IL95230A - Self adjustable gyro arrangement with double axis platform - Google Patents

Description

TECHNICAL FIELD The invention concerns a self adjusting gyro system which includes a double axis gyroscope each having a take off and a torque creator at each intake level at which the gyro is located at the stabilizing element of a du-axial platform, the servo motors of which being active are being controlled by signals of the take off of the gyroscope and at which there are provided inclination causing mean at the stabilizing element, from the signals of which controlling signals can be derived, which latter are applied to the torque creator of the gyroscope.

BASIC STATE OF ART EP 81-0 048 212 Specification describes a course-position-reference device. The course position reference device comprises a du-axial electrically moored gyroscope. The gyroscope has a twist-axis and two axes which are normal relative to each other and to the twist axis. At each of the axes of the gyroscope there are positioned a take off and a torque creator,. The gyroscope is located in the stabilized element of a du-axial platform. The platform includes an azimuth-frame which is turnable by a servo motor about a substantially vertical azimuth axis relative to a housing. Within the azimuth frame is provided an internal frame which is turnable by a servo motor about an axis which is normal to the azimuth axis, substantially horizontal second from axis, also turnable by a servo motor. The stablized element which is attached to the internal frame includes, apart from the gyroscope, two acceleration meters. The sensibility axes of the two acceleration meters are normal relative to one another. The sensibility axis of the first acceleration meter extends parallel relative to the substantially horizontal twist axis of the gyroscope. The sensibility axis of the second acceleration meter extends parallel relative to the single, substantially horizontal entrance axis of the gyroscope. The output signal of the first acceleration meter is directed via amplifiers with appropriately determined transmission function to the torque creation means of the du-axial, electrically moored gyroscope. The second acceleration meter makes correctly signals compensating the influence of the vertical component of earth rotation. Tappings of the gyroscope are directed, via amplifiers, to the servo motors of th platform. With the said arrangement the axis of rotation of the gyroscope is maintained in the horizontal and due to the horizontal component of earth rotation is adjusted toward the north.

The said known arrangement is subject to disturbances. In case that it is tried to suppress such disturbances by means of time delay of the amplifiers, the setting time becomes unduly long.

The DE C2-31 1 342 specification describes a course-position-reference device wherein a du-axial platform is employed. There the azimuth angle is obtained by integration of a commutating gyro signal. Further, there is provided a compensation of influence of earth rotation. As support of the gyroscope acceleration signals are derived from the acceleration meter at an average still standing carrier. These former signals are inserted into the system of an imagined "virtual" platform which is turned in the azimuth towards the north and form deviation signals. From these deviation signals correction signals are formed and are connected to the torque moment creator of the du-axial platform or respectively to an integrator for the gyro-turn signal.

The DE-B-29 03 282 describes the elimination of gyroscope faults by turning the gyroscope about its twin-axis to different angular positions such that differences of measure signals are produced. At that arrangement the turning axis is positioned vertically. The north direction is determined by the relation of signals to two input axes which are staggered relative to one another by 90°, at which components of the horizontal component earth rotation becomes effective.

It is known through specification DE-B-29 03 282 to eliminate gyro faults by turning the gyroscope through a number of different angles about an axis of rotation and so to create different measuring signals. At that arrangement the axis of rotation is vertical. North direction is being determined by relation of two inlet axes which are staggered by 90° relative to one another at which compo¬ nents of the horizontal component of earth rotation are effective.

Similar arrangements are disclosed by the description of DE-CZ-31-43.527, DE-B-29 10 282 and DE-Cl-32 27 568. DE-C3-30 28 649 describes a device for determination of north direction at which an electrically moored two axial gyroscope with horizontal axis of rotation and equally horizontal own axis is turnable to different angular positions about a vertical azimuth axis on an azimuth frame. On the azimuth frame there are two acceleration meters with mutually normally extending horizontal sensibility axes. In a memory there are stored fault parameters. Course and position signal are created by combining gyro- and acceleration signals and the fault parameters .

A similar set scheme is described in DE-C2-30 50 575. Another arrangement with horizontal twist (couple) axis and measuring in various positions is described in DE-C2-30 19 372.

The German specification DE-C2-30 46 822 relates to a method of determination of north direction at which similar to the already mentioned DE-C3-30 28 649 at least three measuring steps are to be effected of the exit magnitude, in turn dependent of the inlet angle speed at different position turned against the zero position about - - definable angles of the pendulum body, so as to calculated the north direction.

DE-C2-32 33 612 relates to a device for determination of north including a du-axial platform and a du-axial gyroscope. The platform uncouples the gyroscope from a carrier about two horizontal platform axes. The twist axis of the gyroscope is in the vertical. Two acceleration meters horizontalize the platform above the gyro. In one embodiment of the said DE-C2-32 33612 is an electrically moored gyroscope. By way of integration of the gyro signals the position angles of a "virtual" platform are obtained. A first transformation - switching transforms the acceleration signals of the acceleration meters into the system of the the virtual platform. In this system the signals are processed for creation of position sizes. The obtained sizes are re-transferred into the system of the gyroscope and are superposed at the entry of the integrators to the gyroscope signals.

DISCLOSURE OF INVENTION It is the object of the invention to provide a self-adjusting gyroscope arrangement of the type referred to above which at minor affectability by external factors assures a quick and exact adjustment of the axis of rotation in the north direction.

This object is being obtained thereby that: a) Signals of the inclination causing means are set at an optimal estimation which gives estimated values for position faults of the stabilized element. b) The estimated values of the position faults are set at an optimal controlling means which gives adjustment signals, and c) The adjustment signals of said controlling means are introduced to each one of the creators of rotation moments of the gyroscope.

The estimation of position calculates estimated values of the fault angles, subject to consideration of distortion due to the surroundings out of signals of gradient sensors. Thereby externally caused disturbances are eliminated. An optimal control means supplies position signals out of these fault angles. The optimal control means permits speedy adjustment . of the platform of the gyroscope.

Practical embodiments of the invention are subjects of ancillary claims.

Short Description of Drawings Figure 1 is a schematical view and shows the mechanical build up of the gyroscope arrangement and the general structure control circuit.

Figure 2 is a schematical layout, similar to Figure 1 and illustrates an embodiment of the circuit.

Figure 3 is a schematical view, similar to Figure 2 and illustrates a second embodiment of the circuit.

Figure shows schematically the stabilized element with the sensor.

Figure 5 is a block diagram showing the basic structure of fault compensation.

Figure 6 is a block diagram and shows a linear temperature pattern for exact compensation.

Figure 7 illustrates the compensation of gyro faults by means of setting the gyroscope once to north and once to south, showing the employed angles.

Figure 8 is a schematical view, similar to Figure 3 and shows fault identification at the embodiment of Figure 3.

Preferred Embodiment of the Invention At Figure 1 the numeral 10 indicates a du-axial platform. Platform 10 includes a first frame in the shape of an azimuth-frame 12. The frame 12 is turnable about a substantially vertical axis 14 in a casing 16. The azimuth frame 10 is turnable by means of a servomotor 18 relative to casing 16. An angle definer 20 fixes the angle of rotation Ψ of the azimuth frame 12 about the azimuth axis.

In azimuth frame 12 is seated a second frame 22 being rotatable about a platform axis 21 which is normal to the azimuth axis 14. The frame 22 can be turned by a servomotor 23 Rela ive to the azimuth axis 12. The second frame 22 constitutes a stabilized element 24. The element 24 is decoupled - in a manner to be described - from the movements of the casing over the substantially vertical azimuth axis 14 and the substantially horizontal platform axis 21. No decoupling occurs relative to a third axis 26 which is normal relative to azimuth axis 14 and platform axis 21. Thus the stabilized element 24 may be inclined about this axis 26. The stabilized element 24 defines a system of coordinates with a coordinate axes x^, y^ and z^. The coordinate axis y^ is of identical location with platform axis 21. Within the stabilized element is placed a du-axial, electrically moored gyroscope 28. The axis of rotation 30 of gyroscope 28 extends in the direction coordinate axis x^. A first entry axis 32 of gyroscope 28 is parallel to coordinate axis y^ and platform axis 21. A second entry axis 34 of gyroscope 28 is parallel with the coordinate axis z^ and thus with the azimuth axis 14. On the first o' · -axis 32 is seated a first take off 36. The first take off 36 supplies a take off signal as determined by the movement (not shown) of the gyroscope 28 from zero position, about the axis 34. On entry axis 32 there is seated also a first creator of moment of rotation 40; The latter could effectively cause rotational movement about axis 32 on the gyro rotor. On the second entry axis 34 there is positioned a second take off 42. The second take off 42 supplier a signal determined by the movement of the gyro rotor about axis 34 relative to the casing 38, from zero position. On the second entry axis is also seated a second creator of moment of rotation 44. The second creator 44 of the moment of rotation could cause a rotary moment about the second axis 34 onto the gyro rotor.

In the stabilized element 24 there is provided an acceleration meter 40. The axis of sensibility (indicated by an arrow) of the first acceleration meter 46 extends parallel to the first entry axis 32 of gyro 28. In the stabilized element 24 there is also provided a second acceleration meter 48. Also indicated by an arrow, the sensibility axis of the second acceleration meter 48 is parallel with the first axis 30 of gyroscope 28.

The take off signal of the first take off 36 is switched onto the servo motor 23 via an amplifier 50 turning the frame 22 relative to azimuth frame 12. The take off signal of the second take off 42 is switched, via an amplifier 52 onto the servo motor 18 which turns the azimuth frame 12 relative to casing 16.

The outgoing signals of the acceleration meters 46 and 48 are adjusted to an optimal condition estimator 54. The outgoing signals of the acceleration meters 46 and 48 are united to a vector z. The condition estimator 54 supplies an optimally estimated condition vector x. The elements of these condition vectors are positive faults of the platform, i.e. the angle faults ey and ez about platform axis 21. The estimate values x are switched to a control means 56. The optimal control means 56 supplies manipulated variables which can be united into a manipulative variable vector U. The manipulated variables are each switched to rotation moment creators 40 and 44.

At the casing 16 there are provided further acceleration meters 58 and 60 which are normal both mutually and rotative to a sensibility axis and the azimuth axis 14.

The mechanical build up of the gyro arrangement of Figure 2 is in accordance with Figure 1. Identical parts are given the same reference indications as in Figure 1.

In Figure 2 the construction of the optimal condition estimator 56 and the optimal control means are shown.

In the embodiment of Figure 2 the optimal condition estimator receives the signals acx from acceleration meter 48. The acceleration signals illustrate acceleration in the direction of coordination axis xc and supply a measure for the indication of the stabilized element 24 about the platform axis 21. The acceleration signals acx from acceleration meter 48 are filtered by a low-pass filter 58. The filtered acceleration signal acx is scanned in tact with a tact period T and is ,digitalized. This is symbolized in Figure 2 by means of a switch 60. There is obtained a course of measurement values zn, in which the index "n" represents the current number of tact.

The optimal condition estimator 54 is a Kalman-Filter . The Kalman-Filter is characterized by a condition - or system equation and a measurement equation. The condition - or system equation supplies the dynamic behaviour of a model of the system, in the present case of the platform 10. In practice the condition equation indicates the change of a condition vector, i.e. of a vector which is formed by the condition variables as a function of the condition vector itself. The measurement equation supplies the co-relation between observed magnitudes of measurements and condition vectors.

In the present case the condition vector xn is represented by the position fault of the stabilized element 24 in relation to azimuth axis 14 and the platform axis 21.

With that condition vector the condition - or system equation (written in the permanent form as a differential equation): Wherein Q is the horizontal component of earth rotation Ucv("fc) and Ucz(t) are the manipulated variables of the optimal control means 56 and dy, dz are the gyro drifts. A change of the position fault is being created thereby that at one entry axis of the gyroscope one component of the horizontal component nc of earth rotation becomes effective, this component being proportional to the position fault at small angle position, further by manipulated variables of optimal control means 56 and the gyro drifts. The measurement equation ties the measured magnitude acx(-) to the condition magnitudes by means Wherein the position angle of the stabilized elements 24 is defined by the non-stabilized third axis which is normal relative to platform axis 21. "bx" indicates the zero point fault of the acceleration meter 48.

The estimate values 0y,n and 0z,n are formed in a alman-Filter in the conventional way show that the pre-estimated value for the condition vector is corrected with a correction term. The said correction term includes the difference between measured measure magnitude (or a measurement magnitude vector) and the measurement magnitude which may be expected according to the measurement equation out of the estimated value of the condition vector. That difference is being multiplicated each time by a Kalman-amplifier Κη· In the present case the elements of the condition vector are position faults. Therefore one could assume the estimated value of condition vector to be zero since the estimated position faults can always be ruled out. This simplifies the equation of the estimated value of the condition vector to (4) ¾ as being defined in equation (1) and Kn being a vector of alman amplificators.

The vector Kn is again derived in a well known way recursively - s stem matrices φ, the measure - matrice H (in this case measurement vector H) and the co-variances Q and R of the operational noise, or measurement noise, respectively. This is shown by Figure 2.

Out of Kn and the estimate value Pn(~) a variance Pn is being calculated of the estimate fault, in accordance with the relation ( 5 ) P =(£ -V_) p (-) , n n n n in which I is the identity matrix. This is shown by block 62. The thus calculated variance of Pn of the estimate fault is delayed by one tact. This is shown by block 64. As a consequence thereof the estimate value for the variance of the estimate fault is being calculated according to the relation Pn(-) - - 1 Pn-1 Vn-1 *Qn-1 - with system matrix φ and the transponated system matrix φ^. This is shown by block 66. From this estimate value for the variance Pn (-) the Kalman amplification Kn is defined again by means of measure matrix H and the transponited matrix and the variance Rn of operation noice according to relation as shown by block 68.

Thus with a Kalman amplification for every cadence, which in the present case is a vector, there is created according to equation (4) the estimate value of the condition vector xn. The multiplication of zn by is shown in Figure 2 by means of block 70. The estimate value provided by the optical condition estimator 54 of condition vector xn is superposed onto the optimal controller 56. The optimal controller 56 provides manipulated variables Uy>n and Uz>n which can be united to a control vector. The control vector un is interrelated with the estimate value of the condition vector xn» i.e. with estimate vaiuo oi.' L u position faults according to tlia rulnt.l.on as illustrated by block 72.

This is and T being the scanning period.

To the manipulated variable Uy>n there is superposed in one point 72 a correction term (sin an ns-3y).

In the. latter cxn is the angle of inclination of the stabilized element relative to the third axis which is normal to the azimuth axis 14 and the platform axis 21, s the vertical component of earth rotation and dv an estimate value for the drift of the gyro and the entry axis 32. The first term takes into consideration the fact that due to an inclination of the gyro 28 about the said third axis a component of the vertical component Ωβ of earth rotation becomes active at the entry axis 32 of the gyro. This influence is being compensated by the first term. The second term does compensate the gyro drift.

The so compensated manipulated variable which is designated by Ty>n is introduced with the tact into a Ty-register 76. The tact is shown again as a switch 78. The contents of the Ty-register 76 is transformed into an analog signal and is switched onto the torque creator 44.

Accordingly a correction term is superposed onto manipulated value uz>n at a point 80, dz being an estimated value for the gyro drift and the entry axis 34. The first term compensates the component which is active at axis 34 of the vertical component s of earth rotation. The second term compensates the gyro drift. The so compensated manipulated variable, designated by T„ „ is introduced with the tact into a T∑ -register. The tact is shown as a switch 84. The contents of the Tz-register 82 is transformed into another signal and is switched onto the torque creator 40.

The acceleration meter 46 has the purpose of identifying the angle an> i e> the angie Qf inclination about the non-stabilized third axis and its trigonometric functions sinus and cosinus. The acceleration signal of the acceleration meter is smoothed through a low pass filter 86. The smoothed acceleration signal aVc ^s being digitalized in the tact and is worked up as result of digital values aCy>n. This is symbolically represented by switch 88. At one point the digital value aCy>n w^tn opposite sign is overlaid by a value which corresponds to an estimated value by of the zero point fault of the acceleration meter 46. The such obtained corrected acceleration measurement value is then divided by earth rotation g. This is illustrated in Figure 2 by block 92. Our of this division there is obtained sin an which is available at exit 94.

From the thus obtained value a value cos an can be obtained by creation of 1 - sin a. which layer is shown in Figure 2 in the form of a block 96. That value of cosctn is available at an exit 98. The values obtained in this key-sinan and cosan serve at creation of compensation signals which at points 74 and 80 respectively are superposed on manipulated variants Uy>n and uz>n respectively.

The above described arrangement functions as follows: The gyro 28 is expected to eventually orientate itself such that the twist axis 30 is positioned horizontally and points in north direction. This condition is attainable with the two degrees of freedom of the two-axial platform 10. Of course, in this final position the stabilized element 24 can be inclined through an angle a in relation to the third non stabilized axis which is normal to azimuth axis 14 and the platform axis 21, whenever the casing 16 is correspondingly inclined. That could be the case when the casing 16 is mounted in a vehicle. The gyro 28 is being pre-orientated into the vicinity of the desired final position.

In the desired final position the gyro 28 is stable: The acceleration meter 48 does not any longer issue an acceleration signal of the earth acceleration, since its sensibility axis, in parallelism with the twist axis 30 is now horizontal. The entry axis 32 does not determine any component of the horizontal component Ωε becomes effective because the entry axis 34 is positioned vertically relative to the north direction. The components of the vertical component c of earth rotation which are effective at the entry axis 23 and 34 are compensated at points 74 and 80 by the compensation signals.

The optimal condition estimator 54 supplies, out of the acce- Λ A. leration values Zn estimate vehicles Θγ,η anc* Θζ,η ^or positive faults, i.e. for angular deviations of the platform 10 from the already mentioned final condition the twist axis 30 is horizontal and is directed to north.

The estimate vakues Qy>n and ^ζ,η are ratner high at the recourse steps at at commencement of the straightening. This is caused by the behaviour of the alman Filter. These initially high estimate values cause via the optimal regulator a quick approximation to the final position. Subsequently, the estimate values become lower, such that the final position is. reached in an exact manner, without overshooting .

The estimate values xn do supply via the optimal regulator 56 the manipulated values un. The manipulated value Ty is switched onto the creator of rotational moment 44. This causes a deflection of the rotor of gyro 28 about the axis 32. The take off 36 which is positioned on that axis 32 takes along, via the servomotor 23, the stabilized element 24 of that deflection of the gyro. That takes place until Τγ^π becomes zero. The manipulated value Tz is switched onto the rotation moment creator 40. This latter causes a deflection of the rotor of gyro 28 about the entry axis 34. The take off 42 which is positioned on that axis 34 takes along, via the servomotor 18, the stabilized element 24 of that deflection. Also that takes place until ZjI1 becomes zero.

The described gyro arrangement permits a quick setting of the axis of rotation 30 towards north, free of influence of external disturbances.

Figure 3 illustrates an automatically orientable gyro arrangement wherein occurs a continuous optimal determination of the angle of deviation of the stabilized element about the non-stabilized third axis. The mechanical basis build up is the same as that of Figures 1 and 2. Corresponding parts are again designated by identical reference numerals.

Additionally to the two acceleration meters 46 and 48, there is provided in Figure 3 a third acceleration meter 100, whose sensibility axis extends normally relative to sensibility axes of acceleration meters 46 and 48. Three acceleration meters 46, 48 and 100 supply three acceleration signals acx(t), acy(t) and -acz(t). The three acceleration signals acx(t), acy(t) and acz(t) are switched onto first transformation means 102. The transformation means 102 receive also an estimate value "a" of the angle of deviation 25. This first transformation means 100 supply according to relation (11) Acceleration signals such as would be supplied by the two horizontally set acceleration meters of a virtual, also about a third axis of casing 16 decoupled platform. In the system of this virtual platform there occurs the successive further processing of the signals.

The transformed acceleration signals avx(t) and avv(t) are switched onto an optimal condition estimator 104 in the shape of a Kalman Filter. The optimal condition estimator 104 supplies estimate values of position faults, in a manner similar to that described in connection with Figure 2, which values - according to equation (1) are continued into an estimated condition vector x(t).

The condition vector (x)(t) is switched onto an optimal controller 106. The optimal controller 106 supplies, in a way similar to controller 56 of Figure 2, in accordance with relation manipulated variables uvx(t), uv„(t) and uvz(t) which are combined into a manipulated vector ux.

At one point 108 onto the manipulated variable uvx(t) there is superposed a signal which corresponds to the horizontal component Qc of earth rotation. At one point 110 there is superlaid onto the manipulated value uvz(t) a signal with inverted sign which corresponds with the vertical components Qs of earth rotation. The manipulated variable uvy(t) remains unchanged. Thus there are present the three signals The three signals are in conformity with the manipulated variables which would be required for the orientation of the "virtual" three axial platform about the three platform axes. The three "virtual" manipulated variables are switched onto two transformation means 112. The second transformation means 112 receives - like the first transformation means 102 - an estimate value a(t) of the deviation of the stabilized element. The second transformation means 112 cause a back transformation of the "virtual" manipulated variables into the coordination system of stabilized element 24. There are obtained manipulated variables (14) cos o. T +· sin a T, = - sin a T »· COS I which are being related to the coordination system of the stabilized element 24.

The signal Tyc from the transformation means 112 is switched onto the rotation moment creator 44 of gyro 28. The signal Tzc from - - transformation means 112 is switched onto the rotation moment creator AO of the gyro 28.

The signal Txc serves the determination of the estimate value (t) of the angle of inclination. The angle of inclination a(t) is calculated from the acceleration signals of the acceleration meters 58 and 60 which are positioned at the casing 16 and the azimuth angle signal (t) of the angle generator 20. The acceleration signals a^Ct) and ay^it) are each smoothed by a low pass filter 114 or 116 respectively. Calculation means 118 create therefrom a signal which represents the inclination angle (t). Onto this signal a finely adjusted correction signal is superposed at a point 120. This finely adjusted correction signal is obtained by integration of manipulated variable Txc(t) from transformation means 112. The integration is illustrated in Figure 3 by block 122.

The manipulated variable Txc(t) came into existence from a straightening fault about the x-axis of the virtual platform. It corrects this fault by integration, such as the manipulated variables TyC and Tzc at the rotation moment creators 44 and 40 are integrated by the gyro-laws. This takes place until Txc becomes zero. Then the virtual platform is positioned horizontally and the inclination angle a of the stabilized element 24 of the real platform is shown realistically.

Apart from that, the orientation of the platform 10 is performed in such a manner that the axis 30 points horizontally to north through manipulated variables Tyc and Tzc as has been described in connection with Figure 2. The compensation of the components of earth rotation is obtained at points 108 and 110 within the system of the virtual platform.

By means of the steps described heretofore a speedy orientation of the gyro to north is being reached suppressing the influence of external distortions, the drift of the gyro being considered as a known, measured value. However, certain difficulties arise whenever a platform is to be orientated quickly immediately after in-switching. Due to heating up within the platform and the gyro the drift of gyro changes immediately after switching. This might hinder the compensation of drift.

In order to solve such problems and to create an exact self adjustment, first the stabilized element 24 with the sensor group 28 and the acceleration meters 46, 48 is realized as a compact thermic-ally enclosed construction assembly 124 in which the sensors are in well conductive thermic condition between each other and speedily assume equal temperature. This group 124 is shown schematically in Figure 4. The construction group 124 includes heating elements 126, 128, 130 at sensors 28, 46, 48 by means of which the group 124 is raised to a temperature slightly above the upper end of the obligatory temperature range. At practical use of the gyroscope the group 124 is heated to 72°C. At such a temperature the group 124 is roughly stabilized by temperature sensors 132, 134, 136. So as to obtain initiatory starting conditions high capacity performance and switching in of servo-circuit are resorted to only when the temperature of the group surpasses a predetermined value of, e.g. 0°C.

Temperature dependent sensor faults are being compensated. Figure 5 shows as block diagram the structure of signal processing with fault compensation for the coordination axis yc which determines the exactitude of adjustment about azimuth axis 14.

The gyroscope 28 affects the servo circuit 138 by means of the amplifier 52 of Figure 2 or 3. The servo circuit 138 does affect via the servo motor 18 and frame 12 the acceleration meter 48. The signal of acceleration meter 48 supplies, via selection electronics 140, velocity increments Δνχ. The velocity increments ανχ are summed up over the cadence period T. This is illustrated by way of block 142 in Figure 5. The summed up velocity elements ΣΔνχ/η are scanned cadence wise. This is shown in Figure 5 by means of switch 144. Further, the sum total of the velocity increments is being divided by the number thereof. The so obtained acceleration datum is then sub-laid to the signal processing described in connection with Figures 2 and 3 which step is symbolized by the block 146 in Figure 5. There is obtained a manipulated variable T„ which in the cadence as T,, „ y y»" in the form of electronics 148 is switched onto the torque creator. The cadence is also symbolized by switch 149. The torque creator 44 produces a moment of torsion at the gyroscope 28 which constitutes the entry of gyroscope 28. This is the structure of normal function of the gyro arrangement as described in connection with Figure 2 and 3.

.The gyroscope 28 has a scale factor fault ASF. This latter does affect at point 150 torsion moment created by the torque creator 44. In addition different drifts ... - m a - aa: * d y _ yo ' come into existence. In the above m is the mass inertia, q is "quadrature" and dyo is the normal bias drift. The first two terms depend of acceleration ayc and azc respectively. These drifts are effective at the exit of the gyroscope at point 152 as indicated by arrows 154. The acceleration meter 48 equally has a scale factor fault ASK. The latter is effective at point 156 on the entry of the acceleration meter, as illustrated in Figure 5 by arrows. The acceleration meter 48 has a zero point fault bx_ which becomes effective at point 160 at the exit from the acceleration meter. That is shown in Figure 5 by arrow 162.

These temperature dependent faults are being compensated. To this end a memory 164 is provided. In memory 164 there are stored the estimate values for temperature dependent fault parameters, i.e. the mass unbalance m(^ , the quadrature q(v), the drift dy0($), the scale factor fault ASF(¾} of gyroscope 28, the zero point fault bxo("0») of acceleration meter 48 and scale factor fault ASK($>) of acceleration meter 48.

Calculation means 166 receive manipulated variable Tv>n from floor 146 "straightening equations" and the estimate value of scale factor fault ASF(v) from memory 164. Therefrom they create the product Δ S F (-¾■ ) . T,.in .

This product is added at point 152 to the gyroscope signal, of gyroscope 28 exist thereby compensating the influence of scale factor fault ASF at point 150.

The calculation means 168 receive the temperature $ from sensors 132, 134, 136 (Fig. 4) and accelerations ayC and azc from acceleration meter=s 46 and 100 and fault parameters m(v), q(v) and dvo(v) and create therefrom The thus obtained signal is also superposed at point 152 to the gyroscope signal, thereby compensating the drift gyroscope 28 as far as effective influences are concerned.

For the faults of acceleration measurement is assumed Aacx - bxQ + (ASK Generally the acceleration meter exit is not approachable directly. Directly approachable are only the velocity increments which are summed up in an accelerator. For the fault of this sum.

+ T Δ (Τ-':< ) = J bxod -isk. aGx(t)dt T + A S ∑ is applicable.

Accordingly the calculation means contain as entry values the temperature from temperature sensors 132, 134 and 136, the total of velocity increments ∑v from block 142 and the fault parameters 1 x bxo($) and ASK(i$) from memory 170. The calculation means form therefrom a correction term. Λ7. sxo So as to obtain the maximal exactitude (limited only by accidental faults of the sensors) the above mentioned compensation described with reference to Figure 5 would often be unsatisfactory. That would be valid also whenever other, known fault types as wrong assembly are to be rectified. Figure 6 therefore shows a compensation arrangement at which "integral" resulting total behaviour of the platform including drift and acceleration meter faults in the environment of a control temperature of the stabilized element are described by means of a linear temperature pattern.

At one entry 174 of the reference temperature is assumed at which the fault ΔΤν of the manipulated variable Ty is zero. It is assumed that in the vicinity of that reference temperature the faults depend linearly from the temperature deviations Δ&. The actual temperature of the group 124 of elements (Fig. 124) which comprises the sensors is applied at an entry 176. At a point 178 there comes into existence the difference Δ9* between actual temperature and reference temperature.

From acceleration meter 42 there arrives an acceleration signal in the shape of a total sum of angle increment signal ∑Δνχ. The acceleration signal is present at an entry 148. The acceleration signal is multiplicated by a scale factor fault which with a coefficient S is proportional relative to the temperature deviation Δ¾Κ This is illustrated by block 180. The block 180 receives via entry 182 the temperature deviation Δν and at the entry 178 it receives the acceleration signal. Furthermore an integral axis-fault is assumed relative to the orientation of the coordination axis yc. The integral axis fault results from the total measuring section for - - the entry signal of orientation equations and comprises shares which are due to gyroscope drift, servo circuit and selection electronics as well as acceleration faults.

This axis fault is assumed to be in proportion with the temperature deviation Δ¾> with proportionality factor B*. The factor B* is shown in Figure 6 by means of block 184 at which - via an "entry" 186 - the temperature deviation is introduced and at whose "exit" 188 the axis fault Β*Δ^ is being delivered. The integral fault is superposed on the acceleration fault at a point 190 onto the acceleration signal which had been multiplicated by the scale factor fault. There is obtained an acceleration fault Δ(φΔνχ). The latter is subjected to the steps to which according to Figure 3 the acceleration signal acx is being subjected. This is shown in Figure 6 by block 192. Thus a fault of the manipulated variable is obtained. However, also the gyroscope 28 has a scale factor fault. This latter fault is assumed as being proportional to the temperature deviation Δ$ by a proportionality factor SF*. With that scale factor fault the fault of the manipulated variable being produced by block 192 is again multiplicated. This is illustrated by block 194. Block 194 receives at "entry" 196 the temperature deviation Δ$. By way of the multiplication there is obtained a correction signal ΔΤν which superposed onto the manipulated variable Ty at the rotation moment creator 44.

The constants SF* and B* and SK* of that linear temperature pattern are determined by appropriate temperature test for the fully assembled and operative platform 10 and are stored in the calculator. With the aid of that temperature pattern platform faults can be compensated, which could be determined by deviations between actual temperature and reference temperature and exert influence onto the entry signals and exit signals of the orientation equations.

In the stationery position the temperature deviation is equal to zero. Thus there disappears also the correction term.

It is expected that by means of the above described steps all faults which could affect the exactitude of orientation should be compensated, thus only incidental faults could have remained. Such incidental faults could be e.g. faults of the platform which could have been caused during current orientation operations (external rumours). Incidental faults would further be such platform faults which occur during time periods from one switch-on step to the next one and which after switch-on are not recognizable but remain constant. As an example a constant but not recognized change dyC of the drift about the coordination axis yc of the platform 10 could cause an orientation fault in the azimuth of .C a, (16) e_ = c wherein Ωβ represents the earth rotation and φ the geographic width.

So as to obtain an exactitude of orientation of 1 milliradiant (1 m rad) the dyc must not be greater than 0.01°/n.

In order to exclude disallowable drift values of the above described kind, the platform 10 is oriented once to the north and once to the south.

Due to a drift fault between switch-on to switch-on an orientation fault occurs at north orientation of ΘζΝ·· Tne coordination axis xc deviates by an angle ΘΖΝ from north. That is shown in Figure 7 in an exaggerated manner. The course indication of platform 10 being freed from cardan fault would be - - (17) - θ ZN . at the end of north orientation.

In the above ΨΝ is the angle defined by the axis of the vehicle (or any other reference direction) and the coordination axis xc of the platform system Ψ being the angle defined by the axis of the vehicle and north. At orientation of platform 10 to south the orientation fault becomes O^S' Again, the indication of the course angle which is freed from cardan fault is, as shown in Figure 7: (18) ψ s = 8 o ° -Φ - e zs .

In angle Ψ, i.e. the angle defined by the longitudinal axis of the vehicle and north. By subtraction of equation (18) from that equation (17), there is obtained (19) z'v = ,J ,j sq + 1 80 ° + cos ( At a comparison with equation (16) one recognizes the advantage that the fault in determining the north direction is no longer due to drift changes dcy between switch-on to switch-on but to differences between drifts between the two positions north and south. Such differences may be due to slow accidental changes of drift.

A further means for identification and reduction of effects of drifts of gyroscope 28 reside therein that platform 10 is moved about the coordination axes yc and zc in the explained way. That is shown in Figure 8.

Figure 8 corresponds to a great part to Figure 3 and corresponding parts are marked by identical reference sign. At the gyroscope arrangement of Figure 8 the manipulated variables TyC(t) and Tzc(t) which are supplied by transformation means 112 are superposed by time dependent signals wy(t) and wz(t). the signal wy(t) is positioned at entry 198. The signal wy(t) is superposed at a point 200 onto manipulated variable Tyc(t). Therefrom is obtained a combined manipulated variable - - (23) X c ., i - * = T, . c ( t ) ÷ The time dependent signal wz(t) is positioned at entry 202. The signal wz(t) is superposed onto manipulated variable Tzc(t)*. Therefrom is obtained a continued manipulated variable The additional signals Wy(t) and wz(t) cause a defined additional movement of the gyroscope 28. From that additional movement and signal parts contained therein the drift dyC of the gyroscope 28 may be identified. The additional signals are so chosen that on the one hand an identification of the drift is possible and on the other hand that the so caused added movement of the gyroscope 28 the exactitude of orientation is not influenced. By identification of the drift dyC a further improvement orientation exactitude can be obtained, as compared with orientation without such "identification" additional movement.

Additional signals are those of the form w, ( t ) = c o s ( ) Ω being appropriate, Ωζ being a special rotation velocity which is different from on case to another.

Claims (14)

1. ) A self adjusting gyro system which includes a du-axial gyro-scop 'e/with a take' off, and a creato'r o·f rotational movement at each input axis at which the gyro is placed in the stabilized element of a du-axial platform, the servo motor of which acts on the platform i axis is controlled by the signal of the take off of the gyroscope and at which at the stabilizing element there are provided means for inducing inclination from the signals of which controlling signals can be derived which latter are switched onto the creators of torque, characterized thereby that a) signals of the inclination causing means are applied to an optimal estimator of condition which supplies estimate values of position faults of the stabilized element, b) the estimate values of the position faults are switched onto an optimal control means which supplies variable signals, and c) the variable signals of the optimal control means are switched onto a creator of rotational movement of the gyro.

2. ) Gyro system aceording to claim 1 characterized thereby that from an inclination signal of the inclination causing means there are created compensation signals which compensate the influence of the vertical component of earth rotation and are superposed onto variable signals of the optimal control means.

3. ) Gyro system according to claim 2, characterised thereby that a) the inclination causing means supply a first inclination signal ^"^cx) corresponding to the inclination of the stabilized element about one of the entry axes of the gyroscope, and b) the optimal condition estimated is being subject from the first inclination signal.

4. ) Gyro system according to claim-.^, characterised thereby that a) the inclination causing means create a second inclination signal (acv) corresponding with the inclination of the stabilized element about the vertical twist axis which is normal relative to the input axis of the gyro, b) the compensation signals for compensation of the influence of the vertical component of earth rotation are derived from the second inclination signal (ayC).

5. ) Gyro system according to claim 4 characterized thereby that the inclination signals are created by one of acceleration meters with a succeeding low pass.

6. ) Gyro system according to claim 1 characterized thereby that a) the inclination sensing means include three acceleration meters which have input axes being normal to one another; b) the signals of the three acceleration meters are switched onto first transformation calculation means - to which there is added an estimate value for the angle of inclination of the stabilized element about the vertical non- stabilized axis which is in parallel with the two platform axes, - which supplies virtual acceleration value (axv, ayv) representing the inclination of a virtual platform, being stabilized also about this latter axis, c) the virtual acceleration values (axv, ayv) are switched only the optimal condition estimator which supplies estimate values (x (t)) of the position faults of the virtual platform d) the estimate values (x (t)) of position fault are switched onto an optimal control supplying manipulated signals (uv) for the stabilization of the virtual platform; e) the manipulated signals (uv) for stabilization of the virtual platform are switched onto a second transformation calculation means - also receive the said estimate value (a) of the angle of inclination, and which supply manipulated signals for stabilization of the real platform are switched onto the creator of rotational movement of the gyroscope.

7. ) Gyro system according to claim 6 characterized thereby that a) at a casing carrying the platform, in a plane which is normal relative to the axis of the external frame of the platform there are provided a pair of inclination sensors having mutually normal intake axes which respond to the inclination of the said plane against the horizontal and supply inclination signals of the casing, b) an angle indicator being provided on the axis of the external frame supply in an angle signal indicating the position of said external frame relative to the casing, c) inclination angle calculation means are provided - onto which the casing inclination signals and the said angle signal are switched, - and which supply a value (a(t)) of the angle of inclination of the stabilized element about the non-stabilized axis.

8. ) Gyro system according to claim 7 characterized thereby that a) a manipulated signal (Txc) according to one turn of the said non- stabilized axis about the platform is switched by the second transformation calculating means onto integrating means, and b) the exit of the integrating means to the said calculated value (ct(t)) for the angle of inclination is superposed by key of correction for the erection of the estimate value (a(t)) for the angle of inclination.

9. ) Gyro system according to claim 7 or 8 characterized thereby that a) a manipulated variable of the optimal control according to a movement of the virtual platform about the virtual platform axis is overlaid by a signal which represents the vertical component of earth rotation, and b) which is overlaid by a signal most of movement of the virtual platform about the manipulated variable (Τχν) which is conform with the non-stabilized axis of the optimal control, the said signal representing the horizontal component of earth rotation.

10. ) Gyro system according to any oen of claims 1-9 characterized thereby that a) the sensors (gyro and inclination sensors) of the stabilized element are in thermic contact in a construction assembly, such that they are of a uniform temperature, b) a temperature sensing means is provided which measures said uniform temperature and supplies temperature signal. c) measured temperature dependent fault coefficients are stored in a memory, and d) fault calculating means are provided, - onto which measurements from sensors and fault coefficients can be switched. - which therefrom supply corrections which are transmitted to signals of the system.

11. ) Gyro system according to claim 10 characterized thereby that the sensors are regularizable by heating means towards a temperature at the upper end of allowable temperature range.

12. ) Gyro system according to any one of claims 1-11, characterized thereby that for the sake of identification of fault a defined movement is imparted to the stabilized element.

13. ) Gyro system according to claim 12 characterized thereby that creation of said defined movement signals of the movement there are superposed onto signals which are applied to the two creators of rotational movement.

14. ) Gyro system according any one of claims 1-11 characterized thereby that platform is adjustable to North and South or vice versa, the north direction being obtained by linear combination of the thereby obtained azimuth angle. vaN-ta j)M TO