EP1152206B1 - Method and device for correcting aiming errors between apparatuses - Google Patents

Abstract

The method involves aligning a sensor device and a target measurement sensor in the actuator with the measurement target, determining the deviation between the positions of the target measurement sensor's sight line and of the measurement target detected by the target measurement sensor and recursively corrected the error vector using the least squares method and the existing alignment error vector as a controller input signal. The method involves correcting an alignment error vector. The sensor device is aligned with a measurement target (Ki), a target measurement sensor in the actuator is aligned with the measurement target, a deviation value (Di) between the position of the target measurement sensor's sight line ands the position of the measurement target detected by the target measurement sensor is determined, an existing alignment error vector is used as a controller input signal and the error vector is recursively corrected on the basis of the deviation using the method of least squares. Independent claims are also included for the following: an arrangement correcting an alignment error between a sensor device and an actuator controlled by the sensor device via a servo controller.

Description

The present invention relates to a method and a device for correcting alignment errors between devices of fire control systems and weapons systems according to the preamble of patent claims 1 and 10, respectively.

EP 0 314 721 B1 discloses a method for correcting alignment errors between carriages and devices arranged thereon, wherein the devices can be fire control systems and weapon devices. The procedure is carried out using device correction values of the coarse position of the installed devices measured in quiet fire control systems and weapons installations and their consideration in the servo controls of the carriages; the device correction values are known ex works and / or are determined from measured values.

It is an object of the present invention to improve such a method and to propose a device for its implementation.

The solution of this object is achieved in an advantageous manner according to the invention by a method according to claim 1 and by a device according to claim 10 .

This allows system deviations from a defined ideal geometry to be taken into account In order to calculate the control variables for the carriage servos the accuracy to increase in the shooting operation.

Other advantageous embodiments of the invention will become apparent from the other dependent Claims.

Fig. 1

eine schematische Darstellung der gegenseitigen Vernetzung von Sensoreneinrichtungen und Effektoreneinrichtungen bezüglich ihrer Lage,

Fig. 2

eine Einzelbeobachtung bei der Feinmessung gemäss der Erfindung,

Fig. 3

das Resultat einer Einzelbeobachtung gemäss Fig. 2,

Fig. 4

eine Darstellung zur Erläuterung des verwendeten Koordinatensystems,

Fig. 5

das Resultat eines ganzen Satzes von Beobachtungswerten, und

Fig. 6

das Resultat der erfindungsgemäss korrigierten Werte.

The invention will be explained in more detail below, for example, with reference to a drawing. Show it:

Fig. 1

a schematic representation of the mutual networking of sensor devices and Effektoreneinrichtungen with respect to their location,

Fig. 2

a single observation in the fine measurement according to the invention,

Fig. 3

the result of a single observation according to FIG. 2,

Fig. 4

a representation for explaining the used coordinate system,

Fig. 5

the result of a whole set of observational values, and

Fig. 6

the result of the values corrected according to the invention.

Fig. 1 shows a system with a total of five devices, namely two sensor devices in the form of Feuerleitgeräten T1, T2 and three computer-controlled Effektoreneinrichtungen in the form of guns G1, G2, G3. The sensor devices and the effector devices can be located on a ship or on land. All these devices T1, T2, G1, G2, G3 are included in Laffeten or bedding and mechanically at least roughly aligned.

2 , for example, a helicopter 10 and a simple system with a sensor device T and an effector device G are shown. The sensor device T can be, for example, a fire control or straightening device also designated T for controlling the gun G. The gun G may for example be provided with a TV sensor Sg . The topping device T controls over data or signal lines 11 the gun G. Both the gun G and the target device T are aimed at a common measuring target K, for example, a likewise designated K ball which is attached to a supporting cable 12 of the helicopter 10 degrees.

With such arrangements, a correction for an alignment _{error vector} B or several alignment _{error vectors} B _jk , in FIG. 1, for example B11, B12, B21, B22, B31, B32, is to be determined. It is assumed here that the alignment error vector B or the alignment error vectors B _{jk are} base vectors which are known and stored from coarse measurement measurements, factory measurements, etc.

By means of the method according to the invention, a fine measurement is carried out in order to improve these known values of the alignment _{error vectors} B and B _jk in several steps or after several measurements. For an alignment error vector B corrected with a calculated correction vector P _n , therefore, it holds after a number of i steps, where i are integer values from 1 to n: B (new) = B (old) + P n

After a number of n measurements it can be assumed that P _n ≈ P _s , where P _s corresponds to the real or correct intrinsically unattainable value for the correction of the system as such.

If, for example, the sensor devices or effector devices are located on a ship, a change in the weight of the ship as a result of change in the payload, of the available fuel or of a change in the shape of the hull, etc. results in a new value for the correction vector P _s , which can be determined by new, with the help of attached to the helicopter 10 ball K , measurements in turn approximately in the form of a new P _n value can be determined. Very small changes in the shape of the hull, for example by bending or torsion, especially after an explosion, cause a relatively large change in the reference angles. An object of the invention is to consider these very small changes.

The display shown in FIG. 3 shows how the TV sensor Sg, for example the measurement target K or the sphere K 'sees', in the actually assumed position generally with a certain offset from a crossing point 0 of a crosshair of the display. This shelf, which can be recognized directly by the TV sensor Sg, is a positional error which is the consequence or sum of all kind of system errors; System errors include, for example, mechanical inaccuracies due to manufacturing tolerances or wear, residual errors in the coarse bearing measurement, changes in the shape of the hull, measurement noise. The clip can be considered as a column vector Di with two components, which can be represented transposed as follows: D i = | dy i 'dz i '| T where dy _i 'and dz _i ' are the components of the gap vector D _i in the axes y 'and z', respectively. The amount d of the length of the gap vector D _i can be calculated according to FIG. 3 d = (dy i '2 + dz i '2 ) 1.2

The display of Fig. 3 is calibrated to a predetermined distance so that the components dy _i 'and dz _i ', which are in reality angles, can be represented by lengths or distances. For the
Gap vector D _i holds the relationship: D i = M i * P s + R i = D ic + R i with R _i = residual error

Factors which influence the residual error R _i are, in addition to the thermal noise, among others, the sea state, inaccuracies of the servo system and the fact that the operator can not bring a mark + shown in FIG. 3 exactly to the measurement target in its instantaneous position K _i ,

In the area of the straightening device T and the gun G , a coordinate system according to FIG. 4 is defined. If straightening device T and gun G are on earth, then, for example, the X- axis is directed to the north, the Y- axis to the east and the Z- axis to the earth center. If straightening device T and gun G are on a ship, then the X- axis is for example the longitudinal axis of the ship, the Y- axis the transverse axis of the ship and the Z- axis a clockwise, orthogonal to X- axis and Y- axis Axis. In the coordinate system defined by the X , Y and Z axes, each position that the measurement target K _i can assume is determined by three coordinates x _k , y _k , z _k . For practical reasons, however, the angle quantities α _k and λ _k are also used as coordinates in the shooting system, α _{k being} the side angle and λ _k the elevation angle; the quantities α _k and λ _k and are therefore redundant. The coordinates x _k, y _k, z _k are considered to be components of a target vector 0K _i, where α and the azimuth from these coordinates or the elevation angle λ can be calculated. The projection of the vector 0K onto the plane XY in FIG. 4 defines a straight line g; a straight line also lying in the plane XY and the straight line g at the zero point 0 perpendicular intersecting straight line is selected as the λ axis .

Δx_i

eine Drehung um die X-Achse,

Δy_i

eine Drehung um die Y-Achse,

Δz_i

eine Drehung um die Z-Achse und

Δλ_i

eine Drehung um die λ-Achse.

The recursively calculated correction vector P _i mentioned at the outset preferably has four components, as follows: P i = | Δx i Dy i Az i Δλ i | wherein Δx _i , Δy _i , Δz _i and Δλ _{i are} small angle values, where:

Δx _i

a rotation about the X-axis,

Δy _i

a rotation about the Y-axis,

Δz _i

a rotation about the Z-axis and

Δλ _i

a rotation about the λ-axis.

These rotations or tilting result from the fact that the plane of rotation of the effector device, ie the gun G, is not parallel to the plane of rotation of the sensor device, ie the straightening device T.

mit i = 1, 2, 3, ..... nThe error that results from this has two degrees of freedom and can therefore be corrected by the two rotations Δx _i about the X axis and Δy _i about the Y axis. The rotation Δz _i about the Z axis, on the other hand, also includes the rotation of the azimuth Δα. For each position of a measurement target defined by a target vector 0K _i or for each method step i, there exists a transformation matrix M _i defined as follows:

with i = 1, 2, 3, ..... n

For each method step i, a covariance matrix Si also exists as follows: S i = S i-1 - S i-1 * M i T * M i * S i-1 (M i * S i-1 * M i T + I) where I is a unitary matrix.

Finally, an error vector E (equation error) is defined by the following equation: e i = D i - M i * P i-1

worin C eine Konstante ist.The calculation is initialized with the following values: P 0 = | 0 0 0 0 | T and

where C is a constant.

The recursion starts with initial values P ₀ and S ₀ , with calculated values of M _i and measured values of D _i = | dy _i 'dz _i ' | ^T , where i starts at 1. From this, the values of E _i and S _{i are determined in} accordance with the recursion formulas given above, followed by P _i according to the following recursion formula: P i = P i-1 + Si * M i T * E i with i = 1, 2, 3, ... n

This recursive algorithm minimizes the following quality index J (p) (performance): J (p) = sum (i = 1, 2, .... n) (D i - M i * P i ) T * (D i - M i * P i )

The algorithm of the present invention is based on a special least squares method of applying the "least expensive" values by taking the sum of the squares of the respective differences between the observed value for D _i and the calculated value for D _ic ≈ M _i * P _n gives a minimum.

By means of the transformation matrix M _i , the calculated correction vector P _{i is transformed} into the vector D _i or the components Δx _i , Δy _i , Δz _i and Δλ _i into the components dy _i ', dz _i '. To avoid ambiguities in the observation plane (Figure 3), a matrix S is used. The matrix S is the covariance matrix listed above, which leads in particular to orthogonal-symmetric measurements to a diagonal-symmetric matrix with decreasing diagonal values, that is, the track Asp or convergence number tends to 0. Experiments with respect to the decimal places of this convergence number have shown that it is advantageous to select the value 49.25 or 492.5 etc. for the constant C, for example. At C = 49.25 the value of the trace of the covariance matrix S _n decreases from initially 99.99 ... to about 0.03 with a sufficiently large number n of measurements or steps. However, the constant C can also be 1 or have any value. After a number of n measurements or recursion steps, for example 25 <n <400, preferably n≈200, the value of P _n strives for the sought value P _s .

Fig. 5 shows, in each case by a cross +, a number of actual positions of the measurement target K borne by the helicopter 10. In Fig. 6 , the corresponding corrected values of these positions are shown. When the XYZ coordinate system exits a ship, the helicopter 10 preferably flies in a circular path with a radius of the order of 1.5 km, but helically or with increasing height α _Ti , λ _Ti , Δ _Ti around the ship. Based on the data obtained by the straightener T and taking into account previously known parameters, in particular parallax between the straightener T and the gun G, the sensor sighting line 0 of the gun G (instead of the small parallax offset firing line or gun barrel axis of the gun G) preferably automatically by the controller best possible directed to a measurement target K _i . The crossing point of the reticule of the sight line of the sensor Sg (Fig. 3) points in the direction in which the measurement target K _{i is} expected.

In FIG. 5 , therefore, each point marked with a cross + refers to a respective measured value of α _Ki or λ _Ki that is to say the side angle or the elevation angle of the gun G corresponding to the respective position K _i of the helicopter 10 Measuring target K correspond. By contrast, Fig. 6 corresponds to the theoretical values of α and λ, respectively, which would be measured under exactly the same conditions after the corrections according to the present method, if such a further measurement were practicable at all. In fact, it is impossible to carry out such further measurements with the helicopter 10 in exactly the same positions as in previous measurements, and under the same vessel conditions and so on.

Theoretically, thanks to the correction made, all points + would have to fall to zero point 0 in FIG . Because of the residual error R _i that is inevitably present in the system, as shown in FIG. 6 , the points + do not fall into the zero point 0, that is, one obtains statistically distributed deviations from the zero point 0, whose distribution, however, is averaging, that is Mean of the deviations of the points in both axes is zero.

Compared to other, multi - run, algorithms for the Computers of similar systems prove the algorithm of the present invention particularly advantageous in view of the fact that the initialization according to the invention completely is unproblematic and that it never comes to singularities (determinant = 0), so that you do not have to fear any 'derailment' of the program. Such 'derailments' could For example, if you try for each run, readings to one preset curve, such as a sinusoidal curve, adapt.

As with the system of patent EP 0 314 721 B1 , the measurement-based correction data with which the alignment error vectors are corrected has a real-time corrective misalignment effect. The measurements may be re-performed from time to time, for example, after four or six weeks, to adjust the correction data to changing conditions, such as a ship. This means that the measured values obtained from time to time can be integrated into the system and that they are therefore system-inherent and thus each correspond to an error that can not be directly observed.

The sensor device T (tracker) can be a sensor, a straightening device, a radar, laser or infrared device, etc., or several such devices can be combined. As effector devices G (guns) not only conventional guns such as cannons but also rocket launchers or laser cannons come into question. The measurements can be carried out for different G / T pairs B11, B12, B21, B22, ... (See FIG. 1 ), wherein a sensor device T can also control a plurality of effector devices G.

The equipment described on the basis of the figures can provide the necessary controls, computer means or hardware and programs or software to the various Method or sub-method according to the claimed variants or in any combination to realize the same.

Claims (10)

1. Method for correcting aiming errors between a sensor device (T; T1,T2) and an effecter device (G; G1,G2,G3) controlled by said sensor device by way of a servo control through correction of an aiming error vector (B), characterized by the following steps:

   a) aiming of the sensor device (T; T1,T2) at a test target (K_i );

   b) aiming of a target test sensor (Sg) present in the effecter device (G; G1,G2,G3) at the test target (K_i ), which therefore becomes the common test target (K_i ) of the sensor device (T; T1,T2) and the effecter device (G; G1,G2,G3);

   c) detection of a deviation (D_i ) between the location of the line of sight (0) of the target test sensor (Sg), resulting from the effecter device (G; G1,G2,G3) controlled by the sensor device (T; T1,T2), and the location of the test target (K_i ), as perceived by the target test sensor (Sg);

   d) use of an existing aiming error vector (B) as an input signal of the control;

   e) performance of a subsequent correction of the aiming error vector (B) on the basis of said deviation (D_i ), recursively by the method of the smallest error squares.
2. Method according to claim 1, characterized in that

   for correction of an aiming error vector (B), a vector (Pn) calculated recursively in steps i = 1 through i = n is obtained that exhibits two or more components or coordinates of the deviation (D_i ) for each measured position of the test target (K_i ), and

   the correction of a calculated vector (P_i ) is performed by multiplication of an initial value or a previously calculated vector by a transformation matrix (M_i ) producing transformation of the coordinates of the test target as a function of the azimuth (α_gi ) and the elevation (λ_gi ) of the target test sensor (Sg).
3. Method according to claim 1 or 2, characterized in that the transformation matrix is defined as follows:

   

   where i = 1, 2, 3 through n.
4. Method according to claim 2 or 3, characterized in that for each process step i a covariance matrix (S_i ) is used as follows: Si = Si-1- Si-1 * Mi T * Mi * Si-1 (Mi * Si-1 * Mi T + I) where I is a unit matrix, an initial value of S_o is used to initialize the recursion, and i = 1, 2, 3 through n.
5. Method according to one of the claims 2 through 4, characterized in that an error vector (E) is obtained by the following recursion formula: Ei = Di - Mi * Pi-1 where D_i = □dy_i' dz_i'□ is a vector with the components of the deviation values (d).
6. Method according to claim 5, characterized in that
   the recursive process steps with freely selectable values for P_o, S_o , with calculated values of M_i and with measured values of D_i = □dy_i' dz_i'□^T are performed starting with i = 1, and from this the error vector (E_i ) is derived by the stated recursion formula: Ei = Di - Mi * Pi-1 and the correction vector (P_i ) by the following recursion formula: Pi = Pi-1 + Si * Mi T * Ei where i = 1, 2, 3 through n.
7. Method according to one of the claims 2 through 5, characterized in that the correction vector (P_i ) is formed by two or more of the following four components: Δx_i, Δy_i, Δz_i and Δλ_i.
8. Method according to one of the claims 3 through 7, characterized in that the calculation is made with the correction vector P_i = □Δx_i Δy_i Δz_i Δλ_i□ and initialized by the following values Po = □0 0 0 0□T and

   

   where C is a constant and preferably 49.25.
9. Method according to one of the claims 1 through 8, characterized in that the common test target (K_i ) is led on prescribed paths, preferably by a helicopter (10).
10. Device for correcting aiming errors between a sensor device (T; T1,T2) and an effecter device (G; G1,G2,G3) controlled by said sensor device (T; T1,T2) by way of a servo control through correction of an aiming vector error (B), whereby said sensor device (T; T1,T2) is designed to be aimed at a test target (K_i );
    a target test sensor (Sg) is present in said effecter device (G; G1,G2,G3) that is designed to be aimed at said test target (K_i ), which can therefore become a common test target (K_i ) of said sensor device (T; T1,T2) and said effecter device (G; G1,G2,G3);
    display means are provided to detect a deviation (D_i ) between the location of the line of sight (0) of said target test sensor (Sg), resulting from said effecter device (G; G1,G2,G3) controlled by said sensor device (T; T1,T2), and the location of said test target (K_i ), as perceived by said target test sensor (Sg);
    computer means are provided
    to produce an input signal for said servo control from an existing aiming error vector (B), and
    to correct said aiming error vector (B) on the basis of said deviation (D_i ), recursively by the method of the smallest error squares.

Images