RADAR METHOD AND RADAR SYSTEM

Abstract

A radar method, in particular a primary radar method, in which at least one first and at least one second transceiver unit (S1, S2), which are in particular spatially separated from one another, and transmit and receive signals simultaneously or overlapping in time, wherein a respective comparison signal, in particular mixed signals s.sub.1k,mix(t) or s.sub.2k,mix(t) are formed from a signal transmitted and received by the respective transceiver unit, wherein a phase correction is formed for each of a plurality of sample values, preferably a phase correction value for each of a plurality of sample values from the comparison signals s.sub.1k,mix(t) or s.sub.2k,mix(t), in particular in such a way that, preferably by a mathematical operation, a measure is formed of a phase difference per sample value between the at least two signals s.sub.1k,mix(t) or s.sub.2k,mix(t).

Claims

1. A radar method, in particular a primary radar method, in which at least one first and at least one second transceiver unit (S1, S2), which are in particular spatially separated from one another, and transmit and receive signals simultaneously or overlapping in time, wherein a respective comparison signal, in particular mixed signals s.sub.1k,mix(t) or s.sub.2k,mix(t) are formed from a signal transmitted and received by the respective transceiver unit, wherein a phase correction is formed for each of a plurality of sample values, preferably a phase correction value for each of a plurality of sample values from the comparison signals s.sub.1k,mix(t) or s.sub.2k,mix(t), in particular in such a way that, preferably by a mathematical operation, a measure is formed of a phase difference per sample value between the at least two signals s.sub.1k,mix(t) or s.sub.2k,mix(t).

Description

[0048] The invention is also described hereinafter with respect to further aspects or features on the basis of the appended figures. In the figures:

[0049] FIG. 1 shows two transceiver units communicating with one another and (passive) radar targets;

[0050] FIG. 2 shows a further illustration of two transceiver units according to the invention;

[0051] FIG. 3 shows an exemplary illustration of beat signals;

[0052] FIG. 4 shows an illustration of the superposition of beat signals of the transceiver units;

[0053] FIG. 5 shows an illustration of signals according to FIGS. 3 and 4 after a Fourier transform;

[0054] FIG. 6 shows an exemplary illustration of a phase interference of the signals according to FIGS. 3 and 4;

[0055] FIG. 7 shows beat signals after estimation of one phase interference per sample value;

[0056] FIG. 8 shows an illustration of an improved phase noise suppression of the respective signals from FIGS. 3 and 4 including compensation.

[0057] FIG. 1 shows a block diagram having a distributed radar system for environment detection, which comprises two spatially separated radar units (transceiver units). The radar units preferably each at least have a common channel for simultaneously transmitting and receiving signals (TRX1/2), thus in particular an antenna element, which is used for both purposes.

[0058] The separation of transmitted and received signal at the shared antenna is carried out by way of example here via a transmission mixer (M1/2). Alternatively, this can also be carried out via a suitable coupling structure or a circulator.

[0059] Moreover, at least one radar unit can have further channels for transmitting and/or receiving.

[0060] Furthermore, each radar unit preferably has a separate clock source (LO1/2), a separate phase-locked loop for generating a high-frequency signal (PLL1/2) and an analog-to-digital converter (ADC1/2) for sampling the down-mixed signals s.sub.1k,mix(t) and s.sub.2k,mix(t), which are preferably to be further processed advantageously according to the invention.

[0061] The subscript k indicates the number of the FMCW (frequency modulated continuous wave) chirp, wherein, for example, either only rising or falling ramps or alternately both rising and also following ramps can be used as a trip sequence [1]. Three passive radar targets (Z1-3) are shown by way of example, wherein the respective signal runtime can be specified by τ.sub.m and the damping factor by α.sub.m.

[0062] A system for radiolocation having two radar units is illustrated by way of example in FIG. 2, wherein the hardware components can be identical to those in FIG. 1. A direct transmission path exists here, thus a line of sight (LOS) between the two radar units, which generally represents the strongest transmission path and supplies the relevant items of information (for example distance from one another and velocity in relation to one another). In the present case, the radar units are located in the vicinity of a reflective wall, which can be the case, for example, when locating in buildings. Therefore, a second signal path exists, caused by the multipath propagation. An expansion of the radar units, for example, by multiple receiving channels, would also be possible here. This would have the advantage, for example, that the received power becomes greater, that interfering multipath propagation can be suppressed, and/or that in addition azimuth and/or elevation angles can be estimated for 2D or 3D locating.

[0063] Without restriction of the generality, the modeling is carried out for two radar units (transceiver units), where an analogous calculation can take place pair by pair in the case of more than two radar units. Preferably, each station has an antenna, which is used simultaneously for transmitting and receiving.

[0064] The transmission channel is preferably identical in both directions, thus behaves reciprocally.

[0065] The modulation of the transmitted and received signals by means of linear frequency modulation, referred to as FMCW or LFM (linear frequency modulation) is advantageous for this invention. This is advantageous due to the low demand on the hardware components and the simple assignment of signal components (for example separation of the signals from various radar units) in the spectrum.

[0066] In general, a transmission channel (cf. FIGS. 1 and 2), in particular an air transmission link, may be described by the channel pulse response

[00001]$\begin{array}{cc}h\left(t\right)=\underset{m=0}{\overset{M-1}{.Math.}}{a}_m\delta \left(i-{\tau }_m\right)& \left(1\right)\end{array}$

[0067] wherein α.sub.m∈C (complex number range) and T.sub.m∈R (real number range) indicate the complex amplitude and the runtime via the signal path m∈{0, . . . , M−1}. The expression δ(t) denotes the Dirac distribution. It is assumed for simplification that the signal duration of a chip is sufficiently short (for example less than 10 ms, preferably less than 1 ms and/or more than 10 μs, preferably more than 100 μs), thus the amplitude and the runtime are not subject to any relevant change during this time due to movement of radar targets or radar units. Thus, if a signal s.sub.TX(t) is transmitted via this channel, the received signal s.sub.RX(t) results via the convolution

[00002]$\begin{array}{cc}{s}_{\mathrm{RX}}\left(t\right)=\underset{-\infty }{\stackrel{\infty }{\int }}{s}_{\mathrm{TX}}\left(t^{\prime }\right)*h\left(t^{\prime }\right){\mathrm{dt}}^{\prime }=\underset{m=0}{\overset{M-1}{.Math.}}{a}_m{s}_{\mathrm{TX}}\left(t-{\tau }_m\right).& \left(2\right)\end{array}$

[0068] the linear operator * denotes the convolution in one dimension. The substantive matter represented in (2) indicates that the same signal is transmitted via M signal paths and received time-shifted and damped.

[0069] Using the signal model from [1] with (2), after a preliminary correction of the interference variables described herein, a mathematical model for the beat signals after the mixing process may be found. These can be modeled for station 1 and 2 as

[00003]$\begin{array}{cc}s\text{?}\left(t\right)=& \left(3\right)\end{array}$$\underset{\text{?}=0}{\overset{M-\text{?}}{.Math.}}{a}_m\exp \left\{j\left(2\pi f_{m}t+{\phi }_{km}+\Psi \text{?}\left(t\right)-{\Psi }_{2k}\left(t-{\text{?}}_m\right)+\Theta \text{?}-{\Theta }_{2k}\right)\right\}\mathrm{und}$$\begin{array}{cc}s\text{?}\left(t\right)=& \left(4\right)\end{array}$$\underset{\text{?}=0}{\overset{M-\text{?}}{.Math.}}{a}_m\exp \left\{j\left(2\pi f_{m}t+\phi \text{?}-\Psi \text{?}\left(t-{\tau }_m\right)+{\Psi }_{2k}\left(t\right)-\Theta \text{?}+\Theta \text{?}\right)\right\}$$\text{?}\text{indicates text missing or illegible when filed}$

[0070] wherein the occurring variables are described hereinafter. The runtime or the distance to radar targets can be calculated via the beat frequency f.sub.m (in the so-called “fast time”) and the velocity can be calculated from multiple FMCW chirps (chirp sequence radar) via the phase shift φ.sub.km (in the so-called “slow time”). In systems having multiple receiving antennas, it is also possible via this phase value to determine the angle of incidence of an electromagnetic wave. In total, K FMCW chirps are transmitted and the chirp number is indicated by k. The phase noise (also nonlinearities) of station 1 and 2 is represented via the random variables Ψ.sub.1k(t) and Ψ2.sub.k(t) and the unknown starting phases of the local oscillators via Θ.sub.1k and Θ.sub.2k.

[0071] As is known from [1] or [Pat1] or [Pat3], the beat signals of the two stations are influenced approximately equally, or with opposite signs, by interference. The influence of phase noise in the fast time is thus correlated to a higher extent, since the runtime (often less than τ.sub.m=1 μs) is very short. The typically occurring phase change (described over the power density spectrum) can therefore be assumed to be identical in absolute value in both mixed signals. The following applies

Ψ.sub.PN,km(t):=Ψ.sub.1k(t)−Ψ.sub.2k(t−τ.sub.m)≈−(−Ψ.sub.1k(t−τ.sub.m)+Ψ.sub.2k(t))  (5)

[0072] Furthermore, the received random sequence is independent in a good approximation from the distance to the targets, since the runtime difference is small in relation to the duration of a FMCW chirp. According to the prior art, the approximation

exp{jΨ.sub.PN,km(t)}≈1+jΨ.sub.PN,km(t)  (6)

[0073] is used by terminating a Taylor development at the linear element. This can practically be insured, for example, via signal sources having sufficiently low phase noise.

[0074] After the estimation and correction of the unknown starting phases according to [Pat3] the signals in (3) and (4) can be represented with the relationship in (6) as

[00004]$\begin{array}{cc}s\text{?}\left(t\right)=\stackrel{M-1}{\underset{\text{?}}{.Math.}}a\text{?}\exp \left\{j\left(2\pi f\text{?}t+\phi \text{?}\right)\right\}\left(1+j\Psi \text{?}\left(t\right)\right)\mathrm{und}& \left(7\right)\end{array}$$\begin{array}{cc}s\text{?}\left(t\right)=\underset{\text{?}}{\overset{\text{?}}{.Math.}}a\text{?}\exp \left\{j\left(2\pi f\text{?}t+\phi \text{?}\right)\right\}\left(1-j\Psi \text{?}\left(t\right)\right)& \left(8\right)\end{array}$$\text{?}\text{indicates text missing or illegible when filed}$

[0075] The interference component can be compensated for by a linear operation in the timer frequency range (by addition here) and the dynamic range of the useful signals can be enlarged.

[0076] This method is shown by way of example using the signals from FIG. 3 in FIG. 4, by which a phase noise suppression of approximately 15 dB can be achieved. This step can also be applied in multiple dimensions, which is shown in FIG. 5 for the estimation of distance and velocity. The achievable improvement is predominantly dependent on how well the small signal approximation in (6) was maintained.

[0077] For an improvement of the phase noise suppression, first the property from (5) was applied to the time signals in (3) and (4), which resulted in

[00005]$\begin{array}{cc}\begin{array}{c}s\text{?}\left(t\right)=\exp \left\{j\left(\Psi \text{?}\left(\text{?}\right)+{\Theta }_k\right)\right\}\underset{\text{?}}{\overset{\text{?}}{.Math.}}{a}_m\exp \left\{j\left(2\pi f_{m}t+{\phi }_{km}\right)\right\}\\ =\text{}\exp \left\{j\left(\Psi \text{?}\left(t\right)+{\Theta }_k\right)\right\}.Math.s\text{?}\left(t\right)\mathrm{und}\end{array}& \left(9\right)\end{array}$$\begin{array}{cc}\begin{array}{c}s\text{?}\left(t\right)=\exp \left\{-j\left(\Psi \text{?}\left(t\right)+{\Theta }_k\right)\right\}\underset{m=0}{\overset{M-\text{?}}{.Math.}}{a}_m\exp \left\{j\left(2\pi f_{m}t+{\phi }_{km}\right)\right\}\\ =\text{}\exp \left\{-j\left(\Psi \text{?}\left(t\right)+{\Theta }_k\right)\right\}.Math.s_{0,k}\left(t\right)\end{array}& \left(10\right)\end{array}$$\text{?}\text{indicates text missing or illegible when filed}$

[0078] wherein Θ.sub.k:=Θ.sub.1k−Θ.sub.2k is used for a simplified representation. It is apparent from this that all signal paths of the transmission channel are influenced similarly by the interference variables, which are before the sigma sign. The deterministic component of the signals in (9) and (10) was designated for clarity as

[00006]$\begin{array}{cc}{s}_{0,k}\left(t\right)=\underset{m=0}{\overset{M-\text{?}}{.Math.}}{a}_m\exp \left\{j\left(2\pi f_{m}t+{\phi }_{km}\right)\right\}& \left(11\right)\end{array}$$\text{?}\text{indicates text missing or illegible when filed}$

[0079] and is identical in both radar stations.

[0080] According to the invention, at this point there is a direct estimation of the interference variables (i.e., phase errors) per sample value (for each point in time t) for each FMCW chirp from the time signals via the operation

[00007]$\begin{array}{cc}\begin{array}{c}{\gamma }_k\left(t\right)=\arg \left\{s\text{?}\left(t\right).Math.s\text{?}\left(t\right)\right\}/2\\ =\arg \left\{\exp \left\{j2\left({\Psi }_{\mathrm{PN},k}\left(t\right)+{\Theta }_k\right)\right\}{.Math.s\text{?}\left(t\right).Math.}^2\right\}/2\\ ={\Psi }_{\mathrm{PN},k}\left(t\right)+{\Theta }_k,\end{array}& \left(12\right)\end{array}$$\text{?}\text{indicates text missing or illegible when filed}$

[0081] by which a phase correction operation (i.e. one phase correction value per sample value or for each signal point in time t) is formed. The operation arg{.} supplies the argument of a complex number. A higher level gain in relation to white noise is achieved here by |s.sub.0,k(t)|.sup.2. The division by 2 can result in phase jumps in the slow time, wherein a suitable method for correction is explained in [1]. If signal sources having comparatively high phase noise are used, this operation can also be used in the slow time. The result of this estimation is shown by way of example in FIG. 6, wherein the different colors can be assigned to the 64 chirps.

[0082] The correction using the values from (12) can be calculated via the multiplication

s.sub.1k,mix(t)=exp{−γ.sub.k(t)}s.sub.1k,mix(t)=s.sub.0,k(t)=exp{γ.sub.k(t)}s.sub.2k,mix(t)=s.sub.2k,mix(t)  (13)

[0083] wherein the resulting signals of both stations are identical if the estimation can be assumed to be ideal. The result of this method is graphically depicted in FIG. 7 (one-dimensionally) and FIG. 8 (two-dimensionally) and the resulting improvement of the dynamic range is approximately 40 dB.

[0084] In multichannel radars, the occurring interference in all receiving channels is (quasi-)identical, since the runtime difference is very small. By means of (12), the interference can be ascertained, preferably via a channel having shared antenna for transmitting and receiving, and the correction can be applied to the time signals of all channels.

[0085] If alternatively only the unknown starting phase is to be ascertained, preferably a Fourier transform is carried out in the fast time, from which the beat spectra

[00008]$\begin{array}{cc}{S}_{1k,\mathrm{mix}}\left(f\right)=F\left\{s_{1k,\mathrm{mix}}\left(t\right)\right\}=& \left(14\right)\end{array}$$\exp \left\{j{\Theta }_k\right\}.Math.F\left\{\exp \left\{j{\Psi }_{\mathrm{PN},k}\left(t\right)\right\}\right\}*\underset{\text{?}}{\overset{M-\text{?}}{.Math.}}{a}_m\delta \left(f-f_m\right)\exp \left\{j\phi \text{?}\right\}\mathrm{und}$$\begin{array}{cc}{S}_{2k,\mathrm{mix}}\left(f\right)=F\left\{s_{2k,\mathrm{mix}}\left(t\right)\right\}=& \left(15\right)\end{array}$$\exp \left\{-j{\Theta }_k\right\}.Math.F\left\{\exp \left\{-j\Psi \text{?}\left(t\right)\right\}\right\}*\underset{\text{?}}{\overset{M-\text{?}}{.Math.}}{a}_m\delta \left(f-f_m\right)\exp \left\{j\phi \text{?}\right\}$$\text{?}\text{indicates text missing or illegible when filed}$

[0086] result, wherein F {.} represents a Fourier transform.

[0087] Furthermore, window functions can be applied before the Fourier transform. These are not also incorporated here, since the calculation can be carried out independently thereof. The position of the maxima in (14) and (15) is only weakly influenced by phase noise, because of which

[00009]$\begin{array}{cc}{S}_{1k,\mathrm{mix}}\left(f\right)\text{?}\exp \left\{j{\Theta }_k\right\}\underset{\text{?}}{\overset{M-\text{?}}{.Math.}}{a}_m\delta \left(f-f_m\right)\exp \left\{j{\phi }_{km}\right\}\mathrm{und}& \left(16\right)\end{array}$$\begin{array}{cc}{S}_{2k,\mathrm{mix}}\left(f\right)=\exp \left\{-j{\Theta }_k\right\}\underset{\text{?}=0}{\overset{M-\text{?}}{.Math.}}{a}_m\delta \left(f-f_m\right)\exp \left\{j{\phi }_{km}\right\}& \left(17\right)\end{array}$$\text{?}\text{indicates text missing or illegible when filed}$

[0088] approximately applies. Instead of the multiplication in the time range, the starting phases Σ.sub.k can also be determined via a multiplication of the existing signals in the frequency range, if no correction is to be carried out in the fast time.

[0089] The above-described method can be used for a high-accuracy locating system (having an accuracy in the range of 10 mm or better; or 1 mm or better), which profits from the improved phase noise suppression due to the high signal level. Furthermore, a displacement or the accurate distance can be produced by measuring the signal phase with significantly higher accuracy. In addition, the accuracy of a possible velocity estimation is improved. An expansion of the radar units by multiple receiving channels according to [2] is supported by the above-described method for multichannel systems for radiolocation.

[0090] This method can also be used for distributed MIMO (multiple-input multiple-output) radars for detecting passive radar targets (for example in the automotive field or for airspace monitoring). These systems can be operated coherently, which supplies additional items of information, such as bistatic or multistatic velocity, different observation perspectives, a vectorial velocity estimation, an item of height information by means of interferometry, or the improvement of resolution and accuracy of an angle measurement. Furthermore, the quality of an image of the environment by means of synthetic aperture radar (SAR) can be improved by an observation from multiple perspectives. With the present improved estimation of the phase interference, this can take place independently of the observed scenario and a suppression of this phase interference can be applied to all channels.

[0091] Typically, distributed SAR systems for remote sensing, for example of the Earth's surface, use at least two satellites [3]. At least one satellite emits signals which are reflected at passive radar targets and are received by the same satellite and at least one further satellite. Due to the separate transmitter and receiver, high-accuracy clock sources are preferably used in this application. In addition, a regular phase adjustment preferably takes place via a dedicated synchronization signal, which is transmitted between the satellites.

[0092] The present invention can advantageously be applied in such a way that a FMCW signal is simultaneously both transmitted between the satellites and used to “illuminate” a corresponding region (for example a defined region of the Earth's surface). This can be carried out, for example, via an antenna having suitable directional characteristic or preferably in each case via two antennas having suitable coupling structure connected upstream. The separation of the respective signals by suitable offsets in time or frequency is described in [1]. Due to the greatly differing runtime, a separation of the signals, which were transmitted directly and reflected from objects, is possible in the beat spectrum. Correction values can be ascertained by means of the phase curve of the directly transmitted signal. Phase interference of the signal can be corrected using these correction values, which is used for bistatic or multistatic Earth observation.

[0093] Summary of independent and refining aspects of the invention: [0094] 1) A method in which at least two spatially separated transceiver units transmit and receive and possibly mix signals simultaneously, wherein in each of the at least two transceiver units, a respective signal, s.sub.1k,mix(t) in the transceiver unit 1 or s.sub.2k,mix(t) in the transceiver unit 2, etc., is formed, characterized in that one phase correction value function y.sub.k(t) per sample value (for each point in time t) is formed from the time signals s.sub.1k,mix(t) or s.sub.2k,mix(t) in such a way that a measure of the phase difference per sample value between the at least two signals s.sub.1k,mix(t) and s.sub.2k,mix(t) is formed by a mathematical operation. [0095] 2) The method according to aspect 1)
wherein to form this phase correction value function y.sub.k(t), preferably either a multiplication of the one signal with the conjugated complex of the respective other signal is carried out or the two signals are represented according to absolute value and phase and the phases of the two signals are subtracted from one another for each sample value or for each point in time t. [0096] 3) The method according to 1), in which the phase correction value function y.sub.k(t) is used to correct at least one of the sampled time signals s.sub.1k,mix(t), s.sub.2k,mix(t), etc. preferably in such a way that an expression is formed from the phase correction value function y.sub.k(t), preferably a function of complex numbers having constant amplitude and the phase y.sub.k(t), wherein this expression is then multiplied by the time signal s.sub.1k,mix(t), s.sub.2k,mix(t), etc. and thus one or more phase-corrected signals s.sub.1k,mix(t), s.sub.2k,mix(t), etc. are formed (cf. equation 13)). [0097] 4) The method and device according to 1) using linear frequency-modulated (FMCW/LFM) transmitted and received signals. [0098] 5) The device according to 1) for radiolocation using at least two distributed radar units, which preferably have visual contact with one another, for precise locating, thus the measurement of distances and/or displacement and/or velocities. [0099] 6) The device according to 1) for environment detection using at least two distributed radar units, which are used for detecting objects in the environment, in particular in road traffic or in airspace monitoring. [0100] 7) The device according to 1) having at least two distributed radar units for remote sensing, in particular for earth remote sensing, which are preferably attached to separate satellite and/or airplanes. [0101] 8) The device and method according to 1) for distributed radar units having an antenna array, consisting of multiple receiving antennas, in which the phase error is estimated using the baseband signals of at least one receiving channel and is compensated for in at least one further receiving channel. [0102] 9) The device according to 1) which uses one antenna element for the common transmitting and receiving in at least one radar unit. [0103] 10) The device according to 1) which is used to reduce the demand on the signal generation, for example, by means of phase-locked loop (PLL) or direct digital synthesis (DDS).

[0104] Cited patent literature:

[0105] [Pat1] WO 2010/019975 A1.

[0106] [Pat2] DE 10 2014 104 273 A1.

[0107] [Pat3] WO 2017/118621 A1.

[0108] [Pat4] WO 2018/158281 A1.

[0109] [Pat5] WO 2018/158173 A1.

FURTHER CITED LITERATURE

[0110] [1] M. Gottinger, F. Kirsch, P. Gulden, and M. Vossiek, “Coherent Full-Duplex Double-Sided Two-Way Ranging and Velocity Measurement Between Separate Incoherent Radio Units,” IEEE Trans. Microw. Theory Tech., vol. 67, no. 5, pp. 2045-2061, 2019. [0111] [2] R. Feger, C. Pfeffer, W. Scheiblhofer, C. M. Schmid, M. J. Lang, and A. Stelzer, “A 77-GHz Cooperative Radar System Based on Multi-Channel FMCW Stations for Local Positioning Applications,” IEEE Trans. Microw. Theory Tech., vol. 61, no. 1, pp. 676-684, 2013. [0112] [3] G. Krieger et al., “TanDEM-X: A Satellite Formation for High-Resolution Radar Interferometry,” IEEE Trans. Geosci. Remote Sens., vol. 45, no. 11, pp. 3317-3341, 2007.