve this problem, a feature extraction and image classification model based on the sparse auto-encoder deep neural network is...

In this paper, a random walk model was established for the EEG signal. The characteristic parameters qt and at were extracted from the original EEG signal by Bayesian inference method. By comparing the global maximum of the autocorrelation function of the characteristic parameter qt between patients...

In digital image processing, classic wavelet algorithm often has difficulties in obtaining accurate wavelet coefficients by using continuous integral formula. Hence, clear and accurate edges of images cannot be detected in the classic wavelet algorithm. To deal with this problem, a new algorithm is proposed...