Radar radiation source signal feature selection method based on membrane particle swarm multi-target algorithm

Abstract

The invention, which belongs to the technical field of data processing, discloses a radar radiation source signal feature selection method based on a membrane particle swarm multi-target algorithm. According to the method disclosed by the invention, a membrane calculation optimization theory is combined with a particle swarm optimization algorithm and the uniformity and diversity of a crowding degree maintenance set are utilized; and a data object is optimized by using two objective functions of correlation and redundancy and application to in-pulse feature selection of a radar radiation source signal is carried out. According to the invention, with non-dominated sorting and a crowding distance mechanism in a skin membrane, rapid convergence of the multi-target particle swarm optimizationalgorithm is kept by the algorithm and the solution set has high diversity. And then with KUT and ZDT series test functions, comparison testing is carried out on the algorithm and MOPSO, SPEA2 and PESA2 algorithms. Therefore, rapid convergence to the real Pareto leading edge is realized and the provided algorithm is feasible and effective.

Description

technical field [0001] The invention belongs to the technical field of data processing, and in particular relates to a radar radiation source signal feature selection method based on a membrane particle swarm multi-target algorithm. Background technique [0002] In reality, many engineering and scientific problems have multiple conflicting problems. How to obtain the optimal solution set of these problems is the focus of engineering and academic research. Different from single-objective optimization problems, the objective functions in multi-objective problems are connected through decision variables. The biggest feature is that the performance improvement of one of the objectives will inevitably lead to the decline of the performance of other objectives. Therefore, there is generally no single optimal solution. Optimal solution, but a set of solutions, the so-called Pareto optimal solution set. Due to the complexity of the mathematical properties of the objective function ...

AI Technical Summary

Problems solved by technology

In addition, the multi-objective feature selection algorithm finally obtains a series of compromise solutions, from which it is necessary to select a solution with excellent performance, but currently there are few unsupervised methods available

The main difficulty lies in: (1) The designed feature subset evaluation function and search strategy fail to consider the redundancy and correlation of the feature subset; (2) The evaluation criteria also do not consider the selection of the feature subset dimension. The impact of classification effectiveness; (3) The unsupervised way to extract the feature dimension and the importance ranking of feature subsets in the Pareto solution set of the multi-objective optimization algorithm has not yet been resolved

[0010] Therefore, in view of the above problems and difficulties, a radar emitter signal feature selection method based on the membrane particle swarm multi-objective optimization algorithm is proposed. , evolution rules and message passing mechanism to solve (1) low convergence efficiency, poor solution set diversity and premature falling into local optimum in multi-objective optimization algorithms; (2) redundancy for radar signal feature subsets Design of two-dimensional objective evaluation function of redundancy and relevance; (3) Unsupervised feature subset dimension and feature subset importance ranking algorithm for Pareto solution set

Images