CN106778840A - Satellite telemetering data time series based on particular point linear segmented represents method - Google Patents

Abstract

A kind of satellite telemetering data time series based on particular point linear segmented represents method.(1) preset time sequence X={ x (t₁),x(t₂),…,x(t_n), the initial local extreme point of X is to meet formula (1) institute a little, and x (t₁) and x (t_n)；The initial local extreme value point set of X is designated asAnd p₁=1, p_m=n (2) introduces local extremum retention time parameter C value, further filters out the point that formula (2) is met in IM sequencesAs characteristic point；Wherein, 1≤j≤m 1；(3) starting point and terminal of X are also added to the set of characteristic points obtained after screening, final characteristic sequence is obtainedThe turning point of X meets formula (3) for all in former sequence, in formula, 1 ＜ i ＜ n.So as to obtain turning point characteristic sequenceMerge M and N, resequenced according to time index size, as the crucial point sequence of sequence X, be indicated with this key point sequence pair satellite telemetry initial data.

Description

Satellite telemetering data time series based on particular point linear segmented represents method

Technical field

Method is represented the present invention relates to a kind of satellite telemetering data time series based on particular point linear segmented, is belonged to and is defended The method for expressing technical field of star telemetry time series.

Background technology

Satellite telemetering data be satellite operation on orbit during the relevant satellite health that can obtain of operation management personnel With unique basis of health status.Satellite telemetering data is automated using methods such as data mining and machine learning, intelligence The analysis of energyization, has important value for the in-orbit operation maintenance of satellite and health control, is current space industry both at home and abroad The focus and challenge of research.However, the analysis of existing satellite telemetering data is set up on the basis of handmarking mostly, lack Weary automatic smart tags means, it is difficult to complete the mark to mass data.

The content of the invention

It is the invention aims to solve the problems, such as above-mentioned prior art, i.e., existing special based on time series form The linear segmented method for expressing for levying segmentation a little is not high to the extraction efficiency of time series key point, and characteristic loss rate is bigger than normal to ask Topic.And then a kind of satellite telemetering data time series based on particular point linear segmented of offer represents method.

The purpose of the present invention is achieved through the following technical solutions：

A kind of satellite telemetering data time series based on particular point linear segmented represents method,

(1) preset time sequence X={ x (t₁),x(t₂),…,x(t_n), the initial local extreme point of X is to meet formula (1) institute a little, and x (t₁) and x (t_n)

Wherein, 2≤i≤n-1.

The initial local extreme value point set of X is designated asWherein m≤n, and p₁=1, p_m=n

(2) local extremum retention time parameter C value is introduced, the point that formula (2) is met in IM sequences is further filtered outAs characteristic point.

p_j+1-p_j-1＞ C (2)

Wherein, 1≤j≤m-1.

(3) starting point and terminal of X are also added to the set of characteristic points obtained after screening, final characteristic sequence is obtainedWherein q₁=1, q_k=n.

The turning point of X has the set for meeting formula (3) more in former sequence：

In formula, 1 ＜ i ＜ n.

So as to obtain turning point characteristic sequenceWherein h≤n.

Merge M and N, resequenced according to time index size, as the crucial point sequence of sequence X, with this key point sequence Row are indicated to satellite telemetry initial data.

The present invention for initial data data volume it is big, there are problems that more, propose a kind of line based on particular point Property segmentation method for expressing (Special Points Segmentation, SPSegmentation) come in extracting original series Particular point, relative to existing method, adds steady pattern to the special turning point of other pattern switchings, and it is right to be effectively improved Adaptability of the satellite segment data in cluster analysis, realizes the feature extraction about subtracted with cluster to initial data, reduces Influence of the noise to follow-up clustering algorithm is reduced while amount of calculation.The crucial point sequence that the method for the present invention is extracted more can Reflect the trend characteristic of original data sequence, while remaining the details of morphology of more initial data.

Brief description of the drawings

Fig. 1 is the flow chart that satellite telemetering data time series of the present invention based on particular point linear segmented represents method.

Fig. 2 is the basic changing pattern figure of data sequence.

Specific embodiment

Below in conjunction with accompanying drawing, the present invention is described in further detail：The present embodiment is being with technical solution of the present invention Under the premise of implemented, give detailed implementation method, but protection scope of the present invention is not limited to following embodiments.

The satellite telemetering data time series based on particular point linear segmented involved by the present embodiment represents method, such as schemes Shown in 1, comprise the following steps that：

(1) preset time sequence X={ x (t₁),x(t₂),…,x(t_n), the initial local extreme point of X is to meet formula (1) institute a little, and x (t₁) and x (t_n)

Wherein, 2≤i≤n-1.

The basic changing pattern of the data sequence that formula (1) is represented is as shown in Figure 2.

The initial local extreme value point set of X is designated asWherein m≤n, and p₁=1, p_m=n

(2) local extremum retention time parameter C value is introduced, the point that formula (2) is met in IM sequences is further filtered outAs characteristic point.

p_j+1-p_j-1＞ C (2)

Wherein, 1≤j≤m-1.

(3) starting point and terminal of X are also added to the set of characteristic points obtained after screening, final characteristic sequence is obtainedWherein q₁=1, q_k=n.

The turning point of X has the set for meeting formula (3) more in former sequence：

In formula, 1 ＜ i ＜ n.

So as to obtain turning point characteristic sequenceWherein h≤n.

Merge M and N, resequenced according to time index size, as the crucial point sequence of sequence X, with this key point sequence Row are indicated to satellite telemetry initial data.

For preset time sequence X={ x (t₁),x(t₂),…,x(t_n), the particular point (Special in sequence Point, SP) extraction algorithm false code it is as follows.

Input：Original series X, local extremum retention time a, 3 points of turnover threshold value c.

Output：Special point sequence SPS={ x (t_s1),x(t_s2),…,x(t_sv)}

The control parameter that SPS is extracted determines the quantity of SP in sequence and represents quality.It is first right that parameter selection is generally required Data progress initial analysis and anticipation parameter are chosen.

The above, preferably specific embodiment only of the invention, these specific embodiments are all based on the present invention Different implementations under general idea, and protection scope of the present invention is not limited thereto, it is any to be familiar with the art Technical staff the invention discloses technical scope in, the change or replacement that can be readily occurred in should all be covered of the invention Within protection domain.Therefore, protection scope of the present invention should be defined by the protection domain of claims.

Claims (2)

1. a kind of satellite telemetering data time series based on particular point linear segmented represents method, it is characterised in that

(1) preset time sequence X={ x (t₁),x(t₂),…,x(t_n), the initial local extreme point of X is to meet formula (1) institute A little, and x (t₁) and x (t_n)

$\begin{array}{c}\{x\left(t_i\right)\&le;x\left(t_{i-1}\right)\&cap;x\left(t_i\right)<x\left(t_{i+1}\right)\}\&cup;\\ \{x\left(t_i\right)<x\left(t_{i-1}\right)\&cap;x\left(t_i\right)\&le;x\left(t_{i+1}\right)\}\&cup;\\ \{x\left(t_i\right)\&GreaterEqual;x\left(t_{i-1}\right)\&cap;x\left(t_i\right)>x\left(t_{i+1}\right)\}\&cup;\\ \{x\left(t_i\right)>x\left(t_{i-1}\right)\&cap;x\left(t_i\right)\&GreaterEqual;x\left(t_{i+1}\right)\}\end{array}---\left(1\right)$

Wherein, 2≤i≤n-1；

The initial local extreme value point set of X is designated asWherein m≤n, and p₁=1, p_m=n

(2) local extremum retention time parameter C value is introduced, the point that formula (2) is met in IM sequences is further filtered outMake It is characterized a little；

p_j+1-p_j-1＞ C (2)

Wherein, 1≤j≤m-1；

(3) starting point and terminal of X are also added to the set of characteristic points obtained after screening, final characteristic sequence is obtainedWherein q₁=1, q_k=n；

The turning point of X has the set for meeting formula (3) more in former sequence：

$|x\left(t_i\right)-\frac{x\left(t_{i+1}\right)-x\left(t_{i-1}\right)}{2}|>\mathrm{\&epsiv;}---\left(3\right)$

In formula, 1 ＜ i ＜ n.

So as to obtain turning point characteristic sequenceWherein h≤n；

Merge M and N, resequenced according to time index size, as the crucial point sequence of sequence X, with this key point sequence pair Satellite telemetry initial data is indicated.

2. the satellite telemetering data time series based on particular point linear segmented according to claim 1 represents method, its It is characterised by, for preset time sequence X={ x (t₁),x(t₂),…,x(t_n), the puppet of the particular point extraction algorithm in sequence Code is as follows：

The control parameter that SPS is extracted determines the quantity of SP in sequence and represents quality.