Method and system for improving signal intensity and identification degree of sensor

Abstract

The invention provides a method and a system for improving signal intensity and identification degree of a sensor. The method comprises the following steps: S1, acquiring data of the sensor; s2, extracting generalized feature pairs of the acquired sensor data by using conjugate gradient optimization and orthogonal projection; and S3, filtering the acquired sensor data by using the generalized feature pair, and then carrying out data enhancement and identification processes. According to the method, the problem of online extraction of multiple generalized feature pairs is effectively solved through conjugate gradient optimization and an orthogonal projection thought, the extraction process has high numerical stability, the convergence speed is high, the extraction process is simpler and more convenient, and the intensity and the identification degree of sensor signals can be effectively improved.

Description

technical field [0001] The invention belongs to the field of self-adaptive signal processing, and relates to a method and system for improving sensor signal strength and recognition. Background technique [0002] In sensor signal processing, the subsequent processing of stream data, such as filtering, enhancement, and identification, has higher requirements on the pre-processing of data. The improvement of signal strength and recognition in the early pre-processing process will help for later application of the data. Generalized eigendecomposition is an unsupervised learning method and a linear dimensionality reduction technique. Generalized eigendecomposition aims to extract generalized eigenvectors / pairs of covariance matrices to implement operations such as data filtering, linear discrimination, and data compression. It plays a pivotal role in statistical signal processing and machine learning, such as in beamforming, Extracting the dominant generalized eigenvector is u...

AI Technical Summary

Problems solved by technology

However, matrix factorization techniques usually require large computational complexity and memory overhead, and are not suitable for observation-extraction data stream extraction scenarios.

Therefore, it is a common idea to use the gradient descent method to extract generalized feature pairs, but the gradient descent algorithm usually faces the problem of slow convergence.

In order to improve the convergence speed, second-order algorithms such as Newton's method and quasi-Newton algorithm are used to extract generalized eigenvectors. These methods involve the calculation of the second-order derivative Hessian matrix. In some special cases, these quantities do not exist so that the algorithm A computational stability issue has occurred

Images