DE3401090C1 - Monitoring aircraft projectile or missile - using two extrapolation units to predict flight path - Google Patents

Abstract

A measuring unit (114) establishes the instantaneous location and movement parameters of a flying object. A first extrapolation unit (118) predicts future values of the parameters, if necessary under the assumption that one of the parameters stays constant. A second extrapolation unit (120,122,124,126,130) considers an extrapolation error as a difference between the parameters calculated for a certain time and the parameters received from the object itself for this time, minimising the error so that it matches an error model in a continuous run of values. USE/ADVANTAGE - Improved accuracy of prediction. Suitable for military use.

Description

The invention relates to a device for monitoring of flying objects according to the preamble of claim 1.

When monitoring flying objects are common Predict the future trajectory of the flying object needed, for example for airspace surveillance, to be able to estimate whether there is a collision of Flying objects comes in, or in the fight against flying objects due to the finite flight time of the control projectiles or the like. It is already known here to use an extrapolation unit, which a Prediction of future location parameters of the concerned Object, usually assuming that  one of the movement parameters of the flying object, in particular whose acceleration remains constant. A Device according to the preamble of claim 1, in which the extrapolation error due to manually entered additional ones Reduced parameters regarding the flying object is to be found in US-PS 41 46 780.

The object of the invention is the prognosis  accuracy for the future location of the flying object to improve structurally simple means.

This task is characterized by the characteristics of the license plate of claim 1 in conjunction with the features of the preamble solved. The first extrapolation unit delivers Forecast values which are derived exclusively from the measured Location coordinates and the possibly measured speed of the flying object to a specific one Depend on the time. The second extrapolation unit takes into account the extrapolation error, i.e. the Difference between those predicted for a point in time Location parameters and at that time location parameters actually taken by the flying object and tries to minimize this extrapolation error. This is done by continually adjusting one temporal course of successive extrapolation errors the underlying error model to the occurring extrapolation errors.

The expansion of the known device into an invention Working device requires relatively little Effort because the well-known first extrapolation unit can be retained and only the second is added must become.

To the respective extrapolation errors in a simple manner to be able to receive the measures according to Claim 2 proposed.

The construction effort continues to decrease if the Extrapolation unit according to claim 3, a prediction filter (Signal Extrapolator) includes instead of, for example an extrapolation calculator.

For example, a polynomial second comes as an error model  or higher degree in question. However, this is preferred Fault model according to claim 4, which is a time invariant linear second order differential equation corresponds and with a relatively simple technical implementation delivers particularly good results.

Another improvement in predictive accuracy is achieved by the features of claim 5. Because of the relatively short time interval in which past Extrapolation errors taken into account when adapting the error model are excluded that already Completed flight maneuvers affect the extrapolation.

The measure according to claim 6 also leads to a Improve predictive accuracy since the Larger absolute amounts of the extrapolation error determined in each case have better informative value than the extrapolation error close to the noise with a smaller prediction time interval.

The invention is based on preferred embodiments explained using the drawing. It shows:

Fig. 1 is a schematic block-like, substantially simplified representation of the apparatus according to the invention for monitoring flying objects;

Fig. 2 is a simplified block diagram of an embodiment used in connection with a fire control computer of the monitoring device according to the invention and

Fig. 3 is a functional diagram for the block diagram of FIG. 2.

With the aid of the device for monitoring flying objects according to FIG. 1, generally designated 10 , the best possible estimate is to be made for a location that will be occupied by a flying object 12 in the future. In the figures , index _m indicates that the indexed parameter value results from a measurement and indicates a momentary value assigned to the measurement time, and index _b indicates the fact that the indexed parameter value has been determined on the basis of an extrapolation calculation. It can be seen in Fig. 1, a measurement (and rake) unit 14 for measurement of the location r _m of the flying object 12, for example by means of radar (radar screen 15) and for estimation of the instantaneous velocity v _m and the instantaneous acceleration b _m.

The values obtained are supplied in a first extrapolation unit, designated 16 , which, for example, assumes location r _b at a future point in time, assuming constant acceleration. A second extrapolation unit 18 serves to improve this precalculated value r _b , which calculates an extrapolation error denoted by Δr _b and supplies it with a subtraction element 20 , which forms = r _b - Δr _b . This corrected value can be processed further, for example to display the future flight object location on a screen. The error model on which the extrapolation in the second extrapolation unit 18 is based is adapted to the time profile of the extrapolation errors Δr _m that actually occur. The actual extrapolation error Δr _m is the deviation of the predicted location r _b from the location r _m then actually occupied by the flying object 12 .

In FIGS. 2 and 3 is a special application of the device according to the invention, namely in connection with a fire control an anti-aircraft unit. The device according to the invention is designated generally by 110 and is connected on the one hand to a radar measuring unit 114 with radar screen 115 and, via an output line 116, to an anti-aircraft battery, not shown. The radar measuring unit 114 supplies measured values r _m to a fire control computer 118 , which calculates the acceleration b _m and directional angle α _m (side) and β _m (height). The fire control computer 118 is connected to a memory 120 as well as to a first and a second subtraction element 122 and 124 . The first subtraction element 122 is connected on the input side to the memory 120 and on the output side to the input of a prediction filter (signal extrapolator) 126 . The output of the prediction filter 126 is in turn connected to the second subtraction element 124 , the output of which is connected to the aforementioned output line 116 of the device 110 . A system clock 130 controls the radar measuring unit 116 as well as the fire control computer 118 and the memory 120 .

The functional sequence of the arrangement according to FIG. 2 is explained below with reference to FIG. 3.

At an arbitrarily selected point in time, for example point in time k-1, a radar measurement is carried out and the current location r _m as well as the current speed v _m and acceleration b _m (in the fire control computer 118 ) are determined. This is indicated in the block 132 indicated in broken lines in FIG. 3. Subsequently, the fire control computer 118 calculates the location r _{b of} the flying object at the time k ′ = k + T _k, assuming constant acceleration b _m . Here, k is the next measurement time following measurement time k-1 and T _{k is} the flight time of the projectile intended to combat the flying object until it reaches location r _b . In a block of the time diagram according to FIG. 3, designated 134 , with a time axis running vertically downward, after calculation of r _b, the calculation of directional angles α _b and β _b (side and height) is carried out. However, the predicted location r _b does not correspond to the location r _m actually reached, which is primarily due to measurement errors of the radar measuring unit 114 and possibly also to a sudden change in movement of the flying object 12 . It has now been shown that in many cases these extrapolation errors have a characteristic course over time, which can be satisfactorily approximated by a stochastically excited damped oscillation with the parameters natural frequency, damping and strength of the excitation. The specified model parameters can be determined with sufficient accuracy on the basis of the respectively measured extrapolation errors in a time interval which in each case goes back a few seconds from the current time. Using the model adapted in this way, the extrapolation errors to be expected for the time t = k can be calculated in advance, which in the present case form directional angle corrections Δα _b and Δβ _b . These correction values formed in block 136 according to FIG. 3 or in prediction filter 126 according to FIG. 2 are subtracted in subtraction element 124 or block 138 from the pre-calculated values α _b and β _b , whereupon the corrected angle values and are obtained. These corrected values are supplied to the anti-aircraft battery via line 116 .

In block 140 of FIG. 3 will now be described with the aid of a corresponding radar measurement at time k 'of the flying object to the assumed results time point k' = k + T actually _k occupied position r _m determined and from this by means of the fire control calculator 118, a calculation of the directivity angle α _m and β _m made, ie the directional angle that would lead to a hit at the location r _m .

Next, the actually occurring extrapolation error Δα _m = α _m - α _b and Δβ _m = β _m - β _{b is} calculated in block 142 according to FIG. 3 with the aid of the subtraction element 122 in FIG. 2. It is clear that the memory 120 has stored the values α _b and β _b already determined at the point in time k ′.

These extrapolation errors Δα _m and Δβ _m are now fed to the block 136 in FIG. 3 in accordance with the prediction filter 126 in FIG. 2 to end this time loop in order to adapt the model parameters. A new loop begins with a further radar measurement in block 132 according to FIG. 3 at a next point in time. This point in time can be later than the point in time k 'or earlier (for example at point in time k), in which case it must then be ensured that corresponding values are α by correspondingly designing the memory 120 and correspondingly controlled by the system clock 130 _b and β _b are temporarily stored in the memory 120 and are then each given to the subtraction element 122 when the assigned angle values α _m and β _m are available.

The prediction filter 126 adapts the parameters of the error model to the respectively measured extrapolation errors Δα _m and Δβ _m and the pre-calculation of future extrapolation errors Δα _b and Δβ _b . This can also be referred to as a signal extrapolator, prediction filter, predictor or data extrapolator. With the help of extrapolation filters, such as those often used in digital-to-analog conversion (see, for example, paperback "Elektrotechnik", Karl Hanser-Verlag, Munich-Vienna 1978, Volume III, from page 898), we can draw from the known past extrapolate the associated continuous analog time-dependent signal from a sequence of numbers and, for example, calculate a future function value. The extrapolation can be based, for example, on polynomials or, as in the present case, a stochastically excited damped oscillation. The design of the prediction filter 126 is based on a state model of a time-invariant, linear, stochastic differential equation of the second order. With the aid of the Kalman filter algorithm (see, for example, Ch. Landgraf: stochastic linear systems, TU Berlin, focus on cybernetics 1979), the prediction filter 126 can be designed with the desired properties. The measured extrapolation errors Δα _m and Δβ _{m are present} at the input of the prediction filter; the pre-calculated extrapolation errors Δα _b and Δβ _b form the output signal of the prediction filter. The bandwidth of the prediction filter 126 is defined in such a way that the influence of measured subtraction errors which have occurred in the past drops so rapidly that essentially only measured extrapolation errors from a time interval in the range of seconds affect the adaptation of the model parameters (angular frequency, damping constant and excitation variable).

In the embodiment according to FIGS. 2 and 3, directional angles α and β are calculated from the measured and extrapolated locations r _m and r _b , which in turn are used to minimize extrapolation errors in blocks 142 and 138 . It is obvious that in other applications, for example in the case shown in FIG. 1, the intermediate calculation (calculation of the straightening angles from the locations) can be dispensed with, with a corresponding simplification of the time diagram according to FIG. 3.

According to the figures, r to understand the local sector and accordingly under v and b each the speed vector and the acceleration vector.

Claims (7)

1. Device for monitoring flying objects ( 12 ) with a measuring unit ( 14 ; 114 ) for determining instantaneous movement parameters (r _m ; α _m , β _m ) of a flying object ( 12 ) and with a first extrapolation unit ( 16 ; 118 ) for predicting future location parameters (r _b ; α _b , β _b ), possibly under the assumption that one of the movement parameters (b) of the flying object ( 12 ) remains constant, characterized in that a second extrapolation unit ( 18 ; 20 ; 120 , 122 , 124 , 126 , 130 ) is provided, which is the extrapolation error (Δr; Δα, Δβ) as the difference between the location parameters (r _b ; α _b , β _b ) calculated in advance for a point in time (k ′) and those actually at that point in time location parameters (r _m ; α _m , β _m ) taken by the flying object ( 12 ) are taken into account and this extra polation error (Δr; Δα; Δβ) is minimized by making an error on the basis of the temporal course of successive extra polation errors adapts continuously. 2. Device according to claim 1, characterized by a buffer ( 120 ) connected to the first extrapolation unit ( 118 ) for the pre-calculated location parameters (α _b , β _b ) and a subtraction element ( 122 ) connected to the buffer ( 120 ) for determining the extrapolation error (Δα _m , Δβ _m ). 3. Apparatus according to claim 1 or 2, characterized in that the second extrapolation unit ( 120 to 130 ) comprises a prediction filter ( 126 ) designed according to the error model and possibly connected to the output of the subtraction element ( 122 ). 4. Device according to one of the preceding claims, characterized in that as A damped vibration, preferably, error model with stochastic excitation. 5. Device according to one of the preceding claims, characterized in that the second extrapolation unit ( 120 to 130 ) to optimize the error model essentially only past measured extrapolation errors in a time interval in the range between 0.1 seconds and 50 seconds, preferably 1 to 10 seconds, if necessary by defining the bandwidth of the prediction filter ( 126 ) accordingly. 6. Device according to one of the preceding claims, characterized in that the Prediction time interval (k′-k) is large compared the time interval between successive location measurements and preferably in the range between 0.1 and 20 seconds. is best between 1 and 10 seconds.   7. Device according to one of the preceding claims with a fire control computer ( 118 ) comprising the first extrapolation unit for the orientation parameters (α _b and β _b ) forming the location parameters, characterized in that the second extrapolation unit ( 120 to 130 ) contains the extrapolation errors (Δα _m , Δβ _m ) minimized as the difference between the calculated straightening angles (α _b and β _b ) for a hit location at a specific point in time and the straightening angles (α _m , β _m ) which are assigned to the location actually occupied by the flying object ( 12 ) at the point in time of the hit.