Improved Particle Swarm Optimization-Based BP Neural Networks for Aero-Optical Imaging Deviation Prediction

Liang Xu,Yuan Yao,Zhenhua Yu,Ziye Zhang

doi:10.1109/access.2021.3102669

Abstract

In this article, an improved particle swarm optimization (IPSO) algorithm based on similarity and random mutation is raised. The diversity of particles in the population is decided by the size of the aggregation. When the aggregation degree of particles in the population surpass a certain threshold, the concept of similarity is used to measure the similarity between particles and global extremum, and the particles with higher similarity are discretized by mutation strategy. By increasing the particle swarm’s diversity, the population’s local and global search ability tend to balance. The weight and threshold of the back propagation (BP) neural networks are optimized by the IPSO algorithm. Then, the model of the improved particle swarm optimization back propagation neural network (IPSO-BP) is applied to the aero-optical imaging deviation prediction. The results show that the prediction accuracy of the IPSO-BP model is superior to the PSO-BP model, the extreme learning machine (ELM) model, and the least square support vector machine (LSSVM) model, and its convergence speed is faster than that of the PSO-BP neural network model. Finally, the application of deep learning in aero-optical imaging deviation prediction is analyzed compared with the IPSO-BP neural network model.

Highlights

When the aircraft flies at high speed in the atmosphere, its head is severely compressed, resulting in a turbulent flow field with high pressure, high density, high temperature and even ionization
According to the optimization results obtained by improved particle swarm optimization (IPSO) algorithm, a prediction model of aero-optical imaging deviation based on IPSO-back propagation (BP) is established
The closer the mean square error approaches 0, the closer the determination coefficient approaches 1.It can be seen from Figs. 4-9 and Table 2 that the IPSO-BP prediction results are more in line with the actual situation of aero-optical imaging deviation, and the result of prediction curve is obviously superior to the extreme learning machine (ELM) model, least square support vector machine (LSSVM) model and PSO-BP neural network model, with strong predictive ability

Summary

INTRODUCTION

When the aircraft flies at high speed in the atmosphere, its head is severely compressed, resulting in a turbulent flow field with high pressure, high density, high temperature and even ionization. There are three main methods to overcome the PSO algorithm’s premature problem: (1) Improvement of PSO algorithm control parameters: maximum speed, acceleration coefficient, initial population, inertia weight, and particle out-of-bounds processing [21]-[25]; (2) Maintain the diversity of the particle swarm in a limited scale to continuously search for new solution spaces in the evolutionary processes [26]-[28]; (3) Mix various heuristic algorithms to increase the local search ability of particles [29][31]. The results show that the proposed algorithm can converge faster and avoid the phenomenon of local optimal solution effectively This model can efficiently, stably, timely and accurately predict the aerooptical imaging deviation, avoiding the disadvantages of traditional geometric optical calculation, such as time consuming and labor consuming.

THE CONTROL PARAMETERS OF PSO

SPACE COMPLEXITY

IPSO OPTIMIZED BP NEURAL NETWORKS

Findings

CONCLUSION