METHOD AND DEVICE FOR NON-COHERENT DISTRIBUTED FULL-DUPLEX TRANSMISSION RADAR SYSTEMS

Abstract

The invention relates to a radar method for exchanging signals between at least two non-coherent transceiver units which respectively have initially non-synchronous, in particular controllable, clock sources, having the following steps: a synchronization in which clock offsets and/or clock rates of the clock sources of the at least two transceiver units are adapted; a full-duplex measuring process in which a first transmission signal of the first transceiver unit is transmitted to the second transceiver unit and a second transmission signal of the second transceiver unit is transmitted to the first transceiver unit via a radio channel; with synchronization prior to the full-duplex measuring process being carried out in such a way that a time offset and/or a frequency offset between the transmission signals at least substantially remain(s) constant during a transmission time of the full-duplex measuring process.

Claims

1. Radar method for exchanging signals between at least two non-coherent transceiver units that respectively have clock sources that are initially non-synchronous having the following: a synchronization in which clock offsets and/or clock rates of the clock sources of the at least two transceiver units are aligned; a full-duplex measuring process, in which via a radio channel, a first transmit signal is transmitted from the first transceiver unit to the second transceiver unit and a second transmit signal is transmitted from the second transceiver unit to the first transceiver unit; wherein the synchronization is performed prior to the full-duplex measuring process, in such a way that a time offset and/or a frequency offset between the transmit signals remain(s) at least substantially constant during a transmission time of the full-duplex measuring process.

2. The method according to claim 1, wherein the synchronization is performed by radio, which is preferably reciprocal, in particular over the radio channel or in particular by cable.

3. The method according to claim 1, wherein exchanged synchronization signals are used for synchronization, in particular similar synchronization signals.

4. The method according to claim 1, wherein exchanged synchronization signals are modulated using a similar frequency modulation, in particular an FMCW modulation or an FSK modulation; and wherein a frequency detuning and/or a frequency drift between the synchronization signals is preferably determined in the transceiver units.

5. The method according to claim 1, wherein for synchronization, individual synchronization values, in particular individual signal parameters such as preferably frequency and/or phase values, are transmitted.

6. The method according to claim 5, wherein the individual synchronization values comprise a first global time of the first transceiver unit and/or a second global time of the second transceiver unit, wherein in particular the first global time is determined based on the second global time and a first local time, and/or the second global time is determined based on the first global time and a second local time.

7. The method according to claim 1, wherein, for synchronization, a chronological drift between the clock rates of the clock sources is determined and exchanged between the at least two transceiver units.

8. The method according to claim 1, wherein for synchronization (VS1) the clock sources are controlled using corresponding control signals, in particular control voltages, in such a way that the clock rates of the clock sources are aligned.

9. The method according to claim 1, wherein in the full-duplex measuring process, a distance and/or a relative speed between the at least two transceiver units are determined based on a signal propagation time of the transmit signals over the radio channel.

10. The method according to claim 1, wherein in the full-duplex measuring process, similar transmit signals, in particular FMCW transmit signals are exchanged, which in particular comprise a sequence of alternating up and down chirps, a sequence of only up chirps or a sequence of only down chirps.

11. The method according to claim 1, wherein, in the full-duplex measuring process, a comparison signal is generated in each transceiver unit by mixing and/or correlating received transmit signals with the respective corresponding transmit signals, and the comparison signals are exchanged between the transceiver units, wherein in particular, in at least one of the two transceiver units, the following are performed: Determining and correcting a center frequency; Correcting a phase shift; and Superimposing to form a synthetic received signal.

12. The method according to claim 12, wherein, in the full-duplex measuring process, a comparison signal is generated in each transceiver unit by mixing and/or correlating received transmit signals with the respective corresponding transmit signals, and evaluation parameters, in particular spectral ones, which are exchanged between the transceiver units, are determined based on the comparison signals in the respective transceiver unit.

13. The method according to claim 12, wherein for each signal chirp, a comparison spectrum of the comparison signal is generated, and the evaluation parameters comprise a frequency value of the maximum in the comparison spectrum and a phase value of the maximum in the comparison spectrum.

14. The method according to claim 12, wherein a two-dimensional comparison signal spectrum is generated in each transceiver station, and the evaluation parameters comprise two frequency values per transceiver station, which are the frequency values of a maximum along each dimension of the two-dimensional comparison signal spectrum.

15. A radar system, in particular a secondary radar system, for determining a distance and/or a relative speed, comprising: at least two, preferably spatially separated, non-coherent transceiver units, that respectively have non-synchronous, in particular controllable, clock sources; a synchronization device for carrying out a synchronization, in which clock offsets and/or clock rates of the clock sources of the at least two transceiver units are aligned; wherein the transceiver units are designed to perform a full-duplex measuring process, in which via a radio channel a first transmit signal of a first transceiver unit is transmitted to a second transceiver unit and a second transmit signal of the second transceiver unit is transmitted to the first transceiver unit, wherein the synchronization device is designed to perform the synchronization prior to the full-duplex measuring process, in such a way that a time offset and/or a frequency offset between the transmit signals remain(s) at least substantially constant during a transmission time of the full-duplex measuring process.

16. The radar system according to claim 15, wherein the clock sources are in particular controllable, preferably voltage-controlled, oscillators, in particular quartz oscillators.

17. Use of the method according to claim 1, for mobile devices, preferably for vehicles, in particular unmanned aerial vehicles or preferably passenger cars and/or trucks.

18. Use of the system according to claim 15, for mobile devices, preferably for vehicles, in particular unmanned aerial vehicles or preferably passenger cars and/or trucks.

19. A radar system, in particular a secondary radar system, for determining a distance and/or a relative speed, in particular for carrying out the method according to claim 1.

20. An apparatus for carrying out the method according to one of claim 1.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In the following description, further principles, aspects and embodiments of the invention are described, also with reference to the accompanying drawings. The drawings show the following:

[0047] FIG. 1 two transceiver units as may be found in the prior art;

[0048] FIG. 2 Frequency responses in the baseband for non-synchronized clock sources of the transceiver units as found in the prior art; and

[0049] FIG. 3 Frequency responses in the baseband for synchronized clock sources of the transceiver units.

DETAILED DESCRIPTION OF THE DRAWINGS

[0050] In the following description, the same reference numerals are used for identical and equivalent parts. In the radar method according to the invention, an aligning of the clock sources, or the local oscillators, takes place prior to the actual measurement, which is the full-duplex measuring process.

[0051] For aligning the clock sources, for example, the chronological drift may first be ascertained by measurement, and subsequently the clock sources of the individual radar units are aligned.

[0052] It should further be mentioned here that the reduction of the drift could also be achieved by selecting very high-quality oscillators (for example atomic clocks). However, this is seldom possible for the use case, in view of the size, complexity, energy consumption and cost of atomic clocks.

[0053] FIG. 3 shows the instantaneous frequency of the down-converted signals after synchronization (aligning of the clock sources). Here f.sub.shift=Δf+μΔτ denotes the unknown center frequency, which consists of a frequency offset Δf and the product of sweep rate and time offset ρΔτ. These two instantaneous frequencies are constant and shifted relative to one another by 2τ.sub.0, namely by twice the propagation time in the transmission channel.

[0054] Derivation of the Signal Model

[0055] In the following, signals are denoted by s.sub.ak,x.sup.u/d(t), where the superscript u/d denotes either up or down chirp, the subscripts a denote the number of the transceiver unit (radar station), k∈{1, . . . , K} denotes the number of the FMCW chirp, and x denotes the type of signal (transmit signal—tx, receive signal—rx, mixed signal—mix and beat signal—beat); “beat signal” refers to the digitized (discrete in time and value) mixed signal.

[0056] The phases of the transmit signals for two transceiver units, which also serve as reference signals for the mixing process, may be mathematically expressed as

Φ.sub.1k,tx.sup.u/d=2π(f.sub.c+Δf/2∓B/2)t±2πμ(t.sup.2−Δτt)+Θ.sub.1k+Ψ.sub.1k(t) and  (1)

Φ.sub.2k,tx.sup.u/d=2π(f.sub.c−Δf/2∓B/2)t±2πμ(t.sup.2+Δτt)+Θ.sub.2k+Ψ.sub.2k(t)  (2)

[0057] where f.sub.c and B correspond to the carrier frequency and bandwidth of the RF signal. The variables θ.sub.ak and Ψ.sub.ak(t) represent unknown start phases and phase noise during an FMCW chirp of transceiver unit a. In principle, all FMCW chirps are chronologically shifted by Δr and the carrier frequency of both transceiver units differs by Δf.

[0058] The associated complex time signals may be described as A.sub.a exp (jΦ.sub.ak.sup.u/d(t)). Both signals are transmitted over a transmission channel with the propagation time τ(t) relative to the respective other transceiver unit and are received there. The propagation time τ.sub.k=τ.sub.0+τ.sub.k′ in the transmission channel is composed of a start time τ.sub.0 and a small change τ.sub.k′. On this basis, distance and speed may be calculated over x.sub.k=c.sub.0τ.sub.k=x.sub.0+kvT.sub.sw. The latter influences the phase position from chirp to chirp, but only influences the resulting sampling frequency to a limited degree. The phase position of these received signals

Φ.sub.1k,rx.sup.u/d(t)=Φ.sub.2k,tx.sup.u/d(t−τ.sub.k) und Φ.sub.2k,rx.sup.u/d(t)=Φ.sub.1k,tx.sup.u/d(t−τ.sub.k)  (3)

may accordingly be represented as a time-delayed variant of the transmit signals. As is generally known, the phases of the signals after the mixing process (IQ mixing or I mixing and subsequent Hilbert transformation) and subsequent filtering with a low-pass filter may be represented as:

Φ.sub.1k,mix.sup.u/d(t)=Φ.sub.1k,tx.sup.u/d(t)−Φ.sub.1k,rx.sup.u/d(t) and Φ.sub.2k,mix.sup.u/d(t)=Φ.sub.2k,rx.sup.u/d(t)−Φ.sub.2k,tx.sup.u/d(t)   (4)

[0059] Substituting (1) and (2) into (4) yields the mixed signals:

[00001]$\begin{array}{cc}{s}_{1k,\mathrm{mix}}^{u/d}\left(t\right)=A\exp \left\{j\left(2\pi \left(\mathrm{\Delta ft}+\left(f_c\mp B/2\right){\tau }_k\pm \mu \left({\tau }_0+\Delta \tau \right)t\right)+{\Psi }_{1k}\left(t\right)-{\Psi }_{2k}\left(t-{\tau }_k\right)+{\Theta }_{1k}-{\Theta }_{2k}\right)\right\}\mathrm{and}& \left(5\right)\\ s_{2k,\mathrm{mix}}^{u/d}\left(t\right)=A\exp \left\{j\left(2\pi \left(\mathrm{\Delta ft}-\left(f_c\mp B/2\right){\tau }_k\pm \mu \left(-{\tau }_0+\Delta \tau \right)t\right)+{\Psi }_{1k}\left(t-{\tau }_k\right)-{\Psi }_{2k}\left(t\right)+{\Theta }_{1k}-{\Theta }_{2k}\right)\right\},& \left(6\right)\end{array}$

where it will be apparent that the phase position of both signals is equally influenced by the interference variables. The influence due to the propagation time or the change of the propagation time results in a complex conjugated phase change. Because it is not feasible to sample the mixed signals at the time t, the beat signals

[00002]$\begin{array}{cc}{s}_{1k,\mathrm{beat}}^{u/d}\left(t\right)=s_{1k,\mathrm{mix}}^{u/d}\left(t-\frac{\Delta \tau }{2}\right)=s_{1k,\mathrm{beat}}^{u/d}\left(t\right)\exp \left\{j{\gamma }_1\right\}\mathrm{and}{s}_{2k,\mathrm{beat}}^{u/d}\left(t\right)=s_{2k,\mathrm{mix}}^{u/d}\left(t+\frac{\Delta \tau }{2}\right)=s_{2k,\mathrm{beat}}^{u/d}\left(t\right)\exp \left\{j{\gamma }_2\right\}& \left(5\right)\end{array}$

are a time-delayed version of the mixed signals, which for a mono-frequency signal may also be represented as a phase shift by γ.sub.1 or γ.sub.2.

[0060] The Full-Duplex Measuring Process

[0061] If an estimation of the distance and speed is to be performed, it is possible for example for only up-chirps to be transmitted and received. Advantageously in this case, the unambiguity range of the detectable speed is doubled. This exemplary embodiment is also possible analogously using down-chirps. After a Fourier transform {.Math.} in the distance direction, the spectra of the sampled signals may be described as

[00003]$\begin{array}{cc}{S}_{1k,\mathrm{beat}}^u\left(f\right)=A\delta \left\{f-\left(\Delta f+\mu \left({\tau }_0+\Delta \tau \right)\right)\right\}*W\left(f\right)*F\left\{\exp \left\{j{\Psi }_{\mathrm{pn},1k}\left(t\right)\right\}\right\}.Math.\hspace{1em}\exp \left\{j{\gamma }_1\right\}.Math.\exp \left\{j{\gamma }_2\right\}.Math.\exp \left\{j\left(-2\pi \left(f_c-B/2\right){\tau }_k+{\Theta }_{1k}-{\Theta }_{2k}\right)\right\}\mathrm{and}& \left(8\right)\\ S_{2k,\mathrm{beat}}^u\left(f\right)=A\delta \left\{f-\left(\Delta f+\mu \left(-{\tau }_0+\Delta \tau \right)\right)\right\}*W\left(f\right)*\hspace{1em}F\left\{\exp \left\{j{\Psi }_{\mathrm{pn},2k}\left(t\right)\right\}\right\}.Math.\exp \left\{j{\gamma }_2\right\}.Math.\exp \left\{j\left(-2\pi \left(f_c-B/2\right){\tau }_k+{\Theta }_{1k}-{\Theta }_{2k}\right)\right\}& \left(9\right)\end{array}$

[0062] For the window function in the spectral domain and the phase noise, the notation W(f), Ψ.sub.pn,1k(t)=Ψ.sub.1k(t)−Ψ.sub.2k(t−τ.sub.k) and Ψ.sub.pn,2k(t)=Ψ.sub.1k(t−τ.sub.k)−Ψ.sub.2k(t) was chosen.

[0063] Exchanging the Transmit Signals and Superimposing the Transmit Signals:

[0064] It is now possible to exchange the sampled signals completely. The calculation in this case is simplified compared to methods of the prior art. First, based on (8) and (9), it may be determined that both sampled signals are arranged around a virtual center frequency f.sub.shift. This is calculated via a search for the two peaks

[00004]$\begin{array}{cc}{\hat{f}}_{\mathrm{shift}}=\frac{1}{K}\underset{k}{.Math.}\left(\underset{f}{\mathrm{argmax}}\left\{S_{1k,\mathrm{beat}}^u\left(f\right)\right\}+\underset{f}{\mathrm{argmax}}\left\{S_{2k,\mathrm{beat}}^u\left(f\right)\right\}\right)\approx \Delta f+\mathrm{\mu \Delta \tau }.& \left(6\right)\end{array}$

[0065] In the next step, this shift is corrected, yielding

[00005]$\begin{array}{cc}{\tilde{S}}_{1k,\mathrm{beat}}^u\left(f\right)\approx A\delta \left\{f-{\mathrm{\mu \tau }}_0\right\}*W\left(f\right)*F\left\{\exp \left\{j{\Psi }_{\mathrm{pn},1k}\left(t\right)\right\}\right\}.Math.\hspace{1em}\exp \left\{j{\gamma }_1\right\}.Math.\exp \left\{j\left(2\pi \left(f_c-B/2\right){\tau }_k+{\Theta }_{1k}-{\Theta }_{2k}\right)\right\}\mathrm{and}& \left(11\right)\\ {\tilde{S}}_{2k,\mathrm{beat}}^u\left(f\right)\approx A\delta \left\{f+{\mathrm{\mu \tau }}_0\right\}*W\left(f\right)*\hspace{1em}F\left\{\exp \left\{j{\Psi }_{\mathrm{pn},2k}\left(t\right)\right\}\right\}.Math.\hspace{1em}\exp \left\{j{\gamma }_2\right\}.Math.\exp \left\{j\left(-2\pi \left(f_c-B/2\right){\tau }_k+{\Theta }_{1k}-{\Theta }_{2k}\right)\right\}& \left(12\right)\end{array}$

[0066] Such a shift may be performed, for example, with the aid of a Fourier transform. Both signals are centered around the beat frequency, which corresponds to a distance of 0 m (or a beat frequency of 0 Hz). The phases of the maximum are then obtained for each chirp and divided by two

[00006]$\begin{array}{cc}{\varphi }_k=\left(\arg \left\{\max \left\{{\tilde{S}}_{1k,\mathrm{beat}}^u\left(f\right)\right\}\right\}+\arg \left\{\max \left\{{\tilde{S}}_{2k,\mathrm{beat}}^u\left(f\right)\right\}\right\}\right)/2\approx {\Theta }_{1k}-{\Theta }_{2k}+\left({\gamma }_1-{\gamma }_2\right)/2+{\varphi }_{0k}^u.& \left(7\right)\end{array}$

[0067] Due to the division, a phase jump by ϕ.sub.0k.sup.u∈lπ with l∈ may potentially occur, which may be corrected by unwrapping (except for the phase φ.sub.0.sup.u of the first chirp). The remaining phase noise during an FMCW chirp may be approximated by {exp{jΨ.sub.pn,ak(t)}}≈{1+jϵ.sub.k(t)}, which corresponds to a Taylor series expansion up to the linear element ϵ.sub.k(t). This approximation holds very well in practical applications because the level of phase noise is necessarily much less than the amplitude of the carrier signal. After correction of the phase values per FMCW chirp with (13), the signals are obtained:

[00007]$\begin{array}{cc}{\stackrel{\approx }{S}}_{1k,\mathrm{beat}}^u\left(f\right)\approx A\delta \left\{f-{\mathrm{\mu \tau }}_0\right\}*W\left(f\right)*F\left\{1+j{\mathrm{.Math.}}_k\left(t\right)\right\}\exp \left\{j{\varphi }_0^u\right\}.Math.\hspace{1em}\exp \left\{j\left({\gamma }_1-{\gamma }_2\right)/2\right\}.Math.\exp \left\{j2\pi \left(f_c-B/2\right){\tau }_k\right\}\mathrm{and}& \left(14\right)\\ {\stackrel{\approx }{S}}_{2k,\mathrm{beat}}^u\left(f\right)\approx A\delta \left\{f+{\mathrm{\mu \tau }}_0\right\}*W\left(f\right)*F\left\{1+j{\mathrm{.Math.}}_k\left(t\right)\right\}\hspace{1em}\exp \left\{j{\varphi }_0^u\right\}.Math.\exp \left\{-j\left({\gamma }_1-{\gamma }_2\right)/2\right\}.Math.\exp \left\{-j2\pi \left(f_c-B/2\right){\tau }_k\right\},& \left(15\right)\end{array}$

the phase shift of which due to interference variables is now exactly complex conjugated. Finally, the time signal pertaining to (15) is complex conjugated and superposed to form the synthetic beat signal.

[00008]$\begin{array}{cc}{S}_{k,\mathrm{synth}}^u\left(f\right)\approx A\delta \left\{f-{\mathrm{\mu \tau }}_0\right\}*W\left(f\right)\exp \left\{j\left({\gamma }_1-{\gamma }_2\right)/2\right\}.Math.\hspace{1em}\exp \left\{j2\pi \left(f_c-B/2\right){\tau }_k\right\}\exp \left\{j{\varphi }_0^u\right\}& \left(8\right)\end{array}$

[0068] The relative speed may be obtained computationally efficiently via a Fourier transform of (16) along the chirp number k, by means of which the relative speed may be determined.

[0069] Determining the Frequency Value and Phase Value Per Chirp:

[0070] The required quantity of data to be transmitted and the number of required computation steps may also be reduced as follows: for each FMCW chirp, the beat frequency of the maximum is respectively determined in all transceiver units.

[0071] Calculating these maxima from both transceiver units via

[00009]$\begin{array}{cc}{\hat{\tau }}_{0,k}=\left(\underset{f}{\mathrm{argmax}}\left\{S_{1k,\mathrm{beat}}^u\left(f\right)\right\}-\underset{f}{\mathrm{argmax}}\left\{S_{2k,\mathrm{beat}}^u\left(f\right)\right\}\right)/\left(2\mu \right)& \left(9\right)\end{array}$

leads directly to the propagation time in the transmission channel. Because the distance changes only (very) slightly during the complete transmission sequence, by averaging

[00010]$\begin{array}{cc}{\hat{\tau }}_0=\frac{1}{K}{.Math.}_k{\hat{\tau }}_{0,k}& \left(10\right)\end{array}$

the accuracy of a propagation time measurement (or distance measurement) may be significantly increased. Likewise, it is possible to estimate the phase change by detecting the phase of the maxima in both radars via

[00011]$\begin{array}{cc}{\hat{\varphi }}_k=\arg \left\{\max \left\{S_{1k,\mathrm{beat}}^u\left\{\right\}\right\}\left\{\right\}\arg \left\{\max \left\{S_{2k,\mathrm{beat}}^u\left\{\right\}\right\}\left\{\right\}{\pi \left(f_c-\frac{B}{2}\right)}_{k_0}\right\}\right\}& \left(11\right)\end{array}$

[0072] Via this phase change, the change in the length of the transmission path may be obtained with great precision and speeds may be measured. The variable ϕ.sub.0 represents an unknown start phase that has no influence on the measurement. Thus, the transmission of 2K real values is necessary for a chirp sequence with KFMCW chirps.

[0073] Determining a Two-Dimensional Spectrum Per Transceiver Unit:

[0074] In this exemplary embodiment, distance and relative speed may be estimated with the transmission of 2 real values per transceiver unit (independent of the length of the chirp sequence). For this, it is assumed that systematic interferences are dominant (phase noise has comparatively little influence). Thus, the clock frequencies of the two transceiver units do not match exactly and thus the chronological drift has not been set exactly to zero. This results in a linear phase change for each FMCW chirp, which may be expressed by a frequency offset Δf.sub.2 along the speed axis, and occurs equally in both transceiver units.

[0075] Thus, the two 2D Fourier transforms .sub.2{.Math.} along the FMCW chirps may be described as

[00012]$\begin{array}{cc}{S}_{k1,2D}^u\left(f,f_2\right)\approx A\delta \left\{f-\left(\Delta f+\mu \left({\tau }_0+\mathrm{\Delta \tau }\right)\right)\right\}*W\left(f\right)*F_2\left\{F\left\{1+j{\mathrm{.Math.}}_k\left(t\right)\right\}\right\}*\delta \left\{f_2-\left(f_c-B/2\right){\tau }_k-\Delta f_2\right\}*W_2\left(f_2\right).Math.\exp \left\{j{\gamma }_1\right\}\mathrm{and}& \left(20\right)\\ S_{k2,2D}^u\left(f,f_2\right)\approx A\delta \left\{f-\left(\Delta f+\mu \left(-{\tau }_0+\mathrm{\Delta \tau }\right)\right)\right\}*W\left(f\right)*\hspace{1em}{F}_2\left\{F\left\{1+j{\mathrm{.Math.}}_k\left(t\right)\right\}\right\}*\delta \left\{f_2+\left(f_c-B/2\right){\tau }_k-\Delta f_2\right\}*W_2\left(f_2\right).Math.\exp \left\{j{\gamma }_2\right\}& \left(21\right)\end{array}$

where γ.sub.1′ and γ.sub.2′ represent unknown and irrelevant phase values. The interference due to the 2D Fourier transform of the phase noise .sub.2{{1+jϵ.sub.k(t)}} is quasi-identical at both stations and cancels out. Calculating the maxima along the distance and speed axes (propagation time and change in propagation time) from the first transceiver unit and the second transceiver unit now directly yields the desired measurement values:

[00013]$\begin{array}{cc}{\hat{\tau }}_0=\left(\underset{f}{\mathrm{argmax}}\left\{S_{1k,2D}^u\left(f,f_2\right)\right\}-\underset{f}{\mathrm{argmax}}\left\{S_{2k,2D}^u\left(f,f_2\right)\right\}\right)/\left(2\mu \right)\mathrm{and}& \left(22\right)\\ {\hat{\tau }}_k=\left(\underset{f_2}{\mathrm{argmax}}\left\{S_{1k,2D}^u\left(f,f_2\right)\right\}-\underset{f_2}{\mathrm{argmax}}\left\{S_{2k,2D}^u\left(f,f_2\right)\right\}\right)/\left(2\left(f_c-B/2\right)\right).& \left(23\right)\end{array}$

LIST OF REFERENCE SIGNS

[0076] 1 First transceiver unit [0077] 2 Second transceiver unit [0078] 4 Radar system [0079] 11 Clock source of the first transceiver unit [0080] 12 Radio frequency (RF) generator of the first transceiver unit [0081] 13 Mixer of the first transceiver unit [0082] 14 Analog-to-digital (A/D) converter of the first transceiver unit [0083] 15 RF antenna of the first transceiver unit [0084] 21 Clock source of the second transceiver unit [0085] 22 Radio frequency (RF) generator of the second transceiver unit [0086] 23 Mixer of the second transceiver unit [0087] 24 Analog-to-digital (A/D) converter of the second transceiver unit [0088] RF antenna of the second transceiver unit [0089] T Radio channel