RU2561252C1 - Selective navigation system - Google Patents

Abstract

FIELD: instrument making.

SUBSTANCE: selective navigation system (SNS) contains an inertia navigation system (INS) with one output, receiver of satellite navigation system (RSNS) with one output, radar station (RS) with one output, unit for determination of degrees of observability and formation of measurements having three inputs and one output, Kalman filter and summation unit having two inputs and one output. INS output is connected with the first input of the unit for determination of degrees of observability and formation of measurements, its second input is connected with RSNS output, and the third input is connected with RS output. INS output is also connected with the second input of the summing unit, and Kalman filter output is connected with the first input of the summing input. SNS is equipped with squaring device and accumulating device, and unit for determination of degrees of observability and formation of measurements is made with the forth input. In the unit for determination of degrees of observability and formation of measurements in the criterion of degree of observability the dispersion of state vector components during flight are calculated.

EFFECT: improvement of accuracy and reliability.

2 cl, 2 dwg

Description

The invention relates to the field of development of navigation and orientation systems, in particular to the correction of their errors, numerical criteria for the degree of observability of navigation systems (NK) with an inertial navigation system (ANN).

Known selective navigation complex (SNK), including ANN with one output, satellite navigation system receiver (PSNS) with one output, radar station (radar) with one output, the unit for determining the degrees of observability and the formation of measurements, having three inputs and one output, filter Kalman and an adder having two inputs and one output, while the ANN output is connected to the first input of the unit for determining the degrees of observability and the formation of measurements, the second input of which is connected to the PSNS output, and the third input is connected to the output LS, the ANN output is also connected to the second input of the adder, and the Kalman filter output is connected to the first input of the adder, the output of which is the SNK output (see K. Neusypin. Modern systems and methods of guidance, navigation and control of aircraft. M., Ed. MGOU, 2009, p. 91; p. 156, p. 158-159, see figure 1).

The ANN signal is fed to the first input of the block for determining the degrees of observability and forming measurements, where it is subtracted from the signal received at the second input from the PSNS to form measurements z1, the signal from the radar, which is subtracted from the signal, is fed to the third input of the block for determining the degrees of observability and forming measurements with an ANN for forming measurements z2, then measuring samples Z1 and Z2 are formed by summing 30 received measurements, respectively, which are further averaged. Then, the total values of Z1 and Z2 go to the criterion of the degree of observability to calculate the values of the degrees of observability of each component of the state vector, which includes errors of the ANN in determining the velocity, the angle of deviation of the GPS and the drift of the GPS. Further, the degrees of observability are compared and the signal for further processing is selected at the highest value, which enters the Kalman filter, where the state vector is estimated. From the Kalman filter output, a signal proportional to the ANN error estimates is fed to the first input of the adder, where it is subtracted from the signal from the ANN to compensate for the ANN errors.

At the output of the adder, the signal is equal to the true navigation information with an estimation error.

In practical applications, the level of measuring noise usually changes, so it is not possible to make an unambiguous choice of the structure of the complex in advance. The effectiveness of SNK can be demonstrated analytically.

When measuring additional sources of information about the angle of deviation of the SHG from the horizon plane S ^-1 has the form:

(δV), имеет вид:Dispersion of noise reduced to ANN error in velocity determination $$r_{\mathrm{one}}^{*}$$

(δV), has the form:

Substituting the numerical values in the expression for the variance of the reduced noise:

T = 2 min, β = 10 ^-2 min ^-1 , R = 6370000 m, g = 35280 m / min ² , r = 10 ^-14 rad ² .

.We get $$r_{\mathrm{one}}^{*}(\delta V)=506\raisebox{1ex}{$m^2$}\!\left/ \!\raisebox{-1ex}{$m\mathrm{and}{n}^2$}\right.$$

.

.When using the DISS as an additional source of information, we accept the variance $$r=2500\raisebox{1ex}{$m^2$}\!\left/ \!\raisebox{-1ex}{$m\mathrm{and}{n}^2$}\right.$$

.

Therefore, for instruments having similar characteristics for calculating ANN errors in determining the velocity, it is advisable to measure the angles of deviation of the GPS from the horizon plane and to estimate the errors in determining the velocity from these measurements. In this case, when measuring simultaneously δV and φ in the estimation algorithm, if the component of the state vector is directly measured, then the estimation error does not change from whether the other components are measured or not.

In this case, when simultaneously measuring the speed of the aircraft and the deviation angles of the GPS in the estimation algorithm, if the component of the state vector is directly measured, then the estimation error does not change from whether other components are measured or not. Therefore, in this case, it is more expedient to use only measurements of the angles of deviation of the SHG.

In the drawing (figure 1) presents a diagram of a known SNK.

The designations in the drawing are as follows:

- оценки погрешностей ИНС; $${\tilde{x}}_k$$

- ошибка оценивания; ФК - фильтр Калмана; λ - блок определения степеней наблюдаемости и формирования измерений z_k.θ _k is the true navigation information; x _k - ANN errors; $${\stackrel{⌢}{x}}_k$$

- estimates of ANN errors; $${\tilde{x}}_k$$

- estimation error; FC - Kalman filter; λ is a block for determining the degrees of observability and the formation of measurements z _k .

A disadvantage of the known SNK is that in the criterion of the degree of observability of the variance of the studied components of the state vector, they are determined a priori under the conditions of preflight preparation, and in flight they can vary significantly, which leads to an erroneous determination of the degrees of observability and, consequently, to the wrong choice of the measuring signal and an increase in errors estimation by the Kalman filter and reducing the accuracy of determining the navigation parameters of aircraft.

The task of the invention is to increase the accuracy of measurements of SNK, increase the reliability of SNK.

The technical result is to reduce errors in the estimation of ANN errors by selecting the optimal SNK structure by increasing the accuracy of calculating the degrees of observability of ANN errors.

The indicated task and technical result are achieved by a selective navigation complex (SNK), including an ANN with one output, a satellite navigation system receiver (PSNS) with one output, a radar station (radar) with one output, a unit for determining degrees of observability and forming measurements, having three inputs and one output, Kalman filter and adder having two inputs and one output, while the ANN output is connected to the first input of the unit for determining the degrees of observability and the formation of measurements, the second input of which is connected to PSNS output, and the third input is connected to the radar output, ANN output is also connected to the second adder input, and the Kalman filter output is connected to the first adder input, while the SNK is equipped with a quadrator and a storage device, and the unit for determining the degrees of observability and the formation of measurements is performed with the fourth input, and the output of the Kalman filter is additionally connected to the input of a quadrator, the output of which is connected to the input of the storage device, the output of the storage device is connected to the fourth input of the step detection unit its observability and the formation of measurements.

And also by the fact that in the block for determining the degrees of observability and the formation of measurements in the criteria for the degree of observability, the variances of the components of the state vector during the flight are calculated.

Figure 1 shows a diagram of the SNK of the prototype.

Figure 2 shows a diagram of the proposed SNK.

The selective navigation complex (SNK) includes ANN 1 with one output 2, a satellite navigation system (PSNS) 3 with one output 4, a radar station (radar) 5 with one output 6, a unit for determining degrees of observability and forming measurements 7, having three inputs 8, 9, 10 and one output 11, while output 2 of ANN 1 is connected to the input 8 of the unit for determining degrees of observability and forming measurements 7, the second input 9 of which is connected to output 4 of the PSNS 3, and the third input 10 of block 7 is connected to output 6 Radar 5, as well as Kalman filter 12, adder 13, kva the radiator 14 and the storage device 15. The output 11 of the block 7 is connected to the input 16 of the Kalman filter 12, the output 17 of which is connected to the first input 18 of the adder 13 and the input 19 of the quadrator 14, the output 20 of the quadrator 14 is connected to the input 21 of the storage device 15, the output of which 22 connected to the fourth input 23 of block 7. Output 2 ANN 1 is also connected to the second input 24 of the adder 13, the output 25 of which is the output of the SNK.

The Kalman filter is made in accordance with an algorithm of the form:

(Kuzovkov N.T., Salychev O.S. Inertial navigation and optimal filtering. M.: Mechanical Engineering, 1982, 215 p.).

The ANN error equations have the form:

where f is the matrix of the model; x — state vector includes ANN errors; W — input noise vector.

Part of the state vector is measured:

Here: Z _{k + 1} is the m-vector of measurements; V _{k + 1} is the m-vector of measurement errors;

H _{k + 1, k} is the (m × n) -matrix of measurements.

In the Kalman filter: R _{k + 1} - covariance matrix of the measuring noise V, Q - covariance matrix of the input noise w, K _{k + 1} - filter gain matrix:

. $$V_{k+\mathrm{one}}=Z_{k+\mathrm{one}}-H_{k+\mathrm{one,}k}{Ô}_{k+\mathrm{one,}k}{x}_k\wedge$$

.

Kalman filter has the form:

Here P _{(k + 1) / k} is the a priori covariance matrix of estimation errors; P _{k + 1} is the posterior covariance matrix of estimation errors. Using the Kalman filter, the entire system state vector is restored and the influence of the measuring noise is suppressed.

The quadrator 14 is made in the form of a matrix amplifier and is a logic circuit operating according to the x-x2 algorithm and assembled on the basic "and-not" logic elements (555LA3).

The quadrator 14 performs the squaring of the signal proportional to the estimates of the ANN errors.

The storage device 15 is made on a PIC1684F microprocessor-based microchip and calculates the variance of ANN error estimates based on 30 measurements of z in accordance with the formula:

The adder is made in the form of a digital logical adder.

The criterion of the degree of observability has the form (Neusypin K.A., Proletarsky A.V., Tsibizova T.Yu. Aircraft control systems and information processing algorithms. M., Publishing House of MGOU, 2006):

Here: M [(x ⁱ ) ² ] is the variance of an arbitrary ith component of the state vector;

M [(y ⁱ ) ² ] is the dispersion of the directly measured state vector.

In the criterion of the degree of observability (4), the measure of observability is a scalar. This feature favorably distinguishes the criterion from the known ones, since it allows a comparison of the degrees of observability of the components of various state vectors.

The equations of errors of the inertial navigation system have the form (Salychev O.S. Scalar estimation of multidimensional dynamic systems. M., Mechanical Engineering, 1987, 216 pp.):

where δV _k is the ANN error in determining the speed; ε _k is the drift velocity of the SHG; φ _k is the angle of deviation of the SHG relative to the accompanying trihedron.

We compose the vector of measurements in the form:

Where

Then, for the direct measurement of the components of the state vector, we obtain the following equations:

Define the variance of the measuring noise reduced to the angle of deviation of the SHG relative to the accompanying trihedron:

where r is the variance of the error in the measurement of speed, which is subject to direct measurement using external information.

In accordance with expression (8), we determine the degree of observability of the SHG deviation angle relative to the accompanying trihedron:

The degree of observability of the SHG drift velocity is determined similarly:

We substitute the numerical values of the parameters obtained as a result of a semi-natural experiment with a real ANN. The ANN error in determining the velocity is 60 m / min, the platform deviation angle of the relative accompanying trihedron is -2.10 ^-4 rad, the drift velocity is 10 ^-5 rad / min, the sampling period is chosen to be 1 minute. As a result, we find that the degree of observability of the GPS deviation angle relative to the accompanying trihedron is 0.01, and the degree of observability of the GPS drift velocity is 0.0001.

The obtained values of the degrees of observability have a clear physical meaning. The relative error in estimating the observed component of the state vector with respect to the estimated nominal value in the case of estimating the deviation angle will be the same as the relative error in estimating the directly measured component after 100 minutes, and in the case of the drift velocity, after 10,000 minutes.

The selected value of the measurement sample for calculating the variance in block 15 is justified by the fact that in practical applications in flight conditions after 30 steps of calculating estimates by the Kalman filter, the external operating conditions of the aircraft can change (the appearance of active and passive interference) and it is advisable to confirm the choice of the composition of measurement systems or determine the new composition of the SNK.

SNK works as follows:

The signal from output 2 of ANN 1, proportional to the true navigation information about the parameters of the aircraft with an ANN error, is fed to the first input 8 of block 7, the signal to output 4 of the PSNS 3, proportional to the true navigation information with an error of PSN, is fed to the second input 9 of block 7, the signal from the output 6 of the radar 5, which is proportional to the true navigation information with the radar error, goes to the third input 10 of block 7. In block 7, the signal from input 8 is subtracted from the signal from input 9, the difference signal is proportional to the mixture of errors ANN and PSNS z1, which is then converted to according Twi with the formula:

In block 7, the signal from the second input 9 is subtracted from the signal from the third input 10, the difference signal is proportional to the error mixture of ANN and radar z2, which is then converted in accordance with the formula:

, который использован в критерии степени наблюдаемости (4) формула (9). Вычисленные степени наблюдаемости сравниваются и по наибольшему значению выбирается сигнал z для дальнейшей обработки, который с выхода 11 блока 7 поступает на вход 16 фильтра Калмана 12, на выходе которого 17 сигнал пропорционален оценке вектора состояния (оценке ошибок ИНС $$x\wedge$$

). С выхода 17 фильтра Калмана 12 сигнал, пропорциональный оценкам ошибок ИНС, поступает на первый вход 18 сумматора 13 и на вход 19 квадратора 14, где преобразуется в соответствии с формулой $$x^2\wedge$$

. С выхода 20 квадратора 14 сигнал, пропорциональный квадрату оценки вектора состояния, включающего ошибки ИНС, поступает на вход 21 накопительного устройства 15, где преобразуется в соответствии с формулой:Then, the values of Z1 and Z2 go to the criterion of the degree of observability to calculate the values of the degrees of observability of each component of the state vector, including the ANN errors in determining the velocity, the angle of deviation of the GPS and the drift of the GPS. The fourth input 23 of block 7 receives a signal proportional to the variance of the ANN error estimate $$M\overline{({\phi }^2)}$$

, which is used in the criterion of the degree of observability (4), formula (9). The calculated degrees of observability are compared and the signal z is selected at the highest value for further processing, which from the output 11 of block 7 is fed to the input 16 of the Kalman filter 12, the output of which 17 is proportional to the estimate of the state vector (ANS error estimate $$x\wedge$$

) From the output 17 of the Kalman filter 12, a signal proportional to the error estimates of the ANN is fed to the first input 18 of the adder 13 and to the input 19 of the quadrator 14, where it is converted in accordance with the formula $$x^2\wedge$$

. From the output 20 of the quadrator 14, a signal proportional to the square of the state vector estimate, including ANN errors, is input to the storage device 15, where it is converted in accordance with the formula:

At the output 22 of the storage device 15, the signal is proportional to the variance of the ANN error estimate.

From the output 2 of the ANN 1, the signal proportional to the true navigation information with the error of the ANN x is also fed to the second input 24 of the adder 13, where it is subtracted from the signal from the first input 18, which is proportional to the error estimate of the ANN x, at the output 25 of the adder 13, the signal is the output of the SNK and equal to true navigation information with an estimation error.

The error in estimating ANN errors is significantly smaller than the ANN error, thereby compensating for ANN errors, which leads to an increase in the accuracy of aircraft navigation information.

In the prototype, in the criterion of the degree of observability, block 7 used the variance of the GPS deviation angle, a priori determined in preflight preparation, and in the proposed SNK, the estimated dispersion of the GPS deviation angle is determined in flight taking into account the current features of the aircraft motion.

In contrast to the prototype, the proposed SNK has greater accuracy due to a more accurate determination of the degree of observability of ANN errors, which makes it possible to include in the composition of the used measuring systems those systems that provide the greatest accuracy in estimating ANN errors and, accordingly, greater accuracy of the SNK.

Claims (2)

1. Selective navigation system (SNK), including an ANN with one output, a satellite navigation system receiver (PSNS) with one output, a radar station (radar) with one output, a unit for determining degrees of observability and the formation of measurements, having three inputs and one output, Kalman filter and adder having two inputs and one output, while the ANN output is connected to the first input of the unit for determining the degrees of observability and the formation of measurements, the second input of which is connected to the PSNS output, and the third input is connected to the radar output d ANN is also connected to the second input of the adder, and the output of the Kalman filter is connected to the first input of the adder, characterized in that it is equipped with a quadrator and a storage device, and the unit for determining the degrees of observability and the formation of measurements is made with the fourth input, and the output of the Kalman filter is connected additionally with the input of the quadrator, the output of which is connected to the input of the storage device, the output of the storage device is connected to the fourth input of the unit for determining the degrees of observability and the formation of measurement . 2. The selective navigation system according to claim 1, characterized in that in the unit for determining the degrees of observability and the formation of measurements in the criteria for the degree of observability, the variances of the components of the state vector during the flight are calculated.

Images