Font Size: a A A

Research On Threat Assessment Method Of Air Combat Target Based On Fully Connected Neural Network

Posted on:2021-05-25Degree:MasterType:Thesis
Country:ChinaCandidate:X Y ZhaiFull Text:PDF
GTID:2392330602469028Subject:Information and Communication Engineering
Abstract/Summary:PDF Full Text Request
With the continuous development of modern air combat weapons and equipment,the air combat situation environment is becoming more and more complex,more and more attribute factors affecting the commander’s decision-making,and the requirements for commander situation understanding are becoming higher and higher.As an important auxiliary decision-making tool to reduce the pressure of commanders,threat assessment technology has received extensive attention in the course of air combat situation confrontation.Air combat target threat assessment is an indispensable part of air combat situation confrontation.To a certain extent,the correct assessment of the enemy’s air war target’s threat to our side can dominate the air war situation and achieve rapid victory.And it is the key to obtain the air supremacy in the environment of air war,which is changeable,dynamic and strong.With the continuous development of information-based equipment,the air combat situation information has exploded,which contains a large number of non-linear laws of air combat.Traditional air combat target threat assessment methods do not have self-learning capabilities and can not use a large amount of air combat situation information to target air combat objectives.The level of intelligence can not match the speed of situation evolution in complex environment,which leads to the problem of inaccurate threat assessment results and directly affects the commander’s understanding and judgment of current air combat situation.In order to solve the above problems,this paper analyzes the different threat attributes of air combat targets,studies the air combat target threat assessment method based on fully connected neural network,and improves the intelligence level of the air combat target threat assessment method.This method fits the data distribution of the threat level of air combat targets from a large number of air combat situation data,and then completes the threat assessment of air combat targets.The main contents of this paper are as follows:(1)Characteristic analysis of air combat target threat attributes and common threat assessment methods: analysis of air combat capability attribute and air combat situation attribute of air combat target,reasonable characterization of air combat target attribute,reasonable quantification or calculation of its corresponding threat factor,establishment of target attribute matrix,comprehensive subjective and objective method to determine the weight of each attribute of target,calculate the relative closeness of target,and evaluate the threat degree of air combat target.(2)Threat assessment method of air combat targets based on fully connected neural network: define air combat scenarios,analyze the target threat attributes and quantify them reasonably.A suitable network was constructed based on the fully connected neural network,and the batch normalization algorithm and residual network were combined to enhance the network learning ability and speed up the convergence.The network is trained with the air combat situation data under this scenario,and the law of air combat situation is learned independently.The obtained threat assessment model has strong information processing ability when it faces the situation information of air war and can be used to realize the target threat assessment of air war.(3)Multi-target threat assessment method of air combat based on fully connected neural network: To study the degree to which air combat multi-target ensemble poses a threat to us,In this paper,a multi-target air combat scenario is set up,the threat attribute of multi-target is analyzed,and an appropriate network is constructed based on the fully connected neural network and the optimization algorithm.The quantitative results of each attribute are taken as the input of the network and the output is the corresponding threat level.By adjusting the network hyper-parameter and network structure,the network was trained and tested,the experimental results were analyzed,and the performance and state of the air combat multi-target threat assessment model were judged.The model with better evaluation effect can be used to evaluate air combat multi-target threat.
Keywords/Search Tags:fully connected neural network, air combat target threat, air combat situation information, threat attribute representation, threat assessment model
PDF Full Text Request
Related items