A target number detection estimation method and device, electronic equipment and storage medium

Abstract

The application discloses a target number estimation method and device, electronic equipment and storage medium. It belongs to the technical field of sensors and can be used for assisted driving and automatic driving. The method comprises: constructing a first rotation matrix of an array direction vector between a first radar and a reference point, and constructing a second rotation matrix of an array direction vector between a second radar and the reference point; obtaining a third direction vector by using the first rotation matrix and a first direction vector of the first radar, and obtaining a fourth direction vector by using the second rotation matrix and a second direction vector of the second radar; and estimating the number of detected targets in a target detection area by using the third direction vector, the fourth direction vector and a source number estimation algorithm. The method can improve the accuracy of target number estimation and the accuracy and efficiency of subsequent target position estimation. Further, the method can be applied to the Internet of Vehicles, such as vehicle external connection V2X, vehicle-to-vehicle V2V and the like.

Description

Technical Field

[0001] This application relates to the field of signal processing technology, and in particular to a method, apparatus, electronic device and storage medium for estimating the number of detected targets. Background Technology

[0002] With the rapid development of the automotive market and the continuous improvement of automotive safety standards, radar sensors, due to their superior performance, are widely used to detect potential collision hazards in the space surrounding a vehicle. As the automotive industry moves towards intelligent driving control, radar sensors have become one of the effective means to improve vehicle safety. A radar sensor is an electronic device that uses electromagnetic waves to detect targets. Its basic working principle is: by emitting electromagnetic waves to illuminate the detection area and receiving the echoes, it obtains the position information of the target within the detection area, thereby estimating the target's location. Before locating the target, it is necessary to determine the number of targets within the detection area.

[0003] Currently, to determine the number of targets within the detection area of ​​a radar sensor, a source number estimation algorithm is typically used to estimate the number of targets within the known beam of the radar.

[0004] However, in real-world driving scenarios, when multiple targets are located in the same distance resolution unit and velocity resolution unit, the aforementioned source number estimation algorithm has low accuracy in determining the number of targets, which also leads to low accuracy and efficiency in subsequent target position estimation. Summary of the Invention

[0005] This application provides a method, apparatus, electronic device, and storage medium for estimating the number of detected targets. By constructing a multi-snap signal pointing towards the target detection area using the array channel signals of each radar within a distributed radar common field of view, and then using the multi-snap signal to estimate the number of detected targets within the target detection area, the accuracy of the target number estimation can be greatly improved, as well as the accuracy and efficiency of subsequent target position estimation.

[0006] In a first aspect, embodiments of this application provide a method for estimating the number of detected targets, the method comprising:

[0007] Construct a first rotation matrix for the array steering vector between the first radar and the reference point, and construct a second rotation matrix for the array steering vector between the second radar and the reference point;

[0008] Using the first rotation matrix and the first steering vector of the first radar, a third steering vector is obtained, and using the second rotation matrix and the second steering vector of the second radar, a fourth steering vector is obtained; wherein, the first steering vector is a single snapshot signal of the first radar pointing to the target detection area, the second steering vector is a single snapshot signal of the second radar pointing to the target detection area, and the third and fourth steering vectors are two snapshot signals of the reference point pointing to the target detection area;

[0009] The number of targets within the target detection area is estimated using the third steering vector, the fourth steering vector, and the source number estimation algorithm.

[0010] In this embodiment of the application, by constructing a multi-snap signal pointing to the target detection area using the array channel signals of each radar within the common field of view of distributed radars, and using the multi-snap signal to estimate the number of targets within the target detection area, the accuracy of the target number estimation can be greatly improved, as well as the accuracy and efficiency of subsequent target position estimation.

[0011] Specifically, based on the position information of the first radar and the position information of the reference point, a first rotation matrix of the array steering vector between the first radar and the reference point is constructed. Then, using the first rotation matrix and the first steering vector of the first radar, a third steering vector can be determined. Here, the reference point can be any point, the first steering vector of the first radar is a single-shot signal of the first radar pointing to the target detection area, and the third steering vector is the signal of the reference point pointing to the target detection area. Thus, it can be seen that by using the position information of the first radar, the array channel signal, and the position information of the reference point, the single-shot signal of the first radar pointing to the target detection area can be converted into a signal of the reference point pointing to the target detection area. Similarly, based on the position information of the second radar and the aforementioned reference point, a second rotation matrix of the array steering vector between the second radar and the reference point is constructed. Then, using this second rotation matrix and the second steering vector of the second radar, a fourth steering vector can be determined. The second steering vector of the second radar is a single-shot signal pointing from the second radar towards the target detection area, and the fourth steering vector is a signal from the aforementioned reference point towards the target detection area. Therefore, using the position information of the second radar, the array channel signal, and the aforementioned reference point, the single-shot signal of the second radar pointing towards the target detection area can be converted into a signal from the aforementioned reference point pointing towards the target detection area. The third and fourth steering vectors obtained after the above signal processing are independent of each other, and thus can be approximately equivalent to two uncorrelated single-shot signals from the aforementioned reference point towards the target detection area. The target detection area includes two or more detection targets. Using the aforementioned third and fourth steering vectors, combined with a source number estimation algorithm, the number of detection targets within the detection area can be estimated. The number of detection targets can also be used for subsequent target position estimation, improving the accuracy and efficiency of target position estimation.

[0012] In practical applications of radar target detection, existing target number estimation methods require the radar to complete the detection of the target detection area in a single snapshot. However, in a single snapshot, since the radar signal has only one independent dimension, it significantly impacts the performance of radar source number estimation algorithms, specifically reducing the probability of resolving multiple targets. Therefore, existing target number estimation methods suffer from low accuracy and efficiency in target number estimation. This application's embodiment, without increasing the number of signal channels or snapshots, utilizes the array channel signals of different radars in a distributed radar system to construct an approximately equivalent, uncorrelated multi-snapshot signal pointing to the target detection area from any reference point. This multi-snapshot signal, combined with a source number estimation algorithm, is then used to estimate the number of targets within the target detection area. This significantly improves the radar system's multi-target resolution performance within the same beam, enhancing the accuracy and efficiency of target number estimation.

[0013] In one possible implementation, the first rotation matrix for constructing the array steering vector between the first radar and the reference point includes:

[0014] Using a first azimuth angle, a third azimuth angle, a first distance, and a third distance, a first rotation matrix is ​​constructed for the array guidance vector between the first radar and the reference point; wherein, the first azimuth angle is the azimuth angle of the first radar relative to the target detection area, the third azimuth angle is the azimuth angle of the reference point relative to the target detection area, the first distance is the distance of the first radar relative to the target detection area, and the third distance is the distance of the reference point relative to the target detection area;

[0015] The second rotation matrix for constructing the array guidance vector between the second radar and the reference point includes:

[0016] Using the second azimuth angle, the third azimuth angle, the second distance, and the third distance, a second rotation matrix is ​​constructed for the array guidance vector between the second radar and the reference point; wherein, the second azimuth angle is the azimuth angle of the second radar relative to the target detection area, and the second distance is the distance of the second radar relative to the target detection area.

[0017] This application provides a specific implementation method for constructing a first rotation matrix of the array steering vector between a first radar and a reference point, and a specific implementation method for constructing a second rotation matrix of the array steering vector between a second radar and the reference point. The method involves obtaining the azimuth angle of the first radar relative to the target detection area as a first azimuth angle, obtaining the azimuth angle of the reference point relative to the target detection area as a third azimuth angle, obtaining the distance of the first radar relative to the target detection area as a first distance, and obtaining the distance of the reference point relative to the target detection area as a third distance. Using the first azimuth angle, the third azimuth angle, the first distance, and the third distance, a first rotation matrix of the array steering vector between the first radar and the reference point can be constructed. This first rotation matrix can be used to subsequently convert a single-shot signal from the first radar pointing to the target detection area into a signal from the reference point pointing to the target detection area. Similarly, the azimuth angle of the second radar relative to the target detection area is obtained as the second azimuth angle, and the distance of the second radar relative to the target detection area is obtained as the second distance. Using the second azimuth angle, the third azimuth angle, the second distance, and the third distance, a second rotation matrix of the array guidance vector between the second radar and the reference point can be constructed. This second rotation matrix can be used to convert the single-shot signal of the second radar pointing to the target detection area into a signal of the reference point pointing to the target detection area.

[0018] In one possible implementation, obtaining the third steering vector using the first rotation matrix and the first steering vector of the first radar includes:

[0019] Multiply the first rotation matrix by the first guide vector to the left to obtain the third guide vector;

[0020] The process of obtaining the fourth steering vector using the second rotation matrix and the second steering vector of the second radar includes:

[0021] Multiply the second rotation matrix by the second guide vector to the left to obtain the fourth guide vector.

[0022] This application provides a specific method for obtaining a third guiding vector using a first rotation matrix and a first guiding vector of a first radar, and a specific method for obtaining a fourth guiding vector using a second rotation matrix and a second guiding vector of a second radar. Multiplying the first rotation matrix by the first guiding vector on the left yields the third guiding vector; similarly, multiplying the second rotation matrix by the second guiding vector on the left yields the fourth guiding vector. This application, through data processing of the first rotation matrix and the first guiding vector, can convert a single-shot signal from the first radar pointing towards a target detection area into a signal from a reference point pointing towards that target detection area. Similarly, through data processing of the second rotation matrix and the second guiding vector, it can convert a single-shot signal from the second radar pointing towards a target detection area into a signal from the reference point pointing towards that target detection area. This effectively solves the problem of low accuracy and efficiency in estimating the number of targets within the target detection area under single-shot conditions.

[0023] In one possible implementation, the method further includes:

[0024] Using the third steering vector, the fourth steering vector, and the super-resolution algorithm, the position of any target within the target detection area is estimated. The super-resolution algorithm includes at least one of the following: multiple signal classification (MUSIC) algorithm, minimum modulus (MNM) algorithm, maximum entropy (MEM) algorithm, rotation-invariant subspace (Esprit) algorithm, and relaxation algorithm.

[0025] In this embodiment, after determining the third and fourth steering vectors, a method for super-resolution angle measurement using the third and fourth steering vectors is also provided. Specifically, the position of the target within the target detection area is estimated using the third and fourth steering vectors and a super-resolution algorithm. The super-resolution algorithm can be a multiple signal classification (MUSIC) algorithm, a minimum modulus (MNM) algorithm, a maximum entropy (MEM) algorithm, a rotation-invariant subspace Esprit algorithm, a relaxation algorithm, etc. This embodiment estimates the position of the target within the target detection area using a super-resolution algorithm, which can improve the probability of successful target position estimation for multiple targets within the same beam of the radar system under single-shot conditions.

[0026] In one possible implementation, estimating the position of any target within the target detection area using the third steering vector, the fourth steering vector, and the super-resolution algorithm includes:

[0027] Using the third guiding vector, the fourth guiding vector, the number of targets within the target detection area, and the super-resolution algorithm, the position of any target within the target detection area is estimated.

[0028] In this embodiment, after determining the number of targets within the target detection area using a source number estimation algorithm, a specific method is provided for estimating the positions of the targets within that area. Specifically, the positions of the targets are estimated using the third and fourth steering vectors, the number of targets within the target detection area, and a super-resolution algorithm. This embodiment, by combining the source number algorithm with the estimation of the number of targets within the target detection area and estimating their positions, can improve the success rate of multi-target source number discrimination within the same beam of the radar system, as well as the multi-target resolution performance, under single-shot conditions.

[0029] In one possible implementation, the source number estimation algorithm includes at least one of the following: Minimum Description Length Criterion (MDL) algorithm and Akaike Information Content Criterion (AIC) algorithm.

[0030] In the embodiments of this application, the source number estimation algorithm may be the Minimum Description Length Criterion (MDL) algorithm, the Akaike Information Content Criterion (AIC) algorithm, etc.

[0031] Secondly, embodiments of this application provide a device for estimating the number of detected targets, the device comprising:

[0032] A construction unit is used to construct a first rotation matrix for the array steering vector between the first radar and the reference point, and to construct a second rotation matrix for the array steering vector between the second radar and the reference point;

[0033] The determining unit is used to obtain a third guiding vector using the first rotation matrix and the first guiding vector of the first radar, and to obtain a fourth guiding vector using the second rotation matrix and the second guiding vector of the second radar; wherein, the first guiding vector is a single snapshot signal of the first radar pointing to the target detection area, the second guiding vector is a single snapshot signal of the second radar pointing to the target detection area, and the third and fourth guiding vectors are two snapshot signals of the reference point pointing to the target detection area;

[0034] The estimation unit is used to estimate the number of targets in the target detection area using the third steering vector, the fourth steering vector, and the source number estimation algorithm.

[0035] In this embodiment of the application, by constructing a multi-snap signal pointing to the target detection area using the array channel signals of each radar within the common field of view of distributed radars, and using the multi-snap signal to estimate the number of targets within the target detection area, the accuracy of the target number estimation can be greatly improved, as well as the accuracy and efficiency of subsequent target position estimation.

[0036] Specifically, based on the position information of the first radar and the position information of the reference point, a first rotation matrix of the array steering vector between the first radar and the reference point is constructed. Then, using the first rotation matrix and the first steering vector of the first radar, a third steering vector can be determined. Here, the reference point can be any point, the first steering vector of the first radar is a single-shot signal of the first radar pointing to the target detection area, and the third steering vector is the signal of the reference point pointing to the target detection area. Thus, it can be seen that by using the position information of the first radar, the array channel signal, and the position information of the reference point, the single-shot signal of the first radar pointing to the target detection area can be converted into a signal of the reference point pointing to the target detection area. Similarly, based on the position information of the second radar and the aforementioned reference point, a second rotation matrix of the array steering vector between the second radar and the reference point is constructed. Then, using this second rotation matrix and the second steering vector of the second radar, a fourth steering vector can be determined. The second steering vector of the second radar is a single-shot signal pointing from the second radar towards the target detection area, and the fourth steering vector is a signal from the aforementioned reference point towards the target detection area. Therefore, using the position information of the second radar, the array channel signal, and the aforementioned reference point, the single-shot signal of the second radar pointing towards the target detection area can be converted into a signal from the aforementioned reference point pointing towards the target detection area. The third and fourth steering vectors obtained after the above signal processing are independent of each other, and thus can be approximately equivalent to two uncorrelated single-shot signals from the aforementioned reference point towards the target detection area. The target detection area includes two or more detection targets. Using the aforementioned third and fourth steering vectors, combined with a source number estimation algorithm, the number of detection targets within the detection area can be estimated. The number of detection targets can also be used for subsequent target position estimation, improving the accuracy and efficiency of target position estimation.

[0037] In practical applications of radar target detection, existing target number estimation methods require the radar to complete the detection of the target detection area in a single snapshot. However, in a single snapshot, since the radar signal has only one independent dimension, it significantly impacts the performance of radar source number estimation algorithms, specifically reducing the probability of resolving multiple targets. Therefore, existing target number estimation methods suffer from low accuracy and efficiency in target number estimation. This application's embodiment, without increasing the number of signal channels or snapshots, utilizes the array channel signals of different radars in a distributed radar system to construct an approximately equivalent, uncorrelated multi-snapshot signal pointing to the target detection area from any reference point. This multi-snapshot signal, combined with a source number estimation algorithm, is then used to estimate the number of targets within the target detection area. This significantly improves the radar system's multi-target resolution performance within the same beam, enhancing the accuracy and efficiency of target number estimation.

[0038] In one possible implementation, the construction unit is specifically used to construct the first rotation matrix of the array steering vector between the first radar and the reference point using a first azimuth angle, a third azimuth angle, a first distance, and a third distance; wherein the first azimuth angle is the azimuth angle of the first radar relative to the target detection area, the third azimuth angle is the azimuth angle of the reference point relative to the target detection area, the first distance is the distance of the first radar relative to the target detection area, and the third distance is the distance of the reference point relative to the target detection area;

[0039] The construction unit is further configured to construct the second rotation matrix of the array guidance vector between the second radar and the reference point using the second azimuth angle, the third azimuth angle, the second distance, and the third distance; wherein the second azimuth angle is the azimuth angle of the second radar relative to the target detection area, and the second distance is the distance of the second radar relative to the target detection area.

[0040] This application provides a specific implementation method for constructing a first rotation matrix of the array steering vector between a first radar and a reference point, and a specific implementation method for constructing a second rotation matrix of the array steering vector between a second radar and the reference point. The method involves obtaining the azimuth angle of the first radar relative to the target detection area as a first azimuth angle, obtaining the azimuth angle of the reference point relative to the target detection area as a third azimuth angle, obtaining the distance of the first radar relative to the target detection area as a first distance, and obtaining the distance of the reference point relative to the target detection area as a third distance. Using the first azimuth angle, the third azimuth angle, the first distance, and the third distance, a first rotation matrix of the array steering vector between the first radar and the reference point can be constructed. This first rotation matrix can be used to subsequently convert a single-shot signal from the first radar pointing to the target detection area into a signal from the reference point pointing to the target detection area. Similarly, the azimuth angle of the second radar relative to the target detection area is obtained as the second azimuth angle, and the distance of the second radar relative to the target detection area is obtained as the second distance. Using the second azimuth angle, the third azimuth angle, the second distance, and the third distance, a second rotation matrix of the array guidance vector between the second radar and the reference point can be constructed. This second rotation matrix can be used to convert the single-shot signal of the second radar pointing to the target detection area into a signal of the reference point pointing to the target detection area.

[0041] In one possible implementation, the determining unit is specifically used to multiply the first rotation matrix by the first guiding vector to the left to obtain the third guiding vector;

[0042] The determining unit is further configured to multiply the second rotation matrix by the second guiding vector to the left to obtain the fourth guiding vector.

[0043] This application provides a specific method for obtaining a third guiding vector using a first rotation matrix and a first guiding vector of a first radar, and a specific method for obtaining a fourth guiding vector using a second rotation matrix and a second guiding vector of a second radar. Multiplying the first rotation matrix by the first guiding vector on the left yields the third guiding vector; similarly, multiplying the second rotation matrix by the second guiding vector on the left yields the fourth guiding vector. This application, through data processing of the first rotation matrix and the first guiding vector, can convert a single-shot signal from the first radar pointing towards a target detection area into a signal from a reference point pointing towards that target detection area. Similarly, through data processing of the second rotation matrix and the second guiding vector, it can convert a single-shot signal from the second radar pointing towards a target detection area into a signal from the reference point pointing towards that target detection area. This effectively solves the problem of low accuracy and efficiency in estimating the number of targets within the target detection area under single-shot conditions.

[0044] In one possible implementation, the estimation unit is further configured to estimate the position of any one of the detected targets within the target detection area using the third steering vector, the fourth steering vector, and a super-resolution algorithm. The super-resolution algorithm includes at least one of the following: multiple signal classification (MUSIC) algorithm, minimum modulus (MNM) algorithm, maximum entropy (MEM) algorithm, rotation-invariant subspace (Esprit) algorithm, and relaxation algorithm.

[0045] In this embodiment, after determining the third and fourth steering vectors, a method for super-resolution angle measurement using the third and fourth steering vectors is also provided. Specifically, the position of the target within the target detection area is estimated using the third and fourth steering vectors and a super-resolution algorithm. The super-resolution algorithm can be a multiple signal classification (MUSIC) algorithm, a minimum modulus (MNM) algorithm, a maximum entropy (MEM) algorithm, a rotation-invariant subspace Esprit algorithm, a relaxation algorithm, etc. This embodiment estimates the position of the target within the target detection area using a super-resolution algorithm, which can improve the probability of successful target position estimation for multiple targets within the same beam of the radar system under single-shot conditions.

[0046] In one possible implementation, the estimation unit is specifically used to estimate the position of any one of the targets within the target detection area using the third guiding vector, the fourth guiding vector, the number of targets within the target detection area, and a super-resolution algorithm.

[0047] In this embodiment, after determining the number of targets within the target detection area using a source number estimation algorithm, a specific method is provided for estimating the positions of the targets within that area. Specifically, the positions of the targets are estimated using the third and fourth steering vectors, the number of targets within the target detection area, and a super-resolution algorithm. This embodiment, by combining the source number algorithm with the estimation of the number of targets within the target detection area and estimating their positions, can improve the success rate of multi-target source number discrimination within the same beam of the radar system, as well as the multi-target resolution performance, under single-shot conditions.

[0048] In one possible implementation, the source number estimation algorithm includes at least one of the following: Minimum Description Length Criterion (MDL) algorithm and Akaike Information Content Criterion (AIC) algorithm.

[0049] In the embodiments of this application, the source number estimation algorithm may be the Minimum Description Length Criterion (MDL) algorithm, the Akaike Information Content Criterion (AIC) algorithm, etc.

[0050] Thirdly, embodiments of this application provide a target number estimation device, which includes a processor and a memory; the memory stores computer execution instructions; the processor executes the computer execution instructions stored in the memory to cause the target number estimation device to perform the method described in the first aspect and any possible implementation thereof. Optionally, the target number estimation device further includes a transceiver for receiving or transmitting signals.

[0051] Fourthly, embodiments of this application provide a computer-readable storage medium for storing instructions or a computer program; when the instructions or the computer program are executed, the method described in the first aspect is implemented.

[0052] Fifthly, embodiments of this application provide a computer program product, the computer program product including instructions or a computer program; when the instructions or the computer program are executed, the method described in the first aspect is implemented.

[0053] In a sixth aspect, embodiments of this application provide a chip including a processor and a communication interface, wherein the processor is configured to call and execute instructions from the communication interface, and when the processor executes the instructions, the chip causes the chip to perform the method described in the first aspect or various embodiments thereof.

[0054] In a seventh aspect, embodiments of this application provide a system comprising at least one target number estimation device as described in the second or third aspect or a chip as described in the sixth aspect.

[0055] In this embodiment of the application, by constructing a multi-snap signal pointing to the target detection area using the array channel signals of each radar within the common field of view of distributed radars, and then using the multi-snap signal to estimate the number of targets within the target detection area, the accuracy of the target number estimation can be greatly improved, as well as the accuracy and efficiency of subsequent target position estimation. Attached Figure Description

[0056] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments of this application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0057] Figure 1 A schematic diagram illustrating a scenario for estimating the number of detected targets, provided as an embodiment of this application;

[0058] Figure 2 A schematic diagram of an architecture for estimating the number of detected targets provided in an embodiment of this application;

[0059] Figure 3 A schematic diagram of another architecture for estimating the number of detected targets provided in an embodiment of this application;

[0060] Figure 4 A flowchart illustrating a method for estimating the number of detected targets provided in an embodiment of this application;

[0061] Figure 5 A schematic diagram of another architecture for estimating the number of detected targets provided in an embodiment of this application;

[0062] Figure 6 This is a schematic diagram illustrating the effect of estimating the number of detected targets in an embodiment of this application.

[0063] Figure 7 This is a schematic diagram illustrating the effect of target location estimation provided in an embodiment of this application;

[0064] Figure 8 This is a schematic diagram of the structure of a target number estimation device provided in an embodiment of this application;

[0065] Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0066] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described below with reference to the accompanying drawings.

[0067] The terms "first" and "second," etc., used in the specification, claims, and drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0068] The term "embodiment" as used herein means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0069] It should be understood that in this application, "at least one (item)" means one or more, "more than one" means two or more, "at least two (items)" means two or three or more, and "and / or" is used to describe the relationship between related objects, indicating that there can be three relationships. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the related objects before and after are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0070] This application provides a method for estimating the location of a target. To describe the solution of this application more clearly, some knowledge related to target location estimation will be introduced below.

[0071] Radar, also known as a radar device, detector, radar sensor, radar detection device, or radar signal transmitting device, works by transmitting signals (or detection signals) and receiving reflected signals from target objects to detect them. The signals transmitted by a radar can be radar signals, and correspondingly, the reflected signals received from target objects can also be radar signals.

[0072] The transmission period of a radar sensor (or, as it is called, the sweep period, sweep time, or sweep duration, etc.) refers to the period during which a radar sensor transmits a radar signal with a complete waveform. A radar sensor typically transmits radar signals for multiple sweep periods over a continuous duration.

[0073] In recent years, advanced driver assistance systems (ADAS) have played a crucial role in intelligent vehicles. Utilizing various sensors installed on the vehicle, they continuously sense the surrounding environment, collect data, identify, detect, and track stationary and moving objects, and combine this data with navigation map data for system calculations and analysis. This allows the driver to anticipate potential dangers, effectively increasing driving comfort and safety. It can be said that true autonomous driving is the culmination of ADAS development. In the autonomous driving architecture, the sensing layer is likened to the car's "eyes," including visual sensors such as onboard cameras and radar sensors such as onboard millimeter-wave radar, lidar, and ultrasonic radar. Millimeter-wave radar, due to its lower cost and more mature technology, has become the primary sensor in autonomous driving systems. Currently, ADAS has developed more than ten functions, among which adaptive cruise control, automatic emergency braking, lane change assist, and blind spot monitoring all rely on onboard millimeter-wave radar.

[0074] Millimeter waves refer to electromagnetic waves with wavelengths between 1 and 10 mm, corresponding to a frequency range of 30–300 GHz. Within this frequency band, the characteristics associated with millimeter waves make them ideal for automotive applications. They offer several advantages: wide bandwidth (abundant frequency domain resources, low antenna sidelobes, facilitating imaging or quasi-imaging); short wavelength (reduced radar equipment size and antenna aperture, resulting in lighter weight); narrow beam (millimeter wave beams are much narrower than microwave beams for the same antenna size, leading to higher radar resolution); and strong penetration (compared to lidar and optical systems, they have a greater ability to penetrate smoke, dust, and fog, enabling all-weather operation).

[0075] The embodiments of this application are described below with reference to the accompanying drawings.

[0076] Please see Figure 1 , Figure 1 This is a schematic diagram illustrating a scenario for estimating the number of detected targets, provided as an embodiment of this application. Figure 1As shown, a car A is driving on a road. The radar sensor system on the car detects several cars (one, two, or more) (cars B and C) traveling in the same direction at the same speed to its left front. The dashed fan-shaped area represents the detection area of ​​the radar sensor on car A. Cars B and C are both within car A's safe detection area and are therefore identified by car A's radar sensor system as potential driving risks. Because these targets are located on the same distance and speed resolution unit of car A's radar sensor, the exact number of targets cannot be estimated; only a rough estimate of the target detection area (dashed circular area) containing one or more targets can be made. In this driving scenario, there are often multiple targets at close range, and the prerequisite for measuring them is the ability to distinguish multiple targets. Although radar sensors can distinguish targets using distance and velocity resolution, there are always situations where multiple targets are in the same distance and velocity resolution unit. In such cases, it is unavoidable to use angular dimension resolution to separate multiple targets. Otherwise, the radar sensor of vehicle A may easily treat vehicle B and vehicle C as a single unit, leading to misjudgment and a series of potential driving safety hazards.

[0077] Angle measurement is a fundamental requirement for radar sensors to estimate target position. Existing vehicle-mounted millimeter-wave radar systems primarily use antenna arrays to measure the target's angle. Common angle measurement methods include sum-difference beamforming, phase comparison, and digital beamforming (DBF). The angular resolution of the measured target is inversely proportional to the radar sensor's antenna aperture; that is, the larger the antenna aperture, the higher the antenna gain, the narrower the beamwidth, and the higher the angular resolution, resulting in stronger radar angular resolution capability. However, due to limitations imposed by vehicle installation requirements and cost, the aperture of existing vehicle-mounted radars is always limited. For example, a conventional three-transmitter, four-receiver short-to-medium-range radar has an angular resolution of only about 10 degrees. Figure 1 In the application scenario shown, if there are two vehicles side by side in front of the radar sensor at the same distance and speed resolution unit, and the angular interval between them relative to the radar sensor is much less than 10 degrees, the radar sensor system of vehicle A will treat the two detected targets (vehicle B and vehicle C) as a single target using conventional angle measurement methods. This will result in a huge error in the measurement of multiple targets, making it impossible to correctly perceive the environment ahead and causing driving hazards. Therefore, it is necessary to perform source number estimation processing and super-resolution processing for multiple targets.

[0078] To address the problem of resolving and measuring multiple targets within a single radar beam, a series of super-resolution algorithms can be employed to perform super-resolution processing on the detected targets. These include algorithms such as Multiple Signal Classification (MUSIC), Estimating Signal Parameter via Rotational Invariance (ESPRIT), Relax, and Iterative Adaptive Approach (IAA). These algorithms are named for their ability to resolve and measure multiple targets within a single beam. Furthermore, these algorithms require a known number of detected targets within the beam before performing super-resolution processing on each target. The number of detected targets within the target detection area can be estimated using source number estimation algorithms. Commonly used source number estimation algorithms include the Akaike Information Criterion (AIC) algorithm and the Minimum Description Length (MDL) algorithm. Therefore, combining these two types of algorithms can be used for resolving and measuring multiple targets within a single beam. However, in real-world driving applications, target detection processing must be performed within a single snapshot, and the aforementioned source number estimation and super-resolution algorithms perform poorly under single-snapshot conditions. For example, conventional three-transmitter, four-receiver short-range radars have a very low probability of successfully identifying source numbers under single-snapshot signals. Existing research shows that in a simulated environment with two targets 5 degrees apart, the target source number distribution obtained under a single-snapshot signal indicates that only about 40% of the targets can be successfully identified as two targets, which is far from sufficient for real-world driving applications. Furthermore, under single-snapshot conditions, since the actual independent dimension of the signal is only one dimension, the performance of super-resolution angle measurement algorithms is significantly affected, specifically resulting in a decrease in the multi-target resolution probability and angle measurement accuracy.

[0079] against Figure 1 To address the issues of low success rate in source number identification and low super-resolution angle measurement accuracy in the scenario shown, this application provides a schematic diagram of an architecture for estimating the number of detected targets. Please refer to [link to relevant documentation]. Figure 2The target number estimation architecture in this application embodiment can construct a "multi-snapshot" signal using echo signals from different radars in a distributed radar system without increasing the number of signal channels or snapshots. Here, "multi-snapshot" refers to a signal with an equivalent snapshot count greater than one. Using this signal in conjunction with a source number estimation algorithm for target number discrimination can significantly improve the success rate of multi-target source number discrimination within the radar system's beam. Furthermore, utilizing different radar signals from the distributed radar system in conjunction with a super-resolution algorithm for multi-radar joint super-resolution processing can greatly improve the multi-target resolution performance within the radar system's beam.

[0080] like Figure 2 As shown, multiple radars (radar 1, radar 2, and radar 3) can simultaneously observe signals within the target detection area (dashed elliptical region) and transmit these signals to a central processor via data transmission lines for joint processing. The two targets within this detection area are located in the same range and velocity resolution unit. The central processor can be independent of each individual radar or can be any one of the main radars in a distributed radar system (radar 1, radar 2, or radar 3). The radar data transmitted to the central processor can be raw echo analog-to-digital converter (ADC) data or array channel data detected individually by each radar; no restrictions are placed here.

[0081] In the above Figure 2 Based on the target count estimation architecture shown, its architecture can be further refined by incorporating the specific implementation process of target count estimation. See reference [link to relevant documentation]. Figure 3 , Figure 3 This is a schematic diagram of another architecture for estimating the number of detected targets provided in an embodiment of this application. Figure 3 As shown, radars 1, 2, and 3 are distributed in different locations on the intelligent vehicle. Specifically, they can be any type of radar sensor designed based on the time-of-flight principle, such as millimeter-wave radar or lidar, each possessing independent ranging, speed, and angle measurement capabilities. Radars 1, 2, and 3 share a common field of view and can simultaneously and independently detect targets within this common field of view with millisecond-level time synchronization accuracy. This millisecond-level time synchronization accuracy ensures that the scene within the common field of view of each radar is identical, and this millisecond-level time synchronization accuracy can be achieved using the Controller Area Network (CAN) bus. Optionally, the number of radars in the architecture shown in this embodiment is M, where M is a positive integer greater than or equal to 2. For the sake of simplicity and convenience, M is taken as 3 in this embodiment, and M radars represent radars 1, 2, and 3 as described above.

[0082] Each radar performs a two-dimensional Fast Fourier Transform (FFT) on the raw echo IQ data it acquires to obtain range-velocity domain (RV-MAP) results. Then, after noncoherent accumulation of all antenna channels (with N antennas) for a single radar, independent constant false alarm rate (CFAR) detection is performed to obtain the array channel data (only N*1 dimensional data) related to the target detection area within the radar's common field of view, as well as the target detection area's range information. The array channel data and range information related to the target detection area obtained by each radar are then transmitted to the central processing unit to obtain N*M dimensional data. The raw echo IQ data is obtained by vector decomposition of the raw echo signal, that is, decomposing the raw echo signal into two components with the same frequency and peak amplitude but a 90° phase difference. Using vectors to represent the signal can completely describe the signal's amplitude, frequency, and phase. A vector is a rotating arrow in a Cartesian coordinate system. The length of the arrow represents the peak amplitude of the signal. The counterclockwise rotation direction is the positive direction. The angle between the arrow and the positive half-axis of the horizontal axis is the phase. The signal period corresponds to the time it takes for the arrow to rotate once. The number of rotations the signal completes per second corresponds to the signal frequency. The projection length of the signal vector onto the vertical axis is equal to the peak amplitude of the signal multiplied by the phase sine value; therefore, if the signal is a sine wave, this projection corresponds to the instantaneous amplitude of the signal. Typically, a sine signal and a cosine signal are used to describe these two components, where the cosine component is called the in-phase component (I component), and the sine component is called the quadrature component (Q component). Furthermore, constant false alarm probability detection is a common method for radar target detection. Its detection function is to determine the presence or absence of a target; therefore, its detection function can be transformed into determining the presence or absence of a signal in the presence of noise. Figure 3 As shown, radar 1 performs FFT processing on the acquired raw echo IQ data to obtain RV-MAP results, and independently performs CFAR detection on the N antenna channels of radar 1 after noncoherent accumulation, which can obtain indistinguishable targets ( Figure 2 The distance R1 to the target detection area in the array and the array channel signal vector S1 are also considered. Similarly, radar 2 performs FFT processing on the acquired raw echo IQ data to obtain RV-MAP results, and independently performs CFAR detection on the N antenna channels of radar 2 after noncoherent accumulation, which can obtain indistinguishable targets ( Figure 2The distance R2 to the target detection area in the array and the array channel signal vector S2 are obtained; Radar 3 performs FFT processing on the acquired raw echo IQ data to obtain RV-MAP results, and performs independent CFAR detection on the N antenna channels of Radar 3 after noncoherent accumulation, which can obtain indistinguishable targets ( Figure 2 The distance R3 of the target detection area and the array channel signal vector S3.

[0083] Optionally, the aforementioned radars perform various data processing steps on the raw echo IQ data acquired during detection, ultimately obtaining the indistinguishable target (…). Figure 2 The process of obtaining distance information and array channel data for the target detection area (in the context of the radar) can be performed by each radar, which then transmits the obtained distance information and array channel data to the central processing unit (CPU). Alternatively, the CPU can directly transmit the raw echo IQ data acquired by each radar to the CPU, which then processes the raw echo IQ data to obtain the multidimensional indistinguishable target (in the context of the radar). Figure 2 The distance information of the target detection area and the array channel data.

[0084] After acquiring the aforementioned N*M dimensional data, the central processing unit (CPU) selects an arbitrary point as a reference point. Based on the position information of this reference point, as well as the spatial geometric position and array tilt angle of each radar, it performs coordinate transformation and corresponding signal processing on the array channel data of different radars. This transforms the simultaneous signals of radars at different positions into "multi-snapshot" signals relative to the same geometric position (i.e., the aforementioned reference point). In this case, it should be a three-snapshot signal relative to the aforementioned reference point. The CPU then uses a source number estimation algorithm, combined with the aforementioned three-snapshot signal, to estimate the number of detected targets within the target detection area, which can greatly improve the probability of successful estimation of the number of multiple target sources within the radar system's beam. Based on determining the number of detected targets within the target detection area, the CPU then uses a super-resolution algorithm, combined with the aforementioned three-snapshot signal, to perform super-resolution angle measurement on each detected target within the target detection area. The aforementioned source number estimation algorithm may include the minimum description length (MDL) algorithm, the AIC algorithm, etc. The aforementioned super-resolution algorithms may include multiple signal classification (MUSIC) algorithms, minimum modulus (MNM) algorithms, maximum entropy (MEM) algorithms, rotation-invariant subspace (estimating signal parameter via rotational invariance, Esprit) algorithms, and relaxation algorithms.

[0085] Based on the above Figure 2 and Figure 3 The target number estimation architecture shown in this application also provides a target number estimation method. This method can construct a "multi-snapshot" signal using echo signals from different radars in a distributed radar without increasing the number of signal channels or snapshots. Then, the signal is used to estimate the number of targets in the target detection area and to perform multi-radar joint super-resolution angle measurement processing. This can greatly improve the probability of successfully estimating the number of multiple target sources within the radar system's beam and greatly improve the super-resolution angle measurement performance of multiple targets within the radar system's beam.

[0086] Please see Figure 4 , Figure 4 The following is a flowchart illustrating a method for estimating the number of detected targets, provided in an embodiment of this application. This method includes, but is not limited to, the following steps:

[0087] Step 401: Construct a first rotation matrix for the array steering vector between the first radar and the reference point, and construct a second rotation matrix for the array steering vector between the second radar and the aforementioned reference point.

[0088] The electronic device constructs a first rotation matrix for the array steering vector between the first radar and the reference point, and a second rotation matrix for the array steering vector between the second radar and the aforementioned reference point. In this embodiment, the electronic device is equipped with a processor for data processing. This processor can be independent of each individual radar, or it can be any one of the distributed radars (either the first radar or the second radar).

[0089] In this embodiment, the first and second radars share a common field of view, forming a distributed radar system. Both radars possess independent ranging, velocity, and angle measurement capabilities, enabling simultaneous and independent detection of targets within the shared field of view with millisecond-level time synchronization accuracy. It should be noted that the use of "first" and "second" to name the radars in this embodiment is solely for distinguishing different radars and does not imply any order of precedence or hierarchy. Furthermore, the distributed radar system is not limited to two radars; it can consist of two or more radars sharing a common field of view. For greater clarity and simplicity, this embodiment uses "first radar" and "second radar" as examples.

[0090] In this embodiment, the reference point can be any spatial coordinate point whose location information is known. This reference point can be a point different from the spatial locations of the first and second radars, or it can coincide with the spatial location of either the first or second radar. This embodiment uses a reference point different from the spatial locations of the first and second radars as an example for illustration. It should be noted that in this embodiment, the location information of the reference point, the first radar, and the second radar are all stored in the electronic device, specifically including location-related information such as range, azimuth, elevation, and array tilt.

[0091] The electronic device constructs a first rotation matrix for the array steering vector between the first radar and the reference point based on the position information of the first radar and the reference point. Similarly, it constructs a second rotation matrix for the array steering vector between the second radar and the reference point based on the position information of the second radar and the reference point. Specifically, it obtains the azimuth angle of the first radar relative to the target detection area as a first azimuth angle, the azimuth angle of the reference point relative to the target detection area as a third azimuth angle, the distance of the first radar relative to the target detection area as a first distance, and the distance of the reference point relative to the target detection area as a third distance. Using the first azimuth angle, the third azimuth angle, the first distance, and the third distance, a first rotation matrix for the array steering vector between the first radar and the reference point can be constructed. This first rotation matrix can be used to subsequently convert a single-shot signal from the first radar pointing to the target detection area into a signal from the reference point pointing to the target detection area. Similarly, the azimuth angle of the second radar relative to the target detection area is obtained as the second azimuth angle, and the distance of the second radar relative to the target detection area is obtained as the second distance. Using the second azimuth angle, the third azimuth angle, the second distance, and the third distance, a second rotation matrix of the array guidance vector between the second radar and the reference point can be constructed. This second rotation matrix can be used to convert the single-shot signal of the second radar pointing to the target detection area into a signal of the reference point pointing to the target detection area.

[0092] In this embodiment of the application, the target detection area refers to the smallest detection area that includes one or more targets located at the same distance and velocity resolution unit, such as the aforementioned Figure 2 The dashed elliptical region in the diagram represents the smallest detection area containing two targets at the same distance and velocity resolution unit; therefore, this dashed elliptical region is the target detection area. At this point, the situation within the target detection area is still unrecognizable under a single snapshot signal (low target count identification accuracy), requiring estimation of the number and position of the detected targets through the embodiments of this application.

[0093] Step 402: Using the first rotation matrix and the first steering vector of the first radar, obtain the third steering vector, and using the second rotation matrix and the second steering vector of the second radar, obtain the fourth steering vector.

[0094] The electronic device uses the first rotation matrix and the first steering vector of the first radar to obtain a third steering vector, and uses the second rotation matrix and the second steering vector of the second radar to obtain a fourth steering vector. The first steering vector of the first radar is a single-shot signal pointing from the first radar to the target detection area, and the third steering vector is a signal pointing from the reference point to the target detection area. Therefore, using the position information and array channel signal of the first radar, as well as the position information of the reference point, the single-shot signal pointing from the first radar to the target detection area can be converted into a signal pointing from the reference point to the target detection area. Similarly, the second steering vector of the second radar is a single-shot signal pointing from the second radar to the target detection area, and the fourth steering vector is a signal pointing from the reference point to the target detection area. Therefore, using the position information and array channel signal of the second radar, as well as the position information of the reference point, the single-shot signal pointing from the second radar to the target detection area can be converted into a signal pointing from the reference point to the target detection area. Since the third and fourth steering vectors obtained after the above signal processing are independent, they can be approximately equivalent to two uncorrelated single-shot signals pointing from the reference point to the target detection area. Specifically, multiplying the first rotation matrix by the first guide vector on the left yields the third guide vector; similarly, multiplying the second rotation matrix by the second guide vector on the left yields the fourth guide vector.

[0095] This application embodiment can convert a single-shot signal of the first radar pointing to the target detection area into a signal of the reference point pointing to the target detection area by processing the first rotation matrix and the first guide vector. By processing the second rotation matrix and the second guide vector, the single-shot signal of the second radar pointing to the target detection area can be converted into a signal of the reference point pointing to the target detection area. This effectively solves the problem of low accuracy and efficiency in estimating the number of multiple targets under single-shot conditions.

[0096] Step 403: Using the third steering vector, the fourth steering vector, and the source number estimation algorithm, estimate the number of targets within the target detection area.

[0097] The electronic device utilizes the aforementioned third and fourth steering vectors, combined with a source number estimation algorithm, to estimate the number of targets within the target detection area. The source number estimation algorithm can be the Minimum Description Length (MDL) algorithm, the AIC algorithm, etc.

[0098] In practical applications of radar target detection, existing target number estimation methods require the radar to complete the detection of the target detection area in a single snapshot. However, in a single snapshot, since the radar signal has only one independent dimension, it significantly impacts the angle measurement performance of the radar super-resolution algorithm. Specifically, this manifests as a decrease in the resolution probability of multiple targets and the accuracy of angle measurement. Therefore, existing target number estimation methods suffer from low accuracy and efficiency in estimating the number of detected targets. This application's embodiment, without increasing the number of signal channels or snapshots, utilizes the array channel signals of different radars in a distributed radar system to construct an approximately equivalent, uncorrelated multi-snapshot signal pointing to the target detection area from any reference point. This multi-snapshot signal, combined with a source number estimation algorithm, is then used to estimate the number of detected targets. This significantly improves the radar system's multi-target resolution performance within the same beam and enhances the accuracy and efficiency of subsequent target position estimation.

[0099] Furthermore, after determining the third and fourth steering vectors through step 402, this embodiment can also utilize the third and fourth steering vectors to perform super-resolution angle measurement on the detected targets within the target detection area. Specifically, by using the third and fourth steering vectors and a super-resolution algorithm, the position of the detected targets within the target detection area can be estimated, which can greatly improve the accuracy and efficiency of target position estimation. The super-resolution algorithm can be a multiple signal classification (MUSIC) algorithm, a minimum modulus (MNM) algorithm, a maximum entropy (MEM) algorithm, a rotation-invariant subspace Esprit algorithm, a relaxation algorithm, etc. This embodiment uses a super-resolution algorithm to estimate the position of the detected targets within the target detection area, which can improve the probability of successful target position estimation for multiple targets within the same beam of the radar system under single snapshot conditions.

[0100] Furthermore, after determining the number of targets within the target detection area using the source number estimation algorithm, the positions of the detected targets are estimated using the aforementioned third steering vector, fourth steering vector, the number of targets within the target detection area, and the super-resolution algorithm. This embodiment of the application, by combining the source number algorithm to estimate the number of targets within the target detection area and to estimate the positions of the detected targets, can improve the success rate of multi-target source number discrimination within the same beam of the radar system, as well as the multi-target resolution performance, under single-shot conditions.

[0101] On the other hand, when the aforementioned reference point coincides with the spatial location of either the first or second radar (taking the coincidence with the spatial location of the first radar as an example), the electronic device still stores the position information of the first radar (reference point) and the second radar. Specifically, this includes position-related information such as range, azimuth, elevation, and array tilt. Based on the position information of the first and second radars, the electronic device constructs a rotation matrix for the array steering vector between the first and second radars. Specifically, using the azimuth of the first radar relative to the target detection area, the azimuth of the second radar relative to the target detection area, the distance of the first radar relative to the target detection area, and the distance of the second radar relative to the target detection area, a rotation matrix for the array steering vector between the first and second radars is constructed. This rotation matrix can be used to subsequently convert the single-shot signal of the second radar pointing towards the target detection area into a signal of the first radar pointing towards the target detection area. The electronic device then uses the aforementioned rotation matrix and the second steering vector of the second radar to obtain a fourth steering vector. In this context, the second steering vector of the second radar is a single-shot signal pointing to the target detection area, and the fourth steering vector is the signal of the first radar pointing to the target detection area. Therefore, by utilizing the position information and array channel signals of the second radar, as well as the position information of the first radar, the single-shot signal of the second radar pointing to the target detection area can be converted into the signal of the first radar pointing to the same target detection area. Furthermore, the first steering vector of the first radar is a single-shot signal pointing to the target detection area. Since the first steering vector and the fourth steering vector obtained after the signal processing are independent, they can be approximately equivalent to two unrelated single-shot signals of the first radar pointing to the target detection area. Specifically, multiplying the rotation matrix to the left by the second steering vector yields the fourth steering vector.

[0102] In this embodiment, when the spatial position of the reference point coincides with the spatial position of the first radar, a rotation matrix is ​​constructed between the first and second radars. Then, data processing is performed on the rotation matrix and the second guide vector of the second radar. This can convert the single-shot signal of the second radar pointing to the target detection area into a signal of the first radar pointing to the target detection area. Combined with the first guide vector of the first radar itself, this becomes a single-shot signal of the first radar pointing to the target detection area. This is approximately equivalent to two unrelated snapshot signals of the first radar pointing to the target detection area, effectively solving the problem of low accuracy and efficiency in estimating the number of multiple targets under single-shot conditions.

[0103] Please see Figure 5 , Figure 5 This is a schematic diagram of another architecture for estimating the number of detected targets provided in an embodiment of this application. For example... Figure 5As shown, joint super-resolution processing is performed using two corner radars (radar 1 and radar 2) on the intelligent vehicle. Both radar 1 and radar 2 have independent range, velocity, and angle measurement capabilities. Radar 1 and radar 2 share a common field of view and can simultaneously and independently detect targets within the common field of view with millisecond-level time synchronization accuracy. The millisecond-level time synchronization accuracy ensures that the scene within the common field of view of each radar is identical.

[0104] The target detection area received by each of the above radars within the common field of view ( Figure 5 After detecting the echo from the dashed elliptical region (as shown in the image), the original echo data undergoes FFT processing. Separate fast-time and slow-time domain signal processing is then performed to obtain the RV spectrum of each radar antenna. Non-coherent accumulation and CFAR detection are then performed on each antenna of a single radar to obtain the array channel data related to the target detection area within the radar's common field of view, as well as the target detection area's range information. The array channel data and range information related to the target detection area within the common field of view obtained from each radar are then uploaded to the central processing unit (CPU). The CPU then obtains the range information R1 and R2, as well as the array channel signals S1 and S2 from both radars.

[0105] In this embodiment, the target detection area refers to the smallest detection area that includes the detection targets A and B located at the same distance and velocity resolution unit. At this time, the situation within the target detection area is still an indistinguishable target under a single snapshot signal, which is the target to be detected in this embodiment. The number and position of the detection targets need to be estimated through this embodiment.

[0106] After the central processing unit obtains the distance information and array channel data detected by the various radars, it selects any point as a reference point, such as... Figure 5 As shown, taking the midpoint O of the line segment containing radars 1 and 2 as the reference point, and with reference point O as the origin and the line containing radars 1 and 2 as the X-axis, a Cartesian coordinate system is established. In this system, the position coordinates of reference point O can be represented as (0, 0), the position coordinates of radar 1 as (x1, y1), the position coordinates of radar 2 as (x2, y2), and the position coordinates of the target detection area as (x...). t y t In this context, the center coordinates of the target detection area can be used as the position coordinates of the target detection area. The azimuth angle of radar 1 relative to the target detection area is obtained as the first azimuth angle θ1, and the distance of radar 1 relative to the target detection area is obtained as the first distance R1. The azimuth angle of radar 2 relative to the target detection area is obtained as the second azimuth angle θ2, and the distance of radar 2 relative to the target detection area is obtained as the second distance R2. The array tilt angle of radar 1 is then determined. The array tilt angle of radar 2 is obtained as follows The azimuth angle of reference point O relative to the target detection area is obtained as the third azimuth angle θ0, and the distance of reference point O relative to the target detection area is obtained as the third distance R0. All of the above geometric position information is stored in the central processing unit (CPU). Based on this geometric position information, the CPU can perform coordinate transformation and signal processing on the array channel signals S1 and S2, converting them into target array signals S01 and S02 at the new reference point O.

[0107] Specifically, the central processing unit (CPU) uses the first azimuth angle θ1, the third azimuth angle θ0, the first distance R1, and the third distance R0 to construct a first rotation matrix for the array steering vector between radar 1 and the reference point O. This first rotation matrix can be used to subsequently convert the single-shot signal of radar 1 pointing to the target detection area into a signal of reference point O pointing to the target detection area. Using the second azimuth angle θ2, the third azimuth angle θ0, the second distance R2, and the third distance R0, a second rotation matrix for the array steering vector between radar 2 and the reference point O can be constructed. This second rotation matrix can be used to subsequently convert the single-shot signal of radar 2 pointing to the target detection area into a signal of reference point O pointing to the target detection area. The CPU then multiplies the first rotation matrix by the steering vector of radar 1 to convert the single-shot signal of radar 1 pointing to the target detection area into a signal S01 of reference point O pointing to the target detection area. Similarly, multiplying the second rotation matrix by the steering vector of radar 2 to convert the single-shot signal of radar 2 pointing to the target detection area into a signal S02 of reference point O pointing to the target detection area. Since the target array signals S01 and S02 obtained after the above signal processing are independent of each other, the target array signals S01 and S02 can be approximately equivalent to two uncorrelated snapshot signals pointing from the reference point O to the target detection area. This effectively solves the problem of low accuracy and efficiency in estimating the number of multiple targets under a single snapshot.

[0108] The central processing unit can use the obtained target array signals S01 and S02, along with a source number estimation algorithm, to estimate the number of targets within the aforementioned target detection area. This source number estimation algorithm can include the Minimum Description Length (MDL) algorithm, the AIC algorithm, etc. The probability results of the aforementioned source number estimation can be found in [reference needed]. Figure 6 , Figure 6 This is a schematic diagram illustrating the effect of estimating the number of detected targets according to an embodiment of this application. Figure 6 As shown above, this demonstrates the situation described above. Figure 5 The comparison results of the estimation probability of the number of single radar signal sources and the estimation probability of the number of distributed dual radar signal sources are presented in the simulation environment when the two targets are 5 degrees apart in the architecture shown (the angular resolution of the radar antenna array is about 10 degrees). Figure 6 The second and third bar charts from left to right represent the target number distribution given by the single radar processing of the left and right corner radars, respectively. It can be seen that there is only a 40% probability of successfully identifying the target within the above target detection area as a dual target. Figure 8 The first bar chart from left to right shows the target number distribution given by the combined left and right corner radar dual radar system. It can be seen that there is about a 95% probability of successfully identifying the target within the above target detection area as a dual target, and the probability of successful target number identification is significantly improved.

[0109] Based on the source number estimation algorithm obtained from the target array signals S01 and S02 to estimate the number of targets within the target detection area, the central processing unit can also jointly utilize the dual radar target array signals S01 and S02 to perform super-resolution angle measurement processing on each target (A and B) within the target detection area. The super-resolution algorithm can be the MUSIC algorithm (Multiple Signal Classification), the Minimum Modulus (MNM) algorithm, the Maximum Entropy (MEM) algorithm, the rotation-invariant subspace Esprit algorithm, the Relax algorithm, etc. The results of the angle measurement using the super-resolution algorithm can be found in [reference needed]. Figure 7 , Figure 7 This is a schematic diagram illustrating the effect of target location estimation provided in an embodiment of this application. Figure 7 As shown, the super-resolution results of each detection target (A and B) within the above-mentioned target detection area are demonstrated using dual radar target array signals S01 and S02, combined with the super-resolution algorithm MUSIC. Figure 7 The crosshair icon represents the actual location of the detected targets (target A, target B). Curves one and two represent the dual-target resolution results of a single radar at the left and right corners, respectively. Curve three represents the dual-target resolution results of a combined super-resolution system of two radars at the left and right corners. Figure 7 It can be seen that in dual-target super-resolution scenarios, single radar signals cannot clearly distinguish between two targets and provide relatively accurate angle information for both targets. However, the dual-radar signal joint super-resolution algorithm can not only distinguish between two targets but also measure the angle information of both targets relatively accurately. Based on the above results, it can be concluded that the dual-radar system joint super-resolution algorithm has a significant improvement in angle super-resolution performance compared to single radar signals.

[0110] Therefore, the embodiments of this application can construct a "multi-snap" signal using the echo signals of different radars in a distributed radar without increasing the number of signal channels and the number of snapshots, and use this signal to determine the number of targets. This can greatly improve the success rate of determining the number of multiple target sources within the beam of the radar system. Furthermore, by using the signals of different radars in the distributed radar to perform multi-radar joint super-resolution processing, the resolution performance of multiple targets within the beam of the radar system can be greatly improved.

[0111] 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.

[0112] Please see Figure 8 , Figure 8 This is a schematic diagram of a target number estimation device provided in an embodiment of this application. The target number estimation device 80 may include a construction unit 801, a determination unit 802, and an estimation unit 803, wherein the descriptions of each unit are as follows:

[0113] The construction unit 801 is used to construct a first rotation matrix of the array guidance vector between the first radar and the reference point, and to construct a second rotation matrix of the array guidance vector between the second radar and the reference point;

[0114] The determining unit 802 is used to obtain a third guiding vector using the first rotation matrix and the first guiding vector of the first radar, and to obtain a fourth guiding vector using the second rotation matrix and the second guiding vector of the second radar; wherein, the first guiding vector is a single snapshot signal of the first radar pointing to the target detection area, the second guiding vector is a single snapshot signal of the second radar pointing to the target detection area, and the third and fourth guiding vectors are two snapshot signals of the reference point pointing to the target detection area;

[0115] The estimation unit 803 is used to estimate the number of detection targets in the target detection area using the third steering vector, the fourth steering vector, and the source number estimation algorithm.

[0116] In this embodiment of the application, by constructing a multi-snap signal pointing to the target detection area using the array channel signals of each radar within the common field of view of distributed radars, and using the multi-snap signal to estimate the number of targets within the target detection area, the accuracy of the target number estimation can be greatly improved, as well as the accuracy and efficiency of subsequent target position estimation.

[0117] In one possible implementation, the construction unit 801 is specifically used to construct the first rotation matrix of the array steering vector between the first radar and the reference point using a first azimuth angle, a third azimuth angle, a first distance, and a third distance; wherein the first azimuth angle is the azimuth angle of the first radar relative to the target detection area, the third azimuth angle is the azimuth angle of the reference point relative to the target detection area, the first distance is the distance of the first radar relative to the target detection area, and the third distance is the distance of the reference point relative to the target detection area;

[0118] The construction unit 801 is further configured to construct the second rotation matrix of the array guidance vector between the second radar and the reference point using the second azimuth angle, the third azimuth angle, the second distance, and the third distance; wherein the second azimuth angle is the azimuth angle of the second radar relative to the target detection area, and the second distance is the distance of the second radar relative to the target detection area.

[0119] In one possible implementation, the determining unit 802 is specifically used to multiply the first rotation matrix by the first guide vector to the left to obtain the third guide vector;

[0120] The determining unit 802 is further configured to multiply the second rotation matrix by the second guide vector to the left to obtain the fourth guide vector.

[0121] In one possible implementation, the estimation unit 803 is further configured to estimate the position of any one of the detected targets within the target detection area using the third steering vector, the fourth steering vector, and a super-resolution algorithm. The super-resolution algorithm includes at least one of the following: multiple signal classification (MUSIC) algorithm, minimum modulus (MNM) algorithm, maximum entropy (MEM) algorithm, rotation-invariant subspace (Esprit) algorithm, and relaxation algorithm.

[0122] In one possible implementation, the estimation unit 803 is specifically used to estimate the position of any one of the targets within the target detection area using the third guiding vector, the fourth guiding vector, the number of targets within the target detection area, and a super-resolution algorithm.

[0123] In one possible implementation, the source number estimation algorithm includes at least one of the following: Minimum Description Length Criterion (MDL) algorithm and Akaike Information Content Criterion (AIC) algorithm.

[0124] According to the embodiments of this application, Figure 8 The various units in the illustrated device can be individually or entirely combined into one or more other units, or some of the units can be further divided into multiple functionally smaller units. This achieves the same operation without affecting the technical effects of the embodiments of this application. The above-mentioned units are based on logical function division. In practical applications, the function of one unit can also be implemented by multiple units, or the function of multiple units can be implemented by one unit. In other embodiments of this application, the network device may also include other units. In practical applications, these functions can also be implemented with the assistance of other units, and can be implemented collaboratively by multiple units.

[0125] It should be noted that the implementation of each unit can also be referenced accordingly. Figure 4 The corresponding description of the method embodiments shown.

[0126] exist Figure 8 In the described target number estimation device 80, by constructing a multi-snap signal pointing to the target detection area using the array channel signals of each radar within the common field of view of distributed radars, and then using the multi-snap signal to estimate the number of targets within the target detection area, the accuracy of the target number estimation can be greatly improved, as well as the accuracy and efficiency of subsequent target position estimation.

[0127] Please see Figure 9 , Figure 9 This is a schematic diagram of the structure of an electronic device 90 provided in an embodiment of this application. The electronic device 90 may include a memory 901 and a processor 902. Optionally, it may also include a communication interface 903 and a bus 904, wherein the memory 901, processor 902, and communication interface 903 are interconnected via the bus 904. The communication interface 903 is used for data interaction with the aforementioned target number estimation device 80.

[0128] The memory 901 provides storage space, which can store data such as the operating system and computer programs. The memory 901 includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or compact disc read-only memory (CD-ROM).

[0129] Processor 902 is a module that performs arithmetic and logical operations. It can be one or a combination of processing modules such as a central processing unit (CPU), a graphics processing unit (GPU), or a microprocessor unit (MPU).

[0130] The memory 901 stores a computer program, and the processor 902 calls the computer program stored in the memory 901 to execute the above-mentioned... Figure 4 The method for estimating the number of detection targets is shown.

[0131] For details regarding the execution method of the aforementioned processor 902, please refer to [link / reference needed]. Figure 4 This will not be elaborated upon here.

[0132] Correspondingly, the processor 902 can call the computer program stored in the memory 901, and can also be used to execute the above-mentioned... Figure 8 The method steps performed by the construction unit 801, determination unit 802, and estimation unit 803 in the target number estimation device 80 shown are detailed in the reference document. Figure 8 This will not be elaborated upon here.

[0133] exist Figure 9 In the described electronic device 90, by constructing a multi-shot signal pointing to the target detection area using the array channel signals of each radar within the common field of view of distributed radars, and then using the multi-shot signal to estimate the number of targets within the target detection area, the accuracy of the target number estimation can be greatly improved, as well as the accuracy and efficiency of subsequent target position estimation.

[0134] This application also provides a computer-readable storage medium storing a computer program. When the computer program is run on one or more processors, it can achieve... Figure 4 The method shown.

[0135] This application also provides a computer program product that, when run on a processor, can achieve... Figure 4 The method shown.

[0136] This application also provides a chip including a processor and a communication interface. The processor is used to call and execute instructions from the communication interface. When the processor executes the instructions, it can achieve... Figure 4 The method shown.

[0137] This application also provides a system that includes at least one target number estimation device 80, electronic device 90, or chip as described above.

[0138] In summary, in this embodiment of the application, by constructing a multi-snap signal pointing to the target detection area using the array channel signals of each radar within the common field of view of distributed radars, and then using the multi-snap signal to estimate the number of targets within the target detection area, the accuracy of the target number estimation can be greatly improved, as well as the accuracy and efficiency of subsequent target position estimation.

[0139] Those skilled in the art will understand that implementing all or part of the processes in the methods of the above embodiments can be accomplished by hardware related to a computer program. The computer program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing computer program code, such as read-only memory (ROM) or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A method for estimating the number of detected targets, characterized in that, include: Construct a first rotation matrix for the array steering vector between the first radar and the reference point, and construct a second rotation matrix for the array steering vector between the second radar and the reference point; Using the first rotation matrix and the first steering vector of the first radar, a third steering vector is obtained, and using the second rotation matrix and the second steering vector of the second radar, a fourth steering vector is obtained; wherein, the first steering vector is a single snapshot signal of the first radar pointing to the target detection area, the second steering vector is a single snapshot signal of the second radar pointing to the target detection area, and the third and fourth steering vectors are two snapshot signals of the reference point pointing to the target detection area; The number of targets within the target detection area is estimated using the third steering vector, the fourth steering vector, and the source number estimation algorithm.

2. The method according to claim 1, characterized in that, The first rotation matrix for constructing the array steering vector between the first radar and the reference point includes: Using a first azimuth angle, a third azimuth angle, a first distance, and a third distance, a first rotation matrix is ​​constructed for the array guidance vector between the first radar and the reference point; wherein, the first azimuth angle is the azimuth angle of the first radar relative to the target detection area, the third azimuth angle is the azimuth angle of the reference point relative to the target detection area, the first distance is the distance of the first radar relative to the target detection area, and the third distance is the distance of the reference point relative to the target detection area; The second rotation matrix for constructing the array guidance vector between the second radar and the reference point includes: Using the second azimuth angle, the third azimuth angle, the second distance, and the third distance, a second rotation matrix is ​​constructed for the array guidance vector between the second radar and the reference point; wherein, the second azimuth angle is the azimuth angle of the second radar relative to the target detection area, and the second distance is the distance of the second radar relative to the target detection area.

3. The method according to claim 1, characterized in that, The process of obtaining the third steering vector using the first rotation matrix and the first steering vector of the first radar includes: Multiply the first rotation matrix by the first guide vector to the left to obtain the third guide vector; The process of obtaining the fourth guidance vector using the second rotation matrix and the second guidance vector of the second radar includes: Multiply the second rotation matrix by the second guide vector to the left to obtain the fourth guide vector.

4. The method according to any one of claims 1 to 3, characterized in that, The method further includes: Using the third steering vector, the fourth steering vector, and the super-resolution algorithm, the position of any target within the target detection area is estimated. The super-resolution algorithm includes at least one of the following: multiple signal classification (MUSIC) algorithm, minimum modulus (MNM) algorithm, maximum entropy (MEM) algorithm, rotation-invariant subspace (Esprit) algorithm, and relaxation algorithm.

5. The method according to claim 4, characterized in that, The step of estimating the position of any target within the target detection area using the third steering vector, the fourth steering vector, and the super-resolution algorithm includes: Using the third guiding vector, the fourth guiding vector, the number of targets within the target detection area, and the super-resolution algorithm, the position of any target within the target detection area is estimated.

6. The method according to claim 1, characterized in that, The source number estimation algorithm includes at least one of the following: Minimum Description Length Criterion (MDL) algorithm and Akaike Information Content Criterion (AIC) algorithm.

7. A device for estimating the number of detected targets, characterized in that, include: A construction unit is used to construct a first rotation matrix for the array steering vector between the first radar and the reference point, and to construct a second rotation matrix for the array steering vector between the second radar and the reference point; The determining unit is used to obtain a third guiding vector using the first rotation matrix and the first guiding vector of the first radar, and to obtain a fourth guiding vector using the second rotation matrix and the second guiding vector of the second radar; wherein, the first guiding vector is a single snapshot signal of the first radar pointing to the target detection area, the second guiding vector is a single snapshot signal of the second radar pointing to the target detection area, and the third and fourth guiding vectors are two snapshot signals of the reference point pointing to the target detection area; The estimation unit is used to estimate the number of targets in the target detection area using the third steering vector, the fourth steering vector, and the source number estimation algorithm.

8. The apparatus according to claim 7, characterized in that, The construction unit is specifically used to construct the first rotation matrix of the array guidance vector between the first radar and the reference point using a first azimuth angle, a third azimuth angle, a first distance, and a third distance; wherein, the first azimuth angle is the azimuth angle of the first radar relative to the target detection area, the third azimuth angle is the azimuth angle of the reference point relative to the target detection area, the first distance is the distance of the first radar relative to the target detection area, and the third distance is the distance of the reference point relative to the target detection area; The construction unit is further configured to construct the second rotation matrix of the array guidance vector between the second radar and the reference point using the second azimuth angle, the third azimuth angle, the second distance, and the third distance; wherein the second azimuth angle is the azimuth angle of the second radar relative to the target detection area, and the second distance is the distance of the second radar relative to the target detection area.

9. The apparatus according to claim 7, characterized in that, The determining unit is specifically used to multiply the first rotation matrix by the first guiding vector to the left to obtain the third guiding vector; The determining unit is further configured to multiply the second rotation matrix by the second guiding vector to the left to obtain the fourth guiding vector.

10. The apparatus according to any one of claims 7 to 9, characterized in that, The estimation unit is further configured to estimate the position of any one of the detected targets within the target detection area using the third steering vector, the fourth steering vector, and the super-resolution algorithm. The super-resolution algorithm includes at least one of the following: multiple signal classification (MUSIC) algorithm, minimum modulus (MNM) algorithm, maximum entropy (MEM) algorithm, rotation-invariant subspace (Esprit) algorithm, and relaxation algorithm.

11. The apparatus according to claim 10, characterized in that, The estimation unit is specifically used to estimate the position of any one of the targets within the target detection area using the third guiding vector, the fourth guiding vector, the number of targets within the target detection area, and the super-resolution algorithm.

12. The apparatus according to claim 7, characterized in that, The source number estimation algorithm includes at least one of the following: Minimum Description Length Criterion (MDL) algorithm and Akaike Information Content Criterion (AIC) algorithm.

13. A device for estimating the number of detected targets, characterized in that, include: Processor and memory; The memory is used to store computer-executed instructions; The processor is configured to execute computer execution instructions stored in the memory to cause the target number estimation device to perform the method as described in any one of claims 1 to 6.

14. A computer-readable storage medium, characterized in that, include: The computer-readable storage medium is used to store instructions or computer programs that, when executed, cause the method as described in any one of claims 1 to 6 to be implemented.

15. A terminal, the terminal comprising the apparatus of claim 13.

Images