CN105893663A - Tri-strapdown inertial measurement unit non-quantization dynamic threshold value interval estimation method - Google Patents

Abstract

The invention provides a tri-strapdown inertial measurement unit non-quantization dynamic threshold value interval estimation method. According to the specific scheme, the method includes the steps that 1, trajectory data with a tri-strapdown inertial measurement unit error model is generated through a Monte Carlo shooting method to serve as sample data; 2, the sample data is subjected to a normal distribution test, if the sample data conforms to normal distribution, the step 3 is executed, or else, it is shown that the sample data does not conform to normal distribution, and the method is withdrawn; 3, the threshold value confidence interval is estimated. Compared with a traditional method, normal distribution verification can be more precisely achieved, threshold value design is more reasonable, the threshold value confidence interval can be rapidly estimated, and simulated flight efficiency can be improved.

Description

A kind of three strapdown inertial measurement unit non-quantized dynamic threshold method of interval estimation

Technical field

The invention belongs to the guidance of large-scale vehicle highly reliable flight navigation and control field, relate to a kind of three strapdowns Used group non-quantized dynamic threshold method of interval estimation.

Background technology

Reliability, safety are had higher requirement by following Large Launch Vehicle.Control system is as fortune Carry the nerve centre of rocket, the performance of whole carrier rocket is had very important effect.Strapdown inertial measurement unit is The important component part of Guidance and control, carries out Redundancy Design to it and can be greatly improved reliability.

For three strapdown inertial measurement unit redundant systems, the design of threshold value is the pith of redundant system design, Dynamically, while rational threshold design method can ensure ease for use as far as possible, improve reliability.Currently, Needed for non-quantized dynamic threshold method for designing, sample is mostly based on tradition normal distribution-test method, it is impossible to accurately Realize normal distribution checking, threshold design has certain deviation.

Summary of the invention

The technical problem to be solved is: overcome the deficiencies in the prior art, it is provided that a kind of three strapdowns are used to Group non-quantized dynamic threshold method of interval estimation, solving non-quantized dynamic threshold method for designing needs accurately The problem realizing normal distribution checking.

The present invention includes following technical scheme: a kind of three strapdown inertial measurement unit non-quantized dynamic threshold method of interval estimation, Step is as follows:

1) ballistic data is generated

The ballistic data with three strapdown inertial measurement unit error models will be generated as sample by Monte-Carlo method method Notebook data；

2) sample data is carried out normal distribution-test

21) to step 1) sample data that obtains carries out ascending sort according to numerical values recited, by all of Sample data is divided into N group, and the sample data number often organized is M_i, i=1,2 ... N represents i-th group；x_imax Represent the data value that in i-th group, sample data is maximum, x_iminRepresent the data value that in i-th group, sample data is minimum；Represent the estimation statistical average of i-th group of sample data；n_iRepresent i-th group of sample number According to data amount check；N represents the total number of sample data；

22) below equation is utilized to try to achieve averageAnd variance

$\widehat{\mathrm{\μ}}=\frac{1}{n}\underset{i=1}{\overset{10}{\Σ}}{n}_i{x}_i^{*}=\stackrel{\‾}{X},$

${\widehat{\mathrm{\σ}}}^2=\frac{1}{n}\underset{i=1}{\overset{10}{\Σ}}{(n_i{x}_i^{*}-\stackrel{\‾}{X})}^2;$

23) set the average of normal distribution, variance is respectively μ and variances sigma², whenAnd Time, represent sample data Normal Distribution, enter step 3)；Otherwise show that sample data disobeys normal state Distribution；Exit this method；Wherein A, B are threshold value；

3) threshold value Estimating Confidence Interval is carried out

The distribution density of overall X normal distribution is:

Solve the inverse function of standard normal distribution, obtain x_max、x_min, finally give threshold interval (x_max, x_min)。

The present invention compared with prior art has the advantage that

(1) by step 2) smoothing processing to sample data, normal distribution can be realized more accurately and test Card, makes threshold design more reasonable；

(2) to meeting the sample data of normal distribution, by step 3) threshold value confidence district can be rapidly completed Between estimate, improve simulated flight efficiency.

Accompanying drawing explanation

Fig. 1 is a kind of non-quantized dynamic threshold method for designing flow chart.

Detailed description of the invention

The most just combine accompanying drawing the present invention is described further.

Below in conjunction with accompanying drawing, the present invention is described in more detail.Specifically include that

1. generation ballistic data:

The ballistic data with three strapdown inertial measurement unit error models will be generated as sample by Monte-Carlo method method Notebook data；

2. pair sample data carries out normal distribution-test

21) to step 1) sample data that obtains carries out ascending sort according to numerical values recited, by all of sample Notebook data is divided into N group, and the sample data number often organized is M_i, i=1,2 ... N represents i-th group；x_imaxTable Show the data value that in i-th group, sample data is maximum, x_iminRepresent the data value that in i-th group, sample data is minimum；Represent the estimation statistical average of i-th group of sample data；n_iRepresent i-th group of sample number According to data amount check；N represents that total number of samples is according to number；

22) below equation is utilized to try to achieve averageAnd variance

$\widehat{\mathrm{\μ}}=\frac{1}{n}\underset{i=1}{\overset{10}{\Σ}}{n}_i{x}_i^{*}=\stackrel{\‾}{X},$

${\widehat{\mathrm{\σ}}}^2=\frac{1}{n}\underset{i=1}{\overset{10}{\Σ}}{(n_i{x}_i^{*}-\stackrel{\‾}{X})}^2$

23) set the average of normal distribution, variance is respectively μ and variances sigma², whenAndTime, Represent sample data Normal Distribution, enter step 3)；Otherwise show that sample data disobeys normal distribution； Exit this method；Wherein A, B are threshold value；

3. carry out threshold value Estimating Confidence Interval

The distribution density of overall X normal distribution is:

Solve the inverse function of standard normal distribution, obtain x_max、x_min, finally give threshold interval (x_max, x_min)。

The content not being described in detail in description of the invention belongs to existing known to professional and technical personnel in the field Technology.

Claims (1)

1. a strapdown inertial measurement unit non-quantized dynamic threshold method of interval estimation, it is characterised in that step is as follows:

1) ballistic data is generated

The ballistic data with three strapdown inertial measurement unit error models will be generated as sample by Monte-Carlo method method Notebook data；

2) sample data is carried out normal distribution-test

21) to step 1) sample data that obtains carries out ascending sort according to numerical values recited, by all of Sample data is divided into N group, and the sample data number often organized is M_i, i=1,2 ... N represents i-th group；x_imax Represent the data value that in i-th group, sample data is maximum, x_iminRepresent the data value that in i-th group, sample data is minimum；Represent the estimation statistical average of i-th group of sample data；n_iRepresent i-th group of sample number According to data amount check；N represents the total number of sample data；

22) below equation is utilized to try to achieve averageAnd variance

$\widehat{\mathrm{\μ}}=\frac{1}{n}\underset{i=1}{\overset{10}{\Σ}}{n}_i{x}_i^{*}=\stackrel{\‾}{X},$

${\widehat{\mathrm{\σ}}}^2=\frac{1}{n}\underset{i=1}{\overset{10}{\Σ}}{(n_i{x}_i^{*}-\stackrel{\‾}{X})}^2;$

23) set the average of normal distribution, variance is respectively μ and variances sigma², whenAnd Time, represent sample data Normal Distribution, enter step 3)；Otherwise show that sample data disobeys normal state Distribution；Exit this method；Wherein A, B are threshold value；

3) threshold value Estimating Confidence Interval is carried out

The distribution density of overall X normal distribution is:

Solve the inverse function of standard normal distribution, obtain x_max、x_min, finally give threshold interval (x_max, x_min)。