DE102018010369A1 - MIMO FMCW RADAR SYSTEM - Google Patents

Abstract

A method for the detection of radar targets is described here. According to one exemplary embodiment, the method comprises transmitting a first partial sequence of a frame with chronologically successive chirps via a first channel of a radar system and transmitting one or more further partial sequences of the frame via different channels, the further partial sequences following the first partial sequence and one each have fewer chirps than the first partial sequence. The method further comprises the detection of one or more radar targets and the determination of a distance value and a speed value for each of the detected radar targets based on only the first partial sequence, and the calculation of an angle for each of the detected radar targets based on the other partial sequences.

Description

TECHNICAL AREA

The present description relates to the field of radar sensors, in particular signal processing methods used in MIMO radar systems for the detection of objects.

BACKGROUND

Radar sensors are used in a large number of applications for detecting objects (the so-called radar targets), the detection usually comprising measuring the distances and speeds of the detected objects. In the automotive sector in particular, there is an increasing need for radar sensors that are used, among other things, in advanced driver assistance systems (ADAS) such as in adaptive cruise control (ACC, Adaptive Cruise Control, or Radar Cruise Control) systems can be used. Such systems can automatically adjust the speed of an automobile to maintain a safe distance from other vehicles in front (as well as from other objects and pedestrians). Further applications in the automotive sector are e.g. Blind spot detection, lane change assist and the like. In the field of autonomous driving, radar sensors will play an important role in the control of autonomous vehicles.

A precise detection of objects in the measuring range of a radar sensor usually requires a radar system with several transmission and reception channels. Radar systems of this type are usually referred to as MIMO (multiple input multiple output) radar systems. MIMO systems are particularly required if not only the distance and the speed of an object, but also its azimuth angle are to be determined. Signal processing techniques, which are generally referred to as beamforming techniques, can be used to determine the azimuth angle assigned to a radar target. The publication DE 102014212284 A1 describes a MIMO radar measurement method in which a modulation pattern of the radar transmission signal for different transmission switching states, which differ in the selection of the antenna elements used for transmission, comprises sequences of ramps nested in one another in time. Ambiguous values for the relative speed of the radar target are determined from a position of a peak in a two-dimensional spectrum, and the phase relationships between spectral values of spectra are checked for agreement with expected phase relationships (for a determined relative speed) in order to resolve the ambiguity.

In MIMO radar systems, the problem arises of separating the signals transmitted by different transmission channels in the reception channels. There are various approaches to this, e.g. Different variants of time division multiplex (TDM) or frequency division multiplex (FDM) methods. The object on which the invention is based can be seen in improving and developing existing methods.

SUMMARY

The stated object is achieved by the method according to claim 1 and the system according to claim 12. Various embodiments and further developments are the subject of the dependent claims.

A method for the detection of radar targets is described below. According to one exemplary embodiment, the radar method comprises the transmission of a first partial sequence of chronologically successive chirps via a first channel of a radar system and the transmission of one or more further partial sequences of chronologically successive chirps via different channels, the further partial sequences chronologically following the first sequence and each have a smaller number of chirps than the first partial sequence. The method further comprises the detection of one or more radar targets and the determination of a distance value and a speed value for each of the detected radar targets based on only the first partial sequence, and the calculation of an angle for each of the detected radar targets based on the other partial sequences.

A radar system is also described. According to one exemplary embodiment, the radar system comprises a radar transceiver which is designed to transmit a first partial sequence of chronologically successive chirps via a first channel of the radar system and further one or to transmit a plurality of further partial sequences of chronologically successive chirps via different channels, the further partial sequences following the first sequence in time and each having a smaller number of chirps than the first partial sequence. The system further comprises a computing unit which is designed to detect one or more radar targets and in each case to determine a distance value and a speed value for which the detected radar targets are based on only the first partial sequence, and which is further designed to to determine an angle for the detected radar targets based on the further partial sequences.

Figure list

• 1 ist eine Skizze zur Illustration des Funktionsprinzips eines FMCW-Radarsystems zur Abstands- und/oder Geschwindigkeitsmessung.
• 2 umfasst zwei Zeitdiagramme zur Illustration der Frequenzmodulation (FM) eines vom FMCW-System erzeugen HF-Signals.
• 3 ist ein Blockdiagramm zur Illustration der grundlegenden Struktur eines FMCW-Radarsystems.
• 4 ist ein Schaltbild zur Illustration eines vereinfachten Beispiels eines Radar-Transceivers sowie eines weiteren Radar-Transceiver, der Interferenzen verursacht.
• 5 illustriert anhand eines Blockschaltbilds ein Beispiel eines HF-Frontends eines MIMO-Radarsystems.
• 6 illustriert exemplarisch die digitale Signalverarbeitung von Radarsignalen bei der bei der Detektion von Radar-Targets im Range-Geschwindigkeits-Raum (Range-Doppler-Analyse).
• 7 illustriert in einem Timing-Diagramm den Betrieb von mehreren Sendekanälen eines MIMO-Radarsystems mittels Time-Division-Mulitplexing.
• 8 illustriert in einem Timing-Diagramm den Betrieb von mehreren Sendekanälen eines MIMO-Radarsystems zur Bestimmung von Abstand (Range), Geschwindigkeit und Azimutwinkel.
• 9 illustriert exemplarisch die digitale Signalverarbeitung von Radarsignalen bei der Detektion von Radar-Targets im Range-Winkel-Raum.
• 10 illustriert die zum Beispiel aus 8 gehörenden Range-Doppler-Map (Abstands-Geschwindigkeits-Karte) und Range-Azimut-Map (Abstands-Winkel-Karte).
• 11 illustriert schematisch ein Beispiel eines Signalerarbeitungstechnik, welche in MIMO-Systemen dazu verwendet werden kann, zu detektierten Radar-Targets den zugehörigen Azimutwinkel zu bestimmen.
• 12 und 13 illustrieren exemplarisch verschiedene Modifikationen des Beispiels aus 11.
• 14 und 15 visualisieren exemplarisch die Ergebnisse der in 11 und 13 dargestellten Techniken.

Exemplary embodiments are explained in more detail below with the aid of illustrations. The illustrations are not necessarily to scale and the exemplary embodiments are not limited to the aspects shown. Rather, it is important to present the principles on which the exemplary embodiments are based. The illustrations show:

• 1 is a sketch to illustrate the functional principle of an FMCW radar system for distance and / or speed measurement.
• 2nd comprises two time diagrams to illustrate the frequency modulation (FM) of an RF signal generated by the FMCW system.
• 3rd Fig. 3 is a block diagram illustrating the basic structure of an FMCW radar system.
• 4th FIG. 12 is a circuit diagram illustrating a simplified example of a radar transceiver and another radar transceiver that causes interference.
• 5 uses a block diagram to illustrate an example of an RF front end of a MIMO radar system.
• 6 exemplifies the digital signal processing of radar signals in the detection of radar targets in the range-speed space (range Doppler analysis).
• 7 illustrates in a timing diagram the operation of several transmission channels of a MIMO radar system by means of time division multiplexing.
• 8th illustrates in a timing diagram the operation of several transmission channels of a MIMO radar system for determining distance (range), speed and azimuth angle.
• 9 exemplifies the digital signal processing of radar signals in the detection of radar targets in the range-angle-space.
• 10th illustrates that from, for example 8th belonging range Doppler map (distance-speed map) and range-azimuth map (distance-angle map).
• 11 schematically illustrates an example of a signal processing technique that can be used in MIMO systems to determine the associated azimuth angle for detected radar targets.
• 12th and 13 illustrate various modifications of the example 11 .
• 14 and 15 visualize the results of the in 11 and 13 techniques shown.

DETAILED DESCRIPTION

1 illustrates in a schematic diagram the use of a frequency-modulated continuous-wave radar system (Freqeuncy-Modulated Continuous-Wave Radar System) - mostly referred to as FMCW radar system - as a sensor for the detection of objects, which are usually referred to as radar targets. The detection usually includes the determination of the distance (range) and speed of the respective radar target. In order to localize the target even more precisely, modern radar systems are designed to also determine the azimuth angle of a target.

In the present example, the radar device points 1 separate transmit (TX) and receive (RX) antennas 5 respectively. 6 on (bistatic or pseudo-monostatic radar configuration). However, it should be noted that a single antenna can also be used, which serves simultaneously as a transmitting antenna and a receiving antenna (monostatic radar configuration). The transmitting antenna 5 emits a continuous RF signal s _RF (t) which, for example, uses a type of sawtooth signal (periodic, linear frequency ramp) is frequency modulated. The radiated RF radar signal s _RF (t) is on the radar target T backscattered and the backscattered / reflected signal y _RF (t) (Echo signal) is from the receiving antenna 6 receive. 1 shows a simplified example; In practice, radar sensors are systems with multiple transmit (TX) and receive (RX) channels, including the direction of arrival (DoA) of the backscattered / reflected signal y _RF (t) , ie the mentioned azimuth angle of the radar target T to be able to determine.

2nd illustrates the frequency modulation of the signal as an example s _RF (t) . As in 2nd shown (upper diagram) is the emitted RF signal s _RF (t) composed of a set of "chirps", ie the signal s _RF (t) comprises a sequence of sinusoidal waveforms with increasing frequency (up-chirp) or falling frequency (down-chirp). In the present example, the instantaneous frequency increases f _LO (t) of a chirp starting at a start frequency f _START within a period of time T _CHIRP linear to a stop frequency f _STOP (see lower diagram in 2nd ). Such chirps are also known as linear frequency ramps. In 2nd three identical linear frequency ramps are shown. However, it should be noted that the parameters f _START , f _STOP , T _CHIRP and the pause between the individual frequency ramps (and thus the chirp repetition rate) can vary. The frequency variation does not necessarily have to be linear (linear chirp). Depending on the implementation, transmission signals with exponential or hyperbolic frequency variation (exponential or hyperbolic chirps) can also be used, for example.

3rd FIG. 4 is a block diagram exemplarily showing a possible structure of a radar device 1 (Radar sensor). Accordingly, one or more transmit antennas 5 (TX antennas) and one or more receiving antennas 6 (RX antennas) with an RF front end integrated in a chip 10th connected, which can include all those circuit components that are required for the RF signal processing. These circuit components include, for example, a local oscillator (LO), RF power amplifier, low-noise amplifier (LNA, low-noise amplifier), directional coupler (eg rat-race coupler, circulators, etc.) and mixer for downconversion of the HF Signals in the baseband or an intermediate frequency band (IF band). The HF front end 10th can be integrated in one or more chips, possibly together with other circuit components, which are usually referred to as monolithically microwave integrated circuits (MMICs).

The example shown shows a bistatic (or pseudo-monostatic) radar system with separate RX and TX antennas. In the case of a monostatic radar system, the same antenna would be used both for radiating and for receiving the electromagnetic (radar) signals. In this case, a directional coupler (for example a circulator) can be used to separate the RF signals to be radiated from the received RF signals (radar echo signals). As mentioned, in practice radar systems usually have several transmission and reception channels with several transmission or reception antennas, which among other things enables a measurement of the direction (also referred to as Direction of Arrival, DoA) from which the radar echoes are received. In such MIMO systems, the individual (physical) TX channels and RX channels are usually each constructed identically or similarly. So-called virtual antenna arrays can be formed from several TX antennas and RX antennas, which can be used for the implementation of beamforming techniques. For example, a system with three TX antennas and four RX channels can be used to create an antenna array 12th to implement (three times four) virtual antenna elements.

In the case of an FMCW radar system, the via the TX antenna (s) can 5 radiated RF signals are, for example, in the range from approx. 20 GHz to 100 GHz (eg approx. 77 GHz in some applications). As mentioned, this includes every RX antenna 6 received RF signal radar echoes (chirp echo signals), ie those signal components that are scattered back on one or more radar targets. The received RF signal is in each RX channel y _RF (t) mixed down into the baseband and can be further processed in the baseband using analog signal processing (see 3rd , analog baseband signal processing chain 20th ). The aforementioned analog signal processing can essentially include filtering and possibly amplification of the baseband signal. The baseband signal is finally digitized (see 3rd , Analog-digital converter 30th ) and processed in the digital area. The digital signal processing chain can at least partially be implemented as software (for example firmware) which is installed on one or more processors, for example a microcontroller and / or a digital signal processor (see 3rd , Arithmetic unit 40 ) can be executed.

The entire system is usually controlled by a system controller 50 controlled, which can also be implemented at least partially by means of firmware or software that is executed on a processor such as a microcontroller. The HF front end 10th and the analog baseband Signal processing chain 20th (optionally also the analog-digital converter 30th and the arithmetic unit 40 ) can be integrated together in a single MMIC (ie an RF semiconductor chip). Alternatively, the individual components can also be distributed over several integrated circuits (MMICs). Distributed radar systems with several coupled MMICs are known per se and are not discussed further here. To simplify the illustration, and since it is not necessary for the further explanations, no distinction is made in the following between "single-chip" radar systems and distributed radar systems.

4th illustrates an exemplary implementation of a radar transceiver 1 according to the example 3rd more detailed. In the present example there is in particular the HF front end 10th of the radar sensor 1 (see. 3rd ). It should be noted that 4th represents a simplified circuit diagram to the basic structure of the RF front end 10th with a transmit channel (TX channel) and a receive channel (RX channel). Actual implementations, which can depend heavily on the specific application, are usually more complex and have multiple TX and / or RX channels. Such systems with several TX and RX channels can be operated as MIMO systems.

The HF front end 10th includes a local oscillator 101 (LO) which is an RF oscillator signal s _LO (t) generated. The RF oscillator signal s _LO (t) is in operation as above with reference to 2nd described, frequency modulated and is also referred to as an LO signal. In radar applications, the LO signal is usually in the SHF (Super High Frequency, centimeter wave) or EHF (Extremely High Frequency, millimeter wave) band, for example in the interval from 76 GHz to 81 GHz in some automotive applications. The LO signal s _LO (t) is both in the transmit signal path TX1 (in the TX channel) as well as in the receive signal path RX1 (in the RX channel) processed.

The RF transmit signal s _RF (t) (see. 2nd ) from the TX antenna 5 is emitted by amplifying the LO signal s _LO (t) , for example by means of the RF power amplifier 102 , is generated and is therefore only an amplified and possibly phase-shifted version of the LO signal s _LO (t) . The output of the amplifier 102 can with the TX antenna 5 be coupled (in the case of a bistatic or pseudo-monostatic radar configuration). The RF receive signal y _RF (t) which is from the RX antenna 6 is received, the receiver circuit in the RX channel and thus directly or indirectly the RF port of the mixer 104 fed. In the present example, the RF receive signal y _RF (t) (Antenna signal) by means of the amplifier 103 (Reinforcement g) preamplified. The mixer 104 therefore receives the amplified RF reception signal g · y _RF (t). The amplifier 103 can be an LNA, for example. The mixer's reference port 104 is the LO signal s _LO (t) fed so that the mixer 104 the (pre-amplified) RF reception signal y _RF (t) mixed down into the baseband.

The down-mixed baseband signal (mixer output signal) is subsequently used y _BB (t) designated. This baseband signal y _BB (t) is first processed analog, with the analog baseband signal processing chain 20th essentially a further amplification and (eg bandpass or lowpass) filtering is effected in order to suppress unwanted sidebands and image frequencies. The resulting analog output signal of the receive channel RX1 , which is an analog-digital converter (see 3rd , ADC 30th ) is fed with y (t) designated. Process for the digital further processing of the digitized baseband output signal (digital radar signal y [n] ) are known per se (for example the range Doppler analysis) and are therefore not discussed further here.

In the present example, the mixer mixes 104 the pre-amplified RF reception signal g · y _RF (t) (ie the amplified antenna signal) down to baseband. Mixing can take place in one stage (i.e. from the HF band directly into the baseband) or via one or more intermediate stages (i.e. from the HF band into an intermediate frequency band and further into the baseband). In this case, the receiving mixer includes 104 effectively several individual mixer stages connected in series. The mixer 104 can be implemented in different ways. In some embodiments, can be used as a mixer 104 an IQ demodulator (in-phase / quadrature-phase demodulator) is used, with the result that the (digital) baseband signal y [n] is a complex signal. In other embodiments, normal mixers are used which have a real baseband signal y [n] deliver. The concepts described here can be applied to both variants.

5 schematically illustrates an RF front end 10th of a MIMO system with three TX antennas 5 and two RX antennas 6 . The HF front end 10th includes three TX channels in the example shown TX1 , TX2 and TX3 as well as two RX channels RX1 and RX2 . The TX antennas 5 are with the exits TA1 , TA2 and TA3 of the TX channels TX1 , TX2 respectively. TX3 connected, and the RX antennas 6 are with the exits RA1 and RA2 of the RX channels RX1 respectively. RX2 connected. The RX and TX channels can be constructed essentially the same as the RX channel RX1 and the TX channel TX1 in the example 4th . The baseband output signals of the RX channels RX1 and RX2 are in the present example with y ₁ (t) and y ₂ (t) designated. The baseband output signal is shown below y (t) or the associated digital radar signal y [n] (without subscript) representative of that of any RX channel.

6 uses an example to illustrate the signal processing in a radar system up to the digitization of the baseband signal y (t) , which represents the chirp echo signals, and the subsequent digital processing of the digital radar signal y [n] . Diagram (a) 6 shows part of a chirp sequence that M includes linear chirps consecutive in time. The solid line represents the waveform (waveform, frequency over time) of the outgoing RF radar signal s _RF (t) and the dashed line represents the corresponding signal course of the incoming RF radar signal y _RF (t) that includes the chirp echoes (if any). According to diagram (a) of 6 the frequency of the outgoing radar signal increases starting at a start frequency f _START linear up to a stop frequency f _STOP on (chirp no. 0 ) and then falls back to the start frequency f _START , increases again to the stop frequency f _STOP on (chirp no. 1 ), and so on.

As mentioned, a chirp sequence comprises a plurality of chirps; in the present case the number of chirps in a sequence is included M designated. Which is a number of M Chirp's chirp sequence can also be referred to as a frame. Typically, an evaluation is carried out after each frame sent in order to determine the distance, speed and direction (angle which represents the DoA) of the targets, as will be explained below. Depending on the application, a sequence can also contain chirps with different parameters (start and stop frequency, duration and modulation pause). During a modulation pause between two successive chirps, the frequency can be, for example, the stop frequency of the previous chirp or the start frequency of the following chirp (or another frequency). The chirp duration T _CHIRP can range from a few microseconds to a few milliseconds, for example in the range from 20 microseconds to 2 ms. The actual values can be larger or smaller depending on the application. The number M the chirps in a sequence can correspond to a power of two, eg M = 256.

The incoming RF radar signal (ie received by an RX antenna) y _RF (t) rushes to the outgoing (ie emitted by a TX antenna) RF radar signal s _RF (t) by a time difference Δt to. This time difference Δt corresponds to the signal transit time from the TX antenna to the radar target and back to the RX antenna and is also known as the Round Trip Delay Time (RTDT). The distance d _Ti of a radar target T _i from the radar sensor is d _T = c · Δt / 2, ie the speed of light c times half the time difference Δt .

How to get in diagram (a) 6 can see has the time difference Δt a corresponding frequency difference Δf result. This frequency difference Δf can be determined by the incoming (and possibly preamplified) radar signal y _RF (t) with the LO signal s _LO (t) of the radar sensor mixed (see 5 , Mixer 104 ), the resulting baseband signal y (t) digitized and then a digital spectral analysis is carried out. The frequency difference Δf then appears in the spectrum of the digitized baseband signal y [n] as the beat frequency. If linear chirps are used, the time difference can Δt can be calculated according to Δt = Δf / k, using the factor k denotes the steepness (Hertz per second) of the frequency _ramp , which can be calculated according to k = B / T _CHIRP , where B the bandwidth of a chirp is (B = | f _STOP -f _START |). With regard to the above explanations follows for the distance sought d _Ti of the Targtes T _i : $${\text{d}}_{\text{Ti}}=\text{c}\cdot \Delta \text{t}/\mathrm{2nd}=\text{c}\cdot \Delta \text{f}\cdot {\text{T}}_{\text{CHIRP}}/\left(\mathrm{2nd}\cdot \text{B}\right),$$

Although only the basic operating principle of an FMCW radar sensor has been summarized above, it should be noted that in practice more sophisticated signal processing is usually used. For example, an additional Doppler shift f _D of the incoming signal due to the Doppler effect affect the distance measurement, since the Doppler shift f _D to the frequency difference explained above Δf added. Depending on the application, the Doppler shift can be estimated / calculated from the outgoing and incoming radar signals and taken into account in the measurement, whereas in some applications the Doppler shift can be negligible for the distance measurement. This can be the case, for example, if the chirp duration T _CHIRP and the speed of the target are small, so frequency difference Δf compared to the Doppler shift f _D is great. In some radar systems, the Doppler shift can be eliminated by determining the distance based on an up-chirp and a down-chirp in the distance measurement. Theoretically, the actual distance d _T are calculated as the average of the distance values obtained from a measurement with up-chirps and a further measurement with down-chirps. The Doppler shift is eliminated by the averaging.

An example of a signal processing technique for processing FMCW radar signals includes the calculation of so-called range Doppler maps, which are also referred to as range Doppler images. In general, FMCW radar sensors determine the target information (ie distance, speed, DoA / azimuth angle) by emitting a sequence of chirps (see 6 , Diagram (a)) and mixing the (delayed) echoes from the radar targets with a “copy” of the transmitted signal (cf. 4th , Mixer 104 ). The resulting baseband signal y (t) is the diagram in diagram (b) 6 shown. This baseband signal y (t) can be divided into several segments, each segment of the baseband signal y (t) is assigned to a particular chirp of the chirp sequence.

The mentioned target information can be from the spectrum of the mentioned segments of the baseband signal y (t) are extracted, which contain the chirp echoes generated by one or more radar targets (in the form of the mentioned beat frequency). A range-Doppler map is obtained, as explained in more detail below, for example by means of a two-stage Fourier transformation. Range Doppler maps can be used as the basis for various methods for the detection, identification and classification of radar targets. The result of the first Fourier transformation stage is referred to as a range map.

In the examples shown here, the calculations necessary for determining the range Doppler maps are carried out by a digital computing unit (computing unit), which, for example, has one or more signal processors (cf. 4th , DSP 40 ) includes. In other exemplary embodiments, in addition or as an alternative to a signal processor, other computing units can also be used to carry out the necessary calculations. Depending on the implementation, the calculations can be carried out by different firmware / software and hardware entities (firmware / software and hardware entities) or combinations thereof. In general, the term computing unit is understood here to mean any combination of firmware / software and hardware which is capable and designed to carry out the calculations which are described in connection with the exemplary embodiments explained here.

According to an exemplary implementation, the calculation of a range Doppler map comprises two stages, with several Fourier transformations being calculated in each stage (for example using an FFT algorithm). According to the present example, the baseband signal y (t) (see. 5 ) scanned in such a way that for a (partial) sequence with M Segments N × M samples, ie M Segments with each N Samples. That is, the sampling time interval T _SAMPLE is chosen so that each of the M Segments (chirp echoes in baseband) by a sequence of N Samples is represented. As in diagram (c) of 6 can represent these M Segments each N Samples in a two-dimensional field (array) Y [n, m] be arranged (radar data field). Each column of the field Y [n, m] represents one of the M considered segments of the baseband signal y (t) , and the nth line of the field Y [n, m] contains the nth sample of the M Chirps. The line index n (n = 0, 1, ... N-1) can thus be viewed as a discrete point in time n · T _SAMPLE on a “fast” time axis. Similarly, the column index m (m = 0, 1, ... M-1) can be viewed as a discrete point in time m · T _CHIRP on a “slow” time axis. The column index m corresponds to the number of the chirp in a chirp sequence or the number of the associated segment of the (digitized) baseband signal.

In a first stage there is a first FFT (usually called Range FFT) applied to each chirp. For each column in the field Y [n, m] the Fourier transform is calculated. In other words, the field Y [n, m] is Fourier transformed along the fast time axis, and the result is a two-dimensional field R [k, m] of spectra called a range map, each of the M Columns of the range map in each case N contains (complex valued) spectral values. The "fast" time axis becomes the frequency axis through the Fourier transformation; the row index k the range card R [k, m] corresponds to a discrete frequency and is therefore also called a frequency bin k designated (k = 0, 1, ...., N-1). Each discrete frequency corresponds to a distance according to equation (1), which is why the frequency axis is also referred to as the distance axis (range axis).

The range card R [k, m] is the in diagram (c) 6 shown. A radar echo caused by a radar target results in a local maximum (peak) at a specific frequency index / frequency bin. This local maximum usually appears in all columns of the range map R [k, m] , ie in the spectra of all considered segments of the baseband signal y (t) that can be assigned to a chirp sequence. As mentioned, the associated frequency index k (eg according to equation (1)) can be converted into a distance value.

In a second stage, a second FFT (usually referred to as Doppler FFT) on each of the N Rows of the range map R [k, m] applied (k = 0, ..., N-1). Each line of the range card R [k, m] includes M Spectral values of a specific frequency bin k , where each frequency bin k a certain distance d _Ti of a specific radar target T _i corresponds (ie distance / range and frequency represent the same information). The Fourier transform of the spectral values in a specific frequency bin (assignable to a radar target) k enables the determination of the associated Doppler frequency fb that corresponds to a speed of the radar target. In other words, the two-dimensional field R [k, m] (the range map) is Fourier-transformed line by line, ie along the "slow" time axis. The resulting Fourier transforms again form a field with N × M spectral values, which is called the range Doppler time X [k, 1] is designated (k = 0, ..., N-1 and l = 0, ..., M-1). By the second FFT the "slow" time axis becomes the Doppler frequency axis on which the (Doppler) frequency bins with the index 1 be referenced. The associated discrete Doppler frequency values each correspond to a certain speed. The Doppler frequency axis can therefore be converted into a speed axis (ie speed and Doppler frequency represent the same information).

Every local maximum (every peak) in the Rage-Doppler map X [k, l] indicates a potential radar target. The row index associated with a local maximum k (on the range axis) represents the distance of the target and the column index assigned to the local maximum 1 (on the speed axis) represents the associated speed of the target. It is understood that the methodology described here for determining the distance and speed of radar targets based on the range Doppler map is a comparatively simple methodology. Depending on the application, more complex and advanced algorithms can be used to evaluate range Doppler maps. In some applications, the calculation of a range Doppler map is not necessary. In these cases, distances and speeds of radar targets can be calculated, for example, based on the range map without the range Doppler map having to be calculated beforehand.

In the case of multiple RX antennas, a range map and a range Doppler map can be used Xa [k, l] are calculated for each RX channel, where a denotes the number of the antenna and the associated RX channel. The Range Doppler cards Xa [k, l] can be "stacked" into a three-dimensional field. Similarly, the output data Ya [m, n] (Radar data fields) can be viewed as a three-dimensional field. This is sometimes referred to as a radar data cube. If there are several TX channels, several Range Doppler maps can be calculated for each RX channel, namely one Range Doppler map for each TX channel. That means you get a range Doppler map for each RX / TX channel pair. As mentioned above, virtual antenna elements of an antenna array can be formed by combining RX and TX channels. In this case, the index a represents the number of the virtual antenna, and a range Doppler map can be made for each virtual antenna Xa [k, l] be determined.

As mentioned, the radar data cubes can be the resulting range maps Ra [k, m] or the Range Doppler cards Xa [k, l] can be used as input data for various other signal processing methods. For example, various peak detection algorithms are known to be in the range maps Ra [n, m] or the Range Doppler cards Xa [k, l] to detect local maxima (peaks), each of which is caused by a radar target in the “field of view” of the radar sensor. Other algorithms are used, for example, to calculate the (azimuth) angle of a radar target (cf. 10th ) or the classification of detected radar targets (e.g. whether a radar target is a pedestrian).

The following is a MIMO radar system with three TX channels TX1 , TX2 , TX3 and considered two or more RX channels. When performing a measurement, a sequence with a certain number of chirps is emitted, as already mentioned above. In MIMO systems with several TX channels, there are various possible approaches for distributing the chirps of a sequence over the different TX channels. 7 exemplifies an approach commonly referred to as time division multiplexing (TDM). The active TX channel is changed after each chirp, so that the first chirp is via the first TX channel TX1 , the second chirp over the second TX channel TX2 , the third chirp over the third TX channel TX3 , the fourth chirp again over the first TX channel TX1 , etc. are emitted. This scheme when sending the chirps over the three TX channels TX1 , TX2 and TX3 has the consequence that the period duration in each TX channel T _REP after which a chirp is repeated is at least three times as long as the chirp duration T _CHIRP . As the number of TX channels increases, the ratio changes T _CHIRP / T _REP worse and the chirp repetition rate 1 / T _REP accordingly smaller.

In each receive channel, each of the three transmit channels can TX1 , TX2 , TX3 a range Doppler map can be calculated, due to the low chirp repetition rate 1 / T _REP for the respective channel maximum clearly detectable speed of a radar target is reduced. With only a single TX channel, T _REP = T _CHIRP and the maximum uniquely detectable speed would be correspondingly greater. With the TDM approach there is consequently a tradeoff between the number of (virtual) antennas and the maximum resolvable speed, whereby a larger number of antennas means a better angular resolution when determining the azimuth angle. At this point it should be noted that the maximum uniquely detectable speed is not necessarily the upper end of the speed measuring range. At higher speeds, aliasing effects occur in the range Doppler map (whereby the uniqueness is lost); However, by tracking an accelerating target, for example, a speed can also be measured (and assigned to the target) that is higher than the maximum uniquely detectable speed.

8th illustrates a further approach which the exemplary embodiments described here also use and in which - within a frame - first a first partial sequence of chirps directly following one another via the first TX channel TX1 is emitted. Subsequently, a chirp via the second TX channel follows immediately after the first partial sequence TX2 and a chirp over the third TX channel TX3 emitted. In this case it is related to the first TX channel TX01 the period T _REP equal to the chirp duration T _CHIRP , and a range-doppler map can be calculated with a higher maximum speed than in the previous example. Based on the last chirp of the first part sequence and the one over the TX channels TX2 and TX3 radiated chirps (ie from the associated digital radar signals y [n] a range angle map (range angle map) can be calculated, which can be used to determine the azimuth angle. 9 illustrates the result of the calculations mentioned in a schematic example. The calculation of a range angle map is then made with reference to 10th discussed.

As will be explained further below, the TX channels TX2 and TX3 more than one chirp can also be sent within a frame, typically via the first channel TX1 number of chirps sent, which are used to calculate distance and speed (ie the range Doppler map), is greater than the number of those chirps in the partial sequences of the channels TX2 and TX3 , which are only used to calculate the angles assigned to the targets. It should also be noted that the order within the frame does not correspond to that in 8th shown order is determined, in which those chirps are first used which are used for the calculation of distance and speed, followed by those chirps which are used only for the angle calculation. Rather, within a frame, the chirps for the angle calculation can first be sent out via the TX channels TX2 and TX3 , followed directly by the chirps over the TX channel TX1 used for distance / speed calculation. It should also be noted that although in 8th only one frame is shown, a plurality of frames are transmitted in succession during operation. In other words, the in 8th shown frame sequence is sent out repeatedly, with the distance, speed and angle (DoA) being calculated after each frame sequence, as shown below.

The left part of the 9 shows a range Doppler map with three detected local maxima, each of which can represent a real radar target. According to the present exemplary embodiment, this range Doppler map can be based on that over the first TX channel TX1 radiated partial sequence can be calculated. In 9 the targets and the corresponding local maxima are included in the range Doppler map T ₁ , T ₂ and T ₃ designated. You can see the position of T ₁ at a distance (range) of x ₁ and a corresponding speed v ₁ . The target T ₂ is at a distance of x ₂ at a speed v ₂ , and the target T ₃ is at a distance of x ₃ ≈ x ₂ at a speed v ₃ . In other words, each target detected in the range Doppler map T ₁ , T ₂ , T ₃ (local maximum), a distance value (range) and a speed value can be assigned. It is understood that discrete values (indices) along the range axis and along the speed (Doppler) axis are discrete frequency values, however these frequency values can be converted directly into distances or speeds. In this respect, the discrete frequency indices k , I are regarded as "distance values" or "speed values".

The right part of the 9 shows the same targets T ₁ , T ₂ and T ₃ in the corresponding range-angle map, which can be calculated based on at least three chirps emitted via different channels. The azimuth angle ϕ _{1 of} the target can be obtained from the range angle map T ₁ can be clearly identified. With the targets T ₂ and T ₃ However, a clear assignment is not possible because both targets T ₂ and T ₃ have approximately the same distance from the sensor (x ₃ ≈ x ₂ ). In the range-angle map you can find the two local maxima at the distance coordinate x ₂ which are the targets T ₂ and T ₃ represent, however The range Doppler map and the range angle map cannot be used to easily decide which local maxima in the range angle map (ie which angle) for the distance x ₂ the target T ₂ and which the target T ₃ is to be assigned. This problem of ambiguous assignability (with respect to the angle) always occurs when several targets with the same distance or at a similar distance are detected, ie when several targets overlap (overlapping) signal components in the same frequency bin k (corresponds to in the example 9 a distance value x ₂ ) cause. Overlapping signal components are thus signal components that are identified as signal components of a potential target according to the algorithm used and are assigned to a same bin in a range-Doppler map, but in which bin are separated from one another, ie are assigned an equal distance but different speeds. The property “overlapping signal components” can therefore be detected by determining whether two or more targets have the same distance value or at least a similar distance value (for example corresponding to the adjacent frequency bin), the targets having different speeds. The overlap therefore takes place in the range dimension (same or similar distance), but not necessarily in the Doppler dimension.

At this point it should be noted that a certain distance value x actually represents a distance interval from x-Δx / 2 to x + Δx / 2, where Δx corresponds to the distance resolution. The distance resolution corresponds to the bandwidth of the transmission signal (chirps) (cf. 3rd and 4th , ADC 30th ). However, these relationships are known to a person skilled in the art and are therefore not explained further here.

Before various approaches to solving the mentioned problem of ambiguous assignability are discussed below with reference to 11 the calculation of a range angle map is explained using an example. 10th illustrates schematically the calculation of a range angle map, on the basis of which the distance and azimuth angle of a radar target can be determined. The calculation is very similar to that in 6 shown calculation of a range Doppler map. However, in the present case - unlike in the example according to 6 - does not consider a sequence of chirps that were transmitted via a specific TX channel and received via a specific RX channel, but rather a sequence of chirps that were received by different (virtual) antennas. As already mentioned, an array of virtual antennas can be formed by a combination of TX and RX antennas n _TX TX antennas of every RX antenna n _TX virtual RX antennas can be assigned. With a system with n _TX TX antennas and n _RX RX antennas, an antenna array with V = n _TX × n _RX virtual RX antennas (and a virtual TX antenna) can be formed.

In the in 10th example considered is a sequence of V RF signal segments considered, each over V different virtual antennas were received. These RF signal segments can contain a real echo of a transmitted chirp. Diagram (a) of 10th shows the sequence mentioned in the HF band and diagram (b) the corresponding sequence in the baseband. The corresponding digitized signal segments in the baseband can be made analogously to the example 6 in a matrix Y '[n, v] can be organized with N × V values (see diagram (c) of 10th ), the signal segment of the vth virtual antennas being arranged in the vth column of the matrix (v = 0, 1, ..., V-1). The row index n denotes a sample within a segment (n = 0, 1, ..., N-1). The index axis along the lines is as in 10th referred to as the "fast timeline".

The range-angle map can be calculated by a two-stage Fourier transformation, similar to the range-Doppler map. In a first stage, the FFTs of the columns are calculated, making the matrix R '[k, v] receives. The fast time axis becomes the range axis by the transformation (cf. 6 ). This first FFT level was already used in the calculation of the range maps 6 executed and usually does not have to be repeated for the calculation of the range angle map. In a second stage, the FFTs of the rows of the matrix R '[k, v] calculated (angle-FFT, angle FFT ). The result is the range angle map X '[k, u] , in which the radar targets can be detected like in a range Doppler map. Instead of the speed information, the index u represents angle information on the index axis along the columns.

As in 9 illustrated situations can occur in which targets that were detected in a range Doppler map cannot be clearly assigned targets that were detected in the range angle map. Various methods known per se can be used for the detection of the targets, for example methods which can recognize local maxima (caused by real radar targets) by means of comparison with a threshold value. For example, so-called CFAR methods are common in radar applications, in which the threshold value depends on the noise contained in the signal (CFAR means constant false alarm rate). Such detection methods are known per se and are not further explained here.

11 illustrates an example of a method that makes it possible to separate two or more targets whose detected distance values are the same or similar and to use the resulting separate values in subsequent beamforming algorithms. In the following, a digital radar signal (in the baseband) is considered, which is composed of a large number of signal segments. The considered sequence (frame) of signal segments contains the digitized data acquired during a measurement, wherein each signal segment of the sequence can be assigned to a chirp of an emitted RF radar signal (cf. 6 and 10th ). Partial sequences of the sequence of signal segments can be assigned to different virtual antennas. This means that the partial sequence of signal segments originates from a specific RX channel and is based on chirps that were emitted via a specific TX channel. For example, in 11 the via the virtual antenna 0 (the couple TX1 / RX1 assigned) received successive signal segments of a certain partial sequence of the signal segments. In the same way they correspond via the virtual antenna 1 , 2nd , etc. received signal segments further partial sequences of the signal segments.

According to 11 are initially based on a first partial sequence of M successive signal segments (each N Samples) one or more radar targets are detected. Which the first part sequence with M successive signal segments is the virtual antenna 0 assigned and corresponds to a transmission of M successive chirps over a first TX channel TX1 that after reflection on an object each in an RX channel RX1 are received, whereby the analog-digital conversion generates N samples per signal segment (chirp). In the exemplary embodiment shown, the signal segments of the first partial sequence become a range Doppler map X [k, l] calculated, in which then, for example, by detection of local maxima radar targets T ₁ , T ₂ and T ₃ can be detected (k = 0, 1, ..., N-1, l = 0, 1, ... M-1, cf. 6 ). Similar to the example 9 it can happen that in the kth frequency bin there are two targets T ₂ and T ₃ can be detected at approximately the same distance value, so that signal components based on the target T ₂ and signal components based on the target T ₃ overlap (see 11 ). As already mentioned, the kth frequency bin corresponds to a specific discrete distance value x which, however, actually, as mentioned, represents a certain range of distance values (eg from x-Δx / 2 to x + Δx / 2). This means that if, for example, the same discrete distance value is determined for two objects in the following, this should be understood to mean that the two distance values lie within the distance value range of the respective bin, which is limited by the bin limit values of a respective bin. The speeds of the targets T ₂ and T ₃ , v ₂ respectively. v ₃ , can also be determined from the range Doppler map. The signal segments of the first partial sequence come from a virtual antenna. For the present example, this means that a corresponding partial sequence (of a frame) of chirps is transmitted via the TX channel TX1 radiated and a corresponding partial sequence of signal segments in the baseband is via the RX channel RX1 receive. The combination TX1 / RX1 represents a virtual antenna (virtual RX channel).

If - like in this example 11 - a group of radar targets T ₂ , T ₃ , with (overlapping) signal components in the same frequency bin k (corresponds to a certain discrete distance value), then, as mentioned, a clear calculation of an azimuth angle is problematic (cf. 9 ). To the signal components of the targets T ₂ , T ₃ (with the speeds v ₂ and v ₃ ) of the group mentioned - as in 11 represented schematically - for the radar targets T ₂ , T ₃ spectral values belonging to the group of radar targets q _{k, 2} , q _{k, 3} calculated. The spectral values q _{k, 2} , q _{k, 3} are complex numerical values based on a second partial sequence of signal segments and those for the targets T ₂ , T ₃ determined speeds v ₂ , v ₃ be calculated. The second partial sequence of signal segments can also be assigned to a virtual antenna (ie a virtual RX channel), the second partial sequence being able to be part of the first partial sequence, but not being necessary. For example, the second partial sequence can be the one in 11 partial sequence shown from 0 to C-1, that of a number C. of consecutive chirps that correspond over the TX channel TX1 were broadcast and from the RX channel RX1 were received (virtual antenna 0 ). Similarly, a further partial sequence can be formed by C successive chirps, which are via the TX channel TX2 were sent out, etc.

The signal segments of the second partial sequence can (analog to 6 or 10th ) as columns of a matrix Y "[n, ic] be organized, the C signal segments each N Contains samples (n = 0, 1, ..., N-1, i _C = 0, 1, ..., C-1). The index ic denotes the respective signal segment, the index n the sample within a signal segment. A column-wise Fourier transformation (range FFT) results in the matrix S [k, ic] , which essentially represents a range map (cf. 6 , Range map R [k, m] ). From the previously mentioned target detection in the range Doppler map X [k, l] it is already known that in the k-th frequency bin signal components of the Targets T ₂ and T ₃ are located. The (complex) spectral values contained in this k-th frequency bin can be used as a line vector s _k ^T be written (the superscript T means "transposed").

The vector s _k contains ic elements (corresponds to the number of segments of the second partial sequence). Because those in the vector s _k contained spectral values essentially from the targets T ₂ and T ₃ the vector can be caused s _k can be broken down as follows: $$s_k=H_k\cdot \underset{qk}{\underbrace{\left(\begin{array}{c}{q}_{k\mathrm{,\; 2nd}}\\ q_{k\mathrm{,\; 3rd}}\end{array}\right)}}.$$

Inscribed H _k a C × 2 matrix (for a group with two targets T ₂ , T ₃ ), which contain the speed-dependent segment-to-segment phase shifts Δϕ ₂ and Δϕ ₃ , the two targets T ₂ , T ₃ can be assigned. The associated speeds v ₂ , v ₃ were previously based on the previously calculated range Doppler map R [k, l] determined. This gives you - with two targets T ₂ , T ₃ - for the matrix H _k following structure $$H_k=\left(\begin{array}{cc}1& 1\\ \text{exp}\left(j\cdot 1\cdot \Delta {\varphi }_{\mathrm{2nd}}\right)& \text{exp}\left(j\cdot 1\cdot \Delta {\varphi }_{\mathrm{3rd}}\right)\\ ⋮& ⋮\\ \text{exp}\left(j\cdot \left(\text{C.}-1\right)\cdot \Delta {\varphi }_{\mathrm{2nd}}\right)& \text{exp}\left(j\cdot \left(\text{C.}-1\right)\cdot \Delta {\varphi }_{\mathrm{3rd}}\right)\end{array}\right).$$

The i _C -th line of the matrix therefore contains the phase terms exp (j · i _C · Δϕ ₂ ) and exp (j · i _C · Δϕ ₂ ) (for i _C = 0.1, ..., C - 1 ). The function exp (·) denotes the exponential function. If there are more than two targets, the matrix points H _k correspondingly more than two columns and the vector q _k correspondingly more than two lines. In essence, the matrix does H _k a reverse transformation of the speed of a target (target T ₂ respectively. T ₃ ) caused phase shifts to the phase of the first signal segment (i _C = 0) of the second partial sequence. This takes into account the fact that the signal segments are recorded at different times and that the phase changes from one signal segment (corresponds to one chirp) to the next due to the speed of the respective target.

Equation (2) represents an overdetermined system of equations, which can be solved approximately. Various approaches for the calculation of approximation solutions are known per se, which are not discussed in more detail here. Usually the approximate solution q̂ _k = (ĝ _{k, 2} , ĝ _{k, 3} ) ^T is calculated so that a norm || e || of the error e becomes minimal, ie $$\Vert e\Vert =\Vert s_k-H_k\cdot \left(\begin{array}{c}{\widehat{q}}_{k\mathrm{,\; 2nd}}\\ {\widehat{q}}_{k\mathrm{,\; 3rd}}\end{array}\right)\Vert =\text{min}\left\{\Vert s_k-H_k\cdot \left(\begin{array}{c}{\widehat{q}}_{k\mathrm{,\; 2nd}}\\ {\widehat{q}}_{k\mathrm{,\; 3rd}}\end{array}\right)\Vert \right\}.$$

The calculation of the approximation solution q̂ _k can be repeated for each virtual RX channel (ie for each RX / TX channel pair). At V You get virtual channels for each target T ₂ , T ₃ the group of targets under consideration (with the range index / frequency bin k ) V Spectral values, on the basis of which (for example by means of Fourier transformation) the azimuth angle of the targets T ₂ , T ₃ can be calculated (e.g. as with reference to 10th explained).

That in relation to 11 The described method allows the (in the frequency domain) overlapping signal components (represented by the approximation solutions q̂ _{k, 2} and ĝ _{k, 3} ), which are generated by the different radar targets (e.g. T ₂ and T ₃ ) a group of radar targets in a certain frequency bin (corresponds to a certain distance value) caused to separate. This process could also be referred to as signal decomposition. The in-phase superposition of the approximate solutions would (approximately) the vector s _k result, which in turn corresponds to the considered kth bin of the range map. The approximate solution q̂ _{k mentioned} contains these separated signal components q̂ _{k, 2} and q̂ _{k, 3} . These can be calculated for each virtual reception channel of the radar system, and these results can be used for calculating the angles assigned to the respective radar targets, this assignment being unambiguous due to the separation mentioned (unlike in the example from 9 ).

organisiert werden. Diese Zeilenvektoren haben im Wesentlichen dieselbe Bedeutung wie eine Zeile der Matrix R'[k, v] in dem Beispiel aus 10. Mittels Fourier-Transformation (Winkel-FFT) der erwähnten Zeilenvektoren und Detektion der Maxima kann auf an sich bekannte Weise ein Azimutwinkel und somit die DoA bestimmt werden. Dieser Winkel ist - in dem Beispiel aus 11 - eindeutig dem jeweiligen Target T₂ bzw. T₃ zuordenbar. The one for each virtual channel v determined separated signal components q̂ _{k, 2} and q̂ _{k, 3} can be used as line vectors $$\left({\widehat{q}}_{k\mathrm{,\; 2nd}}\left(v=0\right),...,{\widehat{q}}_{k\mathrm{,\; 2nd}}\left(v=V-1\right)\right)\text{respectively}.\left({\widehat{q}}_{k\mathrm{,\; 3rd}}\left(v=0\right),...,{\widehat{q}}_{k\mathrm{,\; 3rd}}\left(v=V-1\right)\right)$$

get organized. These row vectors have essentially the same meaning as a row of the matrix R '[k, v] in the example 10th . By means of a Fourier transformation (angle FFT) of the line vectors mentioned and detection of the maxima, an azimuth angle and thus the DoA can be determined in a manner known per se. This angle is - in the example 11 - clearly the respective target T ₂ respectively. T ₃ assignable.

Theoretically, situations can arise in which a target (e.g. Target T ₂ ) two or more angles are detected. This means that the target in question is actually two or more targets, each of which has the same distance and the same speed, but is seen by the radar system at different angles.

At the in 11 In the example shown, the second partial sequence (with C signal segments) is part of the first partial sequence from which the range Doppler map is calculated, on the basis of which distance and speed values of targets are detected. However, this does not necessarily have to be the case. 12th illustrates a modification of the FIG 11 described example, in which the second partial sequence is not part of the first partial sequence, but follows this. The second partial sequence can (but does not have to) be provided by the same virtual channel as the first partial sequence. As already mentioned, the in 11 and 12th algorithm shown for other partial sequences (which are each supplied by other virtual channels) repeated, so that for each target of a group of targets, the signal components in a certain frequency bin k caused, overall V Receives spectral values (at V virtual channels). In exemplary embodiments, the number of available virtual channels (RX / TX combinations) typically corresponds to the number of different partial sequences that are used for the calculation of the angle.

13 illustrates another modification of the FIG 11 described example. In the in 13 shown case are in the range Doppler map the local maxima that the targets T ₂ and T ₃ are assignable, not in the same frequency bin k on the range axis, but in two (immediate or close) adjacent frequency bins, e.g. k or k + 1. However, the local maxima are not necessarily sharply limited to a frequency bin. Rather, the “main lobe” assigned to a local maximum can also be “blurred” over several adjacent frequency bins. Therefore it can also be found in the 13 illustrated case may be useful, referring to the above 11 and 12th explained technique to apply, in which case the two frequency bins k and k + 1 are processed separately. The result is - for each virtual channel v = 0, 1, ..., V-1 - the estimated spectral values q̂ _k = (ĝ _{k, 2} , q̂ _{k, 3} ) ^T and q̂ _{k + 1} = (q̂ _{k + 1,2} , q̂ _{k + 1,3} ) ^T for the frequency bins k and k + 1. With wider main lobes it may be necessary or useful to process more than two neighboring bins. Theoretically, the method can be applied to all frequency bins of a range Doppler map, whereby the calculations described above only make sense if two neighboring radar targets (e.g. T ₂ and T ₃ ) cause significant signal components in the same frequency bin on the range axis of a range Doppler map. If only one target is detected in a particular bin (which can also be the case if two targets are at the same distance and at the same speed), the matrix “degenerates” H _k (cf. equation (3)) into a vector and the system of equations becomes mathematically trivial, since in this case no two targets with different speeds have to be “separated” anymore.

14 and 15 serve to visualize the results obtained by the methods according to 11 and 13 or the method according to 13 receives. The results (ie the approximate solutions q̂ _k for a large number of virtual channels) of the in 11 to 13 The calculations shown can be interpreted as part of a kind of range map, similar to the matrix R '[k, v] in 10th . The Fourier transforms thereof can be interpreted as part of a range angle map, similar to that in FIG 9 (right) shown diagram, with the targets T ₂ and T ₃ (please refer 9 ) with signal components in the same frequency bin k (or in adjacent frequency bins) can be separated. So you get for each of these targets T ₂ and T ₃ a separate range angle map, only those parts of the range angle map are calculated which are relevant for the determination of the angle, that is to say the spectral values in the k-th frequency bin and possibly in neighboring bins. This separation of the range-angle maps enables a clear assignment of distance (range), speed and angle to the individual targets T ₂ and T ₃ possible out in the example 9 not possible.

As mentioned in the 11 to 13 shown calculation (estimate) of the spectral values q̂ _k = (q̂ _{k, 2} , q̂ _{k, 3} ) ^T for each virtual channel (each virtual receiving antenna v = 0, 1, ...., V-1, is repeated - in the considered k-th frequency bin - a first set of V Estimated spectral values {q̂ _{k, 2} (v = 0), q̂ _{k, 2} (v = 1), ..., q̂ _{k, 2} (v = V - 1)} that the target T ₂ are assigned and a second set of V Estimated spectral values {q̂ _{k, 3} (v = 0), q̂ _{k, 3} (v = 1), ..., q̂ _{k, 3} (v = V - 1)} that the target T ₃ assigned. About every set of V Spectral values now become one FFT (Angular FFT) performed. After performing the angular FFT for each of the targets T ₂ , T ₃ , (line by line FFT , please refer 10th ) you get for each target T ₂ , T ₃ each have a vector, typically the component of the vector which has the maximum amount indicating the angle for the respective target. The result of the angle FFT can therefore be viewed as part of a range angle map. The method is carried out for all frequency bins in which targets with the same distance have been recorded. The angle FFT is thus for those lines k calculated for which a group of targets was previously detected. The result is in 14 visualized; you get a (partial) range angle map for each target T ₂ , T ₃ from the considered group of targets that were detected in the k-th frequency bin. Each range-angle map contains exactly one local maximum in the k-th frequency bin, which represents exactly one target. This occurs in connection with this separation of the range-angle maps 9 not mentioned problem of lack of uniqueness. The method can be applied analogously for frequency bins with only one target peak (local maximum of the amount), the matrix being explained as above H _k (cf. equation (3)) becomes a vector and the system of equations for determining the spectral values becomes mathematically simple. As explained above, a is then made via the set of spectral values FFT carried out in order to obtain the corresponding angle information.

In the example 13 become spectral values for the frequency bins k and k + 1 calculated. In a general example, the spectral values are calculated for several adjacent frequency bins (for each virtual channel). The result is in 15 visualized. Several range-angle maps are again obtained, one for each target of the group of targets under consideration, these range-angle maps not just one frequency bin k include, but several, namely the considered amount of adjacent frequency bins, in the present example the frequency bins k and k + 1. As already in relation to 14 mentioned the determined angle values can be clearly assigned to the individual targets due to the separation of the targets.

QUOTES INCLUDE IN THE DESCRIPTION

This list of documents listed by the applicant has been generated automatically and is only included for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent literature cited

• DE 102014212284 A1 [0003]

Claims (11)

A method comprising the following: transmitting a first partial sequence of a frame with chronologically successive chirps via a first channel of a radar system and transmitting one or more further partial sequences of the frame via different channels, the further partial sequences following the first partial sequence and a smaller one in each case Have number of chirps as the first partial sequence; Detecting one or more radar targets (T ₂ , T ₃ ) and determining a distance value and a speed value for each of the detected radar targets (T ₂ , T ₃ ) based on only the first partial sequence; and calculating an angle for each of the detected radar targets based on the further partial sequences. The procedure according to Claim 1 , where the different channels are virtual channels of a multiple-input-multiple-output (MIMO) radar system. The procedure according to Claim 1 or 2nd , wherein the detection of one or more radar targets (T ₂ , T ₃ ) comprises: providing a digital radar signal (Y [n, m]) with a sequence of successive signal segments which is assigned to the transmitted frame, the radar targets ( T ₁ , T ₂ , T ₃ ) are detected based on a first partial sequence of the sequence of successive signal segments. The procedure according to Claim 3 , the calculation of the angles comprising: detecting whether there is a group of radar targets with overlapping signal components among the detected radar targets (T ₂ , T ₃ ), and if a group of radar targets (T ₂ , T ₃ ) was detected with overlapping signal components: calculating a spectral value (ĝ _{k, 2} , q̂ _{k, 3} ) for the radar targets of the group based on further partial sequences of the sequence of successive signal segments and further based on those for the group of radar targets ( T ₂ , T ₃ ) determined speed values. The procedure according to Claim 4 , wherein the calculating the angles comprises: calculating azimuth angles for the radar targets of the group of radar targets (T ₂ , T ₃ ) based on the calculated spectral values (q̂ _{k, 2} , q̂ _{k, 3} ). The procedure according to Claim 4 or 5 , wherein calculating a spectral value (q̂ _{k, 2} , q̂ _{k, 3} ) for the radar targets of the group comprises: transforming the signal segments of the one of the further partial sequences into the frequency domain to obtain corresponding spectra; Extracting those values (s _k ) from a frequency bin (k) of the spectra obtained, which is assigned to the group of radar targets (T ₂ , T ₃ ); Estimating a spectral value (q̂ _{k, 2} , q̂ _{k, 3} ) for each of the radar targets (T ₂ , T ₃ ) of the group based on the extracted values (s _k ). The procedure according to Claim 6 , wherein the estimation of the spectral value (ĝ _{k, 2} , q̂ _{k, 3} ) for each of the two or more radar targets (T ₂ , T ₃ ) further comprises: calculation of an approximate solution of an overdetermined system of equations with a coefficient matrix (H _k ), which depends on the speed values detected for the radar targets of the group and with a vector of input values which correspond to the extracted values (s _k ). The method according to one of the Claims 1 to 7 , the first partial sequence of the frame and the further partial sequences of the frame immediately following one another. The method according to one of the Claims 1 to 8th , with the other partial sequences of the frame being of equal length. The method according to one of the Claims 1 to 9 , wherein the first partial sequence of the frame comprises more chirps than the further partial sequences of the frame. A system that: a radar transceiver which is designed to transmit a first partial sequence of a frame with chronologically successive chirps via a first channel of a radar system and to transmit further one or more further partial sequences of the frame via different channels, the further partial sequences following the first partial sequence and each have a smaller number of chirps than the first partial sequence; a computing unit which is designed to detect one or more radar targets and in each case to determine a distance value and a speed value for the detected radar targets based on only the first partial sequence, and which is further designed to determine an angle for the to detect detected radar targets based on the further partial sequences.

Images