Method and device for identifying rocking object in radar data

Abstract

The invention relates to a method and a device for identifying a rocking object in radar data. A method of determining the presence of one or more wobble objects within a distance interval of a plurality of distance intervals based on a plurality of distance-Doppler maps provided by a radar system is disclosed. Each of the plurality of distance-Doppler maps corresponds to a time interval in the sequence of time intervals and includes respective energy values for a plurality of velocity intervals within each of the plurality of distance intervals. For each time interval, a difference between a statistical measure of energy values of a set of velocity intervals having a positive velocity in a plurality of velocity intervals within a distance interval and a statistical measure of energy values of a set of velocity intervals having a negative velocity in a plurality of velocity intervals within the distance interval is calculated from a distance-Doppler map corresponding thereto. A spectrum is then determined from the calculated sequence of differences. It is determined that one or more wobble objects are present at the distance interval under the condition that there is a peak in the spectrum for frequencies above a frequency threshold.

Description

TECHNICAL FIELD

[0001] The present invention relates to identifying swaying objects in radar data, and in particular to determining, based on a plurality of range-Doppler maps provided by a radar system, that one or more swaying objects are present within a range interval of a plurality of range intervals. BACKGROUND

[0002] When monitoring a scene by a surveillance system using radar, it can be desirable to only detect and / or present moving objects in the scene. However, one problem in the prior art is that there is no distinction made between moving objects such as cars or people and objects such as grass, trees and poles that are swaying due to wind, which are usually not of interest. It is therefore desirable to be able to identify these swaying objects, e.g. such that the swaying objects are not identified as detections, or such that any detections related to the swaying objects are not further processed, such as by a tracking algorithm or a classifier. SUMMARY

[0003] It is an object of the present invention to overcome or at least alleviate the problems and drawbacks of the prior art.

[0004] According to a first aspect, there is provided a method for determining, based on a plurality of range-Doppler maps provided by a radar system, that one or more swaying objects are present within a range interval of a plurality of range intervals. Each range-Doppler map of the plurality of range-Doppler maps corresponds to a time interval of a sequence of time intervals and comprises a respective energy value for a plurality of velocity intervals within each range interval of the plurality of range intervals. The method comprises, for each time interval of the sequence of time intervals, computing, from the range-Doppler map corresponding to the time interval, a difference between a statistical measure of energy values of a set of velocity intervals with positive velocities and a statistical measure of energy values of a set of velocity intervals with negative velocities of the plurality of velocity intervals within the range interval, thereby computing a sequence of differences for the sequence of time intervals. The method further comprises determining a spectrum of the computed sequence of differences for the sequence of time intervals, and determining that one or more swaying objects are present at the range interval on condition that there is a peak in the spectrum for a frequency higher than a frequency threshold.

[0005] The wobbling object typically has a periodic motion, such that its motion varies between positive and negative velocities. Therefore, the energy values in the plurality of range-Doppler plots typically vary, such that in the range interval in which the wobbling object is present, the energy values are periodically higher for positive velocities and periodically higher for negative velocities. Therefore, the difference of the statistical measures of the energy values of the positive and negative velocities of the plurality of range-Doppler plots typically also varies periodically. The difference of the statistical measures varies with the same frequency as the frequency at which the wobbling object is wobbling. Therefore, by computing a sequence of differences of the statistical measures of the energy values of the positive and negative velocities for a sequence of time intervals and then determining a spectrum of the computed sequence of differences, the wobbling object at a particular frequency can be identified as a peak in the spectrum at that particular frequency.

[0006] Identifying that one or more wobbling objects are present within the range interval enables avoiding that the detections related to the one or more wobbling objects are further processed, such as by a tracking algorithm or a classifier. Furthermore, presenting trajectories related to the one or more wobbling objects can be avoided.

[0007] Statistical measure refers to a numerical value that summarizes a characteristic of a data set.

[0008] Spectrum refers to a representation of a signal, in this case, to the computed sequence of differences in the frequency domain. In particular, it can refer to an amplitude spectrum that describes the distribution of the amplitudes of the frequency components in the signal.

[0009] In embodiments, the method according to the first aspect further comprises, on condition that it is determined that one or more wobbling objects are present at the range interval, filtering out any detections in velocity intervals of the plurality of velocity intervals within the range interval that have an absolute velocity that is smaller than a first velocity threshold.

[0010] By filtering out, on condition that it is determined that one or more wobbling objects are present at the range interval, detections in the range interval, detections related to the one or more wobbling objects, which are typically not related to a user of the radar system, such as an operator of a surveillance system comprising the radar system, are not used in the subsequent processing or visualization of the detections. Furthermore, by filtering out only any detections in velocity intervals of the plurality of velocity intervals within the range interval that have an absolute velocity that is smaller than a first velocity threshold, any detections having an absolute velocity that is larger than or equal to the velocity threshold are not filtered out. Such detections can for example relate to other moving objects than the one or more wobbling objects.

[0011] The detection here generally refers to that an energy value in a velocity interval at a distance interval is considered to be related to an object. This can be determined based on the energy value exceeding an energy threshold, which generally means that the energy value is considered to be high enough to be related to an object, for example. The energy threshold can be based on a noise level, and the energy value exceeding the energy threshold can mean that a signal-to-noise ratio of the energy value exceeds an SNR threshold, for example. The energy threshold (SNR threshold) can be different for different velocity intervals, and the energy threshold (SNR threshold) for one velocity interval can depend on energy values of other velocity intervals. The energy threshold (SNR threshold) can be selected to provide a probability of detection of a moving object that is higher than a desired probability of detection and a probability of false detection that is lower than a desired probability of false detection.

[0012] In an embodiment, the method according to the first aspect further comprises, on condition that it is determined that there is a wobbling object at the distance interval and that there is no peak in the frequency spectrum for frequencies below a frequency threshold, filtering out energy values of velocity intervals having an absolute velocity smaller than a first velocity threshold among the plurality of velocity intervals within the distance interval.

[0013] By setting the condition that there is no peak in the frequency spectrum for frequencies below a frequency threshold for detections to be filtered out, it can be prevented that detections related to objects other than the one or more wobbling objects are filtered out. Such other objects can be objects moving at low radial speed relative to the radar, for example. The frequency threshold can be set based on a lowest expected frequency at which the one or more wobbling objects wobble, for example.

[0014] The first velocity threshold can be fixed. The velocity threshold can be set based on a maximum expected velocity of the one or more wobbling objects, for example. Then, detections related to absolute velocities higher than the first velocity threshold will be considered to be related to moving objects that are not wobbling.

[0015] In an embodiment, the method according to the first aspect further comprises, for each velocity interval at the distance interval, aggregating the energy values of at least a subset of the plurality of range-Doppler maps based on an aggregation function, thereby deriving an aggregated energy value for each velocity interval at the distance interval. Then, the maximum absolute velocity is identified that is such that there is a velocity interval with positive velocity and a velocity interval with negative velocity among the plurality of velocity intervals within the distance interval that have the maximum absolute velocity and an aggregated energy value that exceeds an aggregated energy threshold. Then, the first velocity threshold is set based on the maximum absolute velocity.

[0016] By identifying the maximum absolute velocity for which it is determined that there is a peak in the spectrum for frequencies above the frequency threshold, the group with the highest absolute velocity is considered to be related to the highest group, which is related to the one or more wobbling objects, and the highest absolute velocity of the group can be used as an estimate of the maximum absolute velocity of the one or more wobbling objects.

[0017] In embodiments, the method according to the first aspect further comprises: dividing, dependent on their absolute velocities, the velocity intervals within the range interval of each range-Doppler plot of the plurality of range-Doppler plots into a plurality of groups. The groups are ordered such that the velocity intervals in a group of higher order have a larger absolute velocity than the velocity intervals in a group of lower order. The plurality of groups are then processed in order by the following steps: for the group currently processed: for each time interval in the sequence of time intervals, compute the difference between the statistical measure of the energy values of the velocity intervals with positive velocity in the group of velocity intervals within the range interval and the statistical measure of the energy values of the velocity intervals with negative velocity in the group of velocity intervals within the range interval from the range-Doppler plot corresponding to the time interval, thereby computing a sequence of differences for the sequence of time intervals. Then, compute the spectrum of the computed sequence of differences for the sequence of time intervals. In case there is a peak in the spectrum for frequencies above the frequency threshold, the next group in the order is processed. In case there is no peak in the spectrum for frequencies above the frequency threshold, the first velocity threshold is set to a value corresponding to the maximum absolute velocity of the velocity intervals in the previous group in the order.

[0018] By identifying the group with the highest absolute velocity for which it is determined that there is a peak in the spectrum for frequencies above the frequency threshold, the group is considered to be related to the highest group, which is related to the one or more wobbling objects, and the highest absolute velocity of the group can be used as an estimate of the maximum absolute velocity of the one or more wobbling objects.

[0019] In embodiments, in the act of computing the difference, the difference between the statistical measure of the energy values of the group of velocity intervals with positive velocity within the range interval having an absolute velocity smaller than the second velocity threshold and the statistical measure of the energy values of the group of velocity intervals with negative velocity within the range interval having an absolute velocity smaller than the second velocity threshold is computed.

[0020] By computing the difference based on only the energy values of the group of velocity intervals having an absolute velocity smaller than the second velocity threshold, the computational load is reduced.

[0021] The statistical measure can be one of a standard deviation, a variance, a mean, a median, and a sum of squares.

[0022] According to a second aspect, there is provided a non-transitory computer- readable storage medium having stored thereon instructions which, when executed in an apparatus having processing capability, cause the apparatus to perform the method according to the first aspect.

[0023] The above-mentioned optional additional features of the method according to the first aspect also apply to the non-transitory computer-readable storage medium according to the second aspect, when appropriate. For the sake of brevity, reference is made to the above.

[0024] According to a third aspect, there is provided an apparatus for determining, based on a plurality of range-Doppler maps provided by a radar system, that one or more wobble objects are present within a range interval of a plurality of range intervals. Each range-Doppler map of the plurality of range-Doppler maps corresponds to a time interval of a sequence of time intervals and comprises a respective energy value for a plurality of velocity intervals within each range interval of the plurality of range intervals. The apparatus comprises circuitry configured to perform a computing function, a first determining function, and a second determining function. The computing function is configured to, for each time interval of the sequence of time intervals, compute, from the range-Doppler map corresponding to the time interval, a difference between a statistical measure of energy values of a set of velocity intervals with positive velocities of the plurality of velocity intervals within the range interval and a statistical measure of energy values of a set of velocity intervals with negative velocities of the plurality of velocity intervals within the range interval. The first determining function is configured to determine a spectrum of the sequence of computed differences. The second determining function is configured to determine, on condition that a peak is present in the spectrum for a frequency higher than a frequency threshold, that one or more wobble objects are present at the range interval.

[0025] The above-mentioned optional additional features of the method according to the first aspect also apply to the apparatus according to the third aspect, when appropriate. For the sake of brevity, reference is made to the above.

[0026] It should therefore be understood that the application is not limited to the particular combinations of systems described or to the particular acts described for the methods described, as such systems and methods can vary. It should also be understood that the terminology used herein is for the purpose of describing particular embodiments only, and is not intended to be limiting. It must be noted that, as used in the specification and the appended claims, the articles "a," "an," and "the" are intended to mean one or more of the elements to which the articles refer and the articles "a" and "an" are therefore to be construed in the same manner as the definitive articles "one" and "the." As used herein, the terms "comprise," "comprising," "include," "including," "contain," "containing," "have," "having," and the like are not intended to be limiting. It is specifically intended that any such articulation does not, without further limitation, exclude other elements or steps. BRIEF DESCRIPTION OF DRAWINGS

[0027] The above and other aspects of the application will now be described in more detail with reference to the accompanying drawings. The drawings are not intended to be limiting, but are for explanation and understanding.

[0028] Figure 1a and Figure 1b A flowchart is shown in relation to an embodiment of a method for determining, based on a plurality of range-Doppler plots provided by a radar system, that there is one or more swaying object within a range interval of a plurality of range intervals.

[0029] Figure 2 An example of a range-Doppler plot is shown.

[0030] Figure 3a and Figure 3b A plot of the difference between the standard deviation of the energy values of the positive velocity intervals and the standard deviation of the energy values of the negative velocity intervals of a set of frames related to a time interval and a plot of the discrete Fourier transform of these differences are shown, respectively.

[0031] Figure 4 A schematic diagram is shown in relation to an embodiment of an apparatus for determining, based on a plurality of range-Doppler plots provided by a radar system, that there is one or more swaying object within a range interval of a plurality of range intervals. DETAILED DESCRIPTION

[0032] The present invention will now be described with reference to the accompanying drawings, in which the current preferred embodiments of the invention are illustrated. The invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein.

[0033] Embodiments of the invention are applicable in scenarios where a radar system is used for monitoring an area, e.g. in order to identify moving objects based on range-Doppler plots of a sequence of time intervals. Some of the moving objects can be swaying objects such as grass, trees, poles, etc. that sway due to the wind, and these swaying objects are typically of less interest for the monitoring purpose, and it is therefore of interest to identify these swaying objects and related radar data. The identified swaying objects and related radar data can be used to implement avoiding visualizing such swaying objects to a user, in order to avoid that such swaying objects distract the user’s attention when there is low or no interest in them. Additionally or alternatively, such radar data can be excluded from further processing in the monitoring system, in order to reduce the amount of computations in the monitoring system.

[0034] A swaying object is an object that moves to and fro and / or back and forth relative to the radar.

[0035] The radar system is of a type that is capable of determining the velocity of objects detected by the radar apparatus. For example, the radar of the radar system can be a frequency-modulated continuous wave (FMCW) radar that uses a short transmission signal (chirp) whose frequency varies over time, typically the frequency is stepped up or stepped down. As another example, the radar can be a phase-modulated continuous wave (PMCW) radar apparatus.

[0036] The radar system is further of the type that is able to determine a direction vector from a radar device of the radar system to an object detected by the radar. For example, the radar can be a multiple-input multiple-output (MIMO) radar device.

[0037] A range-Doppler plot for a time interval is a two-dimensional matrix of radar data that presents energy values from objects reflected at different ranges (radial distances) along a first axis (dimension), e.g. the y-axis, and having different radial velocities along a second axis (dimension), e.g. the x-axis. More specifically, for each range interval in a sequence of consecutive range intervals, e.g. starting from 0 up to a maximum range, along the first axis, an energy value is provided for each velocity interval in a sequence of consecutive velocity intervals, e.g. starting from a maximum negative velocity up to a maximum positive velocity. For example, the maximum negative velocity and the maximum positive velocity have the same absolute velocity. With respect to the range-Doppler plot, the velocity intervals within a range interval can be denoted as “bins”, and the zero velocity can be denoted as “DC Doppler”. The range intervals are intervals from a lower range to a higher range. Similarly, the velocity intervals are intervals from a lower velocity to a higher velocity. Thus, the energy value of a velocity interval at a range interval relates to all velocities from the lower velocity to the higher velocity within all ranges from the lower range to the higher range of the range interval.

[0038] Furthermore, the energy values are signal values in the range-Doppler plot. Thus, the energy value of a velocity interval at a range interval relates to signal values detected for all velocities in that velocity interval at all ranges in that range interval.

[0039] Since the wobble objects are identified based on the range-Doppler plot, only wobble objects having a wobble component in the radial direction relative to the radar will be identified. On the other hand, wobble objects that do not have any radial component of wobble motion in the range-Doppler plot are considered static objects and thus will not result in any detections. For example, for an FMWC radar, each time interval can correspond to a radar frame, and typically can be 10 ms.

[0040] Reference will now be made to Figure 1a and Figure 1bThe flowchart in Fig. 1 describes an embodiment of a method 100 of determining, based on a plurality of range-Doppler plots provided by a radar system, whether one or more wobble objects are present in a range interval of a plurality of range intervals. Each range-Doppler plot of the plurality of range-Doppler plots corresponds to a time interval i of a sequence of time intervals. For example, the sequence of time intervals can be a sequence of N time intervals preceding a current time, where N is chosen high enough to enable identification of wobble objects in case they are present. Preferably, the time intervals should have the same length. Furthermore, the time intervals can be such that the beginning of each time interval directly follows the end of the preceding time interval in the sequence of time intervals. Alternatively, the beginning of each time interval can follow an intermediate time from the end of the preceding time interval in the sequence of time intervals. In the latter case, the intermediate times between the time intervals in the sequence should have equal length.

[0041] Each range-Doppler plot comprises a respective energy value for a plurality of velocity intervals in each range interval of a plurality of range intervals. The velocity intervals in each range interval can for example be consecutive velocity intervals from a negative velocity with a maximum absolute velocity to a positive velocity with a maximum absolute velocity. Figure 2 An example of a range-Doppler plot is shown with velocity intervals (Doppler bins) along the horizontal axis and range intervals (range bins) along the vertical axis. The energy values for the velocity intervals at each range interval are indicated, such that in the velocity intervals, higher energy values are indicated by brighter white color. The velocity interval at zero velocity indicated brighter to the left of zero velocity represents objects with negative radial velocity relative to the radar, and the velocity interval at zero velocity indicated brighter to the right of zero velocity represents objects with positive radial velocity relative to the radar.

[0042] The method 100 is performed for a range interval to determine whether one or more wobble objects are present in that range interval. To determine whether one or more wobble objects are present in other range intervals, the method 100 is repeated for those range intervals. In the following, a velocity interval with positive velocity will be denoted as positive velocity interval, and a velocity interval with negative velocity will be denoted as negative velocity interval.

[0043] The method 100 comprises computing S120 a difference sequence for the sequence of time intervals (i = 1 to N). Starting S110 from a first time interval (i = 1), a difference is computed S120 based on the range-Doppler plot corresponding to the first time interval (i = 1). The computed difference is the difference between a statistical measure of the energy values of a set of positive velocity intervals in the range interval and a statistical measure of the energy values of a set of negative velocity intervals in the range interval.

[0044] The computed statistical measure can be, for example, a standard deviation, a variance, a mean, a median, or a sum of squares. Preferably, a statistical measure of order two or higher is used.

[0045] When computing S120 the differences for each time interval i, the set of positive velocity intervals can be all positive velocity intervals and the set of negative velocity intervals can be all negative velocity intervals. However, to reduce the required computations, the set of positive velocity intervals can be only a subset of all positive velocity intervals and the set of negative velocity intervals can be only a subset of negative velocity intervals. For example, the subset of all positive velocity intervals can be all or a subset of positive velocity intervals having an absolute velocity less than a second velocity threshold and the subset of all negative velocity intervals can be all or a subset of negative velocity intervals having an absolute velocity less than the second velocity threshold. The second velocity threshold can be selected based on an expected maximum absolute velocity of the wobbling object. If the wobbling object is expected to always have an absolute velocity below the maximum absolute velocity, the second velocity threshold can be set to the maximum absolute velocity because the wobbling object is not expected to contribute to the energy values of positive velocity intervals and negative velocity intervals in cases where the absolute velocity is greater than the expected maximum absolute velocity. The second velocity threshold can also be set to a value lower than the maximum absolute velocity expected for the wobbling object because this will include at least some of the velocity intervals in which the wobbling object will contribute to the energy values of at least some of the range-Doppler plots. The set of positive intervals can also be a subset of positive velocity intervals, such as every other positive velocity interval or every other two positive velocity intervals or other selection, and the set of negative velocity intervals can also be a subset of negative velocity intervals, such as every other negative velocity interval or every other two negative velocity intervals or other selection.

[0046] The computation S120 of the differences is repeated for each next time interval (i = i + 1) S125 as long as the number of the next time interval is not higher than the total number of time intervals (i > N) C128. Under the condition that the next time interval is higher than the total number of time intervals (i > N) C128, the sequence of differences for the sequence of time intervals (i = 1 to N) has been computed.

[0047] Turning to Figure 3a , a sequence of differences of standard deviations of energy values of a set of positive velocity intervals within a range interval and a statistical measure of energy values of a set of negative velocity intervals within the range interval for a plurality of range-Doppler plots each having a respective frame number and relating to a respective one of the sequence of time intervals is shown. As can be seen in the figure, the differences of the standard deviations alternate between positive values and negative values. This indicates that there is motion in the range interval that alternates relative to the radar between positive radial directions and negative radial directions. Such motion that alternates relative to the radar between positive radial directions and negative radial directions would, for example, be a wobbling motion of an object at the range interval.

[0048] Turning back to Figure 1a Then, the method 100 continues with determining S130 a spectrum of the computed sequence of differences of the sequence of time intervals. In this spectrum, an amplitude for each frequency is given based on the computed sequence of differences for the sequence of time intervals. In this plot, a peak of the amplitude related to a difference at a certain frequency indicates an oscillation of the difference at this frequency. The amplitude can also be referred to as the magnitude.

[0049] The spectrum can be computed by a transformation into the frequency domain, e.g. by performing a Fast Fourier Transform (FFT) on the sequence of differences. For example, the spectrum computed by the FFT can be plotted in a plot having frequencies along a first axis, e.g. the x-axis, and amplitudes of the transformed values along a second axis, e.g. the y-axis. In this plot, a peak of the amplitude at a certain frequency indicates an oscillation of the difference at this frequency.

[0050] In a condition C140 that there is a peak in the spectrum for a frequency above a frequency threshold, it is determined S145 that there is one or more wobbling objects at the distance interval.

[0051] A peak in the spectrum at a frequency is present if the amplitude at this frequency related to a noise level of the spectrum is above a threshold, that is, if the signal-to-noise ratio (SNR) at this frequency is above a threshold. The noise level in the spectrum can be determined based on the amplitudes of each frequency. For example, the noise level can be determined as the median of the amplitudes of all frequencies. In embodiments where the spectrum is computed by the FFT, the amplitude of each frequency is the transformed value of the amplitude of this frequency.

[0052] Turning back to Figure 3b , a plot of the FFT of the differences of Figure 3a In this plot, a peak at about 1 Hz corresponding to one or more wobbling objects is shown.

[0053] Turning back to Figure 1a In a condition C140 that there is no peak in the spectrum for a frequency above a frequency threshold, it is determined S148 that there is no wobbling object at the distance interval.

[0054] Once it is determined S145 that there is one or more wobbling objects in the distance interval, the further processing of the energy values of the speed intervals at the distance interval can be modified, e.g. to reduce the amount of computation and / or to reduce confusion in the visualization in the monitoring system. For example, the energy values identified or assumed to be related to one or more wobbling objects can be identified and optionally filtered out, i.e. removed, from the further processing, e.g. such that it is not used for the detection of objects, the visualization in the monitoring system, the tracking of objects, the identification of object types, etc.

[0055] For example, turning to Figure 1bAny detections in the plurality of velocity bins having an absolute velocity less than the first velocity threshold in the condition C150 that there is one or more wobble objects at the distance interval can be filtered out S158.

[0056] Detection here generally refers to that an energy value in a velocity bin at a distance interval is considered to be related to an object. For example, this can be determined based on the energy value exceeding an energy threshold, which generally herein means that the energy value is considered to be high enough to be related to an object. For example, the energy threshold can be based on a noise level, and the energy value exceeding the energy threshold can mean that the signal-to-noise ratio of the energy value exceeds an SNR threshold. The energy threshold (SNR threshold) can be different for different velocity bins, and the energy threshold (SNR threshold) of one velocity bin can depend on energy values of other velocity bins. The energy threshold (SNR threshold) can be chosen such that the probability of a moving object resulting in a detection is higher than a desired detection probability, and the probability of a false detection is lower than a desired false detection probability. The identification of detections is generally performed for all velocity bins within a distance interval, such that any further processing in the monitoring system and any visualization to a user is based on the detections.

[0057] The first velocity threshold can be a fixed threshold. The fixed first velocity threshold can be set based on a maximum velocity that a wobble object is expected to move when wobbling. More specifically, the fixed first velocity threshold can be set without considering a specific range-Doppler plot based on which it is determined that there is one or more wobble objects in the distance interval. For example, the fixed first velocity can be set to a maximum velocity that a wobble object is expected to move when wobbling. The fixed first velocity threshold can be set to a value such that it is equal to a maximum velocity that all types of wobble objects are expected to move when wobbling. Alternatively, a respective maximum velocity that each type of wobble object is expected to move when wobbling can be determined. Then, the fixed first velocity threshold can be set to a value such that it is equal to a highest velocity among the maximum velocities of the various types of wobble objects present in the area to be monitored.

[0058] In addition to the condition C150 that there is a swaying object present at the distance interval, a further condition C155 that there is no peak in the frequency spectrum for frequencies below the frequency threshold value can need to be fulfilled in order to perform the step of filtering S158 the energy values in the velocity intervals of the plurality of velocity intervals within the distance interval that have an absolute velocity less than the first velocity threshold value. The inclusion of the further condition is because a peak in the frequency spectrum for frequencies below the frequency threshold value indicates that there is at least one moving object in the distance interval that is not a swaying object. Thus, filtering detections in the distance interval can risk filtering out such detections related to at least one moving object at the distance interval that is not a swaying object which can not be desired. Thus, the filtering is only performed when the further condition C155 that there is no peak in the frequency spectrum for frequencies below the frequency threshold value is also fulfilled.

[0059] The filtering of detections within the plurality of velocity intervals (denoted as the guard bins) according to the DC Doppler can be performed for all distance intervals in order to remove detections related to static objects. In comparison to determining distance intervals in which there is no swaying object or objects present, a greater number of velocity intervals (i.e. all velocity intervals having an absolute velocity less than the first velocity threshold value) within which detections are typically filtered out in the case where it is determined that there is one or more swaying objects present in a distance interval. Thus, filtering S158 any detections in the velocity intervals of the plurality of velocity intervals within the distance interval that have an absolute velocity less than the first velocity threshold value can also be seen as increasing the number of guard bins around the DC Doppler for that distance interval.

[0060] The first velocity threshold value can be further set taking into account a particular distance-Doppler plot based on which it is determined that there is one or more swaying objects present in the distance interval.

[0061] In one embodiment, when it is determined that there is one or more wobbling objects at a distance interval, at least a subset of the plurality of range-Doppler maps is used to determine a first velocity threshold for that distance interval. Then, an aggregation of energy values is performed on at least the subset of the plurality of range-Doppler maps. Specifically, for each velocity interval at that distance interval, the energy values of at least the subset of the plurality of range-Doppler maps are aggregated based on an aggregation function. This results in an aggregated energy value being derived for each velocity interval at that distance interval. The aggregation function can be, for example, a mean, a number of energy values above an energy threshold, etc. Then, the aggregated energy values of the velocity intervals at that distance interval are examined to identify a maximum absolute velocity for which there is at least one positive velocity interval with an aggregated energy value above an aggregated energy threshold and at least one negative velocity interval with an aggregated energy value above the aggregated energy threshold. If the aggregation function is a mean function, the aggregated energy value of a velocity interval is the average energy value of the velocity interval over at least the subset of the plurality of range-Doppler maps. Then, the aggregated energy value of a velocity interval above the aggregated energy threshold can mean that the signal-to-noise ratio of the mean of the energy values of the velocity interval over at least the subset of the plurality of range-Doppler maps exceeds an SNR threshold. If the aggregation function is a count of the number of range-Doppler maps of at least the subset of the plurality of range-Doppler maps for which the energy value of the velocity interval exceeds an energy threshold, the aggregated energy value of a velocity interval is the value of the count of the velocity interval. Then, the energy value of a velocity interval exceeding the energy threshold of the velocity interval can mean that the signal-to-noise ratio of the energy value of the velocity interval exceeds an SNR threshold. Then, the aggregated energy value of a velocity interval exceeding the aggregated energy threshold means that the count of the velocity interval exceeds a count threshold. This can be achieved, for example, by examining the aggregated energy values of the positive velocity intervals to identify a positive velocity interval for which the aggregated energy value exceeds the aggregated energy threshold and examining the aggregated energy values of the negative velocity intervals to identify a negative velocity interval for which the aggregated energy value exceeds the aggregated energy threshold. Then, a maximum absolute velocity is identified for which there is both a positive velocity interval for which the aggregated energy value exceeds the aggregated energy threshold and a negative velocity interval for which the aggregated energy value exceeds the aggregated energy threshold. Then, the first velocity threshold can be set to, for example, the maximum absolute velocity based on the maximum absolute velocity. This embodiment is based on the assumption that one or more wobbling objects have a maximum positive velocity and a maximum negative velocity during wobbling that are the same or close to each other. At least the subset of the plurality of range-Doppler maps is selected based on a lowest frequency at which an object is expected to wobble. For example, if the number of range-Doppler maps per second is 10 and it should be possible to determine the first velocity threshold for an object wobbling at 1 Hz, the subset of the plurality of range-Doppler maps should cover a time range of at least one second, i.e. one period of the wobbling. For example, 10 consecutive range-Doppler maps can be selected for the subset. Depending on the number of range-Doppler maps per second and the lowest frequency for which it should be possible to determine the first velocity threshold, not all consecutive range-Doppler maps are required.Instead, a subset of the range-Doppler maps, such as every other range-Doppler map, every second range-Doppler map, or other subset, can be used. By selecting the subset of range-Doppler maps based on the lowest frequency of the expected object wobble, it can be ensured that the range-Doppler maps cover at least one period of the lowest frequency, so that a wobbling object wobbling at the lowest frequency will have its highest positive and highest negative velocities in the subset of range-Doppler maps. The covered time and the number of range-Doppler maps selected can further depend on the aggregation function used.

[0062] In another embodiment, when it is determined that there is one or more wobbling objects at a range interval, velocity intervals in a plurality of velocity intervals within the range interval of each range-Doppler map in a plurality of range-Doppler maps are divided into a plurality of groups depending on their absolute velocities. Thus, each group of velocity intervals will include both positive and negative velocity intervals. The groups are ordered such that the velocity intervals in a higher ordered group have larger absolute velocities than the velocity intervals in a lower ordered group. In other words, the positive axis is divided into n positive intervals and the negative axis is divided into n negative intervals which are mirror images of the positive intervals around zero Doppler where each positive interval comprises a plurality of positive velocity intervals and each negative interval comprises a corresponding plurality of negative velocity intervals. Then, the group of velocity intervals of order k is the pair of positive and negative intervals and The plurality of groups is then processed in order by performing the following actions for a current group: computing a difference sequence of the velocity intervals in the group, determining a spectrum of the computed difference sequence, and in case there is a peak in the spectrum for frequencies above a frequency threshold, continuing with the next group in the order. The difference sequence of the velocity intervals in a group is computed by, for each time interval in the sequence of time intervals, computing the difference between a statistical measure of the energy values of the velocity intervals in the group of velocity intervals with positive velocities within the range interval from the range-Doppler map corresponding to the time interval and a statistical measure of the energy values of the velocity intervals in the group of velocity intervals with negative velocities within the range interval. This computation is continued until the condition that there is no peak in the spectrum for frequencies above the frequency threshold is met for the current group of velocity intervals. Then, a first velocity threshold is set to a value corresponding to the maximum absolute velocity of the velocity intervals in the previous group in the order, e.g. equal to the maximum absolute velocity. This embodiment is based on using a subset of the steps of the method of determining that there is one or more wobbling objects at a velocity interval using each of the velocity intervals in a group of velocity intervals at the range interval. By performing these steps on the groups in increasing order, i.e. in order of increasing absolute velocity, the last group in the order for which there is a peak in the spectrum for frequencies above the frequency threshold can be identified. In other words, for each group of velocity intervals (i.e. interval the first group for which there is no peak in the spectrum for frequencies above the frequency threshold. Thus, the last group for which there is a peak in the spectrum for frequencies above the frequency threshold is the group that is just sequentially lower. Since a peak in the spectrum for frequencies above the frequency threshold indicates the presence of one or more wobbling objects, this is the group with the highest absolute speed of the speed interval for which there is a presence of one or more wobbling objects. The statistical measure used in the present embodiment for determining the first speed threshold can for example be one of a standard deviation, a variance, a mean, a median and a sum of squares. Further, it can be the same or different as the statistical measure used in the calculation action S210 in the method 100 described above in connection with Figure 1a The statistical measure used in the calculation action S210 in the method 100 described above in connection with

[0063] Figure 4 A block diagram illustrating embodiments related to an apparatus for determining a presence of one or more wobbling objects within a distance interval of a plurality of distance intervals based on a plurality of range-Doppler plots provided by a radar system is shown. Each range-Doppler plot of the plurality of range-Doppler plots corresponds to a time interval of a sequence of time intervals and comprises a respective energy value for a plurality of speed intervals within each distance interval of the plurality of distance intervals. The apparatus 400 can for example be comprised in a radar system.

[0064] The apparatus 400 comprises a circuitry 410. The circuitry 410 is configured to perform the functions of the image processing system 400. The circuitry 410 can comprise a processor 412 such as a Central Processing Unit (CPU), a Graphics Processing Unit (GPU), a Tensor Processing Unit (TPU), a microcontroller or a microprocessor. The processor 412 is configured to execute program code. The program code can for example be configured to perform the functions of the apparatus 400.

[0065] The apparatus 400 can further comprise a memory 420. The memory 420 can be one or more of a buffer, a flash memory, a hard drive, a removable media, a volatile memory, a non-volatile memory, a random access memory (RAM), and other suitable devices. In a typical arrangement, the memory 420 can include a non-volatile memory for long term data storage and a volatile memory used as a scratchpad memory for the circuitry 410. The memory 420 can exchange data with the circuitry 410 over a data bus. There can also be an accompanying control line and address bus between the memory 420 and the circuitry 410.

[0066] The functionality of the apparatus 400 can be implemented in the form of executable logic programs (e.g., lines of code, software programs, etc.) stored on a non-transitory computer readable medium (e.g., memory 420) of the apparatus 400 and executed by the circuit 410 (e.g., using the processor 412). Furthermore, the functionality of the apparatus 400 can be standalone software applications or be part of a software application that performs additional tasks related to the apparatus 400. The described functionality can be considered a method that a processing unit (e.g., the processor 410 of the circuit 412) is configured to perform. Furthermore, although the described functionality can be implemented in software, such functionality can be carried out via dedicated hardware or firmware, or some combination of hardware, firmware, and / or software.

[0067] The circuit 410 is configured to perform a computing function 431, a first determining function 432, and a second determining function 433. The circuit 410 is further optionally configured to perform a filtering function 434 and a setting function 435.

[0068] The computing function 431 is configured to, for each time interval in the sequence of time intervals, compute, from the range-Doppler map corresponding to the time interval, a difference between a statistical measure of energy values of a set of velocity intervals with positive velocities in the plurality of velocity intervals in the range interval and a statistical measure of energy values of a set of velocity intervals with negative velocities in the plurality of velocity intervals in the range interval.

[0069] The first determining function 432 is configured to determine a spectrum of the sequence of computed differences.

[0070] The second determining function 433 is configured to determine that there is one or more wobble objects at the range interval on the condition that there is a peak in the spectrum for frequencies above a frequency threshold.

[0071] The circuit 410 can be further configured to perform the filtering function 434. The filtering function 434 is configured to, on the condition that there is a wobble object at the range interval, filter out any detections in velocity intervals with absolute velocities less than a first velocity threshold in the plurality of velocity intervals in the range interval.

[0072] The filtering function 434 can be further configured to filter out energy values in velocity intervals with absolute velocities less than the first velocity threshold in the plurality of velocity intervals in the range interval on a further condition that there is no peak in the spectrum for frequencies below the frequency threshold.

[0073] The first velocity threshold can be fixed.

[0074] The circuit 410 can further be configured to perform a setting function 435 configured to, for each velocity interval at the distance interval, aggregate the energy values of at least a subset of the plurality of range-Doppler maps based on an aggregation function, thereby deriving an aggregated energy value for each velocity interval at the distance interval, identify a maximum absolute velocity for which there is, among the plurality of velocity intervals within the distance interval, a velocity interval with positive velocity and a velocity interval with negative velocity having the aggregated energy value that is the largest absolute velocity and exceeds an aggregated energy threshold, and set the first velocity threshold based on the maximum absolute velocity.

[0075] Alternatively, the setting function 435 can be configured to divide the velocity intervals in the plurality of velocity intervals within the distance interval of each range-Doppler map in the plurality of range-Doppler maps into a plurality of groups depending on their absolute velocity, wherein the groups are ordered such that the velocity intervals in a higher ordered group have a larger absolute velocity than the velocity intervals in a lower ordered group, process the plurality of groups in order by the following steps: for a currently processed group: for each time interval in the sequence of time intervals, compute a statistical measure of the energy values of the velocity intervals in the group with positive velocity within the distance interval from the range-Doppler map corresponding to the time interval and a statistical measure of the energy values of the velocity intervals in the group with negative velocity within the distance interval, thereby computing a sequence of differences for the sequence of time intervals; determine a spectrum of the computed sequence of differences for the sequence of time intervals; and in case there is a peak in the spectrum for frequencies above a frequency threshold, proceed to processing the next group in the order, in case there is no peak in the spectrum for frequencies above the frequency threshold, set the first velocity threshold to a value corresponding to the maximum absolute velocity of the velocity intervals in the previous group in the order.

[0076] In the first determining function 432, a difference between a statistical measure of the energy values of a group of velocity intervals with positive velocity within the distance interval having an absolute velocity smaller than the second velocity threshold and a statistical measure of the energy values of a group of velocity intervals with negative velocity within the distance interval having an absolute velocity smaller than the second velocity threshold can be computed.

[0077] The statistical measure can be one of a standard deviation, a variance, a mean, a median and a sum of squares.

[0078] The detailed description of the actions of the method 100 described above in connection with Figure 1a and Figure 1b also applies to the corresponding functions of the apparatus 400. Furthermore, the optional additional features of the method 100 described above in connection with Figure 1a and Figure 1b also apply to the corresponding optional additional features of the apparatus 400.

[0079] Those skilled in the art realize that the application is not limited to the examples described above. Rather, many modifications and variations are possible in light of the above teachings. One skilled in the art can appreciate and implement these modifications and variations without departing from the scope of the claimed application. It is therefore contemplated that the application can encompass modifications and variations provided they come within the scope of the appended claims and their equivalents.

Claims

1. A method for determining, based on a plurality of range-Doppler maps provided by a radar system, that one or more wobble objects are present within a range interval of a plurality of range intervals, wherein, Each of the plurality of range-Doppler maps corresponds to a time interval of a sequence of time intervals and comprises respective energy values for a plurality of velocity intervals within each of the plurality of range intervals, the method comprising: For each time interval of the sequence of time intervals, computing a difference between a statistical measure of energy values of a set of velocity intervals with positive velocities of the plurality of velocity intervals within the range interval and a statistical measure of energy values of a set of velocity intervals with negative velocities of the plurality of velocity intervals within the range interval from the range-Doppler map corresponding to the time interval, thereby computing a sequence of differences for the sequence of time intervals, determining a spectrum of the computed sequence of differences for the sequence of time intervals; and on condition that there is a peak in the spectrum for frequencies above a frequency threshold, determining that there is one or more wobbling objects at the range interval.

2. The method of claim 1, further comprising: on condition that it is determined that there is one or more wobbling objects at the range interval, filtering out any detections in velocity intervals of the plurality of velocity intervals within the range interval that have an absolute velocity less than a first velocity threshold.

3. The method of claim 1, further comprising: on condition that it is determined that there is one or more wobbling objects at the range interval and that there is no peak in the spectrum for frequencies below the frequency threshold, filtering out any detections in velocity intervals of the plurality of velocity intervals within the range interval that have an absolute velocity less than a first velocity threshold.

4. The method of claim 2, wherein, The first velocity threshold is fixed.

5. The method of claim 2, further comprising: for each velocity interval at the range interval, aggregating energy values of at least a subset of the plurality of range-Doppler maps based on an aggregation function, thereby deriving an aggregated energy value for each velocity interval at the range interval; among the plurality of velocity intervals within the range interval, identifying the maximum absolute velocity such that there is a velocity interval with positive velocity and a velocity interval with negative velocity that have the maximum absolute velocity and an aggregated energy value that exceeds an aggregated energy threshold; and setting the first velocity threshold based on the maximum absolute velocity.

6. The method of claim 2, further comprising: grouping velocity intervals of the plurality of velocity intervals within the range interval of each of the plurality of range-Doppler maps into a plurality of groups in dependence on their absolute velocities, wherein the groups are ordered such that velocity intervals in a higher order group have a greater absolute velocity than velocity intervals in a lower order group, processing the plurality of groups in order by: for a currently processed group: for each time interval of the sequence of time intervals, computing a difference between a statistical measure of energy values of velocity intervals with positive velocities of the group of velocity intervals within the range interval and a statistical measure of energy values of velocity intervals with negative velocities of the group of velocity intervals within the range interval from the range-Doppler map corresponding to the time interval, thereby computing a sequence of differences for the sequence of time intervals; determining a spectrum of the computed sequence of differences; and on condition that there is a peak in the spectrum for frequencies higher than the frequency threshold, proceeding to processing a next group in the sequence, on condition that there is no peak in the spectrum for frequencies higher than the frequency threshold, setting the first velocity threshold to a value corresponding to the maximum absolute velocity of the velocity intervals in the previous group in the sequence.

7. The method of claim 1, wherein, in the act of computing differences, the difference between the statistical measure of energy values of a group of velocity intervals within the distance interval having positive velocities and having an absolute velocity less than a second velocity threshold and the statistical measure of energy values of a group of velocity intervals within the distance interval having negative velocities and having an absolute velocity less than the second velocity threshold.

8. The method of claim 1, wherein, the statistical measure is one of a standard deviation, a variance, a mean, a median, and a sum of squares.

9. A non-transitory computer-readable storage medium having stored thereon instructions which, when executed in an apparatus having processing capability, cause the apparatus to perform the method of claim 1.

10. An apparatus for determining presence of one or more wobble objects within a range bin of a plurality of range bins based on a plurality of range-Doppler maps provided by a radar system, wherein, each of the plurality of range-Doppler maps corresponds to a time interval in a sequence of time intervals and comprises respective energy values of a plurality of velocity intervals within each of the plurality of range intervals, the apparatus comprising circuitry configured to perform: a computing function configured to, for each time interval in the sequence of time intervals, compute, from the range-Doppler map corresponding to the time interval, a difference between a statistical measure of energy values of a group of velocity intervals in the plurality of velocity intervals within the range interval having positive velocities and a statistical measure of energy values of a group of velocity intervals in the plurality of velocity intervals within the range interval having negative velocities; a first determining function configured to determine a spectrum of the computed sequence of differences; and a second determining function configured to, on condition that there is a peak in the spectrum for frequencies higher than a frequency threshold, determine that there is one or more wobble objects at the range interval.

11. The apparatus of claim 10, wherein, the circuitry is further configured to perform: a filtering function configured to, on condition that it is determined that there is a wobble object at the range interval, filter out any detections of velocity intervals in the plurality of velocity intervals within the range interval having an absolute velocity less than a first velocity threshold.

12. The apparatus of claim 10, wherein, the circuitry is further configured to perform: a filtering function configured to, on condition that it is determined that there is a wobble object at the range interval and that there is no peak in the spectrum for frequencies lower than the frequency threshold, filter out any detections of velocity intervals in the plurality of velocity intervals within the range interval having an absolute velocity less than a first velocity threshold.

13. The apparatus of claim 11, wherein, the first velocity threshold is fixed.

14. The apparatus of claim 11, wherein, the circuitry is further configured to perform: a setting function configured to: for each velocity interval at the range interval, aggregate energy values of at least a subset of the plurality of range-Doppler maps based on an aggregation function, thereby deriving an aggregated energy value for each velocity interval at the range interval, Among the plurality of velocity intervals within the distance interval, identify the maximum absolute velocity such that there are a velocity interval with positive velocity and a velocity interval with negative velocity having the maximum absolute velocity and an aggregated energy value exceeding an aggregated energy threshold, and set the first velocity threshold based on the maximum absolute velocity.

15. The apparatus of claim 11, wherein, The circuit is further configured to perform: a setting function configured to: divide, depending on their absolute velocity, the velocity intervals of the plurality of velocity intervals within the distance interval of each range-Doppler map of the plurality of range-Doppler maps into a plurality of groups, wherein the groups are ordered such that the velocity intervals of a higher ordered group have a greater absolute velocity than the velocity intervals of a lower ordered group, process the plurality of groups in order by: for a currently processed group: for each time interval of the sequence of time intervals, compute, from the range-Doppler map corresponding to the time interval, a difference between a statistical measure of energy values of the velocity intervals with positive velocity of the group velocity intervals within the distance interval and a statistical measure of energy values of the velocity intervals with negative velocity of the group velocity intervals within the distance interval, thereby computing a sequence of differences for the sequence of time intervals, determine a spectrum of the computed sequence of differences for the sequence of time intervals, and in case there is a peak in the spectrum for a frequency higher than the frequency threshold, proceed to processing a next group in the order, in case there is no peak in the spectrum for a frequency higher than the frequency threshold, set the first velocity threshold to a value corresponding to the maximum absolute velocity of the velocity intervals in a previous group in the order.

Images