Data processing method and related product

Abstract

A data processing method and a related product, which are applied in the technical field of detection. A joint angle measurement architecture is used, in which a plurality of detection devices (1, ..., N) use detection signals to detect an object space to obtain reported data, the reported data comprising velocity information, range information, and channel data. A data processing device (10) determines at least one associated point group on the basis of pose information of at least one detection device (1, ..., N), each associated point group comprising a plurality of associated detection points in a plurality of items of reported data, the detection points reflecting the presence of associated channel data, and the channel data of the detection points of each associated point group satisfying a channel consistency condition. The data processing device (10) performs angle measurement by using channel data of a plurality of detection points in associated point groups, i.e., performing joint angle measurement by jointly using associated channels in the plurality of detection devices (1, ..., N). By using the method, the angle measurement capability of a detection system can be greatly improved, the detection precision can be significantly improved, and the confidence level of a detection result can be significantly improved.

Description

A data processing method and related products Technical Field

[0001] This application relates to the field of detection technology, and in particular to a data processing method and related products. Background Technology

[0002] Radar (radio detection and ranging) uses radio waves to detect objects in space, identify targets, and determine their distance, velocity, and angle. Radar emits electromagnetic waves and receives echo signals. It measures the target's distance based on the time difference (time of flight) between the echo and the detected signal. Because the target moves relative to the radar, the frequency of the echo signal differs from the frequency of the emitted electromagnetic wave signal—a phenomenon known as the Doppler effect. Radar uses the Doppler effect to measure the target's velocity. Regarding angle measurement, radar uses data from multiple channels. The accuracy of angle measurement is affected by the number of radar channels and the density of the antenna array.

[0003] Currently, radar has achieved high accuracy in measuring distance and velocity. However, due to the limited number of antennas and channels in the radar array, the angle measurement accuracy still cannot meet users' performance requirements, making it difficult for the radar to pinpoint the target's angle and resulting in low confidence in the detection results. Some solutions increase the number of antennas and antenna density to increase the number of channels, but these solutions are costly and significantly increase the amount of data to be processed, making them impractical. Summary of the Invention

[0004] This application provides a data processing method and related products that can combine associated channels from multiple detection devices to jointly measure angles, greatly improving the angle measurement capability of a multi-detection device system, significantly increasing detection accuracy, and significantly improving the confidence level of the detection results. Furthermore, the channel data of associated channels meet the channel consistency condition, enabling accurate selection of associated channels, thereby further improving the accuracy of joint angle measurement.

[0005] Firstly, this application provides a data processing method, which can be executed by a data processing device having computing capabilities. The data processing device can be a standalone device or a module (e.g., software and / or hardware module) within a standalone device. For ease of description, the following explanation uses a data processing device as the executing entity.

[0006] The data processing method includes: a data processing device acquiring multiple reported data from multiple detection devices, and determining at least one associated point group based on the pose information of at least one detection device. Each associated point group includes multiple associated detection points, and the multiple detection points in each associated point group originate from at least two sets of reported data. The channel data corresponding to the multiple detection points in each associated point group satisfies the channel consistency condition. Wherein, at least one detection device belongs to multiple detection devices.

[0007] Furthermore, the data processing method also includes: the data processing device determines at least one angle information based on the channel data corresponding to multiple detection points in the first associated point group, and the at least one angle information is used to indicate the angle of the target in the object space.

[0008] The reported data refers to the data obtained by the detection device using detection signals to probe the object space. This includes distance information, velocity information, and channel data corresponding to the detection point. For example, the reported data may include distance-velocity data (or RV data). A channel is an independent path for signal transmission, and channel data is the data obtained by processing the signals received by the channel, including one or more pieces of information such as amplitude, phase, or noise level. In some cases, the data processing device can obtain the target's angle by processing the channel data. It should be understood that each of the above-mentioned reported data items includes distance information, velocity information, and channel data corresponding to the detection point. However, the detection threshold and number of detection points in each set of reported data can be set to the same or different settings.

[0009] Currently, terminals typically incorporate multiple detection devices to probe the object space. When multiple devices probe the object space, their different orientations result in different reported data for the same object space, yet these data can exhibit certain correlations. For example, two detection devices positioned at the left front and directly in front of the vehicle may measure different distances, speeds, and obtain different channel data for the same target, but these data are still correlated. In some cases, the data processing device can utilize the device's orientation to compensate for the deviations in reported data caused by these orientation differences, thereby identifying the correlations between the different reported data. For instance, since channel data reflects angle measurement results, two sets of channel data obtained from angle measurements of the same target show high consistency. Furthermore, the data processing device can transform the reported data from one detection device into the coordinate system of the other, aligning the distance and speed measured by both sets of data for the same target relative to the same coordinate system. Therefore, from a data perspective, the data processing device can process the reported data and detect some related detection points in the reported data of different detection devices. These related detection points can reflect the detection points obtained by different detection devices when detecting the same target (or suspected same target). The same target here is not necessarily the same object in the object space, but a target that the detection device perceives with related characteristics (such as the same distance, the same speed, or the corresponding angle).

[0010] In some cases, the reported data includes channel data corresponding to the detection point. This channel data can be used to detect one or more angles associated with that detection point, thereby distinguishing multiple targets at the same distance and speed. Furthermore, since multiple associated detection points may reflect detection points obtained from detecting the same target, the channel data from these associated detection points can be combined for joint angle measurement to obtain the target's angle information (such as angle values ​​or angle ranges), significantly improving the angle resolution capability of the multi-detection device system.

[0011] In summary, in the above scheme, the data processing device can acquire reported data from multiple detection devices and determine associated detection points within these reported data. These associated detection points reflect the results obtained by different detection devices detecting the same target (or suspected same target), and the channel data corresponding to multiple detection points in an associated point group satisfies the channel consistency condition. The data processing device uses the channel data corresponding to the associated detection points to perform joint angle measurement to obtain the angle information of the target in the object space. The obtained angle information can subsequently be used as the angle information of points or targets in the detection results. Since the number of channels of a single detection device is limited, restricting the angle measurement capability of a single detection device, this application can combine associated channels from multiple detection devices to perform joint angle measurement, greatly improving the angle measurement capability of the multi-detection device system and significantly improving detection accuracy. Moreover, obtaining joint angle measurement results using reported data from multiple detection devices is equivalent to perceiving the target in the object space from multiple positions and angles, which can significantly improve the confidence of the detection results.

[0012] Furthermore, the channel data of the detection points in the associated point group meets the consistency condition, meaning that the channel data of multiple detection points have good channel consistency. In this case, it indicates a high probability that multiple detection points have detected the same target (or suspected to be the same target). Therefore, using these detection points with high channel consistency for joint angle measurement can improve the accuracy of the joint angle measurement results.

[0013] In one possible implementation of the first aspect, the multiple reported data include first reported data and second reported data, and the first correlation point group includes a first detection point derived from the first reported data and a second detection point derived from the second reported data. The channel data of the detection points in the first correlation point group satisfies a consistency condition, including that the channel data of the first detection point and the channel data of the second detection point satisfy a channel consistency condition. For example, the channel data of the first detection point and the channel data of the second detection point satisfying the channel consistency condition includes: a first correlation angle is greater than or equal to a first threshold, and the first correlation angle is correlated with the channel data of the first detection point and the channel data of the second detection point. For example, the first threshold is 0.9. Or, for instance, the first threshold is 0.8.

[0014] In the above scheme, the channel consistency between two detection points can be evaluated by calculating the correlation angle. For example, one method for calculating the correlation angle is as follows:

[0015] Where chanData0 is the channel data of the first detection point, chanData1 is the channel data of the second detection point, corrAngle is the correlation angle, n is used to represent the dimension, i.e. the dimension of the channel, and H is the conjugate transpose.

[0016] Optionally, the number of detection devices can be designed in various ways, but the system must include at least two detection devices. For ease of description, the following explanation uses the first and second detection devices as examples. If more detection devices are included, the other detection devices can be referenced from the first and second detection devices.

[0017] In some schemes, the pose information of the detection device used by the data processing device in determining the associated point group includes the position and / or attitude information obtained by online calibration. Online calibration refers to calibrating the position and / or attitude of the detection device using the data reported by the detection device, thereby accurately determining the position and / or attitude of the detection device.

[0018] In another possible implementation of the first aspect, the plurality of detection devices includes a first detection device and a second detection device, and the pose information of the first detection device includes the calibration orientation angle of the first detection device. The method further includes: a data processing device aligning the reference frames of the first reported data and the second reported data to obtain first calibration data and second calibration data, and determining the calibration orientation angle of the first detection device based on the channel data of the detection points in the first calibration data and the channel data of the detection points in the second calibration data.

[0019] Optionally, when the orientation angle of the first detection device is the calibration orientation angle, the channel data of the detection points in the first calibration data has the highest correlation with the channel data of the detection points in the second calibration data. For example, the first reported data includes distance information, velocity information, and channel data of multiple detection points, and the second reported data also includes distance information, velocity information, and channel data of multiple detection points. The data processing device aligns the reference frames of the first and second reported data, transforms the distance and velocity information in the first and second reported data, and the transformed distance and velocity information in the first and second reported data are coordinate-aligned. The data processing device performs an angle search on the channel data of the detection points after position alignment to obtain the angle with the highest correlation of the channel data, thereby determining the calibration orientation angle of the first detection device. For example, the angle search process is as follows: The data processing device first tries the case where the calibration orientation angle of the first detection device is angle 1, performs correlation angle calculation, and obtains the global correlation angle value, which is used to evaluate the correlation. The data processing device then tries the case where the calibration orientation angle of the first detection device is angle 2, performs correlation angle calculation, and obtains the global correlation angle value. The data processing device continuously tries various angles to perform angle search, and obtains the orientation angle with the highest correlation. The angle with the highest correlation is then used as the calibration orientation angle of the first detection device.

[0020] In the above scheme, the reported data provided by the detection device is affected by the reference frame of the detection device, and the channel data in the reported data is also affected by the difference in the orientation angle of the detection device. Therefore, the data processing device aligns the positions of the first and second reported data, aligning the reported data to the same coordinate system. At this time, the difference between the channel data can determine the orientation angle of the detection device. Therefore, the data processing device uses the channel data of the detection points in the first calibration data and the channel data of the detection points in the second calibration data to determine the calibration orientation angle of the first detection device. Using the first and second reported data, the calibration orientation angle of the first detection device can be accurately obtained, which can improve the accuracy of the detection point correlation results and improve the angle measurement accuracy.

[0021] In some cases, the installation orientation of the detection device may change, and using attitude information from a previous period or offline calibrated attitude information for calculations may cause deviations in angle measurement and fusion. In the above embodiment, using online calibration to accurately determine the orientation angle of the detection device can improve the accuracy of angle measurement and increase the confidence of the detection results.

[0022] Optionally, the calibration orientation angle calculated in the above embodiments is a relative orientation angle relative to the second detection device. Alternatively, the calibration orientation angle can also be a relative orientation angle relative to other reference frames. For example, the relative orientation angle of the second detection device relative to other reference frames is also used in the calculation process, thus enabling the calibration orientation angle of the first detection device to be a relative orientation angle with reference to other reference frames. Alternatively, the calibration orientation angle calculated in the above embodiments can be an absolute orientation angle. For example, the absolute orientation angle information of the second detection device may also be used in the calculation process. For example, if the absolute orientation angle of the second detection device is 2°, and the orientation of the first detection device deviates by 30° relative to the second detection device in a direction away from the absolute reference frame, then the calibration orientation angle of the first detection device is 32°.

[0023] In another possible implementation of the first aspect, the data processing device, based on the channel data of the detection points in the first calibration data and the channel data of the detection points in the second calibration data, includes the following operations: compensating the channel data of the detection points in the first calibration data using the installation position and installation orientation angle of the first detection device, and obtaining the calibration deflection angle of the first detection device based on the compensated channel data of the detection points in the first calibration data and the channel data of the detection points in the second calibration data.

[0024] The first calibration data includes distance and speed information of the calibrated detection points, and the positions of the first calibration data and the second calibration data are aligned. Optionally, the first calibration data also includes channel data of the detection points, in which case the channel data of the detection points in the first calibration data is the same as the channel data of the detection points in the first reported data.

[0025] In some implementations, the calibration orientation angle of the first detection device is related to the calibration deflection angle of the first detection device and the installation orientation angle of the first detection device. The calibration deflection angle represents the deviation between the actual orientation angle and the installation orientation angle of the first detection device. Furthermore, when the deflection angle of the first detection device is the calibration deflection angle, the channel data of the detection point in the compensated first calibration data has the greatest correlation with the channel data of the second calibration data.

[0026] In the above embodiment, the data processing device compensates for the channel data from the first detection device, compensating for data differences caused by the installation position and orientation angle of the first detection device, and calculates the calibration deflection angle based on the compensated channel data. Because the differences in the position and orientation angle of the first detection device are compensated for, the calibration deflection angle calculation becomes more accurate, improving the accuracy of the determined calibration deflection angle and calibration orientation angle, thereby further improving the angle measurement accuracy.

[0027] In some possible implementations, when calculating the calibration deflection angle, the data processing device attempts to match the first calibration data with the second reported data at different calibration angles (i.e., the angle search process), and calculates the correlation angle, taking the angle with the largest correlation angle as the calibration deflection angle.

[0028] In another possible implementation of the first aspect, during the online calibration process, the data processing device also matches the distance and velocity of the detection points to obtain detection points with matched distance and velocity. Further, for these detection points with matched distance and velocity, the data processing device compensates for channel data differences caused by the position (optionally including attitude) of the detection device using the position of the detection device. The data processing device uses the channel data (compensated) of the detection points with matched distance and velocity to determine the calibration orientation angle (or calibration deviation angle) of the first detection device.

[0029] For example, the data processing device compensates for the channel data of detection points in the first calibration data using the installation position and installation orientation angle of the first detection device, including the following operations: The data processing device determines at least one calibration association point group based on the first calibration data and the second calibration data. Each calibration association point group includes distance information and multiple detection points matching the distance information. Each calibration association point group includes detection points from the first calibration data and detection points from the second calibration data. The data processing device compensates for the channel data corresponding to the detection points from the first reported data in the at least one calibration association point group using the installation position and installation orientation angle of the first detection device. The data processing device obtains the calibration deflection angle of the first detection device based on the channel data of the detection points in the compensated first calibration data and the channel data of the detection points in the second calibration data, including the following operations: The data processing device obtains the calibration deflection angle of the first detection device based on the channel data in the at least one calibration association point group after compensation.

[0030] In another possible implementation of the first aspect, the data processing device, based on the channel data of the detection points in the first calibration data and the channel data of the detection points in the second calibration data, includes the following operations: the data processing device aligns the reference frames of the first reported data and the second reported data to obtain the first calibration data and the second calibration data; compensates the channel data of the detection points in the first calibration data using the installation position of the first detection device; and obtains the calibration orientation angle of the first detection device based on the compensated channel data of the detection points in the first calibration data and the channel data of the detection points in the second calibration data.

[0031] Furthermore, when the orientation angle of the first detection device is the calibrated orientation angle, the channel data of the detection point in the compensated first calibration data has the greatest correlation with the channel data of the second calibration data.

[0032] In the above embodiment, the data processing device compensates for the first reported data to compensate for the data difference caused by the installation position of the first detection device. Based on the compensated channel data, the calibration orientation angle of the first detection device can be determined. Because the positional difference of the first detection device is compensated, the calculation of the calibration orientation angle can be more accurate, which can improve the accuracy of the determined calibration deflection angle and calibration orientation angle, thereby further improving the angle measurement accuracy.

[0033] In some schemes, the reported data can include two parts: RV spectrum data and channel data of the detection points. The RV spectrum data includes the velocity and range information of the detection points. When performing position alignment, only the RV spectrum data of multiple radars can be aligned, i.e., aligning the reference frames relative to the velocity and range information of the detection points. When channel data compensation is involved, the channel data corresponding to the detection points is compensated.

[0034] In another possible implementation of the first aspect, the data processing device determines at least one group of associated points based on the pose information of at least one detection device, comprising the following steps: the data processing device aligns multiple reported data reference frames based on the pose information of at least one detection device to obtain multiple aligned data, wherein each aligned data corresponds to one reported data, and each aligned data includes aligned distance information and aligned velocity information of the detection points, and the multiple aligned data are aligned with a first coordinate system. Further, the data processing device associates the detection points based on the multiple aligned data to obtain at least one group of associated points.

[0035] It should be understood that the number of associated point groups can be one or more, and there is no strict limit here.

[0036] In the above embodiments, coordinate system transformation is performed on multiple reported data. For a specific reported data item, the pose information of the detection device corresponding to that set of reported data is used to compensate for the measurement offset caused by the positional differences of the detection devices, aligning the reported data from multiple detection devices to the same reference coordinate system. Furthermore, using the first coordinate system for alignment facilitates the fusion of detection results from multiple detection devices, thereby improving the practicality of the detection results.

[0037] The following description uses the example of aligning the position of the first reported data from the first detection device. Multiple reported data include the first reported data, which originates from the first detection device among multiple detection devices. The data processing device aligns the reference frames of the multiple reported data based on the pose information of at least one detection device to obtain multiple aligned data, including the following operation: The data processing device uses the pose information of the first detection device to perform position alignment on the first reported data to obtain the first aligned data. The first aligned data includes the aligned distance information and aligned velocity information of the detection points.

[0038] In the above embodiments, the reported data may include two parts: RV spectrum data and channel data of the detection points. The RV spectrum data includes the velocity information and distance information of the detection points. When performing position alignment, position alignment can be performed only on the RV spectrum data of multiple radars, that is, aligning the reference frames relative to the velocity and distance information of the detection points.

[0039] In another possible implementation of the first aspect, the first coordinate system is a predefined coordinate system, for example, when multiple detection devices are installed on a vehicle, the first coordinate system can be the vehicle body coordinate system.

[0040] In another possible implementation of the first aspect, the first coordinate system can be the coordinate system of one of a plurality of detection devices. The coordinate system of the detection device is the coordinate system of the data reported by the detection device.

[0041] In one possible implementation of the first aspect, the plurality of detection devices include a first detection device and a second detection device, at least one detection device includes a first detection device, and the first coordinate system is the coordinate system of the second detection device.

[0042] For example, a data processing device determines at least one group of associated points based on the pose information of at least one detection device, including the following operations: The data processing device uses the pose information of the first detection device to perform position alignment on the first reported data to obtain first aligned data, the first aligned data including aligned distance information and aligned velocity information of the detection points. The data processing device associates the detection points based on at least the second aligned data and the first aligned data to obtain at least one group of associated points, the first aligned data and the second aligned data being aligned with the coordinate system of the second reported data. Optionally, the second aligned data is the same as the second reported data, the second aligned data including aligned distance information and aligned velocity information of the detection points.

[0043] In the above embodiment, the first detection device and the second detection device are two detection devices located at different positions. When associating detection points, the pose information of the first detection device is used to transform the first reported data reported by the first detection device into the coordinate system of the second reported data, thereby obtaining the second aligned data. Since the coordinate systems of the second reported data are already aligned, no transformation is required; therefore, the second reported data can be directly used as the coordinate-aligned data, i.e., the second aligned data. The data processing device determines the associated detection points based on the first aligned data and the second aligned data (or the second reported data), thereby determining which detection points in the first reported data and the second reported data are associated.

[0044] Furthermore, using the coordinate system of the data reported by a certain detection device as the transformed coordinate system allows the coordinate systems of multiple reported data to be aligned with that coordinate system. This reduces the amount of data involved in coordinate transformation, lowers the matching deviation caused by coordinate transformation, improves the accuracy of joint angle measurement results, and enhances the efficiency of data processing.

[0045] Optionally, the multiple detection devices include multiple detection devices arranged along a first direction, such as three radars mounted on the front of the vehicle. The second detection device is a detection device located in the middle of the multiple detection devices, or the multiple detection devices are the detection devices closest to the midpoint of the terminal along the first direction, where the first direction is either the front-back direction or the left-right direction of the terminal. In other words, the data processing device can transfer the reported data from the edge detection devices to the coordinate system of the central detection device, which facilitates subsequent point cloud fusion, improves the usability of the detection results, and enhances the intelligence level of the terminal.

[0046] In some cases, the pose information of the detection device can indicate the position and orientation of the detection device. Furthermore, in the above embodiments, the pose information of the first detection device can indicate the position and orientation of the first detection device relative to the second detection device.

[0047] In another possible implementation of the first aspect, the pose information of the detection device includes the installation position of the detection device and the attitude information of the detection device. Taking the first detection device as an example, the installation position of the first detection device may include one or more of the following: the coordinates of the installation position of the first detection device, the distance of the installation position of the first detection device relative to the origin (or coordinate axis) of the first coordinate system, etc. The distance here may include translational distance along one or more directions, and may also include radial distance. The attitude information of the first detection device includes one or more of the following: the calibration orientation angle of the first detection device, the calibration deflection angle of the first detection device, the installation orientation angle of the first detection device, the measurement target deflection angle of the first detection device, etc. In some cases, the calibration deflection angle of the first detection device needs to be used in conjunction with the installation orientation angle of the first detection device to indicate the calibration orientation angle of the first detection device. The installation angle of the first detection device can be used alone.

[0048] The following description continues using the example of aligning the position of the first reported data from the first detection device. In another possible implementation of the first aspect, the pose information of the first detection device includes the installation position and attitude information of the first detection device. The aligned distance information of the detection points in the first aligned data is related to the distance information of the detection points in the first reported data, the attitude information of the first detection device, and the installation position of the first detection device. The aligned velocity information of the detection points in the first aligned data is related to the velocity information of the detection points in the first reported data and the installation position of the first detection device.

[0049] Optionally, the attitude information of the first detection device includes the orientation angle of the first detection device. This orientation angle can be the calibration orientation angle, the installation orientation angle, etc. In some cases, the orientation angle of the first detection device can be calculated from the calibration deflection angle and the installation orientation angle. In this case, the installation orientation angle is the angle obtained from offline calibration.

[0050] In another possible implementation of the first aspect, the transformed distance information R of the detection points in the alignment data i The distance information R between the detection points and the reported data. i It satisfies the following formula: R i ′=M i ·M shift_i ·R i

[0051] Among them, R i To report the distance information of the detection points in the data, M i Let M represent the rotation matrix. shuft_i This represents the translation matrix. The aforementioned rotation and translation matrices are related to the pose information of the detection device. Rotation matrix M i Let be the rotation matrix of the detection device i relative to the first coordinate system.

[0052] The conversion process will now be described using the first detection device as an example. For instance, the rotation matrix M1 of the first detection device (taking i=1 as an example) satisfies the following equation:

[0053] in, Let be the rotation angle of the first detection device relative to the first coordinate system. It satisfies the following formula:

[0054] Where γ1 is the installation orientation angle (or installation angle) of the first detection device, ε mis_1 Let θ1 be the calibration angle of the first detection device and θ2 be the target angle of the first detection device. Taking the first coordinate system as the coordinate system of the second detection device as an example, the target angle θ1 of the first detection device can be calculated by the following formula:

[0055] Where R1 is the distance measured by the first detection device to a point in space, R2 is the distance between the second detection device and that point, and d is the distance between the first detection device and the second detection device.

[0056] For example, the translation matrix M of the first detection device (taking i=1 as an example) i It satisfies the following formula:

[0057] Where d is the translation distance of the first detection device relative to the origin of the first coordinate system, and when the first coordinate system is the coordinate system of the second detection device, d is the distance between the first detection device and the second detection device.

[0058] In another possible implementation of the first aspect, the transformed distance information V of the detection points in the alignment data i Distance information V between the detection points and the reported data i It satisfies the following formula: V i ′=V i *cos θ1.

[0059] In yet another possible implementation of the first aspect, the distance information indicates the distance or range of the detection point. Exemplarily, the distance information includes a distance bin index, also referred to as a distance index or distance bin number.

[0060] Similarly, speed information indicates the speed or speed range of the detection point. For example, speed information includes a speed compartment index, also known as a speed index or speed compartment number.

[0061] In another possible implementation of the first aspect, distance information and velocity information corresponding to multiple detection points in the associated point group are matched.

[0062] As one possible scenario, multiple reported data include first reported data and second reported data. The first associated point group includes a first detection point derived from the first reported data and a second detection point derived from the second reported data. Distance information includes a distance bin index, and speed information includes a speed bin index. The distance and speed information of the first and second detection points match under the following two conditions: Condition 1: The aligned distance bin index of the first detection point is the same as the aligned distance bin index of the second detection point, or the difference between the aligned distance bin index of the first and second detection points is less than a first distance threshold. Condition 2: The aligned speed bin index of the first and second detection points is the same, or the difference between the aligned speed bin index of the first and second detection points is less than a first speed threshold.

[0063] As another possible scenario, the associated point group includes a first detection point and a second detection point, which originate from different reported data. The aligned distance information of the detection points is used to indicate the aligned distance, and the aligned velocity information is used to indicate the aligned velocity. The distance and velocity information of the first and second detection points match if at least two of the following conditions are met: Condition 1: The difference between the aligned distance of the first and second detection points is less than or equal to a second distance threshold. Condition 2: The difference between the aligned velocity of the first and second detection points is less than or equal to a second distance threshold.

[0064] In another possible implementation of the first aspect, the data processing device obtains a group of associated points by associating detection points based on multiple aligned data sets, including the following operations: the data processing device matches the velocity information and distance information of the detection points in the multiple aligned data sets to obtain a candidate group of associated points where the distance and velocity match. The matching conditions can refer to the two cases described above. This candidate group of associated points can be directly used as the group of associated points, or the candidate group of associated points needs to be further filtered to obtain the group of associated points. For example, in the latter case, the data processing device performs a channel consistency check on the detection points in the candidate group of associated points, filters the detection points in the candidate group of associated points that meet the channel consistency condition, and obtains at least one group of associated points.

[0065] In another possible implementation of the first aspect, the data processing device determines at least one angle information based on channel data corresponding to multiple detection points in the first associated point group, including the following steps: the data processing device obtains multiple frequency domain data based on the channel data corresponding to multiple detection points in the first associated point group, obtains summary data based on the multiple frequency domain data, and obtains at least one angle information based on the summary data.

[0066] The above embodiments provide an exemplary joint angle measurement process. The data processing device converts the channel data to the frequency domain and summarizes the frequency domain data. Using the summarized data, one or more angle values ​​(or angle value ranges) are obtained, for example, through spectrum analysis. Using channel data from multiple detection devices for joint angle measurement can improve angle measurement accuracy and increase the confidence level of the detection results.

[0067] In another possible implementation of the first aspect, the data processing method further includes: a data processing device generating joint point cloud data, the joint point cloud data including information of multiple points, the information of each point including the point's velocity information, the point's distance information and the point's angle information, the point's angle information being related to at least one angle information.

[0068] In the above embodiments, the data processing device can generate point cloud data. The velocity information of the points in the point cloud data is obtained from the velocity information of the detection points in multiple reported data sets, the distance information of the points is obtained from the distance information in multiple reported data sets, and the angle information of the points is related to at least one of the aforementioned angle information sets. The point cloud data reflects the coordinates and related information of the points, which facilitates target recognition, fusion perception, and other processing in the backend. Moreover, the point cloud data has high accuracy in ranging, velocity, and angle measurement, making it valuable and readily available.

[0069] Furthermore, point cloud data can be output as a detection result, or the target recognition result can be output as a detection result after the point cloud data is used for target recognition.

[0070] In another possible implementation of the first aspect, multiple detection devices are installed on the terminal. The data processing device generates joint point cloud data, including the following steps: the data processing device generates joint point cloud data based on the absolute pose information of at least one detection device, wherein the reference system of the joint point cloud data is an absolute coordinate system.

[0071] In the above embodiments, the coordinate system of the joint point cloud data is referenced to an absolute coordinate system, such as a geodetic coordinate system. Therefore, when the terminal uses the joint point cloud data, there is no need to convert the joint point cloud data even if the terminal's pose changes, thus improving the usability of the joint point cloud data.

[0072] In another possible implementation of the first aspect, the absolute pose information of at least one detection device includes the absolute orientation angle of at least one detection device. Further, the data processing method further includes: a data processing device determining the absolute orientation angle of a first detection device based on the calibrated orientation angle of at least one detection device, the pose information of the terminal, joint point cloud data, and independent point cloud data of the first detection device, wherein the first detection device belongs to at least one detection device. The independent point cloud data of the first detection device is point cloud data obtained based on the data reported by the first detection device.

[0073] In the above embodiments, the process of determining the absolute orientation angle of the first detection device can utilize point cloud data independently detected by the first detection device, as well as joint point cloud data obtained by the data processing device. This can further improve the accuracy of the calibration angle of the detection device, which is beneficial to improving the accuracy of subsequent detection point association, thereby improving the accuracy of the angle obtained by joint angle measurement. In some cases, the results of target recognition can also be used in the process of determining the absolute orientation angle of the detection device.

[0074] In another possible implementation of the first aspect, the detection signal is a radio electromagnetic wave.

[0075] In another possible implementation of the first aspect, the detection signal is a light beam, such as an FMCW laser beam.

[0076] Secondly, this application also provides a detection system, including a data processing device and multiple detection devices. The multiple detection devices are used to provide reported data to the data processing device, each reported data being data obtained by the corresponding detection device using a detection signal to detect the object space. The data processing device is used to implement the method described in the first aspect or any possible embodiment of the first aspect.

[0077] Furthermore, multiple detection devices are positioned at different locations on the terminal. Optionally, the data processing device is communicatively connected to the multiple detection devices.

[0078] In one possible implementation, the plurality of detection devices includes a first detection device, which utilizes...

[0079] Thirdly, this application also provides a data processing apparatus, which includes a data acquisition module and a processing module. The data acquisition module is used to transmit data with multiple detection devices, and the processing module is used to process the data. The data processing apparatus is used to implement the method described in the first aspect or any possible embodiment of the first aspect.

[0080] Fourthly, this application also provides a data processing apparatus, including a memory and at least one processor, wherein the memory is used to store computer instructions, and the at least one processor is used to invoke the computer instructions stored in the memory to implement the method described in the first aspect or any possible implementation of the first aspect.

[0081] Fifthly, this application also provides a chip including a communication interface and at least one processor, wherein the communication interface is used to input data and the at least one processor is used to execute computer instructions to implement the method described in the first aspect or any possible implementation of the first aspect.

[0082] Sixthly, this application also provides a terminal, which includes the aforementioned detection system, or includes a data processing device, or includes a chip.

[0083] In a seventh aspect, this application also provides a computer-readable storage medium including computer program instructions that, when executed by at least one processor, implement the method described in the first aspect or any possible implementation thereof.

[0084] Eighthly, this application also provides a computer program product containing instructions that, when executed by at least one processor, implement the method described in the first aspect or any possible implementation thereof.

[0085] The beneficial effects of aspects two through eight of this application can be found in the beneficial effects of aspect one. Attached Figure Description

[0086] The accompanying drawings used in the description of the embodiments will be briefly introduced below.

[0087] Figure 1 is a schematic diagram of distance dimension data, velocity dimension data, and RV data;

[0088] Figure 2 is a schematic diagram of a point cloud-level fusion detection system architecture;

[0089] Figure 3 is a schematic diagram of a satellite radar detection system architecture;

[0090] Figure 4 is a schematic diagram of the architecture of a detection system provided in an embodiment of this application;

[0091] Figure 5 is a schematic diagram of a vehicle provided in an embodiment of this application;

[0092] Figure 6 is a flowchart illustrating a data processing method provided in an embodiment of this application;

[0093] Figure 7 is a schematic diagram of the pose of multiple detection devices provided in the embodiments of this application;

[0094] Figure 8 is a schematic diagram of two types of reported data provided in an embodiment of this application;

[0095] Figure 9 is a schematic diagram of the installation location of a radar according to an embodiment of this application;

[0096] Figure 10 is a schematic diagram illustrating the consistency of the two channel data provided in the embodiments of this application;

[0097] Figure 11 is a schematic diagram of two alignment data provided in an embodiment of this application;

[0098] Figure 12 is a schematic diagram of a joint point cloud data provided in an embodiment of this application;

[0099] Figure 13 is a schematic diagram of the structure of a data processing device provided in an embodiment of this application;

[0100] Figure 14 is a schematic diagram of the structure of another data processing device provided in an embodiment of this application. Detailed Implementation

[0101] The following section will introduce some of the technical terms.

[0102] A detection device is a device that uses detection signals to probe the object space. The detection signal is an electromagnetic wave, such as a radio wave or a light beam. When the detection signal is a radio wave, the detection device can be called radar (radio detection and ranging). When the detection signal is a light beam, the detection device can be called laser radar (light detection and ranging). Furthermore, the detection signal can be a frequency-modulated continuous wave (FMCW), where the frequency varies with time, and the variation pattern can include one or more of the following: sawtooth, triangular, or sinusoidal.

[0103] A channel refers to an independent path for signal transmission and processing, such as the path for a detection device to transmit (T) a signal, receive (R) a signal, or a path that includes both transmission and reception. The number of channels is related to the number of transmitting and receiving units in the detection device. Taking radar as an example, if one transmitting antenna and one receiving antenna are set, the number of channels is 1. In some schemes, virtual channels can be formed by using multiple-input multiple-output (MIMO) technology, thereby increasing the number of channels of the detection device. For example, a 3T4R radar can form 12 channels.

[0104] The Fast Fourier Transform (FFT) is the process of converting a time-domain signal into a frequency-domain signal. FFT decomposes a time-varying signal (time-domain signal) into combinations of different frequency components (frequency-domain signal). The receiver of the detection device receives signals from the object space to form a sampled signal. The signal processing module performs an FFT (one-dimensional FFT, or 1DFFT) on the sampled signal (optionally with additional filtering and noise reduction preprocessing) to obtain range data. Performing a second FFT (two-dimensional FFT, or 2DFFT) on this sampled signal yields velocity data, or Doppler data.

[0105] In some solutions, distance and velocity data are fused and combined with channel data from multiple channels to form distance-velocity (RV) data. RV data includes three dimensions: distance, velocity, and the channel data corresponding to the detection point. In other solutions, the three dimensions of RV data can also be separated, for example, into RV spectrum data and channel data. The RV spectrum data includes the distance and velocity dimensions of the detection point, while the channel data is the channel data corresponding to the detection point. For ease of understanding, the three dimensions of RV data are described below:

[0106] In the distance dimension, after the detection device emits a detection signal, targets at different distances will return echoes at different times. Based on the arrival time of the echoes, the entire detection range is divided into multiple distance chambers (or distance cells or distance gates). Each distance chamber records relevant information about the echoes, such as the amplitude and phase of the echoes from multiple channels, forming channel data. As shown in part (a) of Figure 1, this is a type of distance dimension data, with the shaded area representing the distance range where echoes are suspected to exist. This distance data also records the channel data corresponding to the distance chambers where echoes exist. In RV data, the distance dimension is usually a coordinate axis, with data points corresponding to different distance chambers. The data in these detection points may contain information such as the amplitude and phase of the echoes received by multiple channels within that distance chamber.

[0107] In the velocity dimension, when a target in object space moves radially relative to the detection device, the frequency of the received echo changes due to the Doppler effect. By analyzing the frequency changes of the echo signal, the radial velocity information of the target can be obtained. The detection device performs spectral analysis on the echo signal, separating the components of different frequencies, which correspond to targets at different velocities. As shown in part (b) of Figure 1, separating different frequency components yields multiple velocity chambers. The shaded area represents the velocity range where a target is suspected to exist. This velocity data also records the channel data corresponding to the velocity chambers where echoes are present. In RV data, the velocity (or Doppler) dimension is also a coordinate axis, and its data points correspond to different velocity chambers (or velocity units or velocity gates). The data in each velocity chamber may contain information such as the amplitude and phase of the echoes received by multiple channels within that velocity chamber.

[0108] In RV data, if a distinctive feature signal (detected after CFAR processing) clearly distinguishes itself from the background noise at a certain location (i.e., the intersection of a range bin and a velocity bin), it indicates the likely presence of a target. The location of this feature signal can be used to determine the target's range and velocity, thus achieving target localization. As shown in part (c) of Figure 1, the RV data includes multiple detection points that meet the detection conditions. These detection points have corresponding range bin indices and velocity bin indices, as well as corresponding channel data. Alternatively, the channel dimension in Figure 1 can be replaced with the antenna dimension, in which case the channel dimension data can be separated from the data received by the antenna.

[0109] Constant false alarm rate (CFAR) processing is a signal processing algorithm that aims to maintain a relatively stable false alarm probability in the output under different background noise and interference environments, thereby effectively detecting the true target signal. A false alarm occurs when the detection system incorrectly identifies noise or other interference signals as the target signal. The signals received by the detection device include signals reflected from the target in the object space and background noise (as well as any possible interference signals). For example, CFAR algorithms include, but are not limited to, one or more of the following: Cell Average CFAR (CA-CFAR), Ordered Statistical CFAR (OS-CFAR), Minimum Selection CFAR (SO-CFAR).

[0110] As a possible example, CFAR processing is based on analyzing the statistical characteristics of noise and interference in reference cells (also known as background cells) surrounding the received signal. Based on measurement data from these reference cells, such as statistical quantities like signal power or amplitude, the detection threshold is adaptively adjusted. For example, suppose a signal is detected in a certain range cell (i.e., the detected cell) during detection. Using CA-CFAR as an example, CFAR processing selects some cells around the detected cell (such as cells within a ring or rectangular area) as reference cells. It calculates the average power of these reference cells (assuming an average power CFAR algorithm is used), and then determines an appropriate detection threshold based on a preset false alarm rate and some statistical parameters. If the signal power of the detected cell exceeds this threshold, it is determined that a target signal is present, and the detected cell can be used as a detection point. Otherwise, it is determined that there is no target signal.

[0111] Direction of arrival (DOA) processing is primarily used to determine the direction of arrival of signals received by the detection device. Through DOA processing, the device can determine the azimuth and elevation angles of the target relative to the transceiver. In detection systems, DOA processing relies on an array antenna, which consists of multiple antenna elements arranged in a specific geometry (such as a linear array or a planar array). When a target signal arrives at the array antenna, the received signal will differ due to the spatial differences in the individual antenna elements, including phase and / or amplitude differences. These differences can be recorded using channel data, and therefore, the direction of arrival of the target's echo can be calculated based on the channel data, thereby locating the target's angle.

[0112] The foregoing explanation of the technical terms may be used in the embodiments described below.

[0113] Currently, an increasing number of terminals are deploying multiple detection devices in order to comprehensively explore the object space from multiple angles. For detection systems involving multiple detection devices, manufacturers primarily employ point cloud-level fusion architectures or satellite radar architectures, both of which have their own limitations. The following section uses radar as an example to introduce these two architectures and their respective problems.

[0114] Please refer to Figure 2, which is a schematic diagram of a point cloud-level fusion detection system architecture. This detection system includes N radars, where N is a positive integer. Each radar includes a monolithic microwave integrated circuit (MMIC), which integrates a radio frequency (RF) module, a 2DFFT module, a CFAR module, and a DOA module. The RF module transmits and receives signals, the MMIC processes the received signals to obtain sampled signals, the 2DFFT module processes the sampled signals to obtain range and velocity data, and the CFAR detection module processes the range and velocity data to obtain RV data. The RV data includes the range, velocity, and channel data corresponding to the detection point. The DOA uses the RV data to perform angle measurement to obtain the angle information of the detection point. Finally, the radar outputs independent point cloud data to the backend (taking the data processing device as an example). The points in the point cloud include range, velocity, and angle. For a detection system including N (N is an integer and N≥2) radars, each radar performs range, velocity, and angle measurement before outputting point cloud data to the data processing device. The data processing device performs point cloud-level fusion based on point clouds from N radars to obtain detection results, which can be either fused point clouds or target identification results. Currently, due to the limited angle measurement capability of a single radar, fusing data from multiple detection devices in the form of point clouds provides only a slight improvement in angle measurement accuracy, making it difficult to meet the required confidence level of the fused detection results. Moreover, this architecture requires high-performance processing chips in the radar to complete the angle measurement process, resulting in high costs.

[0115] Compared to the architecture shown in Figure 2, the satellite radar architecture retains the MMIC (Micro-Instrument Microcontroller). The MMIC processes the received signals to obtain sampled signals and provides them to the backend. The following explanation uses the backend data processing unit as an example. The data processing module completes other signal and data processing procedures. Referring to Figure 3, the sampled signal can be the signal after analog-to-digital converter (ADC), hence represented as ADC data in Figure 3. In this architecture, the data processing unit obtains an independent point cloud for each radar output from the ADC data, and fuses the independent point clouds of multiple radars to obtain the detection result. Since the angle measurement capability of a single radar is limited, this architecture does not solve the problems of low angle measurement accuracy and poor confidence in the detection result. Furthermore, in this architecture, the radar needs to transmit ADC data to the backend, resulting in a large amount of data transmission and high transmission costs.

[0116] In summary, current detection systems generally suffer from weak angle measurement capabilities and poor confidence levels in detection results, making it difficult to meet the performance requirements of detection devices. This is particularly true in the vehicle field, where the surrounding environment is complex, with targets exhibiting a wide range of distances and speeds. Detection systems need to use angles to distinguish targets at the same distance and speed. If the angle measurement capability of the detection device is weak, it may be difficult to distinguish between two targets at the same distance and speed but located at different angles, leading to poor confidence levels in the detection results. For example, detection systems may struggle to identify stationary objects (including living things) near guardrails, or stationary targets suspended in mid-air, making driving decisions prone to errors, such as sudden braking or failure to brake in time, thus affecting driving comfort and safety.

[0117] In view of this, embodiments of this application provide a data processing method and related products that can jointly measure angles by combining associated channels from multiple detection devices, greatly improving the angle measurement capability of a multi-detection device system, significantly increasing the accuracy and confidence of the detection results. Furthermore, the channel data of associated channels satisfy the channel consistency condition, enabling accurate selection of associated channels, thereby further improving the accuracy of joint angle measurement.

[0118] The architecture and business scenarios of the detection system to which this application's embodiments can be applied are described below. It should be noted that the system architecture and business scenarios described in this application are for the purpose of more clearly illustrating the technical solutions of this application and do not constitute a limitation on the technical solutions provided in this application. It should be understood that as system architectures evolve and new business scenarios emerge, the technical solutions provided in this application are equally applicable to similar technical problems.

[0119] Please refer to Figure 4, which is a schematic diagram of the architecture of a detection system provided in an embodiment of this application. The detection system 100 includes N radars and a data processing device 10, where N is an integer and N≥2.

[0120] Radar can detect objects in space using detection signals, obtain reported data, and provide it to a data processing device. The reported data includes distance information, velocity information, and channel data corresponding to the detection point. Channel data includes the amplitude and phase of the radar channel at the detection point. Processing the channel data yields the target's angle information. For example, the radar may include an MMIC, a 2DFFT module, and a CFAR module (optional). The MMIC module transmits and receives signals and processes the received signals to obtain sampled signals. After 2DFFT and CFAR processing (CRAF processing is optional), the reported data is obtained, including distance information (e.g., represented as R), velocity information (e.g., represented as V), and channel data for the detection point.

[0121] The data processing device 10 is a computing device. Based on reported data from N radars, it can perform channel association and joint angle measurement. Channel association involves identifying associated detection points in the reported data from multiple radars, i.e., establishing the association relationship between detection points. These detection points may have associated relationships with the corresponding channels. Joint angle measurement refers to calculating angles using the channel data corresponding to the associated detection points. In some cases, the data processing device 10 is also used for joint calibration. Joint calibration involves updating the radar's attitude (or pose) information using reported data from at least two radars, combined with existing radar pose information (or position information), to calibrate the radar's attitude (or pose) online. The aforementioned existing radar pose information may include the radar pose information obtained through offline calibration, and / or, the pose information obtained by the data processing device through joint calibration of the radars.

[0122] In some possible implementations, there is a communication connection between the radar and the data processing device 10, such as through a wired link or a wireless link. For example, the radar 1 and the data processing device 10 can be connected via wired connection technologies such as Ethernet (e.g., automotive Ethernet), media-oriented systems transport (MOST), controller area network (CAN), or universal serial bus (USB). Alternatively, the radar 1 and the data processing device 10 can also be connected via wireless connection technologies such as StarFlash, Bluetooth, or Wireless Fidelity (Wi-Fi). Of course, the connection methods between the N radars and the data processing device can be the same or different. Furthermore, this application also applies to cases where there is no direct connection between the radar and the data processing device 10; for example, the N radars can provide their reported data to an intermediate device, which then provides the reported data from multiple radars to the data processing device 10.

[0123] In some possible implementations, N radars can be installed in a terminal, such as a vehicle, robot, or drone. Referring to Figure 5, a vehicle may include 6 radars (this is just an example), with 3 radars installed at the front of the vehicle and the other 3 at the rear. Of course, the installation locations and number of radars shown in Figure 5 are merely examples.

[0124] In some cases, the fields of view of some or all of the multiple detection devices overlap. For example, the aforementioned radar 1, radar 2, and radar 3 are used to detect the forward field of view of a vehicle, and their fields of view overlap. That is to say, radar 1, radar 2, and radar 3 can detect a common detection area.

[0125] Optionally, when N radars are installed in the terminal, the data processing device 10 can be located inside the terminal or outside the terminal, such as in a server. Possible implementations of the data processing device 10 will be described below.

[0126] As mentioned above, the data processing device 10 has computing capabilities, which may include hardware modules with computing capabilities and / or software modules with computing capabilities. Examples are given below based on hardware and software implementations, respectively.

[0127] As an example of hardware implementation, the data processing device 10 may include at least one processor, which is a module with processing capabilities. In one implementation, the processor may include circuitry with instruction read and execute capabilities, such as an arithmetic logic unit (ALU), processor core, central processing unit (CPU), microprocessor, microcontroller unit (MCU), graphics processing unit (GPU), or digital signal processor (DSP). In another implementation, the processor can implement certain functions through the logical relationships of hardware circuitry, which may be fixed or reconfigurable. For example, the processor may be a hardware circuitry implemented using an application-specific integrated circuit (ASIC) or a programmable logic device (PLD), such as a field-programmable gate array (FPGA). In reconfigurable hardware circuitry, the process of the processor loading a configuration document and configuring the hardware circuitry can be understood as the process of the processor loading instructions to implement the corresponding function. Furthermore, the processor can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as a neural network processing unit (NPU), tensor processing unit (TPU), deep learning processing unit (DPU), etc. In some implementations, the data processing device 10 includes at least one processor integrated as a system-on-chip (SOC), which is commonly referred to as an SOC by those skilled in the art. The SOC may include at least one processor, and when the SOC includes multiple processors, the types of processors can be different, such as including a CPU and an NPU.

[0128] For example, the detection system 100 can be applied to vehicle perception scenarios, and the data processing device 10 can be a computing device in the vehicle. For instance, the data processing device 10 is an ECU in the vehicle. More exemplarily, the data processing device 10 includes, but is not limited to, a domain controller (DC), a mobile data center (MDC), an electronic control unit (ECU), a vehicle integrated / integration unit (VIU), etc. The DC may include a cockpit domain controller (CDC), an intelligent driving domain controller, etc.

[0129] As an example of software implementation, the data processing device 10 may include software functional units. As another example of a software functional unit, the data processing device 10 may include one or more of the following: an executable computer program, computer code, or computer instructions, where "executable" means capable of running on a processor or computing instance. As yet another example of a software functional unit, the data processing device 10 may include computing instances, including virtual machines, containers, etc. A virtual machine is a computer system simulated by software, possessing complete hardware system functionality and running in an isolated environment. A container is an isolated environment obtained by packaging applications and their dependencies.

[0130] In some possible implementations, the data processing device 10 is independent of the N radar devices, i.e., it is a standalone device. In other cases, the data processing device 10 may also be located in one of the radars, or it may be divided into multiple modules and located in some radars. In other words, the data processing device 10 may be integrated with the radar design.

[0131] Alternatively, the radar in Figure 4 can be replaced with other detection devices, such as Lidar.

[0132] The methods provided in the embodiments of this application will be described below.

[0133] Please refer to Figure 6, which is a flowchart illustrating a data processing method provided in an embodiment of this application. Optionally, this data processing method can be applied to a detection system, such as the detection system shown in Figure 4 or Figure 5 above. The data processing method shown in Figure 6 may include one or more steps from S601 to S603. It should be understood that, for ease of description, the order of S601 to S603 is used here, and it is not intended to limit the execution to the above order. In addition, the execution entity in the embodiment shown in Figure 6 is also for ease of describing the method; in specific implementation, the names of devices, information, etc., can be replaced. This application embodiment does not limit the order of execution, execution time, number of executions, etc. of the above one or more steps. S601 to S603 are as follows:

[0134] S601, the data processing device acquires multiple reported data from multiple detection devices.

[0135] The detection device is a device that uses detection signals to probe the object space, such as radar and lidar. Radar uses radio waves to detect the object space; these radio waves can be transmitted into the object space via an RF module and received by the RF module. Lidar uses a light beam to detect the object space; the transmitting end includes a laser, and the receiving end is a photodetector. Here, the light beam is FMCW light, which enables the detection of target velocity information.

[0136] The reported data is data obtained by the detection device using detection signals to probe the object space, including distance information of the detection points, velocity information of the detection points, and channel data corresponding to the detection points. Distance information indicates distance or distance range, such as a distance value or a distance index. Velocity information indicates velocity or velocity range, such as a velocity value or a velocity index. A channel is an independent path for signal transmission, and channel data is the characteristic data of the signal received by the channel (including virtual channels), which may include one or more information such as amplitude, phase, or noise level. Processing the channel data allows for the detection of the target's angle. Figure 1(c) is a schematic diagram of the information expressed by the reported data, which includes distance information, velocity information, and channel data corresponding to multiple detection points.

[0137] For example, the reported data can be RV data, such as the data shown in part (c) of Figure 1, which describes the detection point from the perspectives of velocity, distance, and channel. Further, the reported data is data processed by 2DFFT and CFAR. Of course, in some cases, the reported data can be data processed by 2DFFT, while the CFAR processing can be performed by a data processing device. That is, the data processing device can perform CFAR processing on the reported data to obtain the processed reported data. The processed reported data includes the distance information of the detection point, the velocity information of the detection point, and the channel data corresponding to the detection point.

[0138] It should be understood that when a detection device probes an object space, the reported data obtained by different detection devices probing the same object space are usually different due to their different orientations.

[0139] For example, referring to Figure 7, taking the radar as the detection device and the coordinate system of radar 2 as the reference frame, the orientation angle of radar 1 relative to radar 2 is β1, and the orientation angle of radar 3 is β2. The orientation angle is the angle between the radar's normal (or the direction of the 0° azimuth angle and / or 0° elevation angle) and the reference coordinate system. In some schemes, the orientation angle at the time of installation is recorded during offline calibration after the detection device is installed, i.e., the installation orientation angle. For example, the installation orientation angle of radar 1 is γ1. Optionally, the installation orientation angle of the detection device can be used as the orientation angle of the radar. Optionally, the installation orientation angle can be calculated during offline calibration (e.g., written by the user or automatically written by the program). In other schemes, the pose of the detection device may change. In this case, there may be a deviation between the actual orientation angle of the radar and the installation orientation angle, which is the deflection angle of the detection device. As shown in Figure 7, the deflection angle of radar 1 can be expressed as ε. mis_1 Therefore, in some cases, the orientation angle of the detection device can be calculated from the installation orientation angle and the calibration deflection angle. For example, the orientation angle and / or calibration deflection angle can be given during the joint calibration process (described below).

[0140] Besides the difference in orientation angle, the target's position in object space also affects the radar's measurement deviation. For example, for target 1, radar 1 and radar 2 have a measurement deviation angle of θ1, while radar 3 and radar 2 have a measurement deviation angle of θ2. The measurement deviation angle can also be calculated, for example, by measuring the distances of radar 1 and radar 2 to target 1, and by determining the distance between radar 1 and radar 2.

[0141] Please refer to Figure 8, which is a schematic diagram of two reported data sets. Figure 8(a) shows the reported data from radar 1 shown in Figure 7, while Figure 8(b) shows the reported data from radar 2 shown in Figure 7. The detection areas of radar 1 and radar 2 overlap, meaning their fields of view overlap, so they may both detect the same target in the object space. Because radar 1 and radar 2 are installed at different locations on the terminal, the information obtained when detecting the same target will differ, but their detection data in the object space are related. Combining Figures 8(a) and (b), it can be seen that the reported data provided by radar 1 includes detection points L1 and L2, while the reported data provided by radar 2 includes detection points M1 and M2. The detection points in these two reported data sets may include the detection results for the same target.

[0142] S602, the data processing device determines at least one group of associated points based on the pose information of at least one detection device.

[0143] The pose information of the detection device includes its installation position and attitude information. Taking the first detection device as an example, the installation position of the first detection device may include one or more of the following: the coordinates of the installation position of the first detection device, the distance of the installation position of the first detection device relative to the origin (or coordinate axis) of the first coordinate system, etc. This distance may include translational distances along one or more directions, or radial distances. See Figure 9, which is a schematic diagram of the installation position of a radar. Taking the coordinate system of radar 2 as a predefined reference coordinate system as an example, the distance between radar 1 and radar 2 along the first direction (such as the X direction) is d.

[0144] Continuing with the example of the first detection device, the attitude information of the first detection device includes one or more of the following: the calibration orientation angle, the calibration deflection angle, the installation orientation angle, and the measured target deflection angle. As a possible example, the orientation angle of the detection device refers to the angle between the normal (or centerline, or 0° azimuth and / or 0° elevation direction) of the detection device and the reference object (or reference coordinate system). Referring to Figure 9, the orientation angle of radar 2 is 0°, meaning that the coordinate system of radar 2 is used as the reference object, and the orientation angle of radar 1 is β1. The target deflection angle is the angle between the target and the origin of the reference coordinate system, and between the target and the detection device. Taking radar 2 as the reference object, the target deflection angle of radar 2 is 0°, while the target deflection angle of radar 1 is θ1.

[0145] In some possible implementations, the orientation angle of the detection device is the installation orientation angle. The installation orientation angle refers to the angle of deviation of the detection device relative to a reference object during installation, and this installation orientation angle can be a value obtained during offline calibration.

[0146] In some possible implementations, the orientation angle of the detection device is obtained through online calibration. The online calibration process is described below. Online calibration refers to calibrating the position and / or attitude of the detection device using data reported by the detection device. Online calibration can accurately determine the position and / or attitude of the detection device. Online calibration requires data reported by at least two detection devices, and is therefore also called joint calibration. Taking the calibration of the orientation angle of the first detection device as an example, the calibrated orientation angle is called the calibrated orientation angle of the first detection device, which can be included in the device's pose information (or attitude information).

[0147] In an online calibration process, taking the calibration between a first detection device and a second detection device as an example, the data processing device obtains first calibration data and second calibration data based on a reference system that aligns the first and second reported data. It then determines the calibration orientation angle of the first detection device based on the channel data of the detection points in the first and second calibration data. Furthermore, when the orientation angle of the first detection device is the calibration orientation angle, the channel data of the detection points in the first and second reported data have the highest correlation.

[0148] For example, the first reported data includes distance information, velocity information, and channel data of multiple detection points, and the second reported data also includes distance information, velocity information, and channel data of multiple detection points. The data processing device aligns the reference frames of the first and second reported data, and transforms the distance and velocity information in the first and second reported data. The transformed distance and velocity information in the first and second reported data are coordinate-aligned. The data processing device performs an angle search on the channel data of the detection points after position alignment to obtain the angle with the highest correlation among the channel data, thereby determining the calibration orientation angle of the first detection device. For example, the angle search process is as follows: The data processing device first tries the case where the calibration orientation angle of the first detection device is angle 1, performs correlation angle calculation, and obtains the global correlation angle value, which is used to evaluate the correlation. The data processing device then tries the case where the calibration orientation angle of the first detection device is angle 2, performs correlation angle calculation, and obtains the global correlation angle value. By continuously trying multiple angles and performing angle search, the data processing device obtains the orientation angle with the highest correlation, and uses the angle with the highest correlation that was tried as the calibration orientation angle of the first detection device.

[0149] Optionally, the calibration orientation angle obtained from online calibration is a relative orientation angle with respect to the second detection device, i.e., the second detection device is the reference coordinate system. For example, the relative orientation angle of the second detection device is 0°, and the calibration orientation angle of the first detection device is 30°.

[0150] Alternatively, the calibration orientation angle can also be a relative orientation angle with respect to other reference frames. For example, the calculation process also uses the relative orientation angle of the second detection device with respect to other reference frames, thus enabling the calibration orientation angle of the first detection device to be a relative orientation angle with respect to other reference frames. For instance, if the relative orientation angle of the second detection device is 5° and the orientation of the reference frame is 0°, then the orientation of the first detection device deviates by 30° relative to the second detection device in the direction away from the reference frame, and the calibration orientation angle of the first detection device is 35°.

[0151] Alternatively, the calibration orientation angle calculated in the above embodiments is an absolute orientation angle. For example, the absolute orientation angle information of the second detection device may also be used in the process of calculating the calibration orientation angle. For example, if the absolute orientation angle of the second detection device is 2°, and the orientation of the first detection device is deviated by 30° relative to the second detection device in the direction away from the absolute reference frame, then the calibration orientation angle of the first detection device is 32°.

[0152] During the online calibration process described above, the data processing device needs to transform the reported data to the same coordinate system. This is to match the velocity and distance dimensions, ensuring that the distance and velocity information of the detection points are aligned (i.e., relative to the same coordinate system) across multiple sets of reported data. After matching the velocity and distance dimensions, the data processing device uses the installation position of the detection device (optionally including the orientation angle) to compensate for the differences between the channel data of the detection points. Based on the compensated channel data, the calibration orientation angle of the detection device is determined.

[0153] To facilitate understanding, two possible compensation methods are introduced below:

[0154] Method 1: The data processing device compensates for the channel data from the first detection device, compensating for data differences caused by the installation position and orientation angle of the first detection device. The calibration deflection angle is then calculated based on the compensated channel data. Because the differences in the position and orientation angle of the first detection device are compensated for, the calibration deflection angle calculation is more accurate, improving the accuracy of the determined calibration deflection angle and calibration orientation angle, thereby further enhancing the angle measurement accuracy.

[0155] As one possible implementation, the data processing device uses the installation position and orientation angle of the first detection device to compensate for the channel data of the detection points in the first calibration data. Based on the compensated channel data of the detection points in the first calibration data and the channel data of the detection points in the second calibration data, the calibration deflection angle of the first detection device is obtained. The first calibration data includes the calibrated distance information and the calibrated velocity information of the detection points, and the positions of the first calibration data and the second calibration data are aligned. Optionally, the first calibration data also includes the channel data of the detection points; in this case, the channel data of the detection points in the first calibration data is the same as the channel data of the detection points in the first reported data.

[0156] In some implementations, the calibration orientation angle of the first detection device is related to the calibration deflection angle of the first detection device and the installation orientation angle of the first detection device. The calibration deflection angle represents the deviation between the actual orientation angle and the installation orientation angle of the first detection device. Furthermore, when the deflection angle of the first detection device is the calibration deflection angle, the channel data of the detection point in the compensated first calibration data has the greatest correlation with the channel data of the second calibration data.

[0157] In some cases, the angle search and compensation processes are combined, with compensation performed during the conversion of channel data to the sin domain. Referring to Figure 7, taking radar 1 and radar 2 as examples, the data processing device attempts to use α1 as the calibration deflection angle of radar 1. At this time, the orientation angle of the detection device is β1, where β1 = α1 + γ1. During the conversion of radar 1 channel data to the sin domain, the data processing device compensates for the radar 1 channel data. Specifically, the data processing device uses the distance d between radar 1 and radar 2, the calibration deflection angle α1 of radar 1, and the installation orientation angle γ1 of radar 1 to compensate for the channel data obtained by the first detection device. The compensation process can be expressed as: FFT(exp(-i·2π·d·sin(α1+γ1)))

[0158] Where γ1 is the installation orientation angle of radar 1, and d is the distance between radar 1 and radar 2. The installation orientation angle of radar 2 is 0°, and exp is the exponent of the natural logarithm e.

[0159] In some cases, the channel data of Radar 2 is also converted to the sin domain. An exemplary conversion process is as follows: FFT(exp(-i·2π·d·sin(α1)))

[0160] For the transformed sin domain data, the data processing device calculates the correlation angle to obtain a global correlation angle value, which is then used to evaluate the correlation. Similarly, the data processing device continues to attempt to calibrate the cases with deflection angles of α1, α3, etc., and then selects the angle value with the highest correlation, such as α... j, which serves as the calibration angle for radar 1.

[0161] The calibration orientation angle of the first detection device is related to its calibration deflection angle and its installation orientation angle. For example, referring to Figure 7, if the installation orientation angle of the first detection device is 30° and the calibration deflection angle is 0.3°, then the calibration orientation angle is 30.3°. As another example, if the installation orientation angle of the first detection device is 30° and the calibration deflection angle is -0.2°, then the calibration orientation angle is 29.8°. It can be seen that by using the installation orientation angle and calibration deflection angle of the first detection device, the orientation angle value of the detection device can be accurately obtained.

[0162] In some possible implementations, when calculating the calibration deflection angle, the data processing device attempts to match the first calibration data with the second reported data at different calibration angles (i.e., an angle search process), calculates the correlation angle, and takes the angle with the largest correlation angle as the calibration deflection angle. The angle search process can be found in the foregoing description.

[0163] Method 2 involves the data processing device compensating for the first reported data to offset data discrepancies caused by the installation position of the first detection device. Based on the compensated channel data, the calibration orientation angle of the first detection device can be determined. Compared to Case 1, Case 2 only compensates for positional differences and does not utilize the installation orientation angle; therefore, the determined angle value is the calibration orientation angle.

[0164] As one possible implementation, the data processing device aligns the reference frames of the first and second reported data to obtain first and second calibration data. It then compensates for the channel data of the detection points in the first calibration data using the installation position of the first detection device. Based on the compensated channel data of the detection points in the first and second calibration data, it obtains the calibration orientation angle of the first detection device. Furthermore, when the orientation angle of the first detection device is the calibration orientation angle, the channel data of the detection points in the compensated first calibration data has the highest correlation with the channel data of the second calibration data.

[0165] The above two scenarios are merely examples. In actual implementation, other calculation methods can be used to calculate the calibration orientation angle of the detection device using the data reported by the detection device.

[0166] In some schemes, during online calibration, the data processing device also matches the distance and velocity of the detection points to obtain detection points with matched distance and velocity. Further, for these detection points with matched distance and velocity, the data processing device compensates for channel data differences caused by the position (optionally including attitude) of the detection device using the position of the detection device. The data processing device uses the compensated channel data of the detection points with matched distance and velocity to determine the calibration orientation angle (or calibration deflection angle) of the first detection device. Thus, by first determining the associated detection point group and then using the associated detection point group to calculate the calibration orientation deflection angle of the detection device, the accuracy of the deflection angle calculation can be improved.

[0167] Taking Scenario 1 as an example, the data processing device compensates for the channel data of the detection points in the first calibration data using the installation position and orientation angle of the first detection device. This includes the following operations: Based on the first and second calibration data, the data processing device determines at least one calibration association point group. Each calibration association point group includes distance information and multiple detection points that match the distance information. Each calibration association point group includes detection points from the first and second calibration data. The data processing device compensates for the channel data corresponding to the detection points from the first reported data in at least one calibration association point group using the installation position and orientation angle of the first detection device. The data processing device obtains the calibration deflection angle of the first detection device based on the channel data of the detection points in the compensated first and second calibration data. This includes the following operations: The data processing device obtains the calibration deflection angle of the first detection device based on the channel data in the compensated at least one calibration association point group.

[0168] In some solutions, the data processing device can also calculate the absolute orientation angle of the detection device based on its calibrated orientation angle. This absolute orientation angle can be used in the position alignment stage of subsequent channel association processes, or in the stage of generating joint point cloud data. For example, the calibrated orientation angle is determined based on a certain reference coordinate system. By transforming the position of this reference coordinate system, the absolute orientation angle of the detection device can be calculated. For example, the calibrated orientation angle of the first detection device is 30°, and the reference coordinate system is the coordinate system of the second detection device. Through the absolute pose information of the second detection device, which includes the absolute orientation deviation angle, for example, if the absolute orientation deviation angle is 20° relative to the absolute reference system (e.g., the ground reference system), then the calibrated orientation angle of the first detection device is 50°. In some cases, the absolute pose information of the second detection device can be calculated from the attitude information of the terminal on which the detection device is installed (e.g., the yaw angle, pitch angle, roll angle of the vehicle).

[0169] Optionally, the timing of online calibration by the data processing device can be designed in various ways, such as condition-triggered and / or periodically triggered. As one possible example, the data processing device periodically updates the calibration angles of at least two detection devices based on one or more frames of reported data. The updated calibration angles can participate in subsequent calculations, such as position alignment (i.e., coordinate system transformation). For example, each time the vehicle is powered on, the data processing device periodically updates the calibration angles of at least two detection devices based on one or more frames of reported data. Another example is that the data processing device determines the calibration angle used for position alignment in the current frame of reported data based on the previous frame of reported data. Here, the aforementioned one frame of reported data refers to the data detected by the detection device within a detection period. Alternatively, the one or more frames of reported data can be replaced with one or more frames of independent point cloud data, or with one or more frames of joint point cloud data.

[0170] The above describes the online calibration process of the detection device. Online calibration can calibrate the actual pose of the detection device, making the results of the channel association stage more accurate. The following describes the process by which the data detection device obtains the associated point group, also known as the channel association process.

[0171] As mentioned earlier, differences in the pose information of the detection devices lead to different data reported by different devices. Therefore, by combining the positional differences of the detection devices themselves, it is possible to discover the correlation between multiple sets of reported data, thereby obtaining the associated detection points in the multiple sets of reported data and identifying at least one group of associated points. Each group of associated points includes multiple associated detection points, which originate from at least two sets of reported data. For example, referring to Figure 8, by combining the pose information of radar 1, it can be determined that detection points L1 and M1 are two associated detection points, belonging to one group of associated points. Similarly, detection points L2 and M2 are two associated detection points, belonging to another group of associated points. It should be noted that not all detection points can be matched with associated detection points; that is, a single set of reported data may contain detection points that are not associated with other reported data.

[0172] Optionally, multiple detection points in an associated point group may fall into the following categories: Scenario 1: Any two detection points are associated with each other. Scenario 2: A detection point is associated with all other detection points except itself, but the other detection points are not necessarily associated with each other. Scenario 3: Each detection point is associated with at least one other detection point. Of course, other scenarios may exist in specific implementations, which will not be listed here.

[0173] In one possible implementation, the data processing device can utilize the pose of the detection devices to compensate for deviations in reported data caused by differences in the poses of the detection devices. It can transform the reported data from multiple detection devices into the same coordinate system and match the transformed, coordinate-aligned data to determine associated detection points. These associated detection points reflect the detection points obtained by different detection devices detecting the same target (or a suspected same target). Here, the same target is not necessarily the same object in object space, but rather the suspected same target perceived by the detection devices. For example, since channel data reflects the angle measurement results, the consistency between two sets of channel data obtained from angle measurements of the same target is high. Alternatively, the data processing device can transform the reported data from one detection device into the coordinate system of another set of reported data, ensuring that the distance and velocity measured by the two sets of reported data for the same target are aligned relative to the same coordinate system. Therefore, from a data perspective, the data processing device can process the reported data and detect some related detection points in the reported data of different detection devices. These related detection points can reflect the detection points obtained by different detection devices when detecting the same target (or suspected same target). The same target here is not necessarily the same object in the object space, but a target with the same characteristics perceived by the detection device.

[0174] In some cases, the associated points in at least one group of associated points satisfy a certain association relationship. As one possible association relationship, multiple detection points belonging to the same group of associated points have corresponding channel data that satisfy the channel consistency condition. Good channel consistency among the channel data of multiple detection points indicates a high probability that the multiple detection points detected the same target (or a suspected same target). Therefore, using the channel data corresponding to the detection points to detect the channel consistency improves the accuracy of screening associated detection points and enhances the accuracy of joint angle measurement results.

[0175] In some cases, the degree of channel consistency reflects the difference in signal strength of the channel data corresponding to the detection point at different azimuth angles. For example, referring to part (a) of Figure 10, the difference in signal strength between the channel data of detection point 1 and detection point 2 is small within a certain azimuth angle range, that is, their channel consistency is good, and they may belong to related detection points. Conversely, referring to part (b) of Figure 10, the channel consistency of detection point 1 and detection point 3 is poor, and they may not be related detection points.

[0176] For example, the multiple reported data include first reported data and second reported data, and the first correlation point group includes a first detection point derived from the first reported data and a second detection point derived from the second reported data. The channel data of the first detection point and the channel data of the second detection point satisfy the channel consistency condition, including: a first correlation angle is greater than or equal to a first threshold, and the first correlation angle is correlated with the channel data of the first detection point and the channel data of the second detection point. The first threshold can be represented as TH1, where TH1 ≥ 0.5. For example, TH1 is 0.9. Another example is TH1 being 0.8.

[0177] For example, one way to calculate the correlation angle is as follows:

[0178] Where chanData0 is the channel data of the first detection point, chanData1 is the channel data of the second detection point, corrAngle is the correlation angle, H is the conjugate transpose, and n is the dimension of the matrix, such as the number of channels.

[0179] In some other schemes, the distance and velocity information of multiple detection points belonging to the same associated point group are matched. That is, the distance and velocity information of multiple detection points in the associated point group are correlated.

[0180] It should be understood that the matching here considers the difference in pose. For example, referring to Figure 7, if radar 1 detects target T1, the distance information of the detection point L1 is RL1, and the velocity information is VL1. Radar 2 detects target T1, and the distance information of the detection point M1 is RM1, and the velocity information is VM1. The distance and velocity of detection point L1 are obtained with reference to the coordinate system of radar 1, while the distance and velocity of detection point M1 are obtained with reference to the coordinate system of radar 2. The two detection points are equivalent to detecting target T1 from different positions and angles. When observing whether the distance and velocity of the detection points match, the coordinate systems of the reported data from both can be aligned. For example, the data processing device uses the pose information of radar 1 to convert the reference coordinate system of the distance and velocity information of detection point L1 to the coordinate system of radar 2. The converted detection point L1 can be represented as L1', and the distance information of L1' is RL1', and the velocity information is VL1'. After the transformation, the coordinate systems of the two radars are aligned, which is equivalent to detecting the object space from the same location. Based on this, if RL1' and RM1 are the same or similar, and VL1' and VM1 are the same or similar, it means that the two radars may have detected the same target, and detection point L1 and detection point M1 may be associated detection points.

[0181] In one possible implementation, during the determination of the associated point group, the data processing device first performs position alignment on the reported data. The purpose of position alignment is to transform the reported data from multiple detection devices into the same coordinate system, thereby facilitating the matching of distance and velocity information of the detection points. Specifically, based on the pose information of at least one detection device, the data processing device aligns the reference systems of multiple reported data to obtain multiple aligned data, where each aligned data corresponds to one reported data. Each aligned data includes the aligned distance information and aligned velocity information of the detection point, and the multiple aligned data are aligned with a first coordinate system. Further, the data processing device associates the detection points based on the multiple aligned data to obtain at least one associated point group.

[0182] The following example illustrates the position alignment of first reported data from a first detection device. The data processing device acquires multiple reported data, including the first reported data, which originates from the first detection device among multiple detection devices. The data processing device uses the pose information of the first detection device to perform position alignment on the first reported data, obtaining first aligned data. This first aligned data includes aligned distance information and aligned velocity information of the detection points. Similarly, position alignment is performed on second reported data from a second detection device to obtain second aligned data.

[0183] The aforementioned first coordinate system can have several possible configurations. For example, it can be the coordinate system of one of multiple detection devices, i.e., the coordinate system of the data reported by that detection device. Another example is a predefined coordinate system. For instance, when multiple detection devices are installed on a vehicle, the first coordinate system can be the vehicle body coordinate system. These two scenarios will be described in detail below:

[0184] In scenario 1, the first coordinate system can be the coordinate system of one of the detection devices, which is also the coordinate system of the data reported by that detection device. For example, referring to Figures 7 and 8, since the reported data of radar 1, radar 2, and radar 3 are all obtained based on their respective coordinate systems, the data processing device can align the reported data of radar 1 and radar 3 to the coordinate system of radar 2, and determine the associated detection points based on the transformed and aligned reported data.

[0185] The following describes a specific position alignment process. Taking the position alignment of the reported data from the first detection device as an example, the data processing device acquires multiple reported data, including first reported data and second reported data. The first reported data comes from the first detection device, and the second reported data comes from the second detection device. The coordinate system of the first reported data is the coordinate system of the first detection device, and the coordinate system of the second reported data is the coordinate system of the second detection device.

[0186] As a possible example, the data processing device uses the pose information of the first detection device to perform position alignment on the first reported data to obtain first aligned data. The first aligned data includes the aligned distance information and the aligned velocity information of the detection points. Referring to Figure 11(a), the first aligned data includes detection points L1' and L2', where detection point L1' is obtained based on detection point L1, and detection point L2' is obtained based on detection point L2. Therefore, the detection points before and after the position conversion are in one-to-one correspondence. Since the coordinate system of the second reported data is already aligned, no conversion is necessary. Therefore, the second reported data can be directly used as the coordinate system aligned data, i.e., the second aligned data, as shown in Figure 11(b), which is consistent with Figure 8(b). After the conversion, the first aligned data and the second aligned data are aligned with the coordinate system of the second reported data. The data processing device associates the detection points at least based on the second aligned data and the first aligned data to obtain an associated point group. The associated point group includes detection points from the first reported data and detection points from the second reported data (the conditions for detection point association are described below).

[0187] Optionally, when converting the first reported data to the coordinate system of the second reported data, the pose information of the first detection device can indicate the position and attitude of the first detection device relative to the second detection device, such as the distance relative to the first detection device, the fixed deflection angle relative to the second detection device, and the measurement deflection angle.

[0188] Scenario 2: A coordinate system other than the multiple detection devices in the first coordinate system. For example, when multiple detection devices are installed on a vehicle, the first coordinate system can be the vehicle body coordinate system. In this case, multiple reported data are aligned in position, and the aligned reported data all refer to the first coordinate system.

[0189] The above describes two possible scenarios for position alignment. In both scenarios, the data reported by one or more detection devices will be transformed to obtain the aligned data. In some solutions, during position alignment, the data processing device only transforms the distance and velocity information of the detection points to obtain aligned distance and velocity information. For example, the data processing device can use the distance information of the detection points combined with their pose information to calculate the aligned distance information, and then use the same combination to calculate the aligned velocity information. That is, the position transformation process only transforms the velocity information of the detection points, which is equivalent to performing coordinate alignment in the distance and velocity dimensions without changing the channel data of the detection points.

[0190] The above describes the specific data processing procedure for position alignment. The following describes a possible calculation process for position alignment.

[0191] As an example of distance transformation, the aligned distance information of the detection points in the aligned data is related to the distance information of the detection points in the reported data, the attitude information of the detection device corresponding to the reported data, and the installation position of the detection device corresponding to the reported data. For example, the transformed distance information R of the detection points in the aligned data... i The distance information R between the detection points and the reported data. i It satisfies the following formula: R i ′=M i ·M shift_i ·R i

[0192] Among them, R i To report the distance information of the detection points in the data, M i Let M represent the rotation matrix. shift_i This represents the translation matrix. The aforementioned rotation and translation matrices are related to the pose information of the detection device. Rotation matrix M i Let be the rotation matrix of the detection device i relative to the first coordinate system.

[0193] Combining Figure 8(a) and Figure 11(a), if the reported data of radar 1 is the first reported data, the coordinate system after transformation and alignment is the coordinate system of radar 2, and the transformed reported data is the first reported data, then the rotation matrix M1 of radar 1 (taking i=1 as an example, it can be regarded as the first detection device) satisfies the following equation:

[0194] in, Let be the rotation angle of radar 1 relative to radar 2. It satisfies the following formula:

[0195] Where γ1 is the installation orientation angle of radar 1, ε mis_1 Let θ1 be the calibration deflection angle of radar 1, and θ2 be the target deflection angle measured by radar 1. Taking radar 1's coordinate system as the first coordinate system as an example, the target deflection angle θ1 measured by radar 1 can be calculated using the following formula:

[0196] Where R1 is the distance measured by radar 1 to a point in space, R2 is the distance between radar 2 (which can be regarded as a second detection device) and that point, and d is the distance between radar 1 and radar 2.

[0197] For example, the translation matrix M of radar 1 (taking i=1 as an example) i It satisfies the following formula:

[0198] Where d is the translation distance of the first detection device relative to the origin of the first coordinate system, and when the first coordinate system is the coordinate system of the second detection device, d is the distance between the first detection device and the second detection device.

[0199] As an example of velocity conversion, the aligned velocity information of the detection points in the aligned data is related to the velocity information of the detection points in the reported data, as well as the installation position of the detection device corresponding to the aligned data. For example, the converted distance information V of the detection points in the aligned data... i Distance information V between the detection points and the reported data i It satisfies the following formula: V i ′=V i *cos θ1.

[0200] As mentioned above, the distance and velocity information of multiple detection points belonging to the same associated point group are matched. After aligning multiple reported data to the same coordinate system, the data processing device can match the distance and velocity information to determine multiple detection points with matching distance and velocity.

[0201] For example, the associated point group includes a first detection point and a second detection point, which originate from different reported data. The aligned distance information of the detection points includes the aligned distance bin index, and the aligned velocity information includes the aligned velocity bin index. The distance and velocity information of the first and second detection points match if at least two of the following conditions are met: Condition 1: The aligned distance bin index of the first detection point is the same as the aligned distance bin index of the second detection point, or the difference between the aligned distance bin index of the first and second detection points is less than a first distance threshold. Condition 2: The aligned velocity bin index of the first and second detection points is the same, or the difference between the aligned velocity bin index of the first and second detection points is less than a first velocity threshold. For example, referring to Figures 8 and 11, the distance bin index (or distance) of detection point L1 is RL1, and the velocity bin index is VL1. The distance bin index of detection point M1 obtained by radar 2 from detecting target T1 is RM1, and the velocity bin index is VM1. After position alignment, detection point L1 is converted to L1', with distance bin index RL1' and velocity bin index VL1'. If RL1' and RL1 are the same or the difference is less than the first distance threshold, and VL1' and VL1 are the same or the difference is less than the first velocity threshold, then the distance and velocity information of detection point L1 and detection point M1 are matched.

[0202] The above example illustrates how reported data is represented using distance bin indices and speed bin indices, respectively, for distance and speed information. In some cases, distance and speed information may be represented as specific data values. In such cases, the data processing device can also match detection points based on these specific data values.

[0203] For another example, the associated point group includes a first detection point and a second detection point, which originate from different reported data. The aligned distance information of the detection points is used to indicate the aligned distance, and the aligned velocity information is used to indicate the aligned velocity. The distance and velocity information of the first and second detection points match if at least two of the following conditions are met: Condition 1, the difference between the aligned distance of the first and second detection points is less than or equal to a second distance threshold. Condition 2, the difference between the aligned velocity of the first and second detection points is less than or equal to a second distance threshold.

[0204] In one possible implementation, the data processing device matches the velocity and distance information of detection points in multiple aligned data sets to obtain candidate associated point groups with matching distances and velocities. The matching conditions can be found in the two examples described above. Optionally, the candidate associated point groups can be directly used as associated point groups, or the candidate associated point groups may need to be further filtered to obtain associated point groups. For example, the data processing device matches the velocity and distance information of detection points in multiple aligned data sets to obtain candidate associated point groups with matching distances and velocities. The data processing device performs channel consistency checks on the channel data of the detection points in the candidate associated point groups, removing detection points that do not meet the channel consistency conditions to obtain the associated point groups. For example, candidate associated point group 1 includes detection point 1, detection point 2, and detection point 3, which come from different reported data. The channel data of detection points 1 and 2 meet the consistency conditions, while detection point 3 does not meet the consistency conditions with either detection points 1 or 2; therefore, detection point 3 is removed. In some cases, detection point 3 may not meet the consistency condition with detection point 1, but it may meet the consistency condition with detection point 2. In such cases, it is possible to pre-design whether to retain detection point 3.

[0205] In some cases, more or fewer matching conditions may be designed for associated detection points, which will not be explained in detail here.

[0206] S603, the data processing device determines at least one angle information based on the channel data corresponding to the detection points in the first associated point group.

[0207] The first associated point group belongs to at least one associated point group. This explanation uses the first associated point group as an example. In some cases, the data processing device can perform angle measurement processing on each associated point group to obtain one or more angle information corresponding to each associated point group. That is, the data processing device can determine one or more angle information based on the channel data corresponding to multiple detection points in each associated point group.

[0208] The data reported by the detection device includes channel data corresponding to the detection point. This channel data reflects the amplitude and phase data of multiple channels of the detection device at that detection point. The channel data corresponding to the detection point can be used to detect one or more angle information corresponding to that detection point, thereby distinguishing multiple targets at the same distance and speed. Based on this, the data processing device determines the associated detection points of multiple reported data, and combines the channel data of these associated detection points to perform angle measurement to obtain the target's angle information (such as angle value or angle value range). This is equivalent to combining the channels of multiple detection devices, which can multiply the number of channels and thus significantly improve the angle resolution capability of the multi-detection device system.

[0209] In one possible implementation, the data processing device obtains multiple frequency domain data based on the channel data corresponding to multiple detection points in the first associated point group, obtains summarized data based on the multiple frequency domain data, and performs spectral analysis based on the summarized data to obtain at least one angle information. The above implementation provides an exemplary joint angle measurement process. The data processing device converts the channel data to the frequency domain, summarizes the frequency domain data, and uses the summarized data to obtain one or more angle values ​​(or angle value ranges).

[0210] Of course, the above angle measurement process is only an example. In some possible implementations, the algorithm for angle measurement using channel data from associated points can also use existing angle measurement algorithms, such as fast iterative interpolated beamforming (FIIB), multiple signal classification (MUSIC) algorithm, estimating signal parameters via rotational invariance techniques (ESPRIT) algorithm, Capon algorithm (a statistical algorithm), deterministic maximum likelihood (DML) algorithm, RELAX algorithm, etc. However, the input of the angle measurement algorithm is channel data from associated points of multiple detection devices.

[0211] In some possible implementations, the data processing device can generate joint point cloud data, which includes information about multiple points. The information for each point includes its velocity, distance, and angle. Specifically, the velocity of a point in the point cloud data is obtained from the velocity information of detected points in multiple reported data sets; the distance information is obtained from the distance information in multiple reported data sets; and the angle information is related to at least one of the aforementioned angle information. The point cloud data reflects the coordinates and related information of points, facilitating backend processing such as target recognition and fusion sensing. Furthermore, this point cloud data possesses high accuracy in ranging, velocity, and angle measurement, making it valuable and readily available.

[0212] Please refer to Figure 12, which is a schematic diagram of a joint point cloud data provided in an embodiment of this application. This joint data includes points G1, G2, and G3, where each point has corresponding distance information, velocity information, and angle information. The angle information is calculated based on the channel data of the associated point group. For example, taking the associated point group L2 and M2 as an example, two angle information may be calculated, which can be used to obtain the angle information of points G2 and G3. Similarly, for the associated point group L1 and M1 as an example, one angle information may be calculated and used to obtain (e.g., directly as) the angle information of point G2.

[0213] In some cases, point cloud data can be output as the detection result. In other cases, the data processing device performs target identification based on the joint point cloud data and outputs the target identification result as the detection result.

[0214] In some possible schemes, the coordinate system of the joint point cloud data is a relative coordinate system, such as the coordinate system of the second detection device. In other possible schemes, the coordinate system of the joint point cloud data is an absolute coordinate system, such as the coordinate system of the ground.

[0215] In the latter case, the generation of joint point cloud data also requires the use of absolute pose information from some or all of the detection devices, such as absolute position and absolute orientation angle. For example, the data processing device first obtains relatively fused point cloud data based on multiple reported data and angle values ​​obtained from joint angle measurement. The coordinate system of this relatively fused point cloud data is the coordinate system of the second detection device. The data processing device then uses the absolute pose information of the second detection device to transform the coordinate system of the relatively fused point cloud data to an absolute coordinate system, thus obtaining the joint point cloud data.

[0216] Optionally, the absolute pose information includes an absolute orientation angle. The absolute orientation angle can be determined as follows: the data processing device determines the absolute orientation angle of the first detection device based on the calibrated orientation angle of at least one detection device, the terminal's pose information, joint point cloud data, and independent point cloud data of the first detection device, wherein the first detection device belongs to at least one detection device. The independent point cloud data of the first detection device is point cloud data obtained based on the data reported by the first detection device. For example, the terminal's pose information includes one or more of the following: the terminal's position, heading angle, roll angle, pitch angle, etc.

[0217] Optionally, in some cases, the results of target identification can also be used in determining the absolute orientation angle of the detection device.

[0218] In the embodiment shown in Figure 6, the data processing device can combine the associated channels of multiple detection devices to jointly measure angles, which greatly improves the angle measurement capability of the multi-detection device system, significantly improves the detection accuracy of the detection system, and significantly improves the confidence of the detection results.

[0219] The aforementioned scheme introduced several possible implementation methods. Below, two possible implementation designs are introduced.

[0220] In one possible implementation design, the multiple detection devices in the detection system are radars. The detection system can perform the following steps:

[0221] Step 11: Multiple radars transmit detection signals and receive echo signals. ADC data is obtained based on the echo signals.

[0222] Step 12: Multiple radars perform 2D FFT and CFAR processing on the ADC data respectively to obtain RV spectrum data and channel data. The RV spectrum data includes velocity and distance information of the detection points, but excludes angle information. The channel data includes the channel data corresponding to the detection points.

[0223] Step 13: Multiple radars send RV spectrum data and channel data to the data processing unit (or radar host). Correspondingly, the data processing unit receives the RV spectrum data and channel data reported by the multiple radars.

[0224] Optionally, multiple radars and data processing devices can be connected via Ethernet. Of course, this application also applies to situations using other communication methods.

[0225] Step 14: Data processing uses data reported by multiple radars to perform joint calibration and obtain the calibration orientation angle of the radar.

[0226] In one possible implementation, the data processing device performs position transformation on data reported by multiple radars to obtain RV spectrum data and channel data with a consistent reference coordinate system, i.e., multiple calibration data. Based on the multiple calibration data, the data processing device performs range and velocity matching to obtain a calibration association point group, which includes detection points with matched range and velocity.

[0227] For the channel data in at least one set of calibration correlation points obtained through matching, the data processing device compensates for the angle measurement deviation introduced by the radar's installation position and angle. The data processing device performs global channel correlation matching on the channel data of multiple radars (optionally data after sine domain transformation) through angle scanning matching, and uses the angle with the highest global channel correlation as the relative calibration result, such as the radar's calibration orientation angle and / or calibration deflection angle. For example, using the channel data of radar 1 and radar 2, the calibration orientation angle (or calibration deflection angle) of radar 1 relative to radar 2 is matched. Similarly, for example, using the channel data of radar 3 and radar 2, the calibration orientation angle (or calibration deflection angle) of radar 3 is matched. Likewise, using the channel data of radar 1, radar 2, and radar 3, one or more of the calibration orientation angle (or calibration deflection angle) of radar 1, the calibration orientation angle (or calibration deflection angle) of radar 2, and the calibration orientation angle (or calibration deflection angle) of radar 3 are comprehensively matched.

[0228] Furthermore, in some schemes, the data processing device, based on the relative calibration results and the attitude information of the terminal (e.g., vehicle), uses newly added point cloud data (i.e., joint point cloud data) from multiple radar channels, the original point cloud (i.e., independent point cloud data of the radar), and the target (i.e., the data after target identification) to perform joint calibration, thereby obtaining the absolute orientation angle of the radar. This absolute orientation angle can be used as the result of the joint calibration of the radar.

[0229] It should be understood that there can be multiple sets of joint point cloud data, raw point cloud data, or targets here, such as multiple sets of joint point cloud data, raw point cloud data, and targets over a period of time.

[0230] Optionally, step S14 may not be performed every time this embodiment is implemented. For example, step S14 may be performed periodically or non-periodically.

[0231] Step 15: The data processing device performs channel-level correlation on the data reported by multiple radars.

[0232] Specifically, the data processing device converts the RV spectrum data to a first reference frame, such as vehicle coordinates or radar 2 coordinates. The data processing device matches the converted RV spectrum data to obtain a candidate set of associated points, where the velocity and distance information of the detection points are matched. The data processing device performs a channel consistency check on the candidate set of associated points, removing associated points with poor channel consistency to obtain the final set of associated points.

[0233] Step 16: The data processing device uses the channel data of the associated detection points in the associated point group to perform joint angle measurement and obtain the point cloud angle value.

[0234] For example, the data processing device adds up the frequency amplitudes of the channel data of the associated detection points and obtains point cloud angle values ​​with higher confidence and angle resolution based on peak angle measurement.

[0235] Step 17: The data processing device outputs the joint point cloud data.

[0236] The joint point cloud data includes distance information, velocity information, and point cloud angle values ​​for the points.

[0237] In the above embodiments, the data processing device combines channel data from multiple radars with high channel consistency to perform joint angle measurement, which can improve the angle resolution capability of the detection system in both horizontal and pitch angles. Furthermore, combining data from multiple radars for angle measurement enables multi-angle detection of the object space, reducing the impact of multipath propagation in the horizontal and pitch directions and improving point cloud confidence.

[0238] Furthermore, using joint online calibration to determine the location information of the detection device allows the radar calibration scenario to be unrestricted by the type of moving target in the environment, thereby accelerating the calibration convergence speed.

[0239] In one possible implementation design, the multiple detection devices in the detection system are FMCW lidar. The detection system can perform the following steps:

[0240] Step 21: Multiple FMCW lidars emit detection signals (i.e., FMCW beams) and receive echo signals. ADC data is obtained based on the echo signals.

[0241] Step 22: Multiple FMCW lidars perform 2D FFT and CFAR processing on the ADC data to obtain RV spectrum data and channel data. The RV spectrum data includes velocity and distance information for the detection points, but excludes angle information. The channel data includes the channel data corresponding to the detection points.

[0242] Step 23: Multiple FMCW lidars send RV spectrum data and channel data to the data processing unit (or lidar host). Correspondingly, the data processing unit receives the RV spectrum data and channel data reported by the multiple FMCW lidars.

[0243] Optionally, multiple FMCW lidar units and data processing devices can be connected via Ethernet. Of course, this application also applies to situations using other communication methods.

[0244] Step 24: Data processing. Using data reported by multiple FMCW lidars, joint calibration is performed to obtain the calibration orientation angle of the FMCW lidars.

[0245] Optionally, step S24 may not be performed every time this embodiment is implemented. For example, step S24 may be performed periodically or non-periodically.

[0246] Step 25: The data processing device performs channel-level correlation on the data reported by multiple radars.

[0247] Step 26: The data processing device uses the channel data of the associated detection points in the associated point group to perform joint angle measurement and obtain the point cloud angle value.

[0248] Step 27: The data processing device outputs the joint point cloud data.

[0249] The embodiments of this application can also be found in the description of steps 11 to 17 above.

[0250] The methods of the embodiments of this application have been described in detail above. The apparatus of the embodiments of this application is provided below.

[0251] It should be understood that the division of units in the apparatus provided in this application embodiment is only a logical functional division. In actual implementation, they can be fully or partially integrated into a single physical entity, or they can be physically separated. Furthermore, the units in the apparatus can be implemented by a processor calling software. For example, the apparatus includes a processor connected to a memory, which stores instructions. The processor calls the instructions stored in the memory to implement any of the above methods or to implement the functions of each unit of the apparatus. The processor is, for example, a general-purpose processor, such as a CPU or MPU, and the memory is either internal or external to the apparatus.

[0252] Alternatively, the units in the device can be implemented as hardware circuits. The functionality of some or all of the units can be achieved through the design of these hardware circuits, which can be understood as one or more processors. For example, in one implementation, the hardware circuit is an ASIC, and the functionality of some or all of the above units is achieved through the design of the logical relationships between the components within the circuit. In another implementation, the hardware circuit can be implemented using a PLD (Programmable Logic Controller). Taking an FPGA as an example, it can include a large number of logic gates, and the connection relationships between these logic gates are configured through configuration files to achieve the functionality of some or all of the above units.

[0253] In the embodiments of this application, each unit in the device may be one or more processors (or processing circuits) configured to implement the above methods, such as: CPU, GPU, NPU, TPU, DPU, MPU, digital signal processor (DSP), ASIC, FPGA, or a combination of at least two of these processor forms.

[0254] Furthermore, the units in the above devices can be integrated in whole or in part, or they can be implemented independently. In one implementation, these units are integrated together and implemented in the form of a System-on-a-Chip (SoC). The SoC may include at least one processor for implementing any of the above methods or implementing the functions of the units in the device. The at least one processor may be of different types, such as including a CPU and an FPGA, or including a CPU and an MCU, or including a CPU and a GPU, etc. Several possible devices are listed below.

[0255] Please refer to Figure 13, which is a schematic diagram of a data processing apparatus provided in an embodiment of this application. Optionally, the data processing apparatus 130 can be a standalone device, such as a vehicle or a computing device (e.g., a DC). Alternatively, the data processing apparatus 130 can also be a component within a standalone device (such as a vehicle or a DC), such as a chip or an integrated circuit. The data processing apparatus 130 is used to implement the aforementioned data processing method, such as the data processing method and its possible implementations shown in Figure 6.

[0256] For example, the data processing device 130 includes an acquisition unit 1301 and a processing unit 1302. The acquisition unit 1301 is used to perform one or more operations such as receiving and acquiring, and the processing unit 1302 is used to perform one or more operations such as processing, determining, generating, calculating, aligning, and associating.

[0257] In one possible implementation, the acquisition unit 1301 is used to acquire multiple reported data from multiple detection devices, and the processing unit 1302 is used to determine at least one associated point group based on the pose information of at least one detection device, and to determine at least one angle information based on the channel data corresponding to the detection points in the first associated point group. For the specific actions performed by the above units, please refer to the description in the foregoing embodiments.

[0258] In another possible implementation, the multiple reported data include first reported data and second reported data, and the first correlation point group includes a first detection point derived from the first reported data and a second detection point derived from the second reported data. The channel data of the first detection point and the channel data of the second detection point satisfy the channel consistency condition, including: a first correlation angle is greater than or equal to a first threshold, and the first correlation angle is correlated with the channel data of the first detection point and the channel data of the second detection point. When the orientation angle of the first detection device is the calibrated orientation angle, the channel data of the detection points in the first reported data and the channel data of the detection points in the second reported data have the maximum correlation.

[0259] In another possible implementation, the processing unit 1302 is further configured to align the reference frames of the first reported data and the second reported data to obtain the first calibration data and the second calibration data, and to determine the calibration orientation angle of the first detection device based on the channel data of the detection point in the first calibration data and the channel data of the detection point in the second calibration data.

[0260] In another possible implementation, the processing unit 1302 is further configured to compensate the channel data of the detection points in the first calibration data using the installation position and installation orientation angle of the first detection device, and to obtain the calibration deviation angle of the first detection device based on the channel data of the detection points in the compensated first calibration data and the channel data of the detection points in the second calibration data. In some cases, the calibration orientation angle of the first detection device is related to both the calibration deviation angle and the installation orientation angle of the first detection device.

[0261] In another possible implementation, the processing unit 1302 is further configured to determine at least one calibration association point group based on the first calibration data and the second calibration data. Each calibration association point group includes distance information and multiple detection points that match the distance information. Each calibration association point group includes detection points from the first calibration data and detection points from the second calibration data. The processing unit 1302 is further configured to compensate for the channel data corresponding to the detection points from the first reported data in the at least one calibration association point group using the installation position and installation orientation angle of the first detection device.

[0262] In one possible implementation, the processing unit 1302 is further configured to align the reference frames of the first reported data and the second reported data to obtain the first calibration data and the second calibration data, and to compensate for the channel data of the detection points in the first calibration data using the installation position of the first detection device. The processing unit 1302 is further configured to obtain the calibration orientation angle of the first detection device based on the channel data of the detection points in the compensated first calibration data and the channel data of the detection points in the second calibration data.

[0263] In one possible implementation, the processing unit 1302 is further configured to align multiple reported data based on the pose information of at least one detection device, thereby obtaining multiple aligned data. Each aligned data corresponds to one reported data, and each aligned data includes aligned distance information and aligned velocity information of the detection point. The multiple aligned data are aligned with a first coordinate system. The processing unit 1302 is further configured to associate detection points based on at least the multiple aligned data to obtain at least one group of associated points.

[0264] In one possible implementation, the multiple reported data include first reported data, which comes from a first detection device among multiple detection devices. The processing unit 1302 is further configured to use the pose information of the first detection device to perform position alignment on the first reported data to obtain first aligned data. The first aligned data includes aligned distance information and aligned velocity information of the detection points.

[0265] In one possible implementation, the processing unit 1302 is further configured to use the pose information of the first detection device to perform position alignment on the first reported data to obtain first aligned data. The first aligned data includes aligned distance information and aligned velocity information of the detection points. The processing unit 1302 is further configured to associate the detection points based on at least the second aligned data and the first aligned data to obtain at least one associated point group.

[0266] In one possible implementation, the processing unit 1302 is further configured to match the velocity information and distance information of the detection points in the multiple alignment data to obtain a candidate association point group with matching distance and velocity. This candidate association point group can be directly used as an association point group, or the candidate association point group may need to be further filtered to obtain an association point group. For example, in the latter case, the processing unit 1302 is further configured to perform channel consistency checks on the detection points in the candidate association point group, and filter the detection points in the candidate association point group that meet the channel consistency conditions to obtain at least one association point group.

[0267] In one possible implementation, the processing unit 1302 is further configured to obtain multiple frequency domain data based on the channel data corresponding to multiple detection points in the first associated point group, obtain summary data based on the multiple frequency domain data, and obtain at least one angle information based on the summary data.

[0268] In one possible implementation, the processing unit 1302 is further configured to generate joint point cloud data, which includes information on multiple points. The information on each point includes the point's velocity information, the point's distance information, and the point's angle information. The point's angle information is related to at least one angle information.

[0269] In one possible implementation, the processing unit 1302 is further configured to generate joint point cloud data based on the absolute pose information of at least one detection device, wherein the reference system of the joint point cloud data is an absolute coordinate system.

[0270] In one possible implementation, the processing unit 1302 is further configured to determine the absolute orientation angle of the first detection device based on the calibration orientation angle of at least one detection device, the pose information of the terminal, the joint point cloud data, and the independent point cloud data of the first detection device. The first detection device belongs to at least one detection device, and the independent point cloud data of the first detection device is point cloud data obtained based on the reported data of the first detection device.

[0271] The specific operations performed by the data processing device shown in Figure 13 can also be found in the description of the foregoing embodiments.

[0272] Please refer to Figure 14, which is a schematic diagram of another data processing device provided in an embodiment of this application. The data processing device 140 shown in Figure 14 can be an independent device, such as a vehicle or a computing device (e.g., a DC). Alternatively, the data processing device 140 can also be a component within an independent device (such as a vehicle or a DC), such as a chip or integrated circuit. This data processing device 140 is used to implement the aforementioned data processing method, such as the data processing method and its possible implementations shown in Figure 6.

[0273] The data processing device 140 may include at least one processor 1401 and a memory 1403. Optionally, it may also include a communication interface 1402. Further optionally, it may also include a connection line 1404, wherein the processor 1401, the communication interface 1402 and / or the memory 1403 are connected via the connection line 1404, and / or communicate with each other via the connection line 1404 to transmit control signals and / or data signals.

[0274] in:

[0275] Processor 1401 is a module that performs arithmetic and / or logical operations, and may specifically include one or more of the following modules: CPU, application processor (AP), MCU, ECU, GPU, MPU, ASIC, image signal processor (ISP), DSP, FPGA, complex programmable logic device (CPLD), or coprocessor, etc.

[0276] The communication interface 1402 can be used to provide information input or output to at least one processor, or to receive and / or transmit signals to externally transmitted signals. For example, the communication interface 1402 may include interface circuitry. For instance, the communication interface 1402 may include a wired link interface such as an Ethernet cable, or a wireless link interface (Wi-Fi, Bluetooth, general wireless transmission, vehicular short-range communication technology, and other short-range wireless communication technologies, etc.). Optionally, the communication interface 1402 may also include a radio frequency transmitter, an antenna, etc. If the communication interface 1402 includes an antenna, the number of antennas can be one or more.

[0277] As one possible design, if the data processing device 140 is a standalone device, the communication interface 1402 may include a receiver and a transmitter. The receiver and transmitter may be the same component or different components. When the receiver and transmitter are the same component, this component may be referred to as a transceiver.

[0278] As another possible design, if the data processing device 140 is a chip or circuit, the communication interface 1402 may include an input interface and an output interface, which may be the same interface or different interfaces.

[0279] Alternatively, the functionality of the communication interface 1402 can be implemented via transceiver circuitry or a dedicated transceiver chip.

[0280] The memory 1403 provides storage space, in which data such as the operating system and computer programs can be stored. The memory 1403 can be one or a combination of several of the following: random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or compact disc read-only memory (CD-ROM).

[0281] The functions and actions of each module or unit in the data processing device 140 listed above are merely illustrative examples.

[0282] Each functional unit in the data processing apparatus 140 can be used to implement the aforementioned data processing method, such as the data processing method and its possible implementation methods shown in FIG6.

[0283] Optionally, the processor 1401 may be a processor specifically designed to perform the aforementioned methods (for ease of distinction, referred to as a dedicated processor), or a processor that performs the aforementioned methods by calling a computer program (for ease of distinction, referred to as a dedicated processor). Optionally, at least one processor may include both dedicated processors and general-purpose processors.

[0284] Optionally, if the data processing apparatus 140 includes at least one memory 1403, and the processor 1401 implements the aforementioned data processing method by calling a computer program, the computer program may be stored in the memory 1403.

[0285] This application also provides a chip including logic circuitry and a communication interface. The communication interface is used to receive and / or send information, or to input and / or output information. The logic circuitry is used to process the information. This chip is used to implement the aforementioned data processing method, such as the data processing method shown in FIG6 and its possible implementations.

[0286] This application also provides a computer-readable storage medium storing instructions that, when executed on at least one processor (or data processing device), implement the aforementioned data processing method, such as the data processing method and its possible implementations shown in the embodiments of FIG6.

[0287] This application also provides a computer program product, which includes computer instructions for implementing the aforementioned data processing method, such as the data processing method and its possible implementations shown in FIG6.

[0288] In addition, a few additional points need to be made regarding this application:

[0289] I. The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the protection scope of the technical solutions of the embodiments of this application.

[0290] 2. Unless otherwise stated, “multiple” means two or more.

[0291] 3. Unless otherwise specified or in case of logical conflict, the terms and / or descriptions in different embodiments of this application are consistent and can be referenced by each other. Technical features in different embodiments can be combined to form new embodiments based on their inherent logical relationships.

[0292] IV. The various numerical designations used in this application are merely for descriptive convenience and are not intended to limit the scope of protection of this application. The magnitude of the serial numbers used in this application does not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic. For example, the terms "first," "second," "third," "fourth," and other various terminology (if present) in the specification, claims, and drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. Such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein.

[0293] Furthermore, any embodiment or design described in this application as "exemplary" or "for example" should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner for ease of understanding.

[0294] V. The terms “comprising” and “having” and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or modules is not necessarily limited to those steps or modules that are expressly listed, but may include other steps or modules that are not expressly listed or that are inherent to such process, method, product or device.

[0295] VI. In this application, "for indicating" can be understood as "enabling". "Enabling" can include direct enabling and indirect enabling. When describing information for enabling A, it can include whether the information directly enables A or indirectly enables A, but does not necessarily mean that the information carries A.

[0296] The information that enables the information is called the information to be enabled. In the specific implementation process, there are many ways to enable the information to be enabled, such as, but not limited to, directly enabling the information to be enabled, such as the information to be enabled itself or its index. It can also be indirectly enabled by enabling other information, where there is a relationship between the other information and the information to be enabled. It can also enable only a part of the information to be enabled, while the other parts are known or pre-agreed upon. For example, enabling specific information can be achieved by using a pre-agreed (e.g., protocol-defined) arrangement of various pieces of information, thereby reducing enabling overhead to some extent. Simultaneously, common parts of various pieces of information can be identified and enabled uniformly to reduce the enabling overhead caused by individually enabling the same information.

[0297] VII. In this application, "predefined" may include preconfiguration. For example, predefining certain information means that the information is calculated or received in advance before performing an action that uses the information. The "predefined" can be implemented by pre-storing corresponding codes, tables, or other means that can be used to indicate relevant information in the device (e.g., controller or vehicle). This application does not limit the specific implementation method.

[0298] 8. The term "storage" or "preservation" in this application can refer to storage in one or more memory devices. These memory devices can be separately configured or integrated into an encoder or decoder, processor, or communication device. Alternatively, some memory devices can be separately configured, while others can be integrated into a decoder, processor, or communication device. The type of memory can be any form of storage medium, and this is not limited.

[0299] 9. The arrows or boxes indicated by dashed lines in the schematic diagrams in the accompanying drawings of this application represent optional steps or optional modules.

[0300] 10. Unless otherwise stated, " / " indicates that the objects before and after are in an "or" relationship. For example, A / B can mean A or B. In this application, "and / or" is merely a description of the relationship between the related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. A and B can be singular or plural.

[0301] XI. Unless otherwise stated, the names of devices, systems, modules and other information in the embodiments of this application are merely examples, and devices, systems and modules are used to represent possible entities that implement a certain function, and the meanings of the three can be used interchangeably.

Claims

1. A data processing method, characterized in that, The method includes: Multiple reported data from multiple detection devices are acquired. Each reported data is data obtained by the corresponding detection device using detection signals to detect the object space. Each reported data includes distance information of the detection point, speed information of the detection point, and channel data corresponding to the detection point. At least one associated point group is determined based on the pose information of at least one detection device; wherein, the at least one detection device belongs to the plurality of detection devices, each associated point group includes a plurality of associated detection points, the plurality of detection points in each associated point group are derived from at least two sets of reported data, and the channel data corresponding to the plurality of detection points in each associated point group satisfies the channel consistency condition. At least one angle information is determined based on the channel data corresponding to multiple detection points in the first associated point group. The at least one angle information is used to indicate the angle of the target in the object space. The first associated point group belongs to the at least one associated point group.

2. The method according to claim 1, characterized in that, The multiple reported data include first reported data and second reported data, and the first associated point group includes a first detection point derived from the first reported data and a second detection point derived from the second reported data. The channel data at the first detection point and the channel data at the second detection point meet the channel consistency conditions, including: The first correlation angle is greater than or equal to the first threshold, and the first correlation angle is correlated with the channel data of the first detection point and the channel data of the second detection point.

3. The method according to claim 1 or 2, characterized in that, The plurality of detection devices include a first detection device and a second detection device, the plurality of reported data include first reported data from the first detection device and second reported data from the second detection device, and the pose information of the first detection device includes the calibrated orientation angle of the first detection device; The method further includes: Align the reference frames of the first reported data and the second reported data to obtain the first calibration data and the second calibration data; Based on the channel data of the detection points in the first calibration data and the channel data of the detection points in the second calibration data, the calibration orientation angle of the first detection device is determined.

4. The method according to claim 3, characterized in that, The step of using the channel data of the detection points in the first calibration data and the channel data of the detection points in the second calibration data includes: The channel data of the detection points in the first calibration data are compensated by using the installation position and orientation angle of the first detection device. Based on the channel data of the detection points in the compensated first calibration data and the channel data of the detection points in the second calibration data, the calibration deviation angle of the first detection device is obtained. The calibration orientation angle of the first detection device is related to the calibration deflection angle of the first detection device and the installation orientation angle of the first detection device; When the deflection angle of the first detection device is the calibration deflection angle, the channel data of the detection point in the compensated first calibration data has the greatest correlation with the channel data of the second calibration data.

5. The method according to claim 4, characterized in that, The step of compensating for the channel data of the detection points in the first calibration data by utilizing the installation position and orientation angle of the first detection device includes: Based on the first calibration data and the second calibration data, at least one calibration association point group is determined. Each calibration association point group includes distance information and multiple detection points that match the distance information. Each calibration association point group includes detection points in the first calibration data and detection points in the second calibration data. Using the installation position and orientation angle of the first detection device, the channel data corresponding to the detection point from the first reported data in the at least one calibration association point group is compensated; The step of obtaining the calibration angle of the first detection device based on the channel data of the detection points in the compensated first calibration data and the channel data of the detection points in the second calibration data includes: The calibration angle of the first detection device is obtained based on the channel data in the at least one calibration association point group after compensation.

6. The method according to claim 3, characterized in that, The step of using the channel data of the detection points in the first calibration data and the channel data of the detection points in the second calibration data includes: Align the reference frames of the first reported data and the second reported data to obtain the first calibration data and the second calibration data; The channel data of the detection point in the first calibration data is compensated by utilizing the installation position of the first detection device. Based on the channel data of the detection points in the compensated first calibration data and the channel data of the detection points in the second calibration data, the calibration orientation angle of the first detection device is obtained; When the orientation angle of the first detection device is the calibration orientation angle, the channel data of the detection point in the compensated first calibration data has the greatest correlation with the channel data of the second calibration data.

7. The method according to any one of claims 1-6, characterized in that, Determining at least one group of associated points based on the pose information of at least one detection device includes: Based on the pose information of the at least one detection device, the reference system of the multiple reported data is aligned to obtain multiple aligned data, wherein each of the aligned data corresponds to one of the reported data, and each of the aligned data includes the aligned distance information and the aligned velocity information of the detection point. The multiple aligned data are aligned with the first coordinate system. Based on the multiple alignment data, the detection points are associated to obtain at least one group of associated points.

8. The method according to claim 7, characterized in that, The plurality of detection devices includes a first detection device and a second detection device; The at least one detection device includes the first detection device, and the first coordinate system is the coordinate system of the second detection device.

9. The method according to claim 7, characterized in that, The first coordinate system is different from the coordinate systems of the plurality of detection devices.

10. The method according to any one of claims 7-9, characterized in that, The at least one detection device includes a first detection device, the plurality of reported data includes first reported data from the first detection device, and the plurality of aligned data includes first aligned data corresponding to the first reported data. The pose information of the first detection device includes the installation position of the first detection device and the attitude information of the first detection device; The aligned distance information of the detection points in the first alignment data is related to the distance information of the detection points in the first reported data, the attitude information of the first detection device, and the installation position of the first detection device. The aligned velocity information of the detection points in the first aligned data is related to the velocity information of the detection points in the first reported data and the installation position of the first detection device.

11. The method according to claim 10, characterized in that, The attitude information of the first detection device includes one or more of the following: the calibration orientation angle of the first detection device, the installation orientation angle of the first detection device, the calibration deflection angle of the first detection device, the installation orientation angle of the first detection device, and the measurement target deflection angle of the first detection device.

12. The method according to any one of claims 7-11, characterized in that, The distance and speed information corresponding to multiple detection points in the associated point group are matched.

13. The method according to claim 12, characterized in that, The multiple reported data include first reported data and second reported data, and the first associated point group includes a first detection point derived from the first reported data and a second detection point derived from the second reported data; The distance information includes a distance bin index, and the speed information includes a speed bin index. The distance information and speed information of the first detection point match those of the second detection point if the following two conditions are met: Condition 1: The distance bin index after the first detection point is aligned is the same as the distance bin index after the second detection point is aligned; or, the difference between the distance bin index after the first detection point is aligned and the distance bin index after the second detection point is aligned is less than the first distance threshold. Condition 2: The velocity bin index after the first detection point is aligned is the same as the velocity bin index after the second detection point is aligned, or the difference between the velocity bin index after the first detection point is aligned and the velocity bin index after the second detection point is aligned is less than the first velocity threshold.

14. The method according to any one of claims 1-13, characterized in that, The determination of at least one angle information based on channel data corresponding to multiple detection points in the first associated point group includes: Based on the channel data corresponding to multiple detection points in the first associated point group, multiple frequency domain data are obtained respectively; The aggregated data is obtained based on the multiple frequency domain data; The at least one angle information is obtained based on the summarized data.

15. The method according to any one of claims 1-14, characterized in that, The method further includes: Generate joint point cloud data, which includes information about multiple points. The information about each point includes the point's velocity information, the point's distance information, and the point's angle information. The point's angle information is related to at least one angle information.

16. The method according to any one of claims 1-15, characterized in that, The multiple detection devices are installed on the terminal, and the generation of joint point cloud data includes: Based on the absolute pose information of the at least one detection device, joint point cloud data is generated, wherein the reference system of the joint point cloud data is an absolute coordinate system.

17. The method according to claim 16, characterized in that, The absolute pose information of the at least one detection device includes the absolute orientation angle of the at least one detection device, and the method further includes: Based on the calibrated orientation angle of the at least one detection device, the pose information of the terminal, the joint point cloud data, and the independent point cloud data of the first detection device, the absolute orientation angle of the first detection device is determined. The first detection device belongs to the at least one detection device, and the independent point cloud data of the first detection device is point cloud data obtained based on the data reported by the first detection device.

18. The method according to any one of claims 1-17, characterized in that, The detection signal is a radio electromagnetic wave. Alternatively, the detection signal may be a frequency-modulated continuous wave (FMCW) laser beam.

19. A detection system, characterized in that, The detection system includes a data processing device and multiple detection devices; The plurality of detection devices are used to provide reported data to the data processing device, and each of the reported data is data obtained by the corresponding detection device using detection signals to detect the object space; The data processing device is used to implement the method according to any one of claims 1-18.

20. A data processing apparatus, characterized in that, The data processing device includes a data acquisition module and a processing module. The data acquisition module is used to transmit data with multiple detection devices, and the processing module is used to process the data. The data processing device is used to implement the method according to any one of claims 1-18.

21. A data processing apparatus, characterized in that, The data processing apparatus includes a memory and at least one processor, the memory being used to store computer instructions, and the at least one processor being used to invoke the computer instructions stored in the memory to implement the method according to any one of claims 1-17.

22. A chip, characterized in that, The chip includes a communication interface and at least one processor, the communication interface being used to input data, and the at least one processor being used to execute computer instructions to implement the method according to any one of claims 1-18.

23. A terminal, characterized in that, The terminal includes the detection system of claim 19, or the data processing device of claim 20 or 21, or the chip of claim 22.

24. A computer-readable storage medium, characterized in that, It includes computer program instructions, which, when executed by at least one processor, implement the method as claimed in any one of claims 1-18.

25. A computer program product containing instructions, characterized in that, When the instructions are executed by at least one processor, the method as described in any one of claims 1-18 is implemented.

Images