TECHNICAL FIELD

The disclosure relates generally to an automotive radar, and more particularly the disclosure relates to a method of determining an unambiguous radial velocity in an automotive radar system.

BACKGROUND

Automotive radar is a sensor system to detect targets and estimate their position. The position of the targets is estimated in terms of range, angle, and the radial velocity is estimated with respect to a host vehicle. For example, in car manufacturing, the radial velocity of the targets is estimated based on resolution and a certain unambiguous span of the velocity of the sensor system. If a target’s velocity is outside the certain unambiguous span of velocity, the observed velocity sensed by the sensor system would be a wrapped version based on the velocity and the certain unambiguous span of the velocity of the sensor system. For example, if the sensor system has an unambiguous span of the velocity that goes from -50 kilometer per hour (km/h) to +50 km/h and if the relative velocity of the target is 100 km/h, the observed velocity sensed by the sensor system is 0 km/h. In practice, to cover the required unambiguous span of the velocity with the required velocity, the resolution of the sensor system may require a lot of memory, which is unfeasible for automotive products.

Existing approaches enhance the certain unambiguous span of the radial velocity of the sensor system using a waveform with a train of chirps that present an interleaved frequency shift. The odd and even chirps are processed separately and the phase difference for detected targets is linked to the number of wrapped cycles, which results in a signal to noise ratio (SNR) loss of about 3 decibels (dB). This implies a chance of missing some weak targets, for example, pedestrians, especially in automotive radar system.

Further, some existing approaches use a different waveform between scans. However, the existing approaches have to match the velocity of the targets between scans. The targets may appear and disappear between the scans due to noise or scattering properties. Also, in a dense point cloud, there is a problem of matching the same target between different scans. For estimating the right velocity of the target, the sensor system requires at least two or three scans, thereby latency is introduced in the sensor system. The time due to latency may be critical, especially in safety functions for automotive, for example, emergency braking.

The existing approaches provide phase difference quality to determine the number of cycles that depend only on a single channel. Thereby, estimation of the unambiguous velocity may not be robust enough over multiple channels. Further, the estimation quality is not assessed using any processing procedures such as tracker or classification.

Therefore, there arises a need to address the aforementioned technical drawbacks in existing systems or technologies in determining an unambiguous radial velocity.

SUMMARY

It is an object of the disclosure to provide an automotive radar apparatus and a method of determining an unambiguous radial velocity in an automotive radar system while avoiding one or more disadvantages of prior art approaches.

This object is achieved by the features of the independent claims. Further implementation forms are apparent from the dependent claims, the description, and the figures.

The disclosure provides an automotive radar apparatus and a method of determining an unambiguous radial velocity in an automotive radar system.

According to a first aspect, there is provided an automotive radar apparatus. The automotive radar apparatus includes a waveform generator, a transmission antenna, a reception antenna, a mixer, an analogue to digital converter, ADC, and a signal processing unit. The waveform generator is configured to generate a transmission signal including a sequence of frequency-ramped chirps. The sequence of chirps includes a first time series of chirps interleaved with a second time series of chirps. The transmission antenna is configured to emit a radio wave in response to being driven by the transmission signal. The reception antenna is configured to generate a reception signal in response to receiving a reflected radio wave. The mixer is configured to generate an intermediate frequency, IF, signal by mixing the transmission signal with the reception signal. The analogue to digital converter, ADC, is configured to generate samples of the IF signal, by sampling the IF signal within each time window of a sequence of time windows associated with the sequence of chirps. The time windows have equidistant midpoints. The signal processing unit is configured to calculate for each respective time window a range profile by applying a Fourier transform to the samples pertaining to the respective time window. The signal processing unit is configured to calculate a matrix of ambiguous radial velocities by applying a Fourier transform across the range profiles. The matrix of ambiguous radial velocities includes a first sub-matrix and a second sub-matrix corresponding respectively to lower and upper velocities. The signal processing unit is configured to calculate a sum matrix and a difference matrix. The sum matrix and the difference matrix are a sum and a difference of the first and the second sub-matrix, respectively. The signal processing unit is configured to determine a matrix of first phase differences between elements of the sum matrix and elements of the difference matrix. The signal processing unit is configured to determine a matrix of second phase differences between elements of the matrix of first phase differences and a phase component depending on a measured range position and an ambiguous measured velocity, such that the second phase difference is a function depending only on a cycle index for each position. The signal processing unit is configured to determine a cycle index n _cyc (I ,]') for each position of the ambiguous radial velocity. The signal processing unit is configured to calculate an unambiguous radial velocity based on the ambiguous radial velocity and the cycle index.

The automotive radar apparatus improves target velocity estimation on a single radar scan, thereby reducing network latency especially in safety features of automotive such as emergency braking. The automotive radar apparatus estimates target velocity in a single scan over multiple channels, thereby enhancing robustness of the automotive radar apparatus. The automotive radar apparatus estimates target velocity without a signal- noise ratio (SNR) loss, thereby improving the phase quality. The automotive radar apparatus incorporates flexibility in waveform design and on a received signal recording. The automotive radar apparatus may detect possible errors during target velocity estimation and maintains a tracker for the possible errors.

Optionally, the ADC is configured to generate samples RxTime(i,j ) of the IF signal, i=1,...N, j=1,...M, by sampling the IF signal within each time window W(i) of the sequence of time windows, RxTime(i,j ) being the j-th sample in the i-th time window. Optionally, the signal processing unit is configured to calculate the range profile by, for each i from 1 to N, calculating a range profile Range_SlowTime(i,J ), J=1,...,M, by applying a Fourier transform to the samples RxTime(i,j ^' ) , j=1,...,M. The signal processing unit may be configured to calculate ambiguous radial velocities by, for each J from 1 to M, calculating a matrix RgVel(I,J) , 1=1,..., N, by applying the Fourier transform to the range profile Range_SlowTime(i,J ), i=1,...,N. The first sub-matrix RgVel _S1 (I'J') = RgVel(I'J') , I=1,...,N/2, J’=1,...,M, and the second sub-matrix

Optionally, the signal processing unit is configured to calculate the sum matrix RgVel _SUM and the difference matrix RgVel _DIFF such that,

The automotive radar apparatus may detect one or more targets using the sum matrix or the difference matrix using full processing gain. The automotive radar apparatus may estimate unambiguous velocity for the detected targets by analyzing the interferometric phase for a large number of channels, thereby increasing robustness in estimating unambiguous velocity.

Optionally, the signal processing unit is configured to determine the first phase differences Phase ((I',J)'), F=1,...,N/2, J’=1,...,M, first phase difference Phase(I' J') being a phase difference between element RgVel _SUM (I' ,J') of the sum matrix and element RgVel _DIFF (I',J') of the difference matrix. The signal processing unit may be configured to determine the second phase differences Phase _VelCylce (I',J') = Phase(I',J') - Phase _known (I',J'), F=1,...,N/2, F=1,...,M, and Phase _known (I',J') is the phase component depending on the measured range position and the ambiguous measured velocity, such that Phase _known (I' J') is a function of F and J’, Phase _known (I',J') = /(I',J') , and the second phase difference is a function Phase _VelCylce (I',J') = f(n _cyc (l',J ')) depending only on the cycle index n _cyc (I',J ') for each F J’ position. The first phase differences may also be known as an interferometric phase. The interferometric phase is calculated for each moving target and each channel. The phase compensation with residue is derived for each channel and averaged. Thereby, the phase quality becomes equivalent to the coherent gain of all available channels. Hence, the automotive radar apparatus is robust to unambiguous velocity estimation.

Optionally, a frequency-ramping slope is the same for the first time series of chirps and the second time series of chirps, and where a ramp time for the first and second time series, and an idle time between chirps is adjusted such that the windows sampled by the ADC from the first time series of chirps include a different center frequency to the windows sampled from the second time series of chirps.

Optionally, a frequency-ramping slope is different for the first time series of chirps and the second time series of chirps, such that the windows sampled by the ADC from the first time series of chirps include a different center frequency to the windows sampled from the second time series of chirps, and where a size of the windows sampled by the ADC is adjusted such that a constant bandwidth is sampled. The automotive radar apparatus may incorporate flexibility on waveform design and even on a received signal recording.

Optionally, the signal processing unit is further configured to detect a plurality of targets based on the calculated sum matrix or difference matrix. The first phase differences are determined for each target and averaged across a plurality of channels. The signal processing unit is further configured to estimate a direction of arrival for each detected target. The signal processing unit may estimate unambiguous velocity for one or more detected targets by analyzing the interferometric phase over a large number of channels, thereby increasing robustness in estimating unambiguous velocity over a large number of channels and reducing network latency.

Optionally, the signal processing unit is further configured to estimate a direction of arrival based on the calculated sum matrix and difference matrix using all channels. The signal processing unit is further configured to detect a plurality of targets based on the direction of arrival estimated using the sum matrix or the difference matrix. If the direction of arrival (DoA) estimation for the detected target takes place easily, then, the phase quality improves with zero SNR loss. Optionally, a potential error indication flag is raised based on a value of the second phase difference. The possibility of detecting errors in determining radial velocity may be estimated and may be recorded in a tracker.

According to a second aspect, there is provided a method of determining an unambiguous radial velocity in an automotive radar system. The method includes generating, using a waveform generator, a transmission signal including a sequence of frequency-ramped chirps. The sequence of chirps includes a first time series of chirps interleaved with a second time series of chirps. The method includes emitting a radio wave by driving a transmission antenna with the transmission signal. The method includes generating a reception signal using a reception antenna in response to receiving a reflected radio wave.

The method includes mixing the transmission signal and the reception signal to generate an intermediate frequency, IF, signal. The method includes generating samples of the IF signal, by sampling the IF signal using an analogue to digital converter, ADC, within each time window of a sequence of time windows associated with the sequence of chirps. The time windows have equidistant midpoints. The method includes processing the samples of the IF signal using a signal processing unit to (i) calculate for each respective time window a range profile by applying a Fourier transform to the samples pertaining to the respective time window, (ii) calculate a matrix of ambiguous radial velocities by applying a Fourier transform across range profiles, the matrix of ambiguous radial velocities comprising a first sub-matrix and a second sub-matrix corresponding respectively to lower and upper velocities, (iii) calculate a sum matrix and a difference matrix, the sum matrix and the difference matrix being a sum and a difference of the first and the second sub-matrix, respectively, (iv) determine a matrix of first phase differences between elements of the sum matrix and elements of the difference matrix, (v) determine a matrix of second phase differences between elements of the matrix of first phase differences and a phase component depending on a measured range position and an ambiguous measured velocity, such that the second phase difference is a function depending only on a cycle index for each position, (vi) determine a cycle index n _cyc (I ,]') for each position of the ambiguous radial velocity, and (vii) calculate an unambiguous radial velocity based on the ambiguous radial velocity and the cycle index.

The method estimates target velocity without an SNR loss, thereby improving the phase quality. The method in an automotive radar system improves target velocity estimation on a single radar scan, thereby reducing network latency especially in safety features of automotive such as emergency braking. The method estimates target velocity in a single scan over multiple channels, thereby enhancing robustness of the automotive radar apparatus. The method incorporates flexibility in waveform design and on a received signal recording. The method may detect possible errors during target velocity estimation and maintains a tracker for the possible errors.

Optionally, generating samples of the IF signal includes generating samples RxTime(i,j ) of the IF signal, i=1,...N, j=1,...M, by sampling the IF signal within each time window W(i) of the sequence of time windows, RxT ime(i,j ) being the j-th sample in the i-th time window.

Optionally, processing the samples of the IF signal includes (i) calculating the range profile by, for each i from 1 to N, calculating a range profile Range_SlowTime(i, J ), J=1,...,M, by applying a Fourier transform to the samples RxTime(i,j ), j=1,...,M, and (ii) calculating ambiguous radial velocities by, for each J from 1 to M, calculating a matrix RgVel(I ) , 1=1, ...,N, by applying the Fourier transform to the range profile Range_SlowTime(i,J), i=1,...,N. The first sub-matrix RgVel _S1 (/',J') = RgVel(I' '),

I'=1,...,N/2, J’=1,...,M, and the second sub-matrix RgVel _S2 (/',J') = RgVel (/' + ) F=1,...,N/2, F=1,...,M.

Optionally, processing the samples of the IF signal includes calculating the sum matrix RgVel _SUM and the difference matrix RgVel _DIFF such that:

RgVel _SUM (/',J') = RgVel _s1 (/',J') + RgVel _S2 (/',J') and RgVel _DIFF (/',J') = RgVel _s1 (/',J') ^~ RgVel _S2 (/',J')

F=1,...,N/2, F=1,...,M.

Optionally, processing the samples of the IF signal includes (i) determining the first phase differences Phase (/',J') , F=1,...,N/2, J’=1,...,M, first phase difference Phase(I' J') being a phase difference between element RgVel _SUM (/',J') of the sum matrix and element RgVel _DIFF (/',J') of the difference matrix, and (ii) determining the second phase differences Phase _VelCylce (/',J') = Phase (/',J') — Phase _known (I',J') F=1,...,N/2, J’=1,...,M, Phase _known (/',J') is the phase component depending on the measured range position and the ambiguous measured velocity, such that Phase _known (I',J') is a function of G and G, Phase _known (I',J') = /(I',J'), and the second phase difference is a function Phase _VelCylce (I',J') = f(n _cyc (l',J ')) depending only on the cycle index n _cyc (I',J') for each G J’ position.

Optionally, a frequency-ramping slope is the same for the first time series of chirps and the second time series of chirps, and where a ramp time for the first and second time series, and an idle time between chirps is adjusted such that the windows sampled from the first time series of chirps include a different center frequency to the windows sampled from the second time series of chirps.

Optionally, a frequency-ramping slope is different for the first time series of chirps and the second time series of chirps, such that the windows sampled from the first time series of chirps include a different center frequency to the windows sampled from the second time series of chirps, and where a size of the windows sampled is adjusted such that a constant bandwidth is sampled.

Optionally, the method further includes detecting a plurality of targets based on the calculated sum matrix or difference matrix. The first phase differences are determined for each target and averaged across a plurality of channels. The method optionally includes estimating a direction of arrival for each detected target. The method can enhance the estimation of the unambiguous velocity in a single scan without impacting memory needs.

Optionally, the method further includes estimating a direction of arrival based on the calculated sum matrix and difference matrix using all channels. The method optionally includes detecting a plurality of targets based on the direction of arrival estimated using the sum matrix or the difference matrix. Optionally, a potential error indication flag is raised based on a value of the second phase difference.

According to a third aspect, there is provided a computer-readable medium that includes instructions which, when executed by a processor, cause the processor to perform the method.

A technical problem in the prior art is resolved, where the technical problem concerns lack of efficiency unambiguous velocity estimation, lack of chirp transmission flexibility and lack of received chirp recording flexibility, and presence of signal to noise ratio loss, estimation is not robust to multiple channels. Therefore, in contradistinction to the prior art, according to the automotive radar apparatus, and the method for determining an unambiguous radial velocity in the automotive radar system, the estimation of the unambiguous velocity in a single scan is enhanced without impacting memory needs. The automotive radar apparatus excludes signal to noise ratio (SNR) loss, thereby the processing of the data becomes easy and weak targets are also detected easily. Thus, an efficient way of estimating unambiguous velocity is implemented by the automotive radar system over multiple available channels due to its robust nature. The automotive radar apparatus can be used to identify possible wrong velocity estimation, thereby improving the efficiency of the automotive radar system. Thus, computational costs can be minimal and effective.

These and other aspects of the disclosure will be apparent from and the implementation(s) described below.

BRIEF DESCRIPTION OF DRAWINGS

Implementations of the disclosure will now be described, by way of example only, with reference to the accompanying drawings, in which:

FIG. 1 is a block diagram of an automotive radar apparatus in accordance with an implementation of the disclosure;

FIG. 2 is a graphical illustration that illustrates an exemplary sequence of frequency- ramped chirps that represents relevant time parameters with a same frequency-ramping slope in accordance with an implementation of the disclosure;

FIG. 3 is a graphical illustration that illustrates an exemplary sequence of frequency- ramped chirps that represents relevant time parameters with a different frequency- ramping slope in accordance with an implementation of the disclosure;

FIG. 4 is a graphical illustration that illustrates an exemplary comparison plot of phase and phase residues for moving target at different Doppler velocities in accordance with an implementation of the disclosure;

FIG. 5 is an exemplary representation of a Doppler enhancement factor as a sector for the automotive radar apparatus in accordance with an implementation of the disclosure; FIG. 6 is a graphical illustration that illustrates an exemplary comparison plot of estimated unwrapped velocity and a ground truth velocity of the automotive radar apparatus in accordance with an implementation of the disclosure;

FIG. 7 is a flow diagram that illustrates a method for determining an unambiguous radial velocity before detection of direction of arrival in an automotive radar system in accordance with an implementation of the disclosure;

FIG. 8 is a flow diagram that illustrates a method for determining an unambiguous radial velocity after detection of direction of arrival in an automotive radar system in accordance with an implementation of the disclosure; and

FIGS. 9A-9D are flow diagrams that illustrate a method for determining an unambiguous radial velocity in an automotive radar system in accordance with an implementation of the disclosure.

DETAILED DESCRIPTION OF THE DRAWINGS

Implementations of the disclosure provide an automotive radar apparatus and a method of determining an unambiguous radial velocity in an automotive radar system. The automotive radar apparatus may be a frequency modulated continuous wave radar.

To make solutions of the disclosure more comprehensible for a person skilled in the art, the following implementations of the disclosure are described with reference to the accompanying drawings.

Terms such as "a first", "a second", "a third", and "a fourth" (if any) in the summary, claims, and foregoing accompanying drawings of the disclosure are used to distinguish between similar objects and are not necessarily used to describe a specific sequence or order. It should be understood that the terms so used are interchangeable under appropriate circumstances, so that the implementations of the disclosure described herein are, for example, capable of being implemented in sequences other than the sequences illustrated or described herein. Furthermore, the terms "include" and "have" and any variations thereof, are intended to cover a non-exclusive inclusion. For example, a process, a method, a system, a product, or a device that includes a series of steps or units, is not necessarily limited to expressly listed steps or units but may include other steps or units that are not expressly listed or that are inherent to such process, method, product, or device.

FIG. 1 is a block diagram of an automotive radar apparatus 100 in accordance with an implementation of the disclosure. The automotive radar apparatus 100 includes a waveform generator 102, a transmission antenna 104, a reception antenna 106, a mixer

108, an analogue to digital converter, ADC, 110, and a signal processing unit 112. The waveform generator 102 is configured to generate a transmission signal including a sequence of frequency-ramped chirps. The sequence of chirps includes a first time series of chirps interleaved with a second time series of chirps. The transmission antenna 104 is configured to emit a radio wave in response to being driven by the transmission signal.

The reception antenna 106 is configured to generate a reception signal in response to receiving a reflected radio wave. The mixer 108 is configured to generate an intermediate frequency, IF, signal by mixing the transmission signal with the reception signal. The analogue to digital converter, ADC, 110 is configured to generate samples of the IF signal, by sampling the IF signal within each time window of a sequence of time windows associated with the sequence of chirps. The time windows have equidistant midpoints.

The signal processing unit 112 is configured to calculate a range profile by applying a

Fourier transform to the samples for each time window. The signal processing unit 112 is configured to calculate a matrix of ambiguous radial velocities by applying a Fourier transform across range profiles. The matrix of ambiguous radial velocities includes a first sub-matrix and a second sub-matrix corresponding respectively to lower and upper velocities. The signal processing unit 112 is configured to calculate a sum matrix and a difference matrix. The sum matrix and the difference matrix are a sum and a difference of the first and the second sub-matrix, respectively. The signal processing unit 112 is configured to determine a matrix of first phase differences between elements of the sum matrix and elements of the difference matrix. The signal processing unit 112 is configured to determine a matrix of second phase differences between elements of the matrix of first phase differences and a phase component depending on the measured range position and an ambiguous measured velocity, such that the second phase difference is a function depending only on a cycle index for each position. The signal processing unit 112 is configured to determine a cycle index n _cyc (I ,]') for each position of the ambiguous radial velocity. The signal processing unit 112 is configured to calculate an unambiguous radial velocity based on the ambiguous radial velocity and the cycle index. The automotive radar apparatus 100 improves target velocity estimation on a single radar scan, thereby network latency reduces, especially in safety features of automotive such as emergency braking. The automotive radar apparatus 100 estimates target velocity in a single scan over multiple channels thereby enhancing robustness of the apparatus is very high. The automotive radar apparatus 100 estimates target velocity without an SNR loss, thereby improving the phase quality. The automotive radar apparatus 100 incorporates flexibility in waveform design and on a received signal recording. The automotive radar apparatus 100 may detect possible errors in estimation and maintains a tracker for the possible errors.

Optionally, the analogue to digital converter, ADC, 110 is configured to generate samples RxTime(i,j ) of the IF signal, i=1,...N, j=1,...M, by sampling the IF signal within each time window W(i) of the sequence of time windows, RxTime(i,j ) being the j-th sample in the i-th time window.

Optionally, index i in the RxTime(i,j ) of the IF signal indicates i-th chirp, row i is recorded data for each transmitted chirp within its time recording window and the columns of the corresponding samples to its specific time window. The i-th index is also known as slow time dimension.

Optionally, index j in the RxTime(i,j ) of the IF signal indicates j-th acquisition window that corresponds to recorded sample within each specific acquisition window. The j-th index is also known as fast time dimension.

Optionally, the signal processing unit 112 is configured to calculate the range profile by, for each i from 1 to N, calculating a range profile Range_SlowTime(i,J ), J=1,.. ,M, by applying a Fourier transform to the samples RxTime(i,j ), j=1,...,M. Optionally, the signal processing unit 112 is configured to calculate ambiguous radial velocities by, for each J from 1 to M, calculating a matrix RgVel(I,J ), 1=1,..., N, by applying the Fourier transform to the range profile Range_SlowTime(i,J ), i=1,...,N. The first sub-matrix RgVel _s1 (I',J') = RgVel(I'J') , F=1,...,N/2, J’=1,...,M, and the second sub-matrix The Fourier transform to the range profile may also calculate a spectrum, that is a range of frequencies. The spectrum may include two peaks (P ₁ , P ₂ ). The values of P ₁ , P ₂ may be obtained by the following equations, where n is a number of ambiguous cycles, V _m is a measured velocity, R _m is a measured range. The signal processing unit 112 may divide the spectrum into a first sub-spectrum, and a second sub-spectrum. The first sub-spectrum and the second sub-spectrum may be two sub-apertures in a Doppler domain. The first sub-spectrum may include from -0.25 to 0.25 ambiguous radial velocities. The second spectrum may include the rest of the spectrum. The signal processing unit 112 may calculate a sum peak ( P ₁ + P ₂ )and a difference peak ( P ₁ — P ₂ )of a first peak of the spectrum and a second peak of the spectrum. A phase difference between the sum peak and the difference peak includes information about the number of ambiguity cycles. The following equation represents the phase difference between the sum peak and the difference peak: ( P ₁ - P ₂ ) ( P ₁ + P ₂ )*) = Φ(n, v _m , R _m )

The phase difference may depend on the number of cycles (n), measured range (R _m ), and measured velocity (V _m ) of a moving target. After the removal of a phase component that relates to the measured range and velocity of the moving target, a remaining phase may be linked to the number of cycles with respect to an unambiguous velocity span.

Optionally, the signal processing unit 112 is configured to calculate the sum matrix RgVel _SUM and the difference matrix RgVel _DIFF such that:

The automotive radar apparatus 100 may detect one or more targets using the sum matrix or the difference matrix using full processing gain. The automotive radar apparatus 100 may estimate unambiguous velocity for the detected targets by analyzing the interferometric phase for a large number of channels, thereby increasing robustness in estimating unambiguous velocity.

Optionally, the signal processing unit 112 is configured to determine the first phase differences Phase(I' ,J') , I=1,...,N/2, F=1,...,M, first phase difference Phase(I' ,J') being a phase difference between element RgVel _SUM (I' ,J') of the sum matrix and element RgVel _DIFF (I' ,J') of the difference matrix. Optionally, the signal processing unit 112 is configured to determine the second phase differences Phase _VelCylce (I' ,J') = Phase(I' ') - Phase _known (I',J'), G=1,...,N/2, J’=1,...,M, Phase _known (I' ,J') is the phase component depending on the measured range position and the ambiguous measured velocity, such that Phase _known (I',J') is a function of G and G, Phase _known (I',J') = /(I',J') _> and the second phase difference is a function Phase _VelCylce (J' ,/') = /(n _cyc (I',J')) depending only on the cycle index n _cyc I',J ') for each G J’ position.

Optionally, the first phase difference is defined as Phase (i',j') = ^RgVel _SUM (/',J') _*

RgVel _DIFF (i',j') , being z the angle operator, * the complex product operator and the conjugate operator. Using matrix notation: Phase = RgVel _SUM RgVel _DIFF ), being the angle operator, the Hadamard product and the conjugate operator.

Optionally, the second phase difference cleans the phase component from the known terms that depend on Rg and ambiguous velocity that is directly linked to the I' and J’ indices. Thereby, the resulting phase depends only on the required unknown parameter of the cycle index.

Optionally, the signal processing unit 112 is further configured to detect one or more targets based on the calculated sum matrix or difference matrix. The first phase differences are determined for each target and averaged across one or more channels. Optionally, the signal processing unit 112 is further configured to estimate a direction of arrival for each detected target. The signal processing unit 112 may estimate unambiguous velocity for one or more detected targets by analyzing the interferometric phase over a large number of channels, thereby increasing robustness in estimating unambiguous velocity over a large number of channels and reducing network latency.

Optionally, the signal processing unit 112 is further configured to estimate a direction of arrival based on the calculated sum matrix and difference matrix using all channels. The signal processing unit 112 may be configured to detect one or more targets based on the direction of arrival estimated using the sum matrix or the difference matrix. If the direction of arrival (DoA) estimation for the detected target takes place easily, then, the phase quality improves with zero SNR loss. The one or more targets may be detected by estimating all sum matrices or difference matrices. The unambiguous radial velocity is estimated using a phase difference between DoA from a sum of the sub-matrices (i.e. velocity sub-matrices) and the DoA from the difference between the sub-matrices. The first phase differences may also be known as an interferometric phase. The interferometric phase is calculated for each moving target and each channel. The phase compensation with residue is derived for each channel and averaged. Thereby, the phase quality becomes equivalent to the coherent gain of all channels. Hence, the automotive radar apparatus 100 is robust to unambiguous velocity estimation for all channels. Also, the direction of arrival (DoA) estimation for the detected target takes place. Optionally, the value of the second phase difference may be used to raise a potential error indication flag. The automotive radar apparatus 100 can enhance the estimation of the unambiguous velocity in a single scan without impacting memory needs.

FIG. 2 is a graphical illustration that illustrates an exemplary sequence of frequency- ramped chirps that represents relevant time parameters with a same frequency-ramping slope in accordance with an implementation of the disclosure. The graphical illustration depicts a frequency of the frequency-ramped chirps on a Y-axis and a time (t) of the frequency-ramped chirps on an X-axis. The graphical illustration depicts an idle time of a first time series of chirps 214 at 202. The graphical illustration depicts a ramp time (t _{ramp A} ) of the first time series of chirps 214 at 204. The graphical illustration depicts an idle time of a second time series of chirps 216 at 206. The graphical illustration of the exemplary comparison plot depicts a ramp time (t _rampB ) of the second time series of chirps 216 at 208. The graphical illustration of the exemplary comparison plot depicts an ADC time of the first time series of chirps at 210. The

ADC time of the first time series of chirps 214 may be a time taken for a midpoint of a ramp in the first time series of chirps 214. The graphical illustration depicts an ADC time ( t _ADCB = t _midB ) of the second time series of chirps 216 at 212. The ADC time of the second time series of chirps 216 may be a time taken for the midpoint of a ramp in the second time series of chirps 216.

Optionally, the frequency-ramping slope is the same for the first time series of chirps 214 and the second time series of chirps 216. The ramp time for the first and second time series, and an idle time between chirps are adjusted such that windows sampled by an ADC from the first time series of chirps 214 include a different center frequency to the windows sampled from the second time series of chirps 216.

The first time series of chirps 214 and the second time series of chirps 216 may be interleaved. A frequency shift may be introduced for the first time series of chirps 214 And the second time series of chirps 216 during sample recording. The time between the sample recording middle points may be always same for the first time series of chirps 214 and the second time series of chirps 216 that is (Δt _midAB = Δt _midBA ). Thus, the following equation holds for the same frequency-ramping slope:

Based on the relevant time parameters, Doppler Enhancement factor with respect to original unambiguous Doppler span ( DE _f ) may be given as follows:

For example, with an enhancing factor of 5.5, the first time series of chirps 214 may have t _rampA = 58μs, t _{ID A=} 8 μs, t _Dc = 4μs, and the second time series of chirps 216 may have

For example, if a pulse repetition interval (PRI) of the automotive radar apparatus is 66 ps, with a carrier frequency of 76.5 Giga Hertz (GHz) gives a total velocity span of 106.9 km/h. The enhanced unambiguous span with this configuration may be 587.8 kilometer per hour (km/h).

In an exemplary scenario, the two ramps may have the same duration, t _rampA = t _rampB then, the previous equation is simplified to:

FIG. 3 is a graphical illustration that illustrates an exemplary sequence of frequency- ramped chirps that represents relevant time parameters with a different frequency- ramping slope in accordance with an implementation of the disclosure. The graphical illustration depicts a frequency of frequency-ramped chirps on a Y-axis and a time (t) of the frequency-ramped chirps on an X-axis. The graphical illustration depicts an idle time (t _IDA ) of a first time series of chirps 314 at 302. The graphical illustration depicts a ramp time ( t _rampA ) of the first time series of chirps 314 at 304. The graphical illustration depicts an idle time (t _IDB ) of a second time series of chirps 316 at 306. The graphical illustration depicts a ramp time ( t _rampB ) of the second time series of chirps 316 at 308. The graphical illustration depicts a bandwidth ( Bw _A ) of the first time series of chirps 314 at 310. The graphical illustration depicts a bandwidth ( BW _B ) of the second time series of chirps 316 at 312. The first time series of chirps 314 and the second time series of chirps 316 may be interleaved. A frequency shift may be introduced for the first time series of chirps 314 and the second time series of chirps 316 during sample recording. Optionally, the frequency-ramping slope is different for the first time series of chirps 314 and the second time series of chirps 316, such that windows sampled by the ADC from the first time series of chirps 314 include a different center frequency to the windows sampled from the second time series of chirps 316, and where a size of the windows sampled by the ADC is adjusted such that a constant bandwidth is sampled. The time between the sample recording middle points may be always same for the first time series of chirps 314 and the second time series of chirps 316 that is (Δt _midAB = Δt _midBA).

In an exemplary scenario, the first time series of chirps 314 may have a slope k _A , and the second time series of chirps may have a slope k _B . The recording times can be defined for the first time series of chirps 314 and the second time series of chirps 316 as t _REcA = where N _Rg is the number of range samples, and are the sampling frequencies for the first time series of chirps 314 and the second time series of chirps 316 respectively. Thereby, for the first time series of chirps 314 and the second time series of chirps 316 with different frequency-ramping slope the following equation applies, bandwidth of 320 MHz, yielding in a range resolution of 46.8 cm. FIG. 4 is a graphical illustration that illustrates an exemplary comparison plot of phase and phase residues for moving targets at different Doppler velocities in accordance with an implementation of the disclosure. The graphical illustration depicts an original interferometric phase for the moving targets at 402. The graphical illustration depicts a phase estimation for the moving targets according to its estimated range position at 404. The graphical illustration depicts a Doppler velocity phase estimation for the moving targets at 406. The graphical illustration depicts a phase compensation and a Doppler velocity phase compensation for the moving targets according to its estimated range position at 408. Optionally, a phase difference depends on a number of cycles, measured range, and velocity of the moving targets. Thereby, after removing phase component related to the measured range and velocity of the moving targets, the remaining phase may be linked to the number of cycles with respect to unambiguous velocity span. The graphical illustration depicts a phase compensation for the moving targets according to its estimated range and the Doppler velocity at 410. The phase compensation for the moving targets may fall within certain values based on the unambiguous velocity enhancing factor. The unambiguous velocity enhancing factor may be considered in waveform designing.

FIG. 5 is an exemplary representation of a Doppler enhancement factor as a sector for an automotive radar apparatus in accordance with an implementation of the disclosure. The exemplary representation of the circle is divided into sectors with the Doppler enhancement factor of 4.5. The number of cycles (n _cyc ) may be from -4 to 4. Each sector in the exemplary representation depicts the number of cycles. The exemplary representation depicts the number of cycles that may not set in a cyclic order, thereby, a phase noise may produce a jump from one sector to another.

FIG. 6 is a graphical illustration that illustrates an exemplary comparison plot of estimated unwrapped velocity and a ground truth velocity of an automotive radar apparatus in accordance with an implementation of the disclosure. The graphical illustration depicts the estimated unwrapped velocity of the automotive radar apparatus on a Y-axis and the ground truth velocity of the automotive radar apparatus on an X-axis. The graphical illustration depicts an estimated velocity of the automotive radar apparatus at 602, and an unambiguous region at 604.

FIG. 7 is a flow diagram that illustrates a method for determining an unambiguous radial velocity before the detection of direction of arrival in an automotive radar system in accordance with an implementation of the disclosure. At a step 702, a range profile is calculated by applying a Fourier transform to the samples for each time window. At a step 704, a Doppler profile is calculated by applying the Fourier transform for the one or more targets. At a step 706, upper and lower velocity sub-matrices are generated. At a step 708, the generated sub-matrices are summed and subtracted. A sum matrix and a difference matrix are calculated. The sum matrix and the difference matrix are a sum and a difference of a first and a second sub-matrix, respectively corresponding to lower and upper velocities. At a step 710, the one or more targets are detected using all sums or subtractions of the sub-matrix. At a step 712, an interferometric phase is calculated using all channels for the detected targets. At a step 714, a sub-spectrum cycle or an unambiguous radial velocity is calculated based on the ambiguous radial velocity and the cycle index. At a step 716, a direction of arrival, DOA, is estimated for the detected targets. The method derives residual phase for every channel and averaged, thereby the phase quality is equivalent to beamform all the channels, and hence enhancing robustness of estimation of unambiguous velocity of the automotive radar apparatus.

FIG. 8 is a flow diagram that illustrates a method for determining an unambiguous radial velocity after the detection of direction of arrival in an automotive radar system in accordance with an implementation of the disclosure. At a step 802, a range profile is calculated by applying a Fourier transform to the samples for each time window. At a step 804, a Doppler profile is calculated by applying the Fourier transform for the one or more targets. At a step 806, upper and lower velocity sub-matrices are generated. At a step 808, the generated sub-matrices are summed and subtracted. A sum matrix and a difference matrix are calculated. The sum matrix and the difference matrix are a sum and a difference of a first and a second sub-matrix, respectively corresponding to lower and upper velocities. At a step 810, a direction of arrival, DOA, a is estimated based on the calculated sum matrix using all channels and a DOA is estimated based on the calculated difference matrix using all channels. At a step 812, the one or more targets are detected based on the DOA estimated using the sum matrix or the difference matrix. At a step 814, an interferometric phase is calculated using all channels for the detected targets. At a step 816, a sub-spectrum cycle or unambiguous velocity is estimated for the detected targets. The method estimates target velocity in a single scan over multiple channels, thereby enhancing robustness of estimation of unambiguous velocity of the automotive radar system.

FIGS. 9A-9D are flow diagrams that illustrate a method for determining an unambiguous radial velocity in an automotive radar system in accordance with an implementation of the disclosure. At a step 902, a transmission signal including a sequence of frequency- ramped chirps is generated using a waveform generator. The sequence of chirps includes ' first time series of chirps interleaved with a second time series of chirps. At a step 904, a radio wave is emitted by driving a transmission antenna with the transmission signal. At a step 906, a reception signal is generated using a reception antenna in response to receiving a reflected radio wave. At a step 908, the transmission signal and the reception signal are mixed to generate an intermediate frequency, IF, signal. At a step 910, samples of the IF signal are generated, by sampling the IF signal using an analogue to digital converter, ADC, within each time window of a sequence of time windows associated with the sequence of chirps. The time windows have equidistant midpoints.

At a step 912, the samples of the IF signal are processed using a signal processing unit.

At a step 914, a range profile is calculated by applying a Fourier transform to the samples for each time window. At a step 916, a matrix of ambiguous radial velocities is calculated by applying a Fourier transform across the range profiles. The matrix of ambiguous radial velocities includes a first sub-matrix and a second sub-matrix corresponding respectively to lower and upper velocities. At a step 918, a sum matrix and a difference matrix, are calculated respectively. The sum matrix and the difference matrix are a sum and a difference of the first and the second sub-matrix. At a step 920, a matrix of first phase differences between elements of the sum matrix and elements of the difference matrix are determined. At a step 922, a matrix of second phase differences between elements of the matrix of first phase differences and a phase component depending on the measured range position and an ambiguous measured velocity are determined, such that the second phase difference is a function depending only on a cycle index for each position. At a step 924, a cycle index n _cyc (I 'J') is determined for each position of the ambiguous radial velocity. At a step 926, an unambiguous radial velocity is calculated based on the ambiguous radial velocity and the cycle index.

The method estimates target velocity without an SNR loss, thereby improving the phase quality. The method improves target velocity estimation on a single radar scan, thereby reducing network latency especially in safety features of automotive such as emergency braking. The method estimates target velocity in a single scan over multiple channels, thereby enhancing robustness of the automotive radar apparatus. The method incorporates flexibility in waveform design and on a received signal recording. The method may detect possible errors during target velocity estimation and maintains a tracker for the possible errors. Optionally, generating samples of the IF signal includes generating samples RxTime(i,j) of the IF signal, i=1,...N, j=1,...M, by sampling the IF signal within each time window W(i) of the sequence of time windows, RxT ime(i,j ) being the j-th sample in the i-th time window.

Optionally, processing the samples of the IF signal includes (i) calculating the range profile by, for each i from 1 to N, calculating a range profile Range_SlowTime(i, J ), J=1,...,M, by applying a Fourier transform to the samples RxTime(i,j ), j=1,...,M, and (ii) calculating ambiguous radial velocities by, for each J from 1 to M, calculating a matrix RgVel(IJ) , 1=1, ...,N, by applying the Fourier transform to the range profile Range_SlowTime(i,J ), i=1,...,N, the first sub-matrix RgVel _s1 (I',J') = RgVel(I'J'),

Optionally, processing the samples of the IF signal comprises calculating the sum matrix RgVel _SUM and the difference matrix RgVel _DIFF such that:

RgVel _SUM (I',J') = RgVel _s1 (I',J') + RgVel _S2 (I',J') and

RgVel _DIFF (I',J') = RgVel _s1 (I',J') - RgVel _s2 (I',J,')

Optionally, processing the samples of the IF signal includes (i) determining the first phase differences Phase ( (I',J')), F=1,...,N/2, J’=1,...,M, first phase difference Phase(I' J') being a phase difference between element RgVel _SUM (I' ,/') of the sum matrix and element RgVel _DIFF (I',J') of the difference matrix, and (ii) determining the second phase differences Phase _VelCylce (I',J') = Phase(PJ') — Phase _known (I',J') F=1,...,N/2, J’=1,...,M, and Phase _known (I' ,J') is the phase component depending on the measured range position and the ambiguous measured velocity, such that Phase _known (I',J') is a function of F and J’, Phase _known (I',J') = /(I',J'), and the second phase difference is a function Phase _VelCylce (I',J') = f(n _cyc (I',J ')) depending only on the cycle index n _cyc (I',J ') for each F J’ position. The ambiguous velocity is estimated using the second phase difference, which only depends on the cycle index (a number of ambiguity cycles). The quality of this second phase difference can be improved by averaging all the available channels, thereby increasing robustness in estimating unambiguous velocity.

Optionally, a frequency-ramping slope is the same for the first time series of chirps and the second time series of chirps, and where a ramp time for the first and second time series, and an idle time between chirps is adjusted such that the windows sampled from the first time series of chirps include a different center frequency to the windows sampled from the second time series of chirps.

Optionally, a frequency-ramping slope is different for the first time series of chirps and the second time series of chirps, such that the windows sampled from the first time series of chirps includes a different center frequency to the windows sampled from the second time series of chirps, and where a size of the windows sampled is adjusted such that a constant bandwidth is sampled.

Optionally, the method further includes detecting one or more targets based on the calculated sum matrix or difference matrix. The first phase differences are determined for each target and averaged across one or more channels. The method may include estimating a direction of arrival for each detected target. The method can enhance the estimation of the unambiguous velocity in a single scan without impacting memory needs.

Optionally, the method further includes estimating a direction of arrival based on the calculated sum matrix and difference matrix using all channels. The method may include detecting one or more targets based on the direction of arrival estimated using the sum matrix or the difference matrix. Optionally, a potential error indication flag is raised based on a value of the second phase difference.

A computer-readable medium that includes instructions that, when executed by a processor, cause the processor to perform the above method. Although the disclosure and its advantages have been described in detail, it should be understood that various changes, substitutions, and alterations can be made herein without departing from the spirit and scope of the disclosure as defined by the appended claims.