六自由度Stewart平台运动学遗传神经网络正解.pdf

第

39

卷第

2A

期舰

船

科

学

技

术Vol.

39,

No.

2A

2017

年

2

月SHIP

SCIENCE

AND

TECHNOLOGYFeb.

,

2017

六自由度Stewart平台运动学遗传神经网络正解

李

晶

(广东轻工职业技术学院

实验实训管理中心，广东

广州

510300)

摘要:近年来，舰载相控阵雷达发展迅速，对其隔振措施和隔振机构的要求也越来越高。Stewart

平台为一

种通常的空间六自由度并联机器人，在舰载相控阵雷达中具有广泛的应用。为了解析

Stewart

平台工作空间和设置

机构误差步长，必须运动学求解。本文针对当前

Stewart

平台正解法存在的问题，基于传统的正解法和

Stewart

平台

结构，提出一种遗传神经网络正解算法，通过仿真验证可知，该方法具有较高的收敛性和较快的速度，有进一步推

广应用的潜力。

关键词：六自由度；网络模型；位置向量；姿态向量

中图分类号：U665.22

文献标识码：A

文章编号：

1672

–

7619(2017)2A

–

0112

–

03

doi：10.3404/j.issn.1672

–

7619.2017.2A.038

GeneticneuralnetworkpositivesolutionofStewartplatformkinematics

withsixdegreeoffreedom

LI

Jing

(Experimental

and

Training

Management

Center

Guangdong

Industry

Poly

Technic

Institute,

Guangzhou

510300,

China)

Abstract:

In

recent

years,

Ship

borne

Phased

Array

Radar

is

developing

rapidly,

and

the

requirements

of

vibration

isol-

ation

measures

and

vibration

isolation

mechanisms

become

higher

and

higher.

Stewart

platform

is

a

kind

of

usual

space

six

degree

of

freedom

parallel

robot,

which

is

widely

used

in

Ship

borne

Phased

Array

Radar.

In

order

to

analyze

the

working

space

of

Stewart

platform

and

set

the

error

step

of

mechanism,

the

kinematics

must

be

solved.

Based

on

existing

Stewart

plat-

form

solution

of

the

problem,

the

traditional

method

and

the

structure

of

Stewart

platform

based

on

the

proposed

algorithm

is

a

solution

of

genetic

neural

network

in

this

paper.

This

method

has

high

convergence

and

high

speed

proved

by

sim