EP0591049B1 - Method for antenna calibration in the near field for active antenna - Google Patents

Description

The present invention relates to the measurement and manufacture of active antennas, which include a large number N of channels in parallel. Antennas active use this number N of channels to form the antenna radiation pattern, by superimposition fields resulting from the excitation of each element.

When designing an antenna, calculations theoretical allow, from the characteristics desired radio frequencies, determining the geometry of the radiating sources, as well as that of operating parameters of these sources and active modules associated with them; these parameters are in particular the gain of the amplifiers, the dynamics, and / or the relative phase shift necessary to obtain a depointing sought. These calculations are on the one hand made from certain assumptions and relationships mathematics which describe physical principles, and secondly from physical data concerning the antenna and its components. This data must be determined or confirmed by measurements of radio characteristics of the antenna.

The invention relates to a calibration method active antennas, which, from measurements in near field on the antenna and its radiating sources, allows, by a specific calculation, the determination of control parameters to be applied to active modules, and the resulting fields in the far field.

The active antenna, object of the method according to the invention can operate either in transmission or in reception, either in the two configurations alternation (case of radar antennas).

In the case of an antenna operating in transmission, the signal from a low level centralized transmitter is divided into N signals assumed to be identical on N channels at using a power distributor; then on each channel, an active module amplifies the signal according to a controllable gain, and applies a controllable phase shift, before transmitting the amplified signal to the source radiant (see Figure 1).

In the case of an antenna operating in reception, the signal received on each radiating source is amplified and phase shifted in an active module whose gain and phase shift are controllable. The N signals amplified on the N channels are then brought together by a power, and transmitted on a single channel to a centralized receiver (see Figure 2). This layout is the reverse of the previous layout, and, from a theoretical point of view, its treatment is rigorously symmetrical to the latter.

In the case of antennas operating in transmission and in alternating reception, such as radar antennas, a single device is called to serve as a combiner reception and as a dispatcher on the show, and active modules have a switch between a channel receiving station fitted with a low noise amplifier, and a transmission channel fitted with an amplifier power. Depending on the design of the active module, a controllable phase shifter and attenuator are provided for each channel, or if they are of reciprocal type, they can be arranged on a single track, connected in alternation between the two transmission / reception channels by a SPDT switch (see Figure 3).

In the design of an active antenna, the signals necessary for the formation of beams are calculated by computer from assumptions and approximations that make calculations possible but which are not always consistent with measurable reality regarding the antenna performance.

For example, the sources are assumed to be identical, while they are subject to small variations in their radioelectric characteristics, coming from vagaries of their manufacture. The same is true for active modules: impedance, gain, insertion losses and phase shift can vary from module to module, so that an identical control signal does not produce a identical phase shift or amplitude from one source to another other.

On the other hand, gain controls are assumed rigorously exact, and completely independent of those of phase and vice versa, while in practice, they are not, and a weak influence from one on the other is inevitable.

In addition, the position of a source in the network may influence the radio characteristics of this source, by coupling with neighboring sources. For example, the characteristics of a source located at one end of the network are different from that of a more central source, surrounded by neighboring sources.

Finally, we assume, for theoretical calculations, that the source amplifiers are linear. This means that the resulting fields at source are predictable from the order level applied to amplifiers. Now, if the amplifier works close saturation, which is often the case in transmission, the control signals necessary to obtain a amplitude sought differ from theoretical calculations linear.

The calibration method according to the invention allows to take into account, during theoretical field calculations distant, from these gaps between reality and situation ideal theoretical in characterization and design an active antenna. The results obtained are particularly appreciable for antennas having precisely formed radiation patterns, including antennas forming beams by calculation.

The problem raised by the existence of these errors, compared to the ideal antenna made only with ideal components, is not new. We recall for the reader the three types of errors covered by the method according to the invention: the dispersion (resulting from manufacture) of the radioelectric characteristics of components; control errors (in phase and in gain); and the (variable) couplings between sources radiant, according to their position in the network. The solutions provided by the prior art remain unsatisfactory, for the reasons listed below.

To overcome the dispersion of characteristics between the modules, assumed to be identical for identical components, it is known to incorporate to the active antenna of the calibration circuits specific. For an antenna operating in transmission, these circuits take a known fraction of the signal from of each active module, and return it to the unit of antenna control: for an antenna operating in reception, these circuits inject a known signal into the receiving circuit, and recover it after it has followed the normal path of a reception channel of the antenna.

This solution suffers from two drawbacks major: the need for a specific circuit for each module adds a considerable additional cost to the price already high from an active antenna; Similarly, the size, the weight, electrical power consumption, heat dissipation, and the complexity of the whole grow accordingly. On the other hand, the calibration resulting only takes into account dispersions affecting circuits, but neglects the influence of coupling between sources, as well as the influence of the dispersion of manufacture of the radiating sources themselves.

Another known method is to install a test antenna, for example a horn or a dipole at a certain distance from the active antenna to be calibrated. The transfer function between the test antenna and each of channels is determined by successive measurement of field delivered by each channel, according to the following method. All channels except the channel to be measured are switched off during the measurement of a given channel, each take turns.

The implementation of this art solution previous requires modifications in the realization of the basic active module, to incorporate the function of charging suitable for all channels except that to measure, in turn.

A variant consists in keeping the commands of the other channels, while the channel scanned is variably controlled, which rotates the phase. This theoretically allows to characterize the different phase states of this channel.

But this method suffers from the coupling problem between neighboring sources, which is not measured in conditions representative of the operating state normal: by rotating the phase of the channel under calibration, we slightly disturb the radiation of other sources, therefore the measurement of the radiated field.

A reverse method is outlined by the document D1 = Patent Abstracts of Japan, vol.9 no.236 (21.9.1985) publ. no. 60-89766, to help determine the radiation pattern, unknown a priori, of a network antenna. According to this document, a probe is placed in the near field in front of the network, and all of the phase shifters are controlled in an unspecified variable fashion, one after the other. The probe is calibrated from the emission signal, part of which is sampled using a directional coupler.

This D1 method intrinsically takes into account couplings between radiating elements, which are melted in the resulting field thus measured. But he is not planned use of the information collected for determine the phases to be applied to the sources in order to produce a desired radiation pattern.

Document D2 = EP-A2-496 381 teaches a method to check the operating status or not of a transmitting or receiving module within an antenna network, looking at changes in detected fields when the phase of the module to be controlled is switched between 0 ° and 180 °. If there is no difference, the module is Out of order.

The problem of theoretical modeling of these couplings has also been addressed in the prior art. Many models have been proposed, depending on the type of radiating source. The models aim to know by calculation, what will be the true radiation of the source Si when it is surrounded by the N-1 other sources S _j , j ≠ j, which are all excited by waves a _j . However, real sources are very difficult to model correctly, especially printed antennas (known as "patches" in English). However, these "patches" are used more and more often in active antennas, and the couplings between radiating sources of this type are particularly important. The theoretical calculation methods for coupling are often flawed, as are methods for modifying these couplings (to reduce the mismatch induced on the antenna) by coupling holes made between the access guides, by the judicious arrangement of a dielectric radome, etc. The methods of theoretical prediction of couplings could above all make it possible to correct by calculation their disturbing effects in a calibration sequence; but this is always, in the prior art, independent of the measurement of manufacturing dispersions, or of control errors.

The method according to the invention overcomes these disadvantages of the prior art, and what is more, of correct the three types of error recalled simultaneously above.

For these purposes, the invention proposes a method for calibrating an active antenna comprising N radiating sources; comprising the steps: a probe is successively placed in front of each radiating source so as to measure the near field in front of this source; characterized in that : the phases of the phase shifters are recorded at their nominal values for a desired antenna configuration in order to obtain a desired radiation pattern; and in that : during said measurement of the near field in front of said source, a phase shifter on each channel, in turn, is controlled so as to switch the radiation phase of this channel at 180 ° of phase shift relative to its value nominal, all N-1 other sources being controlled according to their nominal operating values in this configuration in order to obtain the desired radiation pattern.

The invention thus proposes a method for calibrating an active antenna comprising N radiating sources arranged in a network, this arrangement in network giving rise to a coupling between said sources, these sources being supplied by active modules, these active modules comprising phase control means and gain control means; these sources, these active modules, these phase control means and these gain control means having small dispersions of characteristics due to the manufacture of these different elements, these phase and gain control means also having inaccuracies in response aroused by a given order; characterized in that : using an appropriate probe, near-field measurements are carried out according to the above method, simultaneously characterizing the effects of said couplings between sources, of said dispersions of characteristics due to manufacturing, and of said inaccuracies of said means of ordered.

In a more precise method according to the invention, the gain and phase control values for a desired antenna configuration in order to obtain a desired radiation pattern are first determined by the aforementioned method, and these values of orders are applied to said means of order; this more precise method being characterized in that : the near field measurements are repeated with these control values, so as to obtain finer corrections to these values. If necessary, this procedure can be repeated; an iteration of this procedure for a sufficient number of cycles to obtain arbitrary precision of the specified parameters.

In another more precise method according to the invention, a calibration table is formulated at from measurements made on active modules before the antenna assembly, and this table then provides the refined values of the phase shift and gain which will be implemented for control after a only near-field measurement, using the basic method presented at the beginning of this text.

The method according to the invention and its variants can be applied to active antennas operating in transmission, reception, or alternately in transmission and in reception.

In the case of a radar antenna, the method according to the invention will be applied twice: once for the antenna operating in transmission, to determine the phase shift and gain control values program ; and the other time for the antenna operating in reception, in order to determine the command values of phase shift and gain in reception.

The method according to the invention provides numerous advantages over prior art methods for the calibration of active antennas. The method according to the invention allows the calibration of the antenna taking account for all dispersions that generate deviations between the real radiation diagram and the diagram theoretical calculated by software.

In the prior art, the characterization of an antenna by near-field measurement requires a much higher number of point readings. For each of the N radiating sources (the others being on a suitable load), it is necessary to take a reading at each point of a grid in squares of side λ / 2, on a surface clearly extending beyond that of the antenna: taking for example a antenna of N = 96 sources, each with a surface of 2.8 λ ² , approximately 100,000 readings are necessary for the complete calibration of the 96 sources; while the method according to the invention requires only N (N + 1) measurements, where N is the number of radiating sources, ie 9312 in the example chosen.

The method according to the invention thus provides a gain considerable time for antenna calibration (gain by a factor of 11 in the example above). Moreover, the proposed method is well suited to an iteration allowing to approach with an arbitrary precision the final performance of the antenna, and under the conditions actual operating conditions.

Since the method according to the invention takes into account variations in the radio characteristics of components of the antenna it is possible to widen the range of tolerances allowed for these components. The cost of components can be reduced by this way, thereby reducing the overall cost of the antenna.

Compared to certain methods of the prior art, the realization of the antenna is also simplified, because the method according to the invention does not require circuits specific sampling or injection of calibration.

la figure 1, déjà citée, montre schématiquement une antenne active fonctionnant en émission ;

la figure 2, déjà citée, montre schématiquement une antenne active fonctionnant en réception ;

la figure 3, déjà citée, montre schématiquement une antenne active de radar, fonctionnant alternativement en émission et en réception ;

la figure 4, montre symboliquement la relation entre différentes quantités vectorielles et matricielles déterminées par la méthode selon l'invention.

Other advantages and characteristics of the method of the invention will emerge from the detailed description which follows, with its appended drawings, in which:

Figure 1, already cited, schematically shows an active antenna operating in transmission;

Figure 2, already cited, schematically shows an active antenna operating in reception;

FIG. 3, already cited, schematically shows an active radar antenna, operating alternately in transmission and in reception;

FIG. 4 symbolically shows the relationship between different vector and matrix quantities determined by the method according to the invention.

In the figures, the same elements bear the same references, and to the intangible functions are assigned symbols to facilitate the explanation of the method of the invention.

Figure 1 schematically shows in section an example of an active antenna arranged in a linear array, operating in transmission. The example shown in this figure is easily generalizable to the case of a two-dimensional network or the like. In the figure, a low-level transmitter 1 supplies, through a passive power distributor 2, the N radiating sources of the linear network, S ₁ ..., S _i , ... S _N. Between the distributor 2 and the sources S _i , active modules M _i perform a phase shift  _i and an amplification with a gain A _i , the values of phase shift and amplification gain being controllable by the control unit 3. A at the output of the active modules M _i , the complex useful signals a _i are routed to the radiating sources S _i , from which they are radiated. In the event of an impedance mismatch between the modules M and the sources S, there will possibly be reflected signals b _i propagating in opposite directions with respect to the useful signals.

The waves radiated by S sources are superimposed with the amplitudes and phases assigned to them, according to a beamforming calculation, to radiate in a desired direction with a lobe formed in the intended application.

Figure 2 schematically shows in section an example of an active antenna arranged in a linear array, operating in reception. The example shown in this figure is easily generalizable to the case of a two-dimensional network or the like. In the figure, a receiver 11 is supplied, through a passive combiner 12, by the N sources of the linear network, S ₁ ..., S _i , ... S _N. Between the combiner 12 and the sources S _i , active modules M _i effect a phase shift  _i and an amplification with a gain A _i , the values of phase difference and amplification gain being controllable, in particular by the control unit 13 The complex useful signals a _i are routed from the radiating sources S _i to the inputs of the active modules M _i , where they are amplified with the controllable gains and phases for each signal. In the event of an impedance mismatch at the input of the sources S, there will possibly be reflected signals b _i propagating in opposite directions with respect to the useful signals.

The waves arriving on S sources, combined with the amplitudes and the phases which are allotted to them by a beam formation calculation, then arrive on the receiver 11, in a coherent manner, coming from a determined direction, with a lobe formed according to the intended application.

Figure 3 schematically shows an antenna active radar, operating alternately in transmission and in reception. The alternation of transmission / reception is ensured by switches 25, 52 controlled by a synchronization clock 24. On the Figure 3, orthogonal polarizations can be selected by switch 26, for reception as for the show. As in the two examples previous, phase and gain are controllable by control means 23, both in transmission and in reception. The order values that will be provided for controlling a given reception channel, are not necessarily the same as for the same channel used for the show.

In the figure, a single active transmission / module reception is shown, including controllable phase shifter 27 and a controllable attenuator 28, to adjust the gain of the module. However, you need one active module per channel, and, in this example, there are m · m 'channels, each channel being connected to a radiating source composed of K patches S 1 / ij to S k / ij, m 'being the number of columns of sources, of which only the first and second are (partially) represented.

In transmission mode, the transmitter 21 supplies its signals to a distributor / combiner 22, which supplies the active I / O modules. The phase and the attenuation of the signal will be determined by the controllable phase shifter 27 and the controllable attenuator 28, according to the instructions given by the control computer 23. Then, the switches 25 and 52 will be controlled by the clock 24 to engage the power transmission channel, and the signal will be amplified by the power amplifier 29, before being sent to the radiating sources S _ij .

In reception mode, the receiver 31 receives the signals from the combiner / distributor 22, which are routed by the active E / R modules. In the E / R modules, the signals coming from the radiating sources S _ij are switched by the switches 25, 52 on the reception channel and pass through a low noise amplifier 30. Then, the phase shift and the attenuation are applied by the controllable phase shifter 27 and the controllable attenuator 28, controlled by the control computer 23.

These antenna architectures according to Figures 1 to 3 are well known to those skilled in the art, and a description more detailed is not necessary to illustrate the principles of the invention.

To better understand the calibration method according to the invention, a description will be given below using a mathematical formalism belonging to matrix algebra. In this presentation, the scalar quantities will be designated by Roman letters, possibly including indices which indicate their position in a vector or in a matrix; the vector quantities will be designated by underlined Roman letters; and the matrix quantities will be designated by capital letters GRAS . All the quantities are complex, comprising an amplitude and a phase. The relationship between these quantities is symbolically shown in Figure 4.

représente les N commandes de l'antenne active, en amplitude et en phase, avec : |c_i|_max = 1 , ce qui veut dire que le gain maximal des voies est pris comme référence, soit 0 dB.

représente les N excitations réelles, qui sont les ondes incidentes sur les sources rayonnantes.For example, the vector:

represents the N commands of the active antenna, in amplitude and in phase, with: | c _i | _max = 1, which means that the maximum gain of the channels is taken as a reference, ie 0 dB.

represents the N real excitations, which are the waves incident on the radiating sources.

We define a matrix of dispersions D which gives us the relation between the commands and the excitations:
A = D C. The matrix D is NxN and diagonal, with the element
d _i = a _i / c _i which represents the difference in amplitude and phase between the commanded excitation and the real excitation of the channel i. This matrix takes into account the effects of dispersion of characteristics due to manufacturing, as well as the inaccuracies of orders.

The vector I represents "the illumination" or "the field on the opening":

• Dans le cas où les sources rayonnantes sont des guides ouverts, I_i représente l'amplitude et la phase du champ électrique à l'endroit où il est maximal, dans le plan médian parallèle aux petits côtés du guide pour une onde en mode fondamental, TE₁₀.
• Si les sources rayonnantes sont des dipôles demi-onde, I_i représente le courant en leur milieu ; si ce sont des "patches", I_i représente la densité de courant en leur centre.
• Si les sources rayonnantes sont des fentes résonnantes dans la paroi d'un guide, I_i représente la tension entre les deux bords de la fente au milieu de sa longueur, c'est à dire à l'endroit ou cette tension est maximale.

These terms are commonly used in the specialized literature to characterize the electromagnetic field on the radiating plane. To simplify the calculations, it is assumed that the radiating sources are monomode and one is interested only in the direction of nominal polarization. The consequence of these hypotheses is that the distribution of this electric field can thus be characterized by a single complex number (amplitude and phase). Some examples will serve to illustrate this point:

• In the case where the radiating sources are open guides, I _i represents the amplitude and the phase of the electric field at the place where it is maximum, in the median plane parallel to the short sides of the guide for a wave in fundamental mode, TE ₁₀ .
• If the radiating sources are half-wave dipoles, I _i represents the current in their middle; if they are "patches", I _i represents the current density at their center.
• If the radiating sources are resonant slots in the wall of a guide, I _i represents the tension between the two edges of the slot in the middle of its length, ie at the place where this tension is maximum.

These last two examples show us that the physical quantity of I _i is not necessarily a magnetic or electric field, but can be another quantity which characterizes the radiation of the source, this quantity being proportional to the field on the opening of the source.

We then define a matrix R characterizing the radiation phenomena, which allows us to obtain the illumination I from the real excitations A by the relation: I = R A. R is an NxN matrix which would be diagonal in the absence of coupling between the sources: I _i = r _i a _i shows that the illumination would in this case only depend on the incident wave on the source i.

qui s'écrit en notation matricielle : I = R A .But as we underlined above, the couplings between the neighboring sources introduce errors of estimation of radiated fields if they are not taken into account in computations. So our matrix R includes non-diagonal elements, which represent contributions, to the illumination of a given place, from neighboring sources. Thus, the field on the source S _i depends on the waves a _j incident on the other sources, by coupling coefficients r _ij :

which is written in matrix notation: I = R A.

Dans ce formalisme, l'élément s_ii représente le coefficient de réflexion de la source S_i entourée, avec toutes les autres sources S_i≠j mises sur charge.If we take the above example of dipoles or "patches", the couplings can be written in the classic form of a diffraction matrix S having the elements [s _ij ]; to the incident wave ai on the source Si, a reflected wave is superimposed:

In this formalism, the element s _ii represents the reflection coefficient of the surrounded source S _i , with all the other sources S _{i ≠ j} loaded.

où : I = (U - S)A, U étant la matrice diagonale unitaire NxN.The radiation is in fact proportional to the normalized current with respect to the line impedance:

where: I = ( U - S ) A , U being the unitary diagonal matrix NxN.

où : I = (U + S)A, U étant la matrice diagonale unitaire NxN, et S étant toujours la matrice "de diffraction" des coefficients de couplage.In a second example, we consider the case of a network of slot guides, with all the slots identical and arranged in the same way in each guide. The illumination then depends on the tensions on the slits, that is:

where: I = ( U + S ) A , U being the unitary diagonal matrix NxN, and S being always the "diffraction" matrix of the coupling coefficients.

When a "near field" probe is placed at a few wavelengths opposite the center of the source S _i , an electric field is collected:

We observe that E _i is thus a linear function of the illuminations of each source, including mainly the source S _i but also the others.

We thus define a column vector "near field" E and the matrix NxN "near field radiation" P such that E = P I with:

The final step in the construction of this formalism is to record the far field diagram of the active antenna. To carry out this measurement, a test receiving antenna is placed at a great distance from 2D ² / λ, where D is the largest dimension of the radiating plane of the antenna, and λ the wavelength of the radiation. By rotating the active antenna, its radiation pattern is sampled in a sufficient number p directions in space, taking note each time of the amplitude and phase of the signal received by the test antenna, to obtain the values F _j of the "far field diagram" which is represented by the column vector:

Then we define a matrix Nxp "far field radiation" L : F = L I.

The formalism exposed so far is strictly linear, and cannot take into account non-linearities due, for example, to amplifiers operating near saturation. For a more exact treatment in this case, the method must be adapted as we will explain below. In any case, this phenomenon only concerns the transformation: A = D C , all the other relationships remaining linear.

In the case of an antenna operating in reception, all the equations remain linear. We first consider the simplest case of a linear transmit antenna to illustrate the principle of the calibration method according to the invention. It is first a question of measuring each of the NxN complex terms q _ij of the matrix: Q = PRD ,
which we get from: E = Q C = PRD C.
Indeed, the near-field probe allows us to directly measure the elements E _i of the vector E as we explained above.

The calibration method according to the invention allows get the values of all the vectors and matrices from a set of field measurements close performed for a number N of positions of the probe equal to the number of radiation sources: for each position, N + 1 readings are carried out (command initial + switching of each of the N bits by 180 °); the total number of measurements is therefore N (N + 1). The initial calibration can be followed by recalibration iteratively, to achieve desired precision given. We first describe the initial calibration.

1) on commande une loi équi-amplitude et équi-phase, c'est-à-dire c_i = 1 pour tout i de = 1 à N ;

2) on place successivement la sonde champ proche devant chaque source S_i ; on relève alors un signal complexe z_i sur le récepteur, proportionnel au champ électrique E_i au niveau de la sonde ;

3) on commute le bit 180° du déphaseur de la voie n° j de l'antenne active ; on relève alors un nouveau signal z'_ij sur le récepteur.

The measurement is carried out as follows:

1) an equi-amplitude and equi-phase law is controlled, that is to say c _i = 1 for all i from = 1 to N;

2) the near field probe is successively placed in front of each source S _i ; a complex signal z _i is then noted on the receiver, proportional to the electric field E _i at the level of the probe;

3) the 180 ° bit of the phase shifter of channel n ° j of the active antenna is switched; a new signal z ′ _ij is then noted on the receiver.

où la constante q caractérise la base de mesure du champ proche (q est une fonction de la sonde et du récepteur). z'ij = q(qi1c1+...+qij(-cj)+...+qiNcN); En faisant la différence complexe entre ces deux équations (1) et (2), on obtient : z_i - z'_ij = 2q(q_ij·c_j) , soit, car nous avons pris c_j = 1 : qij = zi - z'ij 2q . These two measures are characterized by the following equations:

where the constant q characterizes the measurement base of the near field (q is a function of the probe and the receiver). z ' ij = q (q i1 vs 1 + ... + q ij (-vs j ) + ... + q iN vs NOT ); By making the complex difference between these two equations (1) and (2), we obtain: z _i - z ' _ij = 2q (q _ij · c _j ), that is, because we have taken c _j = 1: q ij = z i - z ' ij 2q .

At each position of the probe in front of a source S _i , the measurement of z _i is carried out, then the N measurements of the z ′ _ij are carried out by switching the 180 ° bits of all the channels in turn. By relation (3) above, we immediately obtain the N elements q _ij of line i of the matrix Q.

The receiver used for the probe must be able to measure complex signals with good accuracy, for example a receiver with two mixers and two channels I (in phase) and Q (Quadrature of phase).

The matrix Q thus obtained is called initial calibration matrix, because it allows us to calculate the commands to obtain a desired radiation pattern in the far field. The radiation diagram is characterized by the vector F of p field values recorded or calculated. To perform beamforming, these p values specified by the calculation represent the desired radiation in these main characteristics. The calculation must then determine how to obtain these far field values from control parameters which will be applied at the level of the control of the antenna.

The control values can be obtained from the vector F by matrix transformations using the initial calibration matrix Q. We summarize: Measurements by near-field probe give us: E = P I = Q VS . We also have: F = L I → I = L ^-1 F.
It follows that: C = Q ^-1 E = Q ^-1 P I = Q ^-1 PL ^-1 F
The orders are thus calculated by: C = Q ^-1 PL ^-1 F.

Q ^{-1 is} obtained by inversion of the initial calibration NxN matrix Q , the measurement of which, term by term, has been described above. L ^-1 represents the transformation of the far field to the field on the aperture, and P is the transformation of the field on the aperture to the near field. These last two matrices are governed by the basic formulas of antenna theory, familiar to those skilled in the art; putting them into matrix computer form presents no difficulty.

Thus, it is possible to obtain, by the calibration method according to the invention, the matrix M = Q ^-1 PL ^-1 which specifies the commands necessary for obtaining a desired far field, in the linear case. Manufacturing dispersions are taken into account in this matrix M , because Q = PRD , where D is the dispersion matrix. The couplings are also taken into account in the matrix R.

For an antenna operating in reception, the above relationships remain the same: as single difference, during near-field measurements, the probe transmits and the active antenna receives.

In the case of a non-linear transmit antenna, the command values from this first measurement may not be sufficiently accurate. For the improve it is possible to iterate through from these first values, as described below.

To overcome the control errors originating from the non-linearities of the amplifiers, or from the imperfections of the phase-shifters and controllable attenuators (also non-linearities), the method according to a preferred variant of the invention begins with a first set of measurements such as described above. The control values C from this calibration are applied to their respective phase shifters and attenuators. The measurement procedure is then repeated, to obtain a second calibration matrix Q ' , slightly different from the first Q , since the dispersion matrix D will have changed somewhat for the new commands c _i . The new dispersion matrix D 'will have diagonal terms of the form: d '' i = a '' i / vs '' i.

We then calculate a second value of the commands: VS '= (Q' -1 PL -1 ) F , as well as the quadratic difference between this control law and the previous one:

ou ε_ et ε_a sont les précisions exprimées en radians et en amplitude relative.If the quadratic deviation is less than the targeted precision objective, the iterative process stops there; for example for a precision objective of 1 ° in phase and 0.15 dB in amplitude, we will take:

where ε _ and ε _a are the precisions expressed in radians and in relative amplitude.

If the convergence criterion is not respected, the iteration process continues in the same way, by measuring the new calibration matrix Q '' for commands C ', to obtain the new reoptimized commands:
C '' = ( Q '' ^-1 PL ^-1 ) F. The iteration thus continues up to the level of precision targeted. In practice, it takes only a few iterations.

In some cases the method according to the invention can take into account measurement data that has already been performed before the antenna calibration. By example, an active antenna comprising several hundreds or even thousands of active modules, is usually made with components that have undergone tests before their integration into the antenna. In addition, control characteristics can be noted on active modules, individually, to verify the correct operation of the latter before assembly.

According to a variant of the method of the invention, the control errors are taken into account in the calibration of the antenna, from the data relating to each active module. These data consist of a complex value (amplitude and phase) for each active module considered, depending on the command applied. The measurement of the fields close to the antenna is then carried out as before, for a uniform excitation law c _i = 1, for all i; then for each far field diagram sought, the commands C = Q ^-1 PL ^-1 F are calculated by taking in Q = PRD the dispersion matrix for c _i = 1, for all i; but for C found, D will be slightly different ( D ' ), hence a new value C' ; and so on by a small software working from the module measurement tables. In this variant, therefore, a near field survey is carried out; it is used to calculate the command C adapted to each diagram F , using this additional software. The iteration relates only to the dispersion matrix D , as a component of the matrix transformation Q = PRD . This last equation shows us that the two methods are theoretically equivalent, insofar as the matrices P and R are independent of the control state.

The choice between these two variants will be made according to the criteria of ease of implementation. In the first variant, M iterations are carried out, each comprising N (N + 1) near-field probe measurements, and this for each desired different radiation pattern. In the second variant, each active module must be characterized individually, but then, to determine the elements of the calibration matrix Q , we will only need a single measurement of N (N + 1) values in the near field, every other control laws which can be calculated from Q and the table of measurements carried out on the active modules.

Of course, the method according to the invention can give more precise measurements if one is ready to perform a higher number of near field readings, for example K times more, where K is a multiplicative factor (K = whole number). According to a variant of the method of the invention, when each active module is connected to a "sub-network" of radiant patches, the surface of which is clearly greater than the grid λ / 2 x λ / 2 optimal in precision to raise the near field of the antenna, we perform readings with K positions by sub-network. This allows you to get closer to the ideal grid, at cost an increase in the duration of the calibration. But accuracy is improved by averaging each group of K measurements by a mathematical "projection" of the field close to these K points radiated by a single subnetwork radiant. The way to accomplish this "projection" mathematics will be described in the following paragraphs.

• Avant la calibration de l'antenne active complète, on mesure le diagramme champ proche e :
  
  d'une source rayonnante seule, aux K points d'échantillonage de sa surface choisis plus haut.
• Après avoir relevé les E '  / nk en p = N . K points devant l'antenne active (K points devant chaque source rayonnante, donnant la "maille" du relevé champ proche), on projete sur e les K relevés correspondant à chaque maille ; soit pour la maille de numéro n₀ :
  

The near field measurements are made, noted: E '/ nk at N · K points, which corresponds to an equal number of directions p = N · K of far field sampling. There are thus K times too many E '/ nk measurements to characterize the N commands of the antenna. The "projection on the near field diagram of a source" consists of following the following steps:

• Before calibrating the complete active antenna, the near field diagram e is measured:
  
  from a single radiating source, at the K sampling points of its surface chosen above.
• After having noted the E '/ nk in p = N. K points in front of the active antenna (K points in front of each radiating source, giving the "mesh" of the near field survey), the K readings corresponding to each mesh are projected onto e ; either for the mesh of number n ₀ :
  

• Ensuite on remplace les N.K releves E '  / nk par N valeurs :
  
  qui sont les moyennes des diagrammes champ prodche sur chaque maille, pondérées par le diagramme de la source située en face de cette maille.
• A partir de cette étape, on procède au calcul de la matrice NxN de calibration Q comme précédemment, où l'on a fait correspondre N relevés champ proche E_n à chaque jeu de N commandes c_i .

Mathematically this projection is expressed by the complex dot product: E not 0 = E '' not 0 . e *, where the symbol "*" indicates the complex conjugate

• Then we replace the NK readings E '/ nk by N values:
  
  which are the averages of the field field diagrams on each mesh, weighted by the diagram of the source located opposite this mesh.
• From this step, we proceed to the calculation of the calibration matrix NxN Q as above, where we made N readings near field E _n to each set of N commands c _i .

F est une matrice colonne de p = N·K termes ;

P et L sont des matrices carrées de p x p termes ;

T ramène à N termes seulement ;

Q est toujours une matrice NxN ; et

C est un vecteur colonne de N termes.

The mathematical operation of projection can be represented in a matrix formalism by the equation
E = T · E ', where T is an N x p matrix. The calibration formula then becomes:
column vector C = ( Q ^-1 T · P · L ^-1 ) F. In this equation,

F is a column matrix of p = N · K terms;

P and L are square matrices of pxp terms;

T brings back to N terms only;

Q is always an NxN matrix; and

C is a column vector of N terms.

The advantage of this variant is the increase in the precision of the results by a factor K ^1/2 , by averaging K measurements for the mesh located opposite each source. This advantage comes at the cost of a number of measurements increased by a factor of K, and the increase in the size of the matrices to be calculated in the same proportions.

The choice will thus be optimized according to the number active modules, number of radiation patterns different, the number of iterations necessary to have the required precision.

The method according to the invention and its variants have inter alia the advantages mentioned above. The calibration matrix Q = [q _ij ] directly links the commands of the active antenna to the near-radiated field, and takes into account all the dispersions which generate deviations between the real and theoretical radiation diagrams calculated by software. Manufacturing dispersions are taken into account in the matrix D = [d _i ]. The imperfections of the controllable elements - phase shifters, gain control - are taken into account by the iterative process, or by measurements carried out individually on the active modules. The influence of the couplings between the radiating sources on their elementary diagrams is taken into account by the non-diagonal terms of the matrix R = [r _ij ].

The calibration according to the invention is therefore better than the calibrations of the prior art, because it takes into account errors which are not taken into account in the methods of the prior art. In addition, the measurements of the present method are faster to perform, since a single scanning of the near-field probe is necessary (if there is no need for successive iterations, if for example the active modules are measured individually), with only N measurement positions, where N is the number of active modules. In the prior art, it is necessary to make a mapping of the whole near field, with a maximum spacing of (λ / 2) ² on a surface 2 to 3 times greater than that of the antenna. The switching of N bits 180 ° for each position of the probe is done very quickly for an electronically controlled antenna.

Because the manufacturing dispersions are taken into account by the method according to the invention, the range of values allowed for these parameters (gain, phase shift depending on commands) can be extended. The severe specifications that lead to the rejection of a large number of components, become useless. In in addition, the method according to the invention does not require any specific circuit in the active antenna. The methods in the prior art, however, ask for example integration of a "calibration BFN", a receiver specific by module, or a switch for setting on individual load of each module only for calibration needs.

Claims (9)

1. Method of calibrating an active antenna having N radiating sources comprising the following steps: a probe is placed in front of each radiating source in succession to measure the near field in front of said source, characterised in that: for an antenna configuration required to obtain a required radiation diagram and in that: during said measurement of said near field in front of said source a phase shifter of each channel in turn is caused to shift the phase of radiation by said channel 180° relative to its nominal value with each of the other N - 1 sources operating at their respective nominal value for said configuration in order to obtain the required radiation diagram.
2. Method of calibrating an active antenna having N radiating sources disposed in an array with coupling between said sources which are energised by active modules comprising variable phase shift means and variable gain control means, said sources, said active modules, said phase shift means and said gain control means having close manufacturing tolerances and said phase shift means and said gain control means being subject to inaccuracies of response conditioned by a given control value, characterised in that: near field measurements are carried out using the method claimed in claim 1 and an appropriate probe to characterise simultaneously the effects of said coupling between sources, said manufacturing tolerances and said inaccuracies.
3. Method according to claim 1 or claim 2 wherein the gain and phase control values for a required antenna configuration to obtain a required radiation diagram are first determined by one of the methods of the aforementioned claims, and said control values are applied to said phase shift and gain control means this method being further characterised in that: the near field measurements are repeated with said control values to obtain closer corrections to said values by successive iterations as required.
4. Method according to claim 1 or claim 2 characterised in that: a calibration table is drawn up on the basis of measurements carried out on said active modules before the antenna is assembled and in that: after the control values are determined by one of the methods of the aforementioned claim the control values are further refined by an iterative subroutine using said module calibration table.
5. Method according to claim 1 for calibrating an active antenna whose modules are arranged in a mesh whose size is significantly greater than λ/2 in at least one dimension of the antenna, characterised in that: K near field measurements are carried out in front of each radiating source corresponding to a mesh of substantially λ/2 and said K measurements are then averaged by projection onto the near field diagram of an isolated source.
6. Method according to any one of claims 1 to 4 applied to transmit active antennas.
7. Method according to any one of claims 1 to 4 applied to receive active antennas.
8. Method according to any one of claims 1 to 4 applied to receive active antennas alternately transmitting and receiving.
9. Method according to claim 7 characterised in that the method is applied to a radar antenna and in that the method is applied twice: once for the antenna transmitting to determine the transmit phase shift and gain control values and again for the antenna receiving to determine the receive phase shift and gain control values.

Images