Method for target detection in DDM radar sensor

Abstract

The invention relates to a method for target detection in a DDM radar sensor having a number of transmitting antennas Tx of NTx and a number of NSlotsgt; the invention relates to a method for determining an occupancy of a transmitting antenna in a DDM time slot of NTx by means of a DDM code, comprising the following steps: calculating a distance Doppler matrix (30), which is divided into NSlot multi-valued regions (32) in the Doppler dimension, and determining the occupancy of the transmitting antenna in the DDM time slot by means of a non-coherent cyclic discrete convolution with the DDM code; the distance Doppler matrix (30) is converted into a three-dimensional distance Doppler multi-valued matrix (34), and target detection is carried out by comparing the distance Doppler multi-valued matrix (34) to a threshold matrix on a cell-by-cell basis.

Description

Technical Field

[0001] This invention relates to a method for target detection in a DDM radar sensor, the radar sensor having a number of N. Tx The number of transmitting antennas is N Sots >N Tx The DDM time slot is determined by the DDM code, which determines the occupancy (Belegung) of the transmitting antenna on the DDM time slot.

[0002] In particular, the present invention relates to a DDM (Doppler Division Multiplex) radar sensor, which is used in a driver assistance system of a motor vehicle or in an autonomous driving system to detect the traffic environment. Background Technology

[0003] Multiple transmitting antennas of the radar sensor are spatially offset from each other, thus enabling angularly resolved localization of the radar target. Orthogonal transmitting signals are provided to the transmitting antennas, allowing multiple antennas to be activated simultaneously, and the signals transmitted by different antennas can still be separated from each other again at the receiving end. The transmitting signal can consist of a sequence of linear frequency modulated (LFM) signals. Then, the distance information regarding the located radar target can be obtained by mixing a portion of the transmitted signal with the received radar echo, resulting in a lower frequency signal obtained through a beat (Schwebung), the frequency of which corresponds to the frequency difference between the transmitted and received signals. Because the frequency of the transmitted signal increases linearly during LFM, the frequency difference is related to the propagation time of the signal and therefore to the distance (Abstand) of the radar target. If the radar target has a non-zero radial velocity relative to the radar sensor, an additional frequency shift occurs due to the Doppler effect, but this additional frequency shift is not significantly distorted given a sufficiently steep ramp of the LFM signal. Spacing measurement. However, Doppler shift causes the received signal to have a certain phase progression (Phasenfortschritt) from one linear frequency modulated signal to another, which allows for the measurement of the radial velocity of the target.

[0004] The DDM modulation method guarantees the orthogonality of the transmitted signals and enables high range and high resolution in the velocity dimension for radar sensors at a relatively moderate cost. For each of the simultaneously active transmit antennas, the DDM phase modulator has a phase shifter that rotates the phase of the signal transmitted to the antenna such that the phase of the signal transmitted in the current linear frequency modulated (LFM) signal rotates by a defined angle relative to the phase of the preceding LFM signal, where the angle varies from antenna to antenna. The phase progression from one LFM signal to another has the same effect on the received signal as the Doppler shift caused by the motion of the radar target itself. In this way, each transmit antenna is characterized by a characteristic “virtual” Doppler shift.

[0005] The range-Doppler matrix can be computed using a two-dimensional Fourier transform (FFT). In the range dimension (spacing dimension), the Fourier integral is performed at the sampling times within a single linear frequency modulated (LFM) signal. In the Doppler dimension (relative velocity dimension), the Fourier integral is performed at the corresponding sampling times of successive LFM signals. In the range-Doppler matrix, each located radar target is represented by the peak values ​​in the cells of the range-Doppler matrix, determined by the radar target's range in the spacing dimension and by the target's relative velocity in the Doppler dimension. However, because the sampling frequencies in both dimensions cannot be arbitrarily chosen in practice, so-called undersampling often occurs, resulting in the spacing and velocity information not being definitively unique. Due to the periodicity of radar signals, the peak values ​​in the Doppler dimension exhibit a regular pattern... Velocity range repetition. If the possible velocities of a radar target are determined from the outset to be within a range smaller than the interval between peak values, a clearly unique velocity measurement can be easily achieved. However, if the possible velocities of the target are dispersed over a large range, specialized methods are needed to resolve the multivaluedness.

[0006] In a typical DDM radar, no phase shift occurs in one time slot (which is indexed 0), while in another time slot (which is indexed 1), the phase of the transmitted signal rotates by a phase angle from one linear frequency modulated (LFM) signal to another LFM signal. In the next time slot, the phase shift from one linear frequency modulated (LFM) signal to another LFM signal is: In the time slot with index 3, the phase offset is Wait a minute. Because with index N slotsThe phase offset is 0 in the time slot, so it passes through the minimum phase offset that is not equal to 0. Determine the number of possible time slots N Slots Because these phase shifts simulate additional Doppler shifts, they are factored in DDM radar. This reduces the explicitly unique range in the velocity dimension. This often results in the actual velocity of the radar target exceeding this explicitly unique range, even with relatively small, manageable variations. If the number of transmitting antennas N... Tx Less than N Slots Then, multivaluedness can be distinguished using known algorithms. Examples of such algorithms are described in EP 3 611 538 A1, EP 4009 074A1 and US2021 / 132,187A. Summary of the Invention

[0007] The objective of this invention is to provide a method that enables better resolution of velocity multivalues. The method.

[0008] According to the present invention, the task is solved by a method having the following steps:

[0009] Calculate the range-Doppler matrix, which is divided into N parts in the Doppler dimension. Slots A multi-valued region,

[0010] The range-Doppler matrix is ​​transformed into a three-dimensional range-Doppler multivalued matrix by incoherent cyclic discrete convolution with the DDM code.

[0011] Target detection is performed by comparing the range Doppler multivalue matrix with the threshold matrix cell by cell.

[0012] According to the present invention, the range-Doppler matrix is ​​convolved unit-by-unit with a kernel that is given by a DDM code but progresses periodically with a modulus function (fortgesetzt). By employing a specialized type of convolution operation, not only is the velocity multivaluedness resolved, but the convolution kernel also simultaneously induces [something] in N. Tx Averaging the signals from each transmitting antenna yields approximately 7log. 10 N TX The noncoherent integral gain of dB results in a high signal-to-noise ratio in target detection. Another advantage is that it enables the detection of multiple targets with the same multivalued velocity during the detection step.

[0013] Advantageous configurations and extensions of the invention are derived from the dependent claims.

[0014] In an advantageous implementation, N Slots ≥2N Tx .

[0015] The threshold matrix can be derived from the so-called noise-floor matrix, which is obtained by applying the N values ​​of the cells over the multivalued region. Ts The threshold matrix is ​​obtained by averaging the minimum values. In this way, not only is a usable metric for the noisy background obtained, but the incoherent integral gain is also obtained based on the averaging, in the case of a noisy signal as in the case of a total signal. In one particular implementation, the threshold matrix can also be determined directly from the range Doppler matrix without averaging.

[0016] By definition, the autocorrelation function of a DDM code has its maximum value at index s = 0, but also exhibits sidebeams at other index values. These sidebeams weaken the signal spacing between the range-Doppler multivalued cell corresponding to the true velocity and the corresponding range-Doppler cells in other multivalued cells, thereby hindering detection. However, for a given number N... Tx The transmitting antenna and a given number of time slots N Slots The following DDM code can be found through system simulation: under this DDM code, the side beams in the autocorrelation function are suppressed to the greatest extent.

[0017] Instead of directly using the range-Doppler multivalued matrix for target detection, it's also possible to limit the multivalued dimension to a bin containing truly strong targets. This approach reduces storage requirements.

[0018] In one embodiment of the method, during actual target detection, only elements that are local maxima in both the spacing and Doppler dimensions are considered. Because the range-Doppler multivalue matrix represents not only the multivalued velocity but also the multivalued range of the target's true velocity, the target's true velocity can be determined. Furthermore, the association between the occupied DDM time slots (and the corresponding multivalued regions in the range-Doppler matrix) and the assigned transmitting antenna is known using the DDM code, enabling explicit and unique signal processing for angle estimation.

[0019] In the case of a MIMO radar sensor with multiple receiving antennas, it should be understood that performing the above steps for each receiving channel allows for the consideration of different receiving channels during angle estimation. Then, the actual angle estimation can be performed, for example, using a Fourier transform based on the processed ortssignal.

[0020] For the detected target, the complex signal passes through N Slots The vector formed by corresponding distance Doppler cells in different multivalued regions can also be used as input parameters for the iterative CLEAN algorithm to separate the superimposed target measurements and thus improve dynamics for targets with the same multivalued velocities. Similarly, the CLEAN algorithm can be used for angle estimation along the angular dimension to determine multiple target angles for targets with the same distinct and unique velocity. Alternatively, other estimation methods with multi-target capabilities known from the literature can be used.

[0021] In an advantageous implementation, to determine the convolution kernel, a DDM code is selected as follows: the DDM code is weakened. Significance of the sidebeams in the autocorrelation function

[0022] In an advantageous implementation, a multivalued cell is selected from the range-Doppler multivalued matrix such that a certain minimum signal strength is achieved in at least one range-Doppler cell within the multivalued cell, wherein the target detection is performed on a RoI matrix containing only the selected multivalued cell.

[0023] In one advantageous implementation, for target detection in terms of spacing and velocity, angle estimation is performed based on data obtained from multiple receiving channels.

[0024] In an advantageous implementation, an estimation algorithm with multi-target capability is employed to more clearly distinguish targets with the same multivalued velocity from one another.

[0025] In one advantageous implementation, an estimation algorithm with multi-target capability is used to determine multiple target angles for targets with the same definite and unique velocity. Attached Figure Description

[0026] The embodiments will now be described in more detail with reference to the accompanying drawings. The drawings show:

[0027] Figure 1 A block diagram of the transmitting section of a radar sensor is shown, by which the method according to the invention can be implemented;

[0028] Figure 2 The diagram shows the frequency / time used to transmit the signal;

[0029] Figure 3 A diagram illustrating phase progression in the case of DDM radar;

[0030] Figure 4A block diagram is shown of an analytical processing apparatus configured for use in the method according to the invention;

[0031] Figure 5 An example of a DDM code is shown;

[0032] Figure 6 Showing according to Figure 5 The autocorrelation function of the DDM code;

[0033] Figure 7 Showing the illustrations in Figure 4 The diagram shows how the VAR solver works, which converts the range-Doppler matrix into a range-Doppler multivalued matrix (RDVAR).

[0034] Figure 8 The diagram shows how to reduce the RDVAR matrix to the area of ​​interest.

[0035] Figure 9 A diagram is shown illustrating the functionality used to plot the VAR solver, which converts the range-Doppler matrix into a noise base matrix; and

[0036] Figure 10 A flowchart is shown, illustrating the main steps of the method according to the invention. Detailed Implementation

[0037] exist Figure 1 The transmitting section of a radar sensor is shown, which has four transmitting antennas Txi (0 ≤ i ≤ 3). Although not shown here, radar sensors are typically configured as MIMO radars, by which angle estimation can be performed based on the MIMO (Multiple-Input Multiple-Output) principle. A signal generator 10 generates a transmitting signal having... Figure 2 The frequency variation curve is shown in the figure. The signal consists of a sequence of so-called linear frequency modulated signals 12, which are repeatedly transmitted at regular time intervals, in which the frequency f increases linearly as a function of time t. The transmitted signals are provided to the transmitting antenna Txi via phase shifters 14, which rotate the signal phase by a certain angle for each linear frequency modulated signal. If p (0 ≤ p ≤ 3) is the index for counting the linear frequency modulated signals 12 in the sequence, and i is the index for counting the transmitting antenna, then the complex amplitude of the signal transmitted in the linear frequency modulated signal p in antenna i can, for example, have the following form:

[0038]

[0039] Where a is a real constant, j is the imaginary unit, ω(t) is the (circular) frequency related to time t, and It is the phase angle. The phase shifter 14 can be manipulated such that the phase angle is adjusted, for example, according to the following formula.

[0040]

[0041] exist Figure 3 The diagram shows the phase angle of a signal used in a conventional DDM radar with four antennas. With antenna i = 0, for all four linear frequency modulated signals, The phase angle is zero. With antenna i=1, the phase angle is zero for the first linear frequency modulated (LFM) signal, and increases by 90° for each subsequent LFM signal. With antenna i=2, the phase angle... Initially, the phase angle increases by 180° from one linear frequency modulated (LFM) signal to another. With antenna i = 3, the phase angle also increases from the first LFM signal... Initially, the phase angle increases by 270° from one linear frequency modulated (LFM) signal to another.

[0042] exist Figure 4 The diagram below shows the main components of the radar sensor's analysis and processing system. In the first analysis and processing stage 16, the range-Doppler matrix is ​​calculated in each measurement cycle and in each receive channel by performing a two-dimensional Fourier transform on the baseband signal obtained by downmixing the received signal. Each cell in the range-Doppler matrix is ​​part of both a range bin (rows of the matrix) and a Doppler bin (columns of the matrix), and the signal value of the cell represents the strength of the signal obtained from the radar target whose range and radial relative velocity lie within the range bin and the Doppler bin. A so-called VAR solver 18 transforms the range-Doppler matrix into a range-Doppler VAR matrix, which is cached in memory 20. The abbreviation VAR stands for Velocity Ambiguity Range and represents a multi-valued range consisting of multiple adjacent columns in the range-Doppler matrix, as will be explained in more detail later. Therefore, the German name for the "Range-Doppler VAR matrix" is "Range-Doppler-Mehrdeutigkeits-Matrix".

[0043] Additionally, the VAR solver 18 transforms the range-Doppler matrix into a so-called noise floor matrix, which is stored in a separate memory 22. This matrix represents the noise background to some extent, and the signal stored in the range-Doppler VAR matrix should stand out compared to this noise background.

[0044] In detection stage 24, radar targets are detected by comparing the cell contents of the range-Doppler VAR matrix with a threshold matrix. The threshold matrix can be directly the noise floor matrix or derived from the noise floor matrix using a known method, such as the CFAR (Constant False Alarm Rate) method. Alternatively, the threshold matrix can also be obtained directly from the range-Doppler matrix using a known method. Before actual detection, a coherent summation (e.g., using angular FFT) or incoherent summation of the range-Doppler spectra of each received channel prior to detection is performed.

[0045] The result of the detection step is a detection list, which is stored in a separate memory 26 and contains the location signals of all located radar targets.

[0046] The detection list is then provided to the angle estimation stage 28, where angle estimation is performed using known methods to obtain a point cloud that characterizes the current environment of the radar sensor.

[0047] exist Figure 1 Each phase shifter in phase shifter 14 shown defines a DDM time slot, characterized by phase progression in the associated transmission channel. The minimum phase progression, not equal to 0, is determined according to the formula... Determine the total number N of time slots present. Slot .exist Figure 3 In the example shown, And correspondingly, N Slot =4. Therefore, in the example, N Slot The number N of transmitting antennas that are activated simultaneously Tx Consistent.

[0048] However, in order to achieve the resolution of velocity multivalues, in practice, it operates with the following phase progression: in the case of said phase progression, N Slot N Tx Largest and preferred are N Tx At least twice that of the previous time slots. In this case, some time slots remain unoccupied because there are not enough transmit antennas. The time slots occupied by the transmit antennas are represented by a so-called DDM code. It is a binary value vector with components d. s ,(s=0,…,(Nslot -1)). In Figure 5 The example of a DDM code is shown graphically.

[0049] d=(1,1,0,1,0,0,0,1,1,0,0,0,0,1,0,0)

[0050] In the example, N Slots =16, and N Tx =6, so that only six time slots are occupied by the antenna.

[0051] Figure 6 The associated autocorrelation function f(m) is shown. The variable m is an index that counts the multivalued regions (also known as spectral slots) in the range-Doppler matrix. If the DDM code (the components of which are represented here by index i) is considered as a function d(s), then f(m) is the result of the convolution of the function d(s) with a kernel that is given in terms of the DDM code d but progresses periodically with a modulus function. Figure 6 The index of the second d in the formula has the following meaning:

[0052] (s+m)modulo N Slots .

[0053] By definition, the correlation function f(m) has a maximum value when m ≠ 0. Then, the sum in the given formula simply equals the number N of non-zero components of the DDM code. Tx This results in a value of 1 or 0 dB after normalization.

[0054] However, when m is not equal to 0, the correlation function exhibits more or less noticeable sidebeams. For example, for m = 6, we obtain:

[0055] NT x f(6)=d0+d6+d1+d7+d3+d9+d7+d 13 +d8+d 14 +d 13 +d3

[0056] =0+1+0+1+0+1

[0057] =4

[0058] Based on the modulus function, the component d3 in the last sum is obtained.

[0059] exist Figure 7 The example for a range-Doppler matrix is ​​illustrated graphically. In this simplified example, the matrix has eight cells, or bins, in the range dimension r, and the number of bins is N in the Doppler dimension v. largeIf the transmitted signal is not phase-modulated, the range Doppler matrix covers the entire theoretically possible range of speeds in the Doppler dimension (e.g., -150 km / h to +150 km / h in the case of radar used in passenger cars), and the sampling frequency is so high that the distance measurement within this range is definitively unique. However, due to phase modulation, N is induced. Slots The result is that, even at relative velocity v = 0, a single peak cannot be obtained in the Doppler spectrum; instead, a peak is obtained for each DDM time slot. Therefore, the range-Doppler matrix is ​​divided into N parts in the Doppler dimension. Slots There are 32 multivalued regions, also known as spectral slots, and are counted here by means of an index m, which also appears in [the context of the index]. Figure 6 Therefore, the number of Doppler cells or v-cells within each multivalued region is N. Small =N Large / N Slots Each matrix cell contains an entry x. RD The entries represent the signal strength of the following radar targets: the spacing and velocity of the radar targets are given by the position of the element in the matrix.

[0060] The function of VAR solver 18 is to, based on, Figure 7 The formula shown converts the two-dimensional range Doppler matrix 30 into a three-dimensional range Doppler VAR matrix 34, which in turn... Figure 7 The middle part is shown graphically on the right. Here, the third dimension is the dimension in which the multivalued index m varies. In the velocity dimension v amb In this context, the matrix has only N. Small Units. In V amb The index "amb" in the text indicates that the velocity is multivalued in the dimension, because in N... Slots The true velocity of the target in each of the warehouses may lie in a multi-valued dimension m. The velocities v in matrix 30 and v in matrix 34... amb The relationship between them is given as follows:

[0061] v = v amb +mv small

[0062] Among them, v small It is the width of a single multivalued region of 32.

[0063] The entry x in each matrix element of the 3D matrix 34 RDVAR It is the distance r and the velocity v amb A function of the multivalued index m. An entry x in a two-dimensional matrix 30.RD It can also be understood as a function of these three variables, but among them, in Figure 7 In the formula, the index m is replaced by the summation index s.

[0064] According to Figure 7 The formula shown in the image indicates that the VAR solver 18 is used for the function x. RD (r,v amb ,s) Implementation and Figure 6 The kernel d(mod(s+m,M) Slots The non-coherent discrete convolution is used. Through this convolution, different multivalued regions 32 are correlated with each other, such that a maximum value is obtained at the point of maximum correlation in the three-dimensional matrix 34. Thus, not only is the multivalued nature of velocity distinguished, but also the integral gain is derived based on the summation over index s.

[0065] exist Figure 7 In the 3D distance-Doppler VAR matrix 34, the number of VAR cells in the VAR dimension m is N. Slots However, in practice, the following situation often occurs: any one of these VAR bins at any point, i.e., at r and v amb In any combination of these cases, no noteworthy signal is contained. Therefore, to reduce storage requirements and computational overhead, it is appropriate to "dilute" the matrix by limiting it to bins where significant signals actually exist. This is in Figure 8 This is symbolized by replacing matrix 34 with a three-dimensional RoI matrix 36 (Range of Interest). The number of VAR bins in this matrix is ​​reduced to N. VAR ,valid.

[0066] Another function of the VAR correlator 18 is to calculate the same two-dimensional noise base matrix 38 based on the two-dimensional distance Doppler matrix 30, wherein the noise base matrix is ​​in the velocity dimension v amb The middle and distance Doppler VAR matrix 34 are the same, only having N small Individual V-warehouse, such as in Figure 9 As shown in the diagram. For each cell, the entry m in the noise base matrix 38 represents the noise background, and according to... Figure 9 The formula shown is used for calculation. Here, the function "mink" is used to identify the N-values ​​in the distance-Doppler matrix 30. Tx The units of the smallest entries are calculated, and the average is taken over these entries in the multivalued dimension s. Based on the averaging, the previously mentioned integral gain is also derived here again.

[0067] Then, for target detection, the noise floor matrix 38 can be used directly as the threshold matrix, or a suitable threshold matrix can be calculated from the matrix using a known method, such as the CFAR method.

[0068] Alternatively, the threshold matrix can also be obtained directly from the distance-Doppler matrix using known methods.

[0069] exist Figure 10 The main steps of the method are summarized in the flowchart. In step S1, the range-Doppler matrix is ​​calculated for each received channel. In step S2, the range-Doppler VAR matrix (range-Doppler multivalued matrix) is calculated respectively using the VAR solver 18. In an optional step S3, the matrix calculated in step S2 is reduced to the RoI matrix 36. Then, in step S4, the noise floor matrix 38 is calculated using the VAR solver 18. Then, in step S5, the threshold matrix is ​​formed from the noise floor matrix, for example, by simply using the noise floor matrix as the threshold matrix. Then, in step S6, using the VAR solver 18... Figure 4 Target detection is performed in detection stage 24. For this purpose, entries in the RoI matrix are compared with entries in the threshold matrix. Here, according to a variation of the method, only elements exhibiting local maxima in terms of dimensional spacing and explicitly unique velocities are considered. The main advantage of the proposed method is that it can also detect targets with the same multivalued velocity v in the detection step. amb .

[0070] Then, in step S7, angle estimation is performed using known methods. Here, the location signals are processed in such a way that they can be compared with each other in terms of angle estimation.

[0071] Then, in the optional step S8, the known iterative CLEAN algorithm can be used to further process the data that underwent convolution operations in step S2, so as to better distinguish and characterize targets with the same multivalued velocities. Here, the CLEAN algorithm can be used not only in the multivalued dimension but also in the angular dimension to determine multiple target angles of targets with the same clearly unique velocity.

[0072] The proposed iterative range-Doppler CLEAN method can be summarized as follows: After successfully estimating the first multivalued range and one or more associated target angles, the complex location signals of these radar targets (i.e., the complex amplitudes of the transmit and receive channels) are calculated. Then, before starting further iterations, the range-Doppler matrix N is coherently derived from the N-axis. SlotsThe location signal is subtracted from the following units of the large input vector: the units are occupied by radar targets according to their multi-valued range. In the next iteration, the DDM code is used again. Figure 7 The formula in [the formula] determines the range of multivalues. By subtracting the signal share of the already detected target from the previous one, weaker targets can now also be detected. Again, according to [the formula]... Figure 4 , Figure 7 , Figure 8 and Figure 9 The detection process is underway.

[0073] The CLEAN algorithm terminates once the interruption criteria are met, such as excessively low amplitude of the remaining positioning signal. This allows for dynamic range greater than 30 dB or even 40 dB with high-quality angle estimation.

[0074] Alternatively, other estimation algorithms with multi-objective capabilities, such as RELAX, can be used instead of the CLEAN algorithm.

Claims

1. A method for target detection in a DDM radar sensor, the DDM radar sensor having a number N Tx of transmit antennas Tx and a number N Slots >N Tx of DDM slots, the occupation of the DDM slots by transmit antennas being determined by DDM codes (d s ), the method having the following steps: A range-Doppler matrix (30) is computed, which is divided in N Slots valued regions (32) in the Doppler dimension, by non-coherent cyclic convolution with the DDM code (d s ) and converting the range-Doppler matrix (30) into a three-dimensional range-Doppler multivaluedness matrix (34), and Target detection is performed by comparing the range-doppler multi-valuedness matrix (34) with a threshold matrix cell by cell.

2. The method of claim 1, wherein, N Slots ≥2N Tx .

3. The method of claim 1, wherein, The threshold matrix is derived from a noise floor matrix (38) which is computed from the range-Doppler matrix (30) in such a way that, in each range-Doppler bin, the N Tx smallest entries of the matrix cells over the multivaluedness region (32) are averaged.

4. The method of any of the above claims, wherein, To determine the convolution kernel, a DDM code (d s ) is selected that attenuates the prominence of sidelobes in the autocorrelation function.

5. The method of any of the above claims, wherein, From the range-doppler multi-valuedness matrix (34) those multi-valuedness bins are selected in which a certain minimum signal strength is achieved in at least one range-doppler cell, wherein the target detection is performed on a RoI matrix which only contains the selected multi-valuedness bins.

6. The method of any of the above claims, wherein, Only those matrix cells are considered for the target detection which are local maxima in range dimension and in Doppler dimension.

7. The method of any of the above claims, wherein, For target detection in range and velocity, an angle estimation is performed based on the data obtained in multiple receive channels.

8. The method of any of the above claims, wherein, An estimation algorithm with multi-target capability is employed to separate targets with the same multi-valuedness velocity from each other more clearly.

9. The method of claim 7 or 8, wherein, An estimation algorithm with multi-target capability is employed to determine multiple target angles for targets with the same unambiguously unique velocity.

10. A radar system in which the method according to any one of the preceding claims is implemented.

Images