Multi-radar joint detection system

Abstract

A radar system and method includes a first data set received from a first radar device. The first data set includes data peaks associated with a first object, wherein the data peaks are associated with distance values and direction of arrival values represented in a first coordinate space. A second data set is received from a second radar device, wherein the second data set includes values associated with distance values and direction of arrival values represented in a second coordinate space. The second data set is modified using a coordinate transformation function to generate a third data set including second values associated with distance values and direction of arrival values represented in the first coordinate space. The third data set is processed to determine that a second peak is associated with valid detection of the first object.

Description

Multi-radar joint detection system Technical Field

[0001] This disclosure generally relates to automotive radar systems, and more specifically, to an automotive radar system comprising a plurality of individual radar units configured to detect objects near a vehicle. Background Technology

[0002] Radar systems, such as automotive radar systems used in civilian vehicle applications, transmit electromagnetic signals and receive back reflections of the transmitted signals. A time delay between the transmitted and received signals can be determined, and this time delay is used to calculate the distance and / or speed of the object causing the reflection. For example, in automotive applications, automotive radar systems can be used to determine the distance and / or speed of oncoming vehicles and other obstacles.

[0003] Automotive radar systems enable implementations of Advanced Driver Assistance Systems (ADAS) that could lead to increasingly safer driving and ultimately fully autonomous driving platforms. Such systems can rely on the output data of the automotive radar system to provide various driver assistance functions, such as collision avoidance assistance.

[0004] The effectiveness of an automotive radar system largely depends on its ability to accurately detect objects with a relatively small number of false positives (i.e., incorrectly determining that an object is near the vehicle) and false negatives (i.e., failing to detect the object). Therefore, in automotive radar applications, it is generally desirable to maximize the probability of object detection while minimizing false object detection. Failure to detect an object in automotive radar applications can lead to poor performance of the vehicle's ADAS (Advanced Driver Assistance Systems). Summary of the Invention

[0005] The summary portion of this invention is neither intended nor should be construed as representing the full extent and scope of this disclosure. Additional benefits, features, and embodiments of this disclosure are set forth in the accompanying drawings and the description below, and as described in the claims. Therefore, it should be understood that the summary portion may not include all aspects and embodiments claimed herein.

[0006] Furthermore, the disclosure herein is not intended to limit or constrain in any way. Moreover, this disclosure is intended to provide those skilled in the art with an understanding of one or more representative embodiments supporting the claims. Therefore, it is important that the claims be considered within the scope of constructions having various features including those of this disclosure, provided that such constructions do not depart from the scope of the methods and apparatus consistent with this disclosure (including the originally filed claims). Furthermore, this disclosure is intended to cover and include obvious improvements and modifications to this disclosure.

[0007] In some aspects, the technology described herein relates to a system comprising: a first radar device including: a first plurality of transmitter modules configured to transmit a first plurality of transmitted radar signals; a first plurality of receiver modules configured to receive a first reflection of the first plurality of transmitted radar signals and generate a first signal based on the first received reflection; and a first processor configured to process the first signal to generate a first dataset including a first two-dimensional data frame, wherein the first two-dimensional data frame includes a first data peak associated with a first object, the first data peak being associated with a first distance value and a first direction of arrival value representing a position relative to the first radar device in a first coordinate space; and a second radar device including: a second plurality of transmitter modules configured to transmit a second plurality of transmitted radar signals; and a second plurality of receiver modules configured to receive the second plurality of transmitted radar signals. The system comprises: a second reflection of a transmitted radar signal, and a second signal generated based on the second received reflection; a second processor configured to process the second signal to generate a second dataset, the second dataset including a first value representing the position relative to the second radar device in a second coordinate space; and a third processor configured to: receive the first dataset from the first radar device; receive the second dataset from the second radar device; modify the second dataset using a coordinate transformation function to generate a third dataset, wherein the third dataset includes the second value represented in the first coordinate space; process the third dataset using a constant false alarm rate (CFAR) algorithm to identify a second data peak associated with a second distance value and a second direction of arrival; and determine that the second data peak is associated with the effective detection of the first object by comparing the second direction of arrival value with the first direction of arrival value.

[0008] In some aspects, the technology described herein relates to a system comprising: a first radar device configured to process a first received signal using a constant false alarm rate (CFAR) algorithm utilizing a first detection threshold to generate a first dataset including data peaks associated with a first object, wherein the data peaks are associated with distance and direction of arrival values ​​represented in a first coordinate space; a second radar device configured to process a second received signal using the CFAR algorithm utilizing the first detection threshold to generate a second dataset including first values ​​associated with distance and direction of arrival values ​​represented in a second coordinate space; and a processor configured to: receive the first dataset from the first radar device, receive the second dataset from the second radar device, modify the second dataset using a coordinate transformation function to generate a third dataset including second values ​​associated with distance and direction of arrival values ​​represented in the first coordinate space, and process the third dataset using a CFAR algorithm utilizing a second detection threshold to determine the presence of a second data peak at the distance and direction of arrival values ​​within the third dataset, and to determine that the second peak is associated with valid detection of the first object, wherein the second detection threshold is less than the first detection threshold.

[0009] In some aspects, the techniques described herein relate to a method comprising: receiving a first dataset from a first radar device, wherein the first dataset includes data peaks associated with a first object, wherein the data peaks are associated with distance values ​​and direction-of-arrival values ​​represented in a first coordinate space; receiving a second dataset from a second radar device, wherein the second dataset includes values ​​associated with distance values ​​and direction-of-arrival values ​​represented in a second coordinate space; modifying the second dataset using a coordinate transformation function to generate a third dataset, the third dataset including second values ​​associated with distance values ​​and direction-of-arrival values ​​represented in the first coordinate space; and processing the third dataset to determine that a second data peak exists within the third dataset at the distance values ​​and the direction-of-arrival values, thereby determining that the second peak is associated with valid detection of the first object. Attached Figure Description

[0010] The accompanying drawings are included to provide a further understanding of this disclosure, and are incorporated in and form a part of this specification. The drawings illustrate embodiments of the disclosure and, together with the description, serve to illustrate the principles of the disclosure.

[0011] In the attached diagram:

[0012] Figure 1A depicts a simplified schematic block diagram of an automotive radar device 100 including a radar unit connected to a radar controller processor.

[0013] Figure 1B illustrates, in a high-level graphical manner, the processing steps that can be implemented by the processor of the radar device in Figure 1A to process digital signals.

[0014] Figure 2 is a simplified schematic block diagram of a radar system including two separate radar units configured according to this disclosure.

[0015] Figure 3 is a block diagram depicting additional details of the sensor exchange block in Figure 2.

[0016] Figure 4 depicts a basic vehicle with two radar units, each of which is in a known position relative to the vehicle.

[0017] Figure 5 depicts two example frames of radar data generated by the first radar device and the second radar device.

[0018] Figure 6 is a graph depicting an example dataset that can be processed by the enhanced CFAR data processing block using a reduced CFAR threshold. Detailed Implementation

[0019] This disclosure generally relates to automotive radar systems, and more specifically, to an automotive radar system comprising a plurality of individual radar units configured to detect objects near a vehicle.

[0020] During normal operation, an automotive radar system transmits electromagnetic signals to a region of interest and receives the back reflections of those transmitted signals. The reflections are received at one or more antennas and processed by a signal processing system to generate output data. The data output by the signal processing system of such radar systems can include, for example, point clouds identifying the positions of objects relative to the radar system. Therefore, in automotive applications, point clouds can represent the positions of objects near the vehicle. The data can then be used by various vehicle subsystems (e.g., advanced driver assistance systems (ADAS)) to provide various notifications to the driver or to enable direct control of one or more vehicle subsystems (e.g., braking and steering systems).

[0021] In many such systems, the point cloud generated by the automotive radar system can also indicate the probability associated with each detected object, representing the confidence level of the object at that location. For example, objects with high signal reflectivity (e.g., high radar cross-section) reflect strong signals, making them detectable with a high confidence level. However, other objects reflecting lower-power signals, with amplitudes closer to ambient noise levels (e.g., objects with small radar cross-sections or objects that absorb some of the emitted signal), can be detected with lower confidence. Furthermore, some objects whose reflected signal amplitudes are below the radar system's predefined noise level may not be detected at all, resulting in poor performance of the automotive radar system and, in some cases, a risk of collision.

[0022] A radar system may also fail to detect a particular object for many other reasons. For example, if the object is too far away and the received signal has a low signal-to-noise ratio (SNR), if the object is near the edge of the radar system's field of view (FOV), where the system's antenna gain and sensitivity are lower compared to other areas of the FOV, if there is a problem with the radar system's hardware or calibration, or if the observed angle to the object causes the reflected signal to interfere with itself destructively, or if the object is obscured by another object, then the object may not be detected.

[0023] To remedy some of these problems, this disclosure provides a radar system comprising multiple individual radar units. Each individual radar unit is configured to independently transmit, receive, and process its own corresponding signals. Then, as provided herein, the various outputs of each individual radar unit are combined, making the resulting output of the entire radar system more likely to detect objects that might have been missed or misclassified by the individual radar units.

[0024] In this way, the radar system can improve the accuracy of object detection and increase the likelihood of detecting smaller objects with low-power reflected signals. In some configurations, the field of view (FOV) of each individual radar unit can overlap to mitigate the problem of poor object detection at the edges of the radar unit's FOV. Furthermore, multiple radar units can serve as backups for each other, allowing other radar units to compensate if one unit suffers a hardware defect or improper calibration.

[0025] Therefore, this disclosure provides an automotive radar system that utilizes two or more individual radar units that may have overlapping fields of view (FOV). The output signal (e.g., a point cloud) of each individual radar unit can be transformed into a region of interest (ROI) on the point cloud of another radar unit to perform a joint object detection algorithm, as described herein.

[0026] To illustrate the design and operation of a single vehicle radar device that can be incorporated into the radar system of this disclosure, reference is now made to FIG1A, which depicts a simplified schematic block diagram of an automotive radar device 100 including a radar unit 10 connected to a radar controller processor 20. In a chosen embodiment, the radar unit 10 may be embodied as a field-replaceable unit (LRU) or modular component designed for quick replacement in an operating location. Similarly, the radar controller processor 20 may be embodied as a field-replaceable unit (LRU) or modular component. The radar device 100 may be implemented as an integrated circuit, wherein the radar unit 10 and the radar controller processor 20 are formed as separate integrated circuits (chips) or a single chip, depending on the application.

[0027] Within the radar device 100, each radar unit 10 includes one or more transmitting antenna elements 102 and receiving antenna elements 104 respectively connected to one or more radio frequency (RF) transmitter (TX) units 11 and receiver (RX) units 12. For example, each radar unit 10 is shown to include individual antenna elements 102, 104 (e.g., TX1,i, RX1,j) respectively connected to three transmitter modules 11 and four receiver modules 12, but these numbers are not limiting, and other numbers are also possible, such as four transmitter modules 11 and six receiver modules 12, or a single transmitter module 11 and / or a single receiver module 12.

[0028] Each radar unit 10 also includes a chirp generator 112 configured and connected to supply a chirp input signal to the transmitter module 11. For this purpose, the chirp generator 112 is configured to receive separate and independent local oscillator (LO) signals and chirp start trigger signals. The operation of the transmitter module 11 can be controlled by a controller 110, which can be fully or partially implemented by the processor 20. A chirp signal 113 is generated and typically transmitted to the transmitter module 11 according to a predefined transmission schedule, wherein the chirp signal 113 is filtered at the RF conditioning module 114 and amplified at the power amplifier 115 before being fed to the corresponding transmit antenna 102 (TX1,i) for radiation.

[0029] Radar signals transmitted by transmitter antenna elements 102 (TX1,i, TX2,i) can be reflected by an object, and a portion of the reflected radar signal reaches receiver antenna elements 104 (RX1,i) at radar unit 10. At each receiver module 12, the received RF antenna signal is amplified by a low-noise amplifier (LNA) 120 and then fed to a mixer 121, where the received signal is mixed with a transmitted chirped signal generated by an RF conditioning module 114. The resulting intermediate frequency signal is fed to a first high-pass filter (HPF) 122. The resulting filtered signal is fed to a first variable gain amplifier 123, which amplifies the signal before feeding it to a first low-pass filter (LPF) 124. This further filtered signal is fed to an analog-to-digital converter (ADC) 125 and output as a digital signal 126 (D1) by each receiver module 12. The receiver modules compress the object signal echoes with various delays into multiple sinusoidal frequencies, the frequencies of which correspond to the round-trip delay of the echo signal.

[0030] The radar device 100 includes a radar controller processing unit 20, which is connected (e.g., via controller 110) to supply input control signals to the radar device 10 and to receive digital output signals (e.g., digital signal 126) generated by receiver module 12.

[0031] In the selected embodiment, the radar controller processing unit 20 may be embodied as a microcontroller unit (MCU) or other processing unit configured and arranged for signal processing tasks, such as, but not limited to, object identification; calculation of object range, object velocity, and object orientation; and generation of control signals. For example, the radar controller processing unit 20 may be configured to generate calibration signals, receive data signals, receive sensor signals, generate spectrum shaping signals (e.g., generating ramps in the case of FMCW radar), and / or enable sequence registration programming or state machine signals for RF (radio frequency) circuitry. Additionally, the radar controller processor 20 may be configured to program the transmitter module 11 to operate in a time-division manner by sequentially transmitting chirps for coordinated communication between the transmit antenna elements 102 TX1,i, RX1,j.

[0032] The radar controller processor 20 is configured to process digital signals 126 to ultimately identify the distance to objects and the angular position and velocity of those objects relative to the radar device 100. These data points can be output as a point cloud, which identifies the distance to detected objects, the radial velocity of detected objects, and the confidence level for each object detection in three-dimensional space. Typically, the digital signals 126 comprise a series of digital values ​​representing the magnitude of radar signals captured over time and received by the receiving antenna element 104. Typically, each digital value is associated with a specific number of chirps and a number of samples.

[0033] Figure 1A illustrates a series of signal processing steps implemented by processor 20 to properly process the digital signal 126 received from radar unit 10 to identify potential nearby objects. To complement Figure 1A, Figure 1B depicts in a high-level graphical manner the processing steps that processor 20 can implement to process the digital signal 126.

[0034] Specifically, the content of digital signal 126 consists of a series of data frames comprising (e.g., captured by ADC 125 of receiver unit 12) multiple digital sample values, wherein the sample values ​​are arranged in a two-dimensional matrix generated based on a pulse signal sequence. The data structure constituting a single captured frame is depicted by matrix 150 in FIG. 1B. As depicted, the single-frame data in matrix 150 comprises a two-dimensional matrix having a first dimension, referred to as the “fast-time” dimension, and representing data values ​​captured from different pulse signals. A second dimension of matrix 150 is referred to as the “slow-time” dimension and represents data values ​​captured in response to different chirped signals, which may be included within a specific pulse signal emitted by transmitter module 11. As shown in FIG. 1B, signal processing may involve processing multiple data frames represented by several matrices 150. In this disclosure, this may involve processing individual data frames, as described herein, to identify a set of peaks within the frame. Alternatively, and as described herein, this may involve processing different sub-segments of the ADC data stream to identify a set of candidate peaks, which are ultimately combined across the entire radar cube to perform final peak detection. Typically, during such signal processing, the data frames represented by matrix 150 are captured for each receive channel. Thus, Figure 1B depicts multiple matrices 150, each associated with a different receive channel and received as input data to a signal processing chain.

[0035] For a sub-segment of radar cube data that may include data represented by matrix 150, radar controller processor 20 initially performs a fast time-range fast Fourier transform (FFT) 21 (Figure 1A) to generate new frame data represented by matrix 152. FFT 21 is performed on the 1D data array (i.e., signal) associated with each different chirp in the original input matrix 150 to generate 1D transformed signals of the same length. The FFTs of each chirp in the original input frame represented by matrix 150 are combined to generate the transformed frame indicated by matrix 152. This process is repeated for each frame associated with each receive channel. The resulting data frame, representing the range map, is represented as matrix 152 in Figure 1B and can be used to determine the distance to a specific object, as reflected in the range map.

[0036] In the next step, the radar controller processor 20 performs an additional Fast Fourier Transform (FFT) 22 (Figure 1A) (referred to as slow-time or Doppler FFT) on the range map to generate new range-Doppler frame data represented by matrix 154. However, in this step, FFT 22 is applied along the opposite dimension to FFT 21. Thus, FFT 22 is performed on the 1D data array (i.e., signals) associated with each range partition in matrix 152 to generate 1D transformed signals of the same length. The FFTs of each signal in the frames of matrix 152 are combined to generate the range-Doppler data frames indicated by matrix 154. This process is repeated for each frame associated with each receiving channel. The range-Doppler data frames associated with matrix 154 provide information about potential objects moving from one sample number to the next over time. Using the data frames associated with the generated matrix 154, it is possible to process the encoded data therein to begin identifying potential objects, and, in the case of detected objects, to determine their velocity and direction of arrival.

[0037] Therefore, the radar controller processor 20 performs constant false alarm rate (CFAR) object detection 23 (Figure 1A) and 156 (Figure 1B).

[0038] If a potential object is detected, the radar controller processor 20 executes a multiple-input multiple-output (MIMO) array measurement construct 24 (Figure 1A), 158 (Figure 1B) to determine the direction of arrival (DoA) 25 (Figure 1A), 160 (Figure 1B) for each object. A final object information dataset, which may include the object identifier DoA and other relevant information (e.g., object speed), is then passed by the radar controller processor 20 (in step 26 of Figure 1A, 162 of Figure 1B) to the ADAS or other systems configured to utilize the object information to control one or more vehicle systems.

[0039] In this radar system, instead of relying on a single radar device 100, multiple radar devices 100 are combined into a single radar system. This single radar system utilizes object detection data from each individual radar device 100 to improve overall object detection and reduce the accuracy of FOV edge object detection in the individual radar devices. Specifically, this radar system implements a joint detection method using data from multiple individual radar devices 100 with overlapping FOVs.

[0040] In the aforementioned conventional signal processing chain, analog signals captured by a single radar device are processed into digital output signals, which are further manipulated to generate a dataset (e.g., a point cloud) including discrete object detection information (e.g., range, DOA, and velocity). If the magnitude of the signal being processed is below a certain threshold, this likely indicates that no reflecting object exists at the corresponding location, and the portion of the signal can be discarded. However, as explained herein, in some cases, those discarded signals may be wasted useful data. Therefore, this radar system uses prior knowledge of the potential existence of objects detected by one radar device (e.g., object detection with a low confidence interval) to recover object detection in noisy analog signals captured by another radar device, which might otherwise have been discarded as useless noise. Thus, this joint detection method effectively “fuses” the output data of multiple radar devices at the radar cube data level to improve object detection on all radar devices.

[0041] In describing the technical benefits of this invention, the process of object detection using this multi-radar device can be summarized as follows. Generally, the radar device will detect objects with a probability... Detect the target object, and will use probability. Missing the detection of the same object. In this radar system, for example, if implemented with two separate radar units, the performance of the dual radar system has four possible outcomes:

[0042] d / d: The first radar device detects an object, and the second radar device also detects an object.

[0043] d / m: The first radar device detected the object, but the second radar device missed (i.e., failed to detect) the object.

[0044] m / d: The first radar device missed the target, but the second radar device detected the target.

[0045] m / m: Both radar devices missed detecting the same object.

[0046] As described herein, this radar system executes a signal processing pipeline that can convert some d / m and m / d results (when one radar device detects an object, but the object is missed by another radar device) into d / d results (when both radar devices detect the object).

[0047] Therefore, compared to conventional automotive radar signal processing methods, this radar system can increase the number of detections and provide more information. Furthermore, the novel Direction of Arrival (DOA) verification method (which determines the angle between the automotive radar system and the detected object) can minimize false positive detections.

[0048] Figure 2 is a simplified schematic block diagram of a radar system 200 or radar system signal processor configured according to the present disclosure, the radar system 200 or radar system signal processor including a signal processing chain for dual separate radar devices 202-1 and 202-2 (collectively referred to as radar devices 202). Although simplified in Figure 2 so that only a subset of the components of each radar device 202 are shown, each radar device 202 may be configured similarly to the radar device 100 of Figure 1A.

[0049] Therefore, each radar device 202 includes input terminals 204-1 and 204-2 configured to receive input digital signals ADC1 and ADC2, respectively, which are equivalent to signals received by the antenna (not shown) of each radar device 202. Signals ADC1 and ADC2 can be generated, for example, by the ADC of each radar device 202 (e.g., ADC 125 of radar device 100 in FIG. 1A).

[0050] Within each radar device 202, the received digital signals are processed by a signal processing pipeline comprising FFT blocks 206-1 and 206-2 (collectively referred to as FFT blocks 206), which are configured to perform range and Doppler FFTs on the received signals ADC1 and ADC2, respectively. FFT blocks 206-1 and 206-2 are equivalent to the fast time (range) FFT 21 and slow time (Doppler) FFT 22 of the radar device 100 of FIG. 1A.

[0051] Once processed, the dataset generated by each FFT block 206-1, 206-2 (e.g., similar to matrix 154 in Figure 1B) is processed through CFAR blocks 207a, 207b (e.g., CFAR detection 23 of radar device 100 in Figure 1A) and DOA estimation blocks 208a, 208-2 (e.g., DOA estimation block 25 of radar device 100 in Figure 1A).

[0052] Therefore, the output of DOA estimation blocks 208a, 208-2 or radar devices 202-1, 202-2 is a dataset including point clouds Pcl1 and Pcl2 generated by each independently operating radar device 202. As described above, for various objects detected by each corresponding radar device 100, each point cloud Pcl1, Pcl2 includes the object's identifier, its relative velocity, its DOA (relative to the radar system that detected the object), and may include additional data describing the object identified by the radar device 100.

[0053] In a specific example, such as if radar device 202 is a MIMO radar device configured with N virtual receivers, then radar device 202 generates ADC data for N channels (i.e., signals ADC1 and ADC2). In this case, the FFT operation performed by FFT block 206 is performed on the ADC data from each of these N channels. Specifically, within each FFT block 206, a first FFT1 performs an FFT along the range bin of each channel, and a second FFT2 performs an FFT along the Doppler bin of each channel. Therefore, the output of FFT block 206 is N 2D complex matrices with range and Doppler (radial velocity) dimensions. These matrices may be referred to herein as a dataset including 'spectrums', and the set of N spectra represents the 'data cube' generated by FFT block 206 for a specific radar device 202.

[0054] Within those spectral datasets, object detection manifests as peaks in other data values ​​at specific distance and Doppler velocity (e.g., radial velocity) values ​​associated with the particular detection. To detect these peaks (also referred to as “data peaks”), the output of FFT block 206 is fed as input to CFAR blocks 207a and 207b (described in more detail below), respectively, to process the input data using the CFAR detection algorithm to detect peaks in the range-Doppler input data, thereby identifying potentially detected objects.

[0055] At each peak identified by enhanced CFAR blocks 212-1, 212-2, the set of N complex values ​​associated with the peak position from each channel is called a 'snapshot' of the peak position. DOA blocks 208a, 208-2 (e.g., DOA estimation block 25 of the radar device 100 in Figure 1A) further use the relative phase between channels in the snapshot to determine the DOA of each peak (i.e., the detected object). Conventional DOA algorithms can be applied to estimate the direction of arrival of the signal in azimuth (and elevation (if available)). DOA algorithms can determine the number of objects at the same distance-Doppler bin and their corresponding DOA.

[0056] In Figure 2, blocks 206, 208, and 210 operate independently within each radar device 202 in a manner substantially unchanged compared to the configuration of a single radar device 100 in Figure 1A. However, within radar system 200, as described below, additional functional components are provided and described to enable radar devices 202-1 and 202-2 to operate in combination as a multi-radar device radar system.

[0057] Therefore, the datasets Pcl1 and Pcl2 output by the DOA block 208 of each radar device 202 are combined by the sensor exchange block 210, so that the point cloud Pcl1 generated by radar device 202-1 can be used by the signal processing system to analyze the output data of radar device 202-2, and conversely, the point cloud Pcl2 generated by radar device 202-2 can be used by the signal processing system to process the output data of radar device 202-1.

[0058] For illustration, Figure 3 is a block diagram depicting additional details of the sensor exchange block 210 of Figure 2. The primary function of the sensor exchange block 210 is to implement the translation of data received from each radar device 202, enabling proper combination and processing of the data. Specifically, because the output dataset of each radar device 202 includes point cloud data generated by each radar device 202, the point cloud data is represented in polar coordinates with an origin set at the location of the antenna of the respective radar device 202. Therefore, the data received from each radar device 202 must be translated within a specific coordinate space to be represented in the same reference system (e.g., with the same origin), allowing the data to be combined into a single usable dataset in a single, consistent coordinate space. Thus, the dataset received by the sensor exchange block 210 from the first radar device 202-1 (and specifically, the data associated with the overlapping area of ​​the FOV of each radar device) is transformed from the polar coordinate space of the first radar device 202-1 to the polar coordinate space of the second radar device 202-2 using a coordinate transformation function. The reverse is also correct, such as converting the data received from the second radar device 202-2 into the polar coordinate space of the first radar device 202-1 via a coordinate transformation function.

[0059] As shown in Figure 3, these coordinate space transformations can be performed in two steps. First, block 304-1 transforms the point cloud (Pcl1) received from the first radar device 202-1 at input terminal 302-1 (which is represented in the coordinate space of the radar device) into a general or "vehicle" coordinate system. The vehicle coordinate system can be represented, for example, by a coordinate system in which the center of the vehicle is determined as the origin of a polar coordinate system.

[0060] To illustrate these different coordinate spaces, Figure 4 depicts a basic vehicle 402 with two radar units 202-1 and 202-2. The position of vehicle 402 (indicated by origin 404) can be defined on the XY axes as the origin of its own vehicle coordinate system. Additionally, each radar unit 202 has a known position for determining its own origin (i.e., origin 406 of radar unit 202-1 and origin 408 of radar unit 202-2). Therefore, the relative position of nearby objects can be represented using polar coordinates relative to any of these origins—origin 406 of the polar coordinate system of radar unit 202-1, origin 408 of the polar coordinate system of radar unit 202-2, and origin 404 of the polar coordinate system of vehicle 402, which can be considered the vehicle coordinate system of the radar system.

[0061] Returning to Figure 3, once transformed to the vehicle's coordinate system, block 306-1 transforms the point cloud to the coordinate system associated with the second radar device 202-2. As mentioned above, this transformation can be performed only on a subset of the data elements of the original point cloud data (i.e., Pcl1) falling within the FOV of the second radar device 202-2.

[0062] Similarly, block 304-2 transforms the point cloud (Pcl2) received from the second radar device 202-2 at input terminal 302-2 (which is represented in the coordinate space of the radar device) to the same vehicle coordinate system. Once transformed to the vehicle coordinate system, block 306-2 transforms the point cloud to the coordinate system associated with the second radar device 202-2. As described above, this transformation can be performed only on data elements of the original point cloud data (i.e., Pcl2) that fall within the FOV of the first radar device 202-1.

[0063] In various implementations, the transformation of various point cloud datasets to different coordinate spaces may involve the transformation of range, azimuth (and optionally elevation) data elements from the point cloud of one radar device to the coordinate system of another radar device.

[0064] Furthermore, this transformation process can involve converting the radial velocity of a point cloud dataset into a desired radial velocity in another coordinate space, as would be observed by another radar device. This is more difficult than a simple transformation of position data between various radar devices, which may only measure a single component of the velocity vector. That is, the radar device may not measure or determine the tangential velocity components in the azimuth and elevation directions. Therefore, one approach to performing the transformation of velocity data from one coordinate space to another could involve assuming that the tangential velocity component in the first radar frame is zero. This can introduce a bias or error into the translated velocity data, but in typical applications, this bias or error is usually negligible in the far field (i.e., when the range-to-baseline ratio is relatively large).

[0065] As described above, the point cloud data processed by sensor exchange block 210 may include confidence data for each peak identified in the point cloud dataset. Therefore, during processing, sensor exchange block 210 can be configured to represent the confidence data for each detected peak using a covariance matrix. For the confidence value associated with the tangential velocity component, sensor exchange block 210 is configured to base it on a reasonable maximum range assumed experimentally. Once determined, the covariance matrix is ​​transformed between radar device coordinate spaces. An example method for such a transformation involves first generating a covariance matrix representing the measurement uncertainty, where each point in the covariance matrix is ​​associated with a specific range-azimuth-elevation angle of the point cloud of radar device 202-1. The uncertainty is generated by the sensor characteristics of the components of radar device 202-1. Reasonable values ​​are assumed for the tangential velocity component when generating the covariance matrix.

[0066] In a specific method for generating the covariance matrix, the polar coordinates of the first point in the point cloud of radar device 202-1 can be expressed as: In this case, the corresponding points in the covariance matrix (i.e., in the reference frame of radar device 202-2) can be defined as follows: In this case, the transformation of the point is established as follows:

[0067] (1)

[0068] In expression (1), These are the corresponding polar coordinates in radar device 202-2. Let be a rotation matrix, and Let be the translation vector between radar frame 1 and radar frame 2. The overall expression (1) can be summarized as follows: It is nonlinear due to the transformation between polar and Cartesian coordinate systems. If expression (1) is linear, for example... Then the covariance matrix can be propagated exactly as The Jacobian matrix J of the function f(.) is the partial derivative matrix that linearizes the nonlinear equation f(.) around the point. The covariance can then be approximately propagated as... Because f(.) can be decomposed into linear operations (and...). and The linearization involves multiplication and nonlinear operations (polar2cart(.) and cart2polar(.) functions), so in practice, it is only feasible to linearize the nonlinear part, and the linearization is divided into two or more distinct steps due to the propagation of the covariance matrix.

[0069] Once the data received from each radar device 202 has been shifted as described above, the resulting dataset can be correlated in correlation block 308 (e.g., combined into a single dataset), and further enhanced object detection analysis can be performed, as described below via tracking enhancement direct blocks 310-1 and 310-2.

[0070] During the processing of sensor exchange block 210 and as described above, these transformations between coordinate spaces are expected to increase detection uncertainty in the near field. However, as mentioned above, this uncertainty decreases in the far field. For this reason, in some applications, the multi-radar enhanced object detection scheme of the present invention can be reliably used only for object detection in the far field (e.g., at distances several times (such as three or more times) the distance between radar devices). Meanwhile, it is precisely in these far-field object detections that the probability of object detection tends to be lowest (compared to near-field detection), therefore even with this constraint, the multi-radar object detection method of the present invention can have significant benefits for automotive radar systems.

[0071] In cases where data generated by each radar device 202 is transformed between coordinate spaces, as described above, data from one radar device (e.g., radar device 202-1) can be used to detect potential missed detections in data generated by another radar device (e.g., radar device 202-2). In other words, data identifying object detections (e.g., peaks) in the point cloud data generated by the first radar device can be used to increase the likelihood of identifying corresponding object detections or true peaks (otherwise missed) in data detected by the second radar device, and vice versa. Specifically, the data processing in blocks 304, 306, and 308 transforms object detection data from the point cloud data of the first radar device to the data and coordinate system of the second radar device, as well as the object detection confidence level, and vice versa. The transformed data is taken from a specific region of interest (ROI) in the data representing a geometric region of volume directly observed by each radar device (e.g., where the FOVs of the radar devices overlap). In various embodiments, the ROI is determined by a propagated covariance matrix Σ (as described above), which inherently describes an ellipse in the radar range-Doppler point cloud space. The position of the ellipse can be used to define the ROI.

[0072] Thus, within association block 308, for the version of point cloud data of the first radar device that has been translated into the coordinate space of the second radar device, sensor exchange block 210 is configured to identify detections (i.e., peaks) at locations in the transformed data of the first radar device that do not correspond to peaks detected at the same locations in the dataset of the second radar device. Similarly, the reverse operation is performed to identify peaks at locations in the transformed data of the second radar device that do not correspond to peaks detected at the same locations in the dataset of the first radar device.

[0073] This analysis yields a set of peaks detected by one radar device but not the other (e.g., missed d / m or m / d detections as described above). To illustrate this process, Figure 5 depicts two example frames of radar data generated by a first radar device 202-1 (frame 502-1) and a second radar device 202-2 (frame 502-2). As shown, the first radar device 202-1 detects two potential objects indicated by peak 504 at a given range and Doppler position. However, in frame 502-2, those peaks 504 are missing or otherwise not detected, indicating that the objects associated with peak 504 were not detected by the second radar device 202-2.

[0074] Similarly, as shown in Figure 5, the second radar device 202-2 detects a potential object peak indicated by peak 508 at a given range and Doppler position. However, in frame 502-1, peak 508 is missing, indicating that the object associated with peak 508 was not detected by the first radar device 202-1.

[0075] Returning to Figure 3, this analysis identifies a set of potential misses representing specific locations (i.e., range-Doppler locations) in the output data of radar device 202, excluding peaks (i.e., potential objects) detected by another radar device 202. This set of potential misses is provided as the output of sensor exchange block 210 and as the input to orbit augmentation detection block 310 (which together represent orbit augmentation data processing chain 216 of Figure 2).

[0076] In addition to a set of missed detections, the output of sensor exchange block 210 may also include the identification of peaks (i.e., potential objects) successfully identified and detected by the two radar devices 202.

[0077] Returning to Figure 3, the mere fact that the processing performed by the sensor exchange block 210 can identify a missed detection where a signal peak is detected by one radar device 202 while another radar device 202 does not detect a similar peak at the same location does not necessarily mean that the first radar device correctly detected the object while the second radar device failed to detect it. The output can also indicate a false positive detection by the first radar device. As an output of the radar system 200, a false positive is certainly undesirable because it could certainly cause the vehicle system connected to the radar system 200 to take unnecessary actions (e.g., warn the driver, apply vehicle braking or steering input) when it mistakenly believes that an object is near the vehicle.

[0078] Therefore, as described below, the tracking enhancement data processing chain 216 is configured to recover potential object omissions through the radar system 200 based on the output of the sensor exchange block 210, regardless of whether the omission is an erroneous negative (e.g., missed object) or an erroneous positive (e.g., ghosted object).

[0079] Therefore, referring to Figure 3, for each potential missed detection received from sensor exchange block 210 (i.e., for each peak observed by one radar system but not detected by another radar system), the missed detection associated with radar device 202-1 is passed to enhanced CFAR block 212-1, as described below, to identify the potential missed detection of radar device 202-1. Similarly, for each potential missed detection received from sensor exchange block 210, the missed detection associated with radar device 202-2 is passed to enhanced CFAR block 212-2, as described below, to identify the potential missed detection of radar device 202-2.

[0080] Specifically, enhanced CFAR blocks 212-1 and 212-2 are configured to perform CFAR analysis to detect peaks and potential objects in the output of their respective radar device 202 using data from another radar device 202, as described herein. The following discussion explains the operation of enhanced CFAR block 212-1; however, it should be understood that enhanced CFAR block 212-2 operates in a substantially similar manner to enhanced CFAR block 212-1 to detect missed detections by radar device 202-2.

[0081] Specifically, for each potential miss by the first radar device 202-1, the enhanced CFAR block 212-1 performs CFAR analysis of the point cloud (Pcl1) of radar device 202-1 at the location of the potential miss using a reduced CFAR threshold compared to the threshold used in conventional automotive radar applications. As described below, the reduced CFAR threshold may be determined at least in part by a probability or confidence value associated with the detection performed by the first radar device 202-1. Any peak detected in the point cloud of radar device 202-1 at the location of the potential miss, identified using the reduced CFAR threshold, represents a potential object that was not detected when radar device 202-1 performed conventional CFAR using a higher CFAR threshold. To increase the likelihood that a peak detected in the point cloud of radar device 202-1 is a true peak (i.e., associated with a real object and representing accurate object detection), translation data received from another radar device 202-2 is analyzed to determine whether radar device 202-2 also detected a peak at the same or nearby location. If a peak is found in the data of both radar devices 202-1 and 202-2 (where a lower CFAR threshold is used to analyze the point cloud of radar device 202-1 at a potential missed detection location), then this indicates a missed detection by radar device 202-1 and that the object should have been properly detected by both radar devices 202.

[0082] When implementing this algorithm, the reduced detection threshold used by the enhanced CFAR block 212-1 to analyze the point cloud data generated by the radar device 202-1 can be expressed as:

[0083] (2)

[0084] Expression (2) represents the adaptive CFAR threshold Thr for a specific distance gate. new Therefore, expression (2) is used to develop different reduced CFAR thresholds for different distances, so that the distance is fixed for a specific threshold and for each Doppler cell j (from 1 to ...). A threshold is set, therefore, each warehouse This actually represents the range of Doppler values ​​falling inside the chamber. In this case, Represented as To distinguish the thresholds used for different Doppler chambers, similarly, the object and noise probabilities are expressed as... and .

[0085] Noise can be modeled as uniform. Therefore, That is, all warehouses have equal probability. The probability is determined by the center of the detected object being within the specific Doppler bin j, and the probability is obtained by targeting the corresponding bin. For the distance gate It is obtained by integrating the probability density of the objects in the middle, that is . Modeled as a Gaussian distribution, i.e. It is uniquely characterized as the mean. and differences . and It is a subset of values ​​obtained by propagating points and covariances from other radars. This is the hyperparameter that determines the maximum allowable reduction of the default threshold. The centroid of the Gaussian distribution is the average parameter described by the output of sensor exchange block 210. The covariance, where a uniform noise distribution can be assumed across the velocity chamber, i.e., It is uniform. In calculating the reduced CFAR threshold... In the middle, hyperparameters It is configured to control the scaling ratio of the decreasing threshold.

[0086] To illustrate how a reduced CFAR threshold can detect previously missed true peaks, Figure 6 is a graph depicting an example dataset represented by trace 602, which can be processed by enhanced CFAR block 212-1 using a reduced CFAR threshold determined according to the method described above. In Figure 6, the horizontal axis represents the Doppler bins of the dataset being processed, while the vertical axis represents the signal amplitude. During normal processing (e.g., at CFAR block 207), a conventional CFAR threshold (e.g., straight horizontal line 604) is used such that the CFAR threshold remains at a predetermined default value across all bins. However, using a variable threshold determined by expression (2) above, the threshold can be reduced in certain ROIs (e.g., areas where the target is detected by another radar device) during processing by enhanced CFAR blocks 212-1 and 212-2, as indicated by line 606, which includes a pair of drops 608 indicating bins whose associated data are being processed using a CFAR algorithm with a reduced CFAR threshold. Therefore, by using a reduced threshold, the enhanced CFAR block 212-1 is able to detect peaks 610 in the data generated by radar device 202-1, which would otherwise be obscured (and missed) due to the noise level surrounding those peaks. If those peaks align with peaks detected by another radar system, this is likely to indicate a missed detection and can confirm the detection of a real object.

[0087] After the new CFAR threshold has been determined, the enhanced CFAR block 212-1 performs regular CFAR processing on the data received from the sensor exchange block 210. This reduced threshold makes it possible to detect more objects (i.e., peaks) in the data generated by the radar device 202-1 (which might be ignored during normal processing), potentially uncovering objects that would be missed if only the raw data generated by the radar device 202-1 were processed using the regular CFAR threshold. With the reduced threshold, the enhanced CFAR block 212-1 is likely to detect more peaks at higher thresholds than with regular processing.

[0088] The peaks detected by the enhanced CFAR block 212-1 (i.e., potential objects) are then processed by the DOA block 214 to determine the DOA of each of the potential objects.

[0089] When determining the DOA for each newly identified potential object, further procedures may be required to confirm that the newly identified potential object was correctly identified as an object or whether it represents a false positive resulting from a reduced CFAR detection threshold during processing of data from radar device 202-1. To provide this confirmation, a process is used to analyze data from another radar device 202-2 (which has been translated to the same coordinate space used by the first radar device 202-2) to determine whether the second radar device 202-2 also detected an object peak at the same location or at least within the ROI surrounding the location of the potential object in the point cloud of the first radar device 202-1. If both radar devices 202 observe peaks at the same or approximately the same location (e.g., within a certain DOA, such as within 2 degrees), then it can be determined that both the first radar device 202-1 and the second radar device 202-2 detected a true object. However, if no such peak is found at or near the location of a candidate peak in the data of the second radar device 202-2 within the data of the first radar device 202-1, then it can be determined that the peak in the data of the first radar device 202-1 is a false positive and can be discarded.

[0090] To implement this process, DOA verifier block 218a is configured to determine the validity of the calculated DOA of the potential object identified by enhanced CFAR block 212-1 by determining whether each of the determined DOAs falls within a threshold degree of the expected DOA, where the expected DOA is the DOA of an object translated from the coordinate space of another radar device. During CFAR processing, DOA information is unavailable, and therefore, the processing step relies solely on the expected range value and expected Doppler value to identify detections made by the two radar devices (e.g., using the reduced CFAR threshold described above). These results are processed using a DOA algorithm to obtain azimuth and elevation estimates for each detection peak. In this step, these DOAs are compared to verify whether the new detection is indeed correct. For example, the threshold could require the DOA determined by DOA estimate 214-1 to be within 3σ of the expected DOA (…). ).

[0091] In an example application of this process, the first radar unit of the radar system can detect a first peak value (which can be represented by a 4x1 vector) associated with four data points, the four data points including The data is processed to generate a corresponding (4x4) covariance matrix. The covariance matrix is ​​mapped to the coordinate space of the second radar device to generate the same four data points represented in the coordinate space of the radar device: and corresponding covariance matrix .use Data points are used to determine the association between detections performed by the first radar device and detections performed by the second radar device. For detections present in a dataset of only one radar device, a reduced CFAR threshold is calculated (e.g., using expression (2) above), and the data is processed using the reduced CFAR threshold to potentially utilize propagation points. The newly identified peak value is partially determined. The newly identified peak value can be represented by a 4x1 vector. Limitations. The following DOA validation process identifies newly discovered peaks detected using a reduced CFAR threshold. The value is the expected peak value identified by another radar device. Whether the values ​​match. These peaks may not match precisely (e.g., attributed to noise), but can be determined using the covariance matrix. The 2x2 portion determines the threshold for finding a match, and the covariance matrix, as described above, limits the threshold to... If the DOA determined by DOA validator block 218a is not within the threshold amount, then the candidate detection will definitely be discarded as an error.

[0092] In some cases, if the signal-to-noise ratio (SNR) of the corresponding signal is too low or the peak established by the snapshot of the signal is not accurately captured at the maximum of the corresponding analog signal peak, then the DOA estimate generated by the DOA verifier block 218a may be noisy. Therefore, in this configuration, the DOA verifier block 218a is configured to reject most false positive peaks, because error detection (typically noise-type) rarely has a correct estimated DOA.

[0093] The enhanced CFAR block 212-2, DOA estimation 214-2 and DOA verifier block 218-2 perform a similar process to determine a set of detection peaks (i.e., objects) of the radar device 202-2.

[0094] For each radar device 202, a set of detected peaks from conventional signal processing (i.e., the output of block 208) and a set of enhanced detected peaks (i.e., the output of DOA verifier block 218) are combined by concatenators 220-1 and 220-2, respectively. These concatenated datasets, representing the union of the output datasets of block 208 and DOA verifier block 218, will include a more complete set of detected objects, such that if radar device 202-1 detects an object but radar device 202-2 fails to detect the same object, the object detection will be recovered in the concatenated dataset output by 220-2. However, if both radar device 202-1 and radar device 202-2 fail to detect a particular object, then the process will not recover object detection. Therefore, generally, the number of detected objects increases without causing detected objects to be lost or ignored.

[0095] Although one or more embodiments of the radar system 200 of the present invention have been described in conjunction with first and second radar devices 202, it should be understood that the radar system can be implemented according to this disclosure, wherein three or more individual radar units may be utilized. In such systems, the output generated by a subset of the radar devices can be used to verify and confirm object detection performed by another radar device using reduced threshold CFAR processing.

[0096] Because the multi-radar processing of this invention can increase the number of potential detections (e.g., by enabling CFAR processing with a reduced threshold), the efficiency of the radar system can be increased to some extent (in terms of processing volume and memory requirements) by using the object detection algorithm of this invention only for far-field objects and relying solely on various radar devices for near-field object detection. Furthermore, processing can be limited to data associated with a specific area around the vehicle, such as at the edge of the radar device's FOV, where the object detection accuracy of a single radar device may be limited.

[0097] In some embodiments, the multi-radar system of the present invention can be used to recover an object when all radar devices fail to detect it. In this case, a reduced CFAR detection threshold can be used by two radar devices during normal signal processing. The detected object can be verified by determining whether the other radar device also detects a peak at the same distance and Doppler location (which may include some false affirmations due to noise peaks exceeding the reduced threshold).

[0098] In some aspects, the technology described herein relates to a system comprising: a first radar device including: a first plurality of transmitter modules configured to transmit a first plurality of transmitted radar signals; a first plurality of receiver modules configured to receive a first reflection of the first plurality of transmitted radar signals and generate a first signal based on the first received reflection; and a first processor configured to process the first signal to generate a first dataset including a first two-dimensional data frame, wherein the first two-dimensional data frame includes a first data peak associated with a first object, the first data peak being associated with a first distance value and a first direction of arrival value representing a position relative to the first radar device in a first coordinate space; and a second radar device including: a second plurality of transmitter modules configured to transmit a second plurality of transmitted radar signals; and a second plurality of receiver modules configured to receive the second plurality of transmitted radar signals. The system comprises: a second reflection of a transmitted radar signal, and a second signal generated based on the second received reflection; a second processor configured to process the second signal to generate a second dataset, the second dataset including a first value representing the position relative to the second radar device in a second coordinate space; and a third processor configured to: receive the first dataset from the first radar device; receive the second dataset from the second radar device; modify the second dataset using a coordinate transformation function to generate a third dataset, wherein the third dataset includes the second value represented in the first coordinate space; process the third dataset using a constant false alarm rate (CFAR) algorithm to identify a second data peak associated with a second distance value and a second direction of arrival; and determine that the second data peak is associated with the effective detection of the first object by comparing the second direction of arrival value with the first direction of arrival value.

[0099] In some respects, the techniques described herein relate to a system in which a first dataset and a second dataset are generated using a constant false alarm rate (CFAR) algorithm with a first detection threshold, and a third dataset is processed using a CFAR algorithm with a second detection threshold, wherein the second detection threshold is less than the first detection threshold.

[0100] In some respects, the techniques described herein relate to a system in which a second detection threshold is determined at least in part by a probability value associated with a first object.

[0101] In some respects, the techniques described herein relate to a system in which a second detection threshold is calculated according to the following expression, where is a first detection threshold, is a probabilistic detection associated with a first object, is a noise probability, and is a hyperparameter configured to control the scaling factor of the second detection threshold:

[0102] In some respects, the techniques described herein relate to a system in which a constant false alarm rate algorithm utilizing a second detection threshold is used to process a third dataset, including processing only a subset of the third dataset that falls within a region of interest at least partially determined by a first direction of arrival value.

[0103] In some respects, the technology described herein relates to a system in which a first field of view of a first radar device at least partially overlaps with a second field of view of a second radar device.

[0104] In some respects, the techniques described herein relate to a system in which a third processor is configured to determine a second direction of arrival value within a threshold of the first direction of arrival.

[0105] In some respects, the techniques described herein relate to a system in which a third processor is configured to determine that a second direction of arrival value is equal to a first direction of arrival.

[0106] In some respects, the technology described herein relates to a system in which a third processor is configured to communicate with a driver assistance system based on determining that a second data peak is associated with the effective detection of a first object.

[0107] In some aspects, the technology described herein relates to a system comprising: a first radar device configured to process a first received signal using a constant false alarm rate (CFAR) algorithm utilizing a first detection threshold to generate a first dataset including data peaks associated with a first object, wherein the data peaks are associated with distance and direction of arrival values ​​represented in a first coordinate space; a second radar device configured to process a second received signal using the CFAR algorithm utilizing the first detection threshold to generate a second dataset including first values ​​associated with distance and direction of arrival values ​​represented in a second coordinate space; and a processor configured to: receive the first dataset from the first radar device, receive the second dataset from the second radar device, modify the second dataset using a coordinate transformation function to generate a third dataset including second values ​​associated with distance and direction of arrival values ​​represented in the first coordinate space, and process the third dataset using a CFAR algorithm utilizing a second detection threshold to determine the presence of a second data peak at the distance and direction of arrival values ​​within the third dataset, and to determine that the second peak is associated with valid detection of the first object, wherein the second detection threshold is less than the first detection threshold.

[0108] In some respects, the techniques described herein relate to a system in which a second detection threshold is determined at least in part by a detection probability value associated with a first object.

[0109] In some respects, the techniques described herein relate to a system in which a second detection threshold is calculated according to the following expression, where is a first detection threshold, is a probability value associated with a first object, is a noise probability, and is a hyperparameter configured to control the scaling factor of the second detection threshold:

[0110] In some respects, the techniques described herein relate to a system in which a processor is configured to determine that a first object is associated with valid object detection by determining that a first direction of arrival of a first object determined using a first dataset is equal to a second direction of arrival determined using a second data peak in a third dataset.

[0111] In some respects, the techniques described herein relate to a system in which a processor is configured to determine that a first object is associated with valid object detection by determining a first arrival direction of a first object determined using a first dataset and a second arrival direction determined using a second data peak in a third dataset within a predetermined threshold.

[0112] In some aspects, the techniques described herein relate to a method comprising: receiving a first dataset from a first radar device, wherein the first dataset includes data peaks associated with a first object, wherein the data peaks are associated with distance values ​​and direction-of-arrival values ​​represented in a first coordinate space; receiving a second dataset from a second radar device, wherein the second dataset includes values ​​associated with distance values ​​and direction-of-arrival values ​​represented in a second coordinate space; modifying the second dataset using a coordinate transformation function to generate a third dataset, the third dataset including second values ​​associated with distance values ​​and direction-of-arrival values ​​represented in the first coordinate space; and processing the third dataset to determine that a second data peak exists within the third dataset at the distance values ​​and the direction-of-arrival values, thereby determining that the second peak is associated with valid detection of the first object.

[0113] In some respects, the techniques described herein relate to a method in which a first dataset and a second dataset are generated using a constant false alarm rate (CFAR) algorithm with a first detection threshold, and a third dataset is processed using a CFAR algorithm with a second detection threshold.

[0114] In some respects, the techniques described herein relate to a method that further includes determining that a second detection threshold is less than a first detection threshold.

[0115] In some respects, the techniques described herein relate to a method that further includes determining a second detection threshold using a detection probability value associated with a first object.

[0116] In some respects, the techniques described herein relate to a method that further includes determining a second detection threshold using the following expression, where is a first detection threshold, is a detection probability value, is a noise probability, and is a hyperparameter configured to control a scaling factor of the second detection threshold:

[0117] In some respects, the techniques described herein relate to a method that further determines the association of a first object with valid object detection by determining a first arrival direction of a first object determined using a first dataset and a second arrival direction within a threshold determined using a second data peak in a third dataset.

[0118] As those skilled in the art will understand, aspects of this disclosure can be embodied as systems, processes, methods, and / or program products. Therefore, aspects of this disclosure can take the form of entirely hardware embodiments, entirely software embodiments (including firmware, resident software, microcode, etc.), or embodiments combining software and hardware aspects (which may be generally referred to herein as "circuit," "circuit system," "module," or "system"). Furthermore, aspects of this disclosure can take the form of program products embodied in one or more computer-readable storage media on which computer-readable program code is embodied. (However, any combination of one or more computer-readable media may be utilized. A computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium.)

[0119] The block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of circuit systems, systems, methods, processes, and program products according to various embodiments of the present disclosure. In this regard, certain blocks in the block diagrams may represent portions of modules, segments, or code comprising one or more executable program instructions for implementing specified logical functions. It should also be noted that in some implementations, the functions mentioned in the various blocks may occur in a different order than those shown in the figures. For example, depending on the functionality involved, two blocks shown consecutively may actually be executed substantially simultaneously, or these blocks may sometimes be executed in reverse order.

[0120] Modules implemented in software for execution by various types of processors may, for example, comprise one or more physical or logical blocks of computer instructions, which may be organized, for example, as objects, programs, or functions. However, the executable files of the identified modules need not be physically located together, but may include distinct instructions stored in different locations, which, when logically joined together, constitute the module and implement its stated purpose. In practice, a module of executable code can be a single instruction, or many instructions, and may even be distributed across several different code segments in different programs and span several memory devices. Similarly, operational data (e.g., the knowledge base of the adaptation weights and / or biases described herein) can be identified and represented within the modules herein, and can be embodied in any suitable form and organized within any suitable type of data structure. Operational data may be collected as a single dataset or may be distributed across different locations, including across different storage devices. The data may provide electronic signals on a system or network.

[0121] These program instructions may be provided to one or more processors and / or controllers of a general-purpose computer, a special-purpose computer, or other programmable data processing equipment (e.g., a controller) to produce a machine such that the instructions, which are executed via the processor of the computer or other programmable data processing equipment, create a circuit system or component for implementing the functions / actions specified in one or more block diagram frames.

[0122] It should also be noted that each box in the block diagram, and combinations of boxes in the block diagram, can be implemented by a dedicated hardware-based system (e.g., which may include one or more graphics processing units) or a combination of dedicated hardware and computer instructions that performs the specified function or action.

[0123] The above specific embodiments are merely illustrative in nature and are not intended to limit the subject matter or the application and use of such embodiments.

[0124] As used herein, the term “exemplary” means “serving as an example, instance, or illustration.” Any embodiment described herein as exemplary is not necessarily to be construed as preferred or advantageous over other embodiments. Furthermore, there is no intention to be bound by any express or implied theory presented in the foregoing technical field, background art, or specific embodiments.

[0125] The connecting lines shown in the various figures included herein are intended to represent exemplary functional relationships and / or physical couplings between various elements. It should be noted that many alternative or additional functional relationships or physical connections may exist in embodiments of the subject matter. Furthermore, certain terms may be used herein for reference only, and therefore these terms are not intended to be limiting, and unless the context clearly indicates otherwise, the terms “first,” “second,” and other such numerical terms referring to structures do not imply order or sequence.

[0126] As used herein, a “node” means any internal or external reference point, connection point, interface, signal line, conductive element, etc., where a given signal, logic level, voltage, data mode, current, or quantity exists. Furthermore, two or more nodes can be implemented with a single physical element (and although receiving or outputting at a common node, two or more signals can still be multiplexed, modulated, or distinguished).

[0127] The foregoing description refers to elements, nodes, or features being "connected" or "coupled" together. As used herein, unless otherwise explicitly stated, "connected" means that one element is directly linked to (or directly connected to) another element, and not necessarily mechanically. Similarly, unless otherwise explicitly stated, "coupled" means that one element is directly or indirectly engaged to (or directly or indirectly electrically connected to, or otherwise connected to) another element, and not necessarily mechanically. Therefore, while the schematic diagrams shown depict an exemplary arrangement of elements, additional intervening elements, devices, features, or components may be present in embodiments of the depicted subject matter.

[0128] While at least one exemplary embodiment has been presented in the foregoing detailed descriptions, it should be understood that numerous variations exist. It should also be understood that the one or more exemplary embodiments described herein are not intended to limit the scope, applicability, or configuration of the claimed subject matter in any way. In fact, the foregoing detailed descriptions will provide a convenient guide for those skilled in the art to implement the one or more described embodiments. It should be understood that various changes can be made to the function and arrangement of the elements without departing from the scope defined by the claims, which includes known and foreseeable equivalents at the time of filing of this patent application.

Claims

1. A system, characterized in that, include: A first radar device includes: a first plurality of transmitter modules configured to transmit a first plurality of transmitted radar signals; a first plurality of receiver modules configured to receive first reflections of the first plurality of transmitted radar signals and generate a first signal based on the first received reflections; and a first processor configured to process the first signal to generate a first dataset including a first two-dimensional data frame, wherein the first two-dimensional data frame includes a first data peak associated with a first object, the first data peak being associated with a first distance value and a first direction of arrival value representing a position relative to the first radar device in a first coordinate space; and a second radar device includes: a second plurality of transmitter modules configured to transmit a second plurality of transmitted radar signals; and a second plurality of receiver modules configured to receive second reflections of the second plurality of transmitted radar signals. The system includes a second signal generated based on the second received reflection, a second processor configured to process the second signal to generate a second dataset including a first value representing the position of the second radar device in a second coordinate space, and a third processor configured to: receive the first dataset from the first radar device; receive the second dataset from the second radar device; modify the second dataset using a coordinate transformation function to generate a third dataset, wherein the third dataset includes the second value represented in the first coordinate space; process the third dataset using a constant false alarm rate (CFAR) algorithm to identify a second data peak associated with a second distance value and a second direction of arrival; and determine that the second data peak is associated with the effective detection of the first object by comparing the second direction of arrival value with the first direction of arrival value.

2. The system according to claim 1, characterized in that, The first dataset and the second dataset are generated using the constant false alarm rate algorithm with a first detection threshold, and the third dataset is processed using the constant false alarm rate algorithm with a second detection threshold, wherein the second detection threshold is less than the first detection threshold.

3. The system according to claim 2, characterized in that, The second detection threshold is determined at least in part by the probability value associated with the first object.

4. The system according to claim 3, characterized in that, The second detection threshold is calculated according to the following expression, where It is the first detection threshold. It is a probability detection associated with the first object. It is the noise probability, and These are hyperparameters configured to control the scaling ratio of the second detection threshold: 。 5. The system according to claim 2, characterized in that, Processing the third dataset using the constant false alarm rate algorithm that utilizes the second detection threshold includes processing only a subset of the third dataset that falls within the region of interest at least partially determined by the first direction of arrival value.

6. The system according to claim 1, characterized in that, The first field of view of the first radar device at least partially overlaps with the second field of view of the second radar device.

7. The system according to claim 1, characterized in that, The third processor is configured to determine that the second arrival direction value is within a threshold of the first arrival direction.

8. The system according to claim 1, characterized in that, The third processor is configured to determine that the second arrival direction value is equal to the first arrival direction.

9. A system, characterized in that, include: A first radar device is configured to process a first received signal using a constant false alarm rate algorithm that utilizes a first detection threshold to generate a first dataset including data peaks associated with a first object, wherein the data peaks are associated with distance values ​​and direction of arrival values ​​represented in a first coordinate space. A second radar device configured to process a second received signal using the constant false alarm rate (CFAR) algorithm utilizing the first detection threshold to generate a second dataset including a first value associated with a distance value and a direction of arrival value represented in a second coordinate space; and a processor configured to: receive the first dataset from the first radar device, receive the second dataset from the second radar device, modify the second dataset using a coordinate transformation function to generate a third dataset including the second value associated with the distance value and the direction of arrival value represented in the first coordinate space, and process the third dataset using the CFAR algorithm utilizing the second detection threshold to determine the presence of a second data peak at the distance value and the direction of arrival value within the third dataset, and to determine that the second peak is associated with valid detection of the first object, wherein the second detection threshold is less than the first detection threshold.

10. A method, characterized in that, include: Receive a first dataset from a first radar device, wherein the first dataset includes data peaks associated with a first object, wherein the data peaks are associated with distance values ​​and direction of arrival values ​​represented in a first coordinate space; receive a second dataset from a second radar device, wherein the second dataset includes values ​​associated with distance values ​​and direction of arrival values ​​represented in a second coordinate space; The second dataset is modified using a coordinate transformation function to generate a third dataset, the third dataset including second values ​​associated with distance values ​​and direction of arrival values ​​represented in the first coordinate space; And process the third dataset to determine that there is a second data peak at the distance value and the direction of arrival value within the third dataset, so as to determine that the second peak is associated with the effective detection of the first object.

Images