CN103715507A - Antenna tracking method based on synchronous disturbance stochastic approximation algorithm - Google Patents

Abstract

The invention discloses an antenna tracking method based on a synchronous disturbance stochastic approximation algorithm. The method includes the first step of initializing parameters of a tracking algorithm and determining an angle rotation initial value of an antenna when the antenna enters a main lobe, and generating a synchronous disturbance vector through a MonteCarlo method, and the second step of processing AGC signals of a receiver system according to the synchronous disturbance stochastic approximation algorithm and carrying out optimizing processing on an AGC extreme value containing noisy signals to obtain the antenna directional angle at the AGC extreme value. If the number of preset maximum iteration times is reached in the iterative process, the tracking algorithm stops iteration, the antenna points at the AGC extreme value, and consequently accurate objective tracking of the antenna can be achieved. The antenna tracking method based on the synchronous disturbance stochastic approximation algorithm has the advantages of being simple in equipment, low in cost, high in tracking accuracy, good in speedability and the like.

Description

A kind of antenna tracking method based on simultaneous perturbation stochastic approximation algorithm

Technical field

The present invention relates in the satellite antenna communications field a kind of fast, high accuracy, satellite antenna automatic tracking method cheaply.

Background technology

Antenna system is one of important component part of satellite communication earth station, is to realize normally receiving signal that satellite forwards and to the equipment of satellite emission signal.Due to reasons such as drift, antenna and the mounting of celestial body, angular transducer demodulating errors, antenna usually can not be to satelloid.Therefore, satellite communication earth station need to be equipped with antenna tracking system conventionally, guarantee antenna can all-the-time stable reliably to satelloid, realize satellite communication station and normally work.

Satellite antenna automatic tracking system has comprised the multinomial technology such as Computer Control Technology, data acquisition and signal processing technology, sensor application technology, Precision Machinery Design technology, emulation technology, motor servo control technology, satellite communication technology, that electromechanical integration, automatic control technology are main body, the product that a plurality of subjects organically combine.

In order to make antenna automatically point to fast and accurately satellite, need to adopt automatic track algorithm.In the automatic track algorithm of current satellite antenna, mainly contain stepping tracking, conical scanning tracking, single-pulse track etc., but all there is weak point separately in these several algorithms: (1) stepping is followed the tracks of: equipment is simple, cheap, but conflicting between tracking accuracy and tracking velocity, tracking accuracy is not high; (2) conical scanning tracking: system equipment is simple, but feed departs from paraboloidal focus forever, and antenna gain declines; (3) single-pulse track: tracking accuracy is high, but system equipment is complicated, and cost is high.

Summary of the invention

The technical problem to be solved in the present invention is to provide a kind of antenna tracking method based on simultaneous perturbation stochastic approximation algorithm, have be easy to realize, algorithm is simple, optimizing is quick, high accuracy, the feature such as with low cost.

For solving the problems of the technologies described above, the technical solution used in the present invention is:

An antenna tracking method based on simultaneous perturbation stochastic approximation algorithm, comprises the following steps:

S1. antenna rotates search target, and the receiver AGC value detecting and the AGC threshold value of setting are compared, if receiver AGC value is less than AGC threshold value, antenna is proceeded to rotate with search target; When receiver AGC value is more than or equal to the AGC threshold value of setting, antenna enters main lobe, execution step S2;

S2. set the initial value of antenna directional angle sensing angle while entering main lobe for antenna; The parameter of initialization simultaneous perturbation stochastic approximation algorithm: set gain sequence a _k=a/ (k+A+1) ^αwith disturbance sequence c _k=c/ (k+1) ^γin non-negative coefficient a, c, α, γ, A, wherein, k is iterations, k=0,1,2,

S3. utilize MonteCarlo method to produce one 2 dimension random perturbation vector Δ _k;

S4. according to definite disturbance sequence c _kwith 2 dimension random perturbation vector Δs _k, determine that antenna is in current antenna directional angle

under simultaneous perturbation vector

antenna turns to respectively antenna directional angle

with

and gathering respectively current antenna AGC signal, it is worth respectively as two loss function measured values

with

According to loss function measured value

with

execution step S5;

S5. by formula

${\hat{g}}_k\left({\widehat{\mathrm{\&theta;}}}_k\right)=\left[\begin{array}{c}\frac{y_k^{(+)}-y_k^{(-)}}{2c_k{\mathrm{\&Delta;}}_{k1}}\\ \frac{y_k^{(+)}-y_k^{(-)}}{{2c}_k{\mathrm{\&Delta;}}_{\mathrm{<figure-callout\; id="2"\; label="k"\; filenames="HDA0000449441880000011.png"\; state="\{\{state\}\}">k</figure-callout>}2}}\end{array}\right],$ Calculate and produce gradient estimated value

Wherein, $y_k^{(+)}=y({\widehat{\mathrm{\&theta;}}}_k+c_k{\mathrm{\&Delta;}}_k),y_k^{(-)}=y({\widehat{\mathrm{\&theta;}}}_k-c_k{\mathrm{\&Delta;}}_k);$

S6. by formula

calculate the antenna directional angle while estimating antenna AGC extreme value as antenna directional angle next time;

If reach predefined maximum iteration time in iterative process, track algorithm termination of iterations, antenna direction

execution step S7; Otherwise, upgrade k=k+1, turn to execution step S3;

S7. driven antenna moves to antenna directional angle.

Wherein, in described step S3, random perturbation vector Δ _kfor each element be according to Bernoulli distribute to produce ± 1.

Wherein, in described step S2, α=0.602, γ=0.101, A is normal value, rule of thumb chooses A≤greatest expected iterations * 10%; A is constant, rule of thumb chooses; C is constant, is rule of thumb chosen for less positive number.

The beneficial effect that adopts technique scheme to produce is:

(1) the present invention adopts simple device as hardware system platform, forms simultaneous perturbation stochastic approximation antenna tracking system, has the advantages such as equipment is simple, cost is low, tracking accuracy is high, rapidity is good.System mainly comprises the execution units such as antenna surface, receiver, simultaneous perturbation stochastic approximation tracking module, driver, motor-driven chain.The AGC signal collecting according to receiver system, antenna directional angle when simultaneous perturbation stochastic approximation tracking module is estimated AGC extreme point, servomechanism driven antenna forwards the antenna directional angle at AGC extreme value estimated value place to, if met the demands, antenna stops operating, otherwise, the sensing angle while adopting simultaneous perturbation stochastic approximation algorithm to continue to estimate AGC optimal value, driven antenna is rotated.

(2) antenna pointing angle when the present invention adopts simultaneous perturbation stochastic approximation algorithm to estimate antenna AGC signal extreme value.By driven antenna, move and gather instant AGC signal, obtain antenna pointing angle accurately, can to stack, measure the AGC signal extreme value optimizing of noise signal, avoided the interference of noise signal, improve tracking accuracy and tracking velocity, avoided the contradiction between tracking accuracy and speed.

(3) the present invention, when antenna automatic tracking is found AGC optimal value, arranges constraints, has avoided antenna not stop judgement at Near The Extreme Point, and ceaselessly heterodromous problem, has improved mechanical structure wearing and tearing.

Accompanying drawing explanation

Fig. 1 is the structural representation of antenna tracking system of the present invention;

Fig. 2 is the control flow chart of antenna tracking method of the present invention;

In figure: 1, simultaneous perturbation stochastic approximation tracking module, 2, receiver system, 3, pitch axis azimuth axis control module, 4, pitching driver, 5, azimuth driver, 6, pitching motor, 7, azimuth-drive motor, 8, antenna surface.

Embodiment

Below in conjunction with the drawings and specific embodiments, the present invention is further detailed explanation.

As shown in Figure 1, antenna tracking system based on simultaneous perturbation stochastic approximation algorithm of the present invention, comprises simultaneous perturbation stochastic approximation tracking module 1, receiver system 2, pitch axis azimuth axis control module 3, pitching driver 4, azimuth driver 5, pitching motor 6, azimuth-drive motor 7 and antenna surface 8; Receiver system 2 is used for obtaining AGC signal; Pitch axis azimuth axis control module 3 is for carrying out the closed-loop control of elevation axis of antenna and azimuth axis; Pitching driver 4 and azimuth driver 5 are accepted external command and are controlled respectively pitching motor 6 and azimuth-drive motor 7 and then 8 rotations of driven antenna face; Simultaneous perturbation stochastic approximation tracking module 1, for according to simultaneous perturbation stochastic approximation algorithm, is processed the AGC signal of receiver system 2, and by the AGC extreme value optimizing containing noise signal is processed, the antenna directional angle while obtaining AGC extreme value.

The concise and to the point course of work of antenna tracking system of the present invention is as follows:

During antenna tracking system work, azimuth driver 4 driven antenna are rotated according to certain orientation, and when the AGC of receiver system 1 signal is more than or equal to setting threshold value, antenna enters simultaneous perturbation stochastic approximation tracing mode.Antenna initialization track algorithm parameter, determines antenna corner initial value, adopts MonteCarlo method to produce simultaneous perturbation vector.Pitching motor 6 and azimuth-drive motor 7 driven antenna faces 8 rotate, and gather the AGC signal of aerial receiver system 2, obtain two loss function measured values, according to the antenna AGC signal of the aerial receiver system 2 collecting, by simultaneous perturbation stochastic approximation tracking module 1, estimate gradient, and upgrade antenna direction θ value and estimate.If reach predefined maximum iteration time in iterative process, track algorithm termination of iterations, antenna is realized and is pointed to AGC extreme value, realizes antenna and accurately follows the tracks of.

As shown in Figure 2, the antenna tracking method based on simultaneous perturbation stochastic approximation algorithm of the present invention comprises the following steps:

S1. antenna rotates search target, and the receiver AGC value detecting and the AGC threshold value of setting are compared, if receiver AGC value is less than AGC threshold value, antenna is proceeded to rotate with search target; When receiver AGC value is more than or equal to the AGC threshold value of setting, antenna enters main lobe, execution step S2;

S2. set the initial value of antenna directional angle

sensing angle while entering main lobe for antenna; The parameter of initialization simultaneous perturbation stochastic approximation algorithm: set gain sequence a _k=a/ (k+A+1) ^αwith disturbance sequence c _k=c/ (k+1) ^γin non-negative coefficient a, c, α, γ, A, wherein, k is iterations, k=0,1,2, α=0.602, γ=0.101; A is normal value, rule of thumb chooses A≤greatest expected iterations * 10%; A is constant, rule of thumb chooses; C is constant, is rule of thumb chosen for less positive number;

S3. utilize MonteCarlo method to produce one 2 dimension random perturbation vector Δ _k; Can select Δ _keach element be according to Bernoulli distribute to produce ± 1;

S4. according to definite disturbance sequence c _kwith 2 dimension random perturbation vector Δs _k, determine that antenna is in current antenna directional angle

under simultaneous perturbation vector antenna turns to respectively antenna directional angle

with

and gathering respectively current antenna AGC signal, it is worth respectively as two loss function measured values

with

According to loss function measured value

with

execution step S5;

S5. by formula

${\hat{g}}_k\left({\widehat{\mathrm{\&theta;}}}_k\right)=\left[\begin{array}{c}\frac{y_k^{(+)}-y_k^{(-)}}{2c_k{\mathrm{\&Delta;}}_{k1}}\\ \frac{y_k^{(+)}-y_k^{(-)}}{{2c}_k{\mathrm{\&Delta;}}_{\mathrm{<figure-callout\; id="2"\; label="k"\; filenames="HDA0000449441880000011.png"\; state="\{\{state\}\}">k</figure-callout>}2}}\end{array}\right],$ Calculate and produce gradient estimated value

Wherein, $y_k^{(+)}=y({\widehat{\mathrm{\&theta;}}}_k+c_k{\mathrm{\&Delta;}}_k),y_k^{(-)}=y({\widehat{\mathrm{\&theta;}}}_k-c_k{\mathrm{\&Delta;}}_k);$

S6. by formula

calculate the antenna directional angle while estimating antenna AGC extreme value

as antenna directional angle next time;

If reach predefined maximum iteration time in iterative process, track algorithm termination of iterations, antenna direction

execution step S7; Otherwise, upgrade k=k+1, turn to execution step S3;

S7. driven antenna moves to antenna directional angle.

Claims (3)

1. the antenna tracking method based on simultaneous perturbation stochastic approximation algorithm, is characterized in that: comprise the following steps:

S1. antenna rotates search target, and the receiver AGC value detecting and the AGC threshold value of setting are compared, if receiver AGC value is less than AGC threshold value, antenna is proceeded to rotate with search target; When receiver AGC value is more than or equal to the AGC threshold value of setting, antenna enters main lobe, execution step S2;

S2. set the initial value of antenna directional angle

sensing angle while entering main lobe for antenna; The parameter of initialization simultaneous perturbation stochastic approximation algorithm: set gain sequence a _k=a/ (k+A+1) ^αwith disturbance sequence c _k=c/ (k+1) ^γin non-negative coefficient a, c, α, γ, A, wherein, k is iterations, k=0,1,2,

S3. utilize MonteCarlo method to produce one 2 dimension random perturbation vector Δ _k;

S4. according to definite disturbance sequence c _kwith 2 dimension random perturbation vector Δs _k, determine that antenna is in current antenna directional angle

under simultaneous perturbation vector

antenna turns to respectively antenna directional angle

with and gathering respectively current antenna AGC signal, it is worth respectively as two loss function measured values

with

According to loss function measured value

with

execution step S5;

S5. by formula

${\hat{g}}_k\left({\widehat{\mathrm{\&theta;}}}_k\right)=\left[\begin{array}{c}\frac{y_k^{(+)}-y_k^{(-)}}{2c_k{\mathrm{\&Delta;}}_{k1}}\\ \frac{y_k^{(+)}-y_k^{(-)}}{{2c}_k{\mathrm{\&Delta;}}_{k2}}\end{array}\right],$ Calculate and produce gradient estimated value

Wherein, $y_k^{(+)}=y({\widehat{\mathrm{\&theta;}}}_k+c_k{\mathrm{\&Delta;}}_k),y_k^{(-)}=y({\widehat{\mathrm{\&theta;}}}_k-c_k{\mathrm{\&Delta;}}_k);$

S6. by formula

calculate the antenna directional angle while estimating antenna AGC extreme value

as antenna directional angle next time;

If reach predefined maximum iteration time in iterative process, track algorithm termination of iterations, antenna direction

execution step S7; Otherwise, upgrade k=k+1,

turn to execution step S3;

S7. driven antenna moves to antenna directional angle.

2. a kind of antenna tracking method based on simultaneous perturbation stochastic approximation algorithm according to claim 1, is characterized in that: in described step S3, and random perturbation vector Δ _kfor each element be according to Bernoulli distribute to produce ± 1.

3. a kind of antenna tracking method based on simultaneous perturbation stochastic approximation algorithm according to claim 1, is characterized in that: in described step S2, and α=0.602, γ=0.101, A is normal value, rule of thumb chooses A≤greatest expected iterations * 10%; A is constant, rule of thumb chooses; C is constant, is rule of thumb chosen for less positive number.

Images