Radar system for a vehicle and target detection method

Abstract

A radar system for a vehicle and a target detection method, the target detection method comprising: a transmitting section of a radio frequency module of the radar system forming a transmission wave with respect to a transmission signal T x and transmitting the transmission wave through a transmitting antenna; a receiving section of the radio frequency module receiving reflection waves with receiving signals R x from a plurality of receiving antennas respectively, the reflection waves being generated by a target reflecting the transmission wave, and forming difference frequency signals respectively from each of the receiving signals R x and the transmission signal T x ; and a signal processor performing two-dimensional fast Fourier transform (2D-FFT) processing on the difference frequency signals to obtain approximate position information of the target including a distance value, determining whether the distance value is greater than a preset distance threshold value, and if so, performing two-dimensional multiple signal classification (2D-MUSIC) calculation based on the approximate position information as a calculation range to obtain an accurate distance value and an accurate angle value of the target.

Description

Technical Field

[0001] This invention relates to a radar system and target detection method for vehicles, and particularly to a radar system and target detection method for vehicles that performs different analyses for distant and near objects. Background Technology

[0002] Radar systems have been used in transportation for many years. For example, radar systems installed in vehicles (hereinafter referred to as automotive radar) can be used to detect objects around the vehicle and are an indispensable sensing component for autonomous driving technology.

[0003] Generally, automotive radar includes an antenna, a radio frequency (RF) module, and a digital signal processor. The RF module transmits millimeter waves through the antenna and receives reflected signals from surrounding objects. The digital signal processor then calculates the relative distance and angle between the radar and surrounding objects, and sends the data back to the vehicle's central processing unit for judgment, processing, and decision-making.

[0004] The aforementioned digital signal processor (DSP) typically uses the FFT algorithm to calculate values ​​such as distance and angle based on the transmitted and received signals. This DSP's calculations must be fast and accurate to allow the vehicle's central processing unit to make immediate decisions. Currently, most automotive radar antennas on the market use the 2T4R type, meaning two antennas for transmitting and four for receiving, which can virtually create a 1T8R antenna array (hereinafter referred to as eight virtual antennas). Taking a 2T4R antenna installed at the front of a vehicle as an example, its resolvable angle θ is 14.32°. Assuming a lane width of 2.8 meters, the radar's detection range is only about 10 meters (see...). Figure 1 However, the radar has poor recognition accuracy for objects more than 10 meters in front of the vehicle.

[0005] Since the angular resolution calculated by the FFT algorithm used in automotive radar is related to the number of virtual antennas, the more virtual antennas there are, the more precise the resolvable angle. Therefore, to achieve greater accuracy, higher-specification hardware can be used, such as upgrading the aforementioned 2T4R antenna to a 4T8R (four antennas for transmitting and eight for receiving, creating a virtual 1T32R antenna array) or even higher-specification antennas. With a 4T8R antenna, the resolvable angle θ can be lower than 4°, significantly increasing the resolvable distance. However, using higher-specification antennas increases product costs. Summary of the Invention

[0006] One objective of this invention is to provide a radar system for vehicles that can calculate high-resolution angle values ​​for distant objects without upgrading antenna hardware specifications.

[0007] This invention relates to a radar system for vehicles, suitable for detecting targets. The system includes an antenna module comprising at least one transmitting antenna and multiple receiving antennas, a radio frequency module including a transmitting section connected to the transmitting antennas and a receiving section connected to the receiving antennas, and a signal processor connected to the radio frequency module.

[0008] The radio frequency module includes a transmitting section connected to the transmitting antenna and a receiving section connected to the receiving antenna. The transmitting section is used to generate a signal T. x The transmitting wave is transmitted through the transmitting antenna, and the receiving segment receives the reflected wave generated by the target object reflecting the transmitted wave from the receiving antenna. The reflected wave carries a received signal R. x The receiving segment will receive each of the received signals R x With the transmitted signal T x These are used to form difference frequency signals.

[0009] The signal processor performs two-dimensional fast Fourier transform (2D-FFT) processing on the difference frequency signal to obtain approximate location information of the target object, including distance value. It then determines whether the distance value is greater than a preset distance threshold. If so, it performs two-dimensional multi-signal classification (2D-MUSIC) calculation based on the approximate location information as the calculation range to obtain the precise distance value and precise angle value of the target object.

[0010] Another object of the present invention is to provide a target detection method implemented by the radar system. The method includes the transmitter segment of the radio frequency module generating a signal T. x The transmitted wave is transmitted through the transmitting antenna; the receiving section of the radio frequency module receives the reflected wave generated by the target object reflecting the transmitted wave from the receiving antenna, the reflected wave carrying the received signal R. x and each of the received signals R x With the transmitted signal T x The signal processor generates difference frequency signals, and performs two-dimensional fast Fourier transform (2D-FFT) processing on the difference frequency signals to obtain approximate location information of the target object, including distance values. The signal processor also determines whether the distance value is greater than a preset distance threshold; if so, it performs two-dimensional multi-signal classification (2D-MUSIC) calculations based on the approximate location information as the calculation range to determine the precise distance and precise angle values ​​of the target object.

[0011] In the radar system of the present invention, the receiving section of the radio frequency module includes a plurality of channels, the same number as the receiving antenna, each channel receiving the received signal R from the connected receiving antenna. xThe signal processor processes the difference frequency signal and digital samples to generate multiple digital data. It performs two FFT processes on the multiple digital data from the channel and builds a radar data cube including distance, speed and angle information.

[0012] The radar system of the present invention allows the signal processor to read approximate location information of the target object, including distance and angle values, from the radar data cube through peak detection.

[0013] In the radar system of the present invention, the signal processor reads that the target object is located within a target grid in the range-angle matrix diagram of the radar data cube, and uses the coordinates of the target grid as the approximate location information; the signal processor performs 2D-MUSIC calculation to obtain the range correlation matrix and angle correlation matrix within the target grid range, and calculates the probability of the target object being located in multiple detailed grids within the target grid, and selects the detailed grid with the highest probability of occurrence as the precise range value and precise angle value of the target object.

[0014] In the radar system of the present invention, the signal processor controls the transmit segment of the radio frequency module to adjust the transmit signal T. x This enhances the gain based on either the angle value or the precise angle value in the approximate location information of the target object, thereby increasing the reflection intensity of the target object.

[0015] The target detection method of the present invention includes a radio frequency module whose receiving segment comprises a plurality of channels, the same number as the receiving antenna, and each channel is connected to the receiving antenna.

[0016] The steps performed by the receiving section of the radio frequency module include: each channel receiving the received signal R from the connected receiving antenna. x The difference frequency signal is obtained through processing, and multiple digital data are generated through digital sampling; and

[0017] The steps by which the signal processor obtains approximate location information about the target object include:

[0018] Perform two FFT processes on multiple digital data from the channel;

[0019] Establish a radar data cube that includes distance, velocity, and angle information;

[0020] Peak detection reads the target object from the radar data cube, showing that it is located within a target grid in the range-angle matrix of the radar data cube. The coordinates of the target grid are used as the approximate location information, including range and angle values.

[0021] The target detection method of the present invention, wherein the step of the signal processor performing 2D-MUSIC calculation based on the approximate location information as the calculation range is implemented using the phased.MUSICEstimator2D tool of Matlab, calculates the probability of the target object being located in multiple fine grids within the target grid, selects the fine grid with the highest probability of occurrence and outputs it as the precise distance value and precise angle value of the target object.

[0022] The target detection method of the present invention further includes: the signal processor controlling the transmitter segment of the radio frequency module to adjust the transmission signal T. x This enhances the gain based on either the angle value or the precise angle value in the approximate location information of the target object, thereby increasing the reflection intensity of the target object.

[0023] The beneficial effects of this invention are as follows: by using 2D-FFT to initially estimate the distance and angle of the target object, and then performing 2D-MUSIC analysis on the locked range to obtain the precise distance and angle values ​​of the target object, high-resolution position information can be obtained without upgrading antenna hardware specifications or significantly increasing the amount of computation. Attached Figure Description

[0024] Other features and effects of the present invention will be clearly presented in the embodiments with reference to the accompanying drawings, wherein:

[0025] Figure 1 This is a schematic diagram illustrating the relationship between radar angular resolution and distance;

[0026] Figure 2 This is a block diagram illustrating the architecture of one embodiment of the radar system for vehicles according to the present invention;

[0027] Figure 3 This is a waveform diagram, illustrating the transmitted signal (T) emitted by the radar system. x ) and the received signal (R) x );

[0028] Figure 4 This is a spectrum diagram, illustrating the beat frequency distribution of a digital signal after FFT processing.

[0029] Figures 5(a) to (d) are 2D-FFT matrix diagrams used for illustrative purposes;

[0030] Figure 6 This is a flowchart illustrating the steps performed by the digital signal processor 3 of the radar system for vehicles according to the present invention. Detailed Implementation

[0031] Before the invention is described in detail, it should be noted that similar elements are represented by the same numbers in the following description.

[0032] An embodiment of the radar system 100 for vehicles of the present invention is as follows: Figure 2 As shown, the radar system 100 is installed at the front of a vehicle (not shown) and is used to detect target 9. The radar system 100 includes an antenna module 1, a radio frequency module 2, and a digital signal processor 3.

[0033] In this embodiment, antenna module 1 is a 2T4R type antenna, including two transmitting antennas 11 for transmitting signals and four equally spaced receiving antennas 12 for receiving signals. The antenna module 1 of this invention is not limited to a specific specification; the number of transmitting antennas 11 and receiving antennas 12 is not limited, but the number of receiving antennas 12 must be multiple. When there are multiple transmitting antennas 11, the beam patterns of each transmitting antenna can be different.

[0034] RF module 2 includes a transmitting section and a receiving section. The transmitting section is used to generate a signal T. x The transmitted wave is transmitted through one of the transmitting antennas 11, and the receiving segment receives the reflected wave generated by the target object 9 reflecting the transmitted wave from the receiving antenna 12. The reflected wave carries the received signal R. x Specifically, in this embodiment, the transmitting segment includes a control signal C connected to the digital signal processor 3 and controlled to output. t The transmission control unit 21 receives the control signal C. t And generate the modulation signal S mod Signal generator 22 receives the modulated signal S mod And by controlling its oscillation frequency, frequency modulation is performed to form the transmitted signal T. x The oscillator 23, and a switching unit 24 connected to the transmission control unit 21 and controlled to switch the connection between the oscillator 23 and one of the transmitting antennas 11 at predetermined periods. Thereby, the transmitted signal T... x The signal is transmitted by the oscillator 23 to one of the transmitting antennas 11 and then transmitted outward through the transmitting antenna 11.

[0035] In this embodiment, the digital signal processor 3 controls the transmit control unit 21, signal generator 22, and oscillator 23 of the radio frequency module 2, causing the transmit antenna 11 to emit signals such as... Figure 3 As shown in period T CA continuously variable frequency signal (chirps, also known as a chirped signal) with an internal frequency variation and a bandwidth of Δf. Furthermore, this embodiment utilizes switching unit 24 to switch beam patterns to improve adaptability; however, in other embodiments, the RF module 2 may not have switching unit 24, meaning the transmitting antenna 11 only transmits a single beam pattern signal T. x .

[0036] The receiving antenna 12 receives the transmitted signal T from the target 9 due to reflection of the transmitted signal T. x The generated reflected waves will respectively generate received signals R. x In this embodiment, the receiving section of the RF module 2 includes several channels, the same number as the receiving antenna 12, each connected to the receiving antenna 12. Each channel includes a mixer 25, a low-pass filter 26, and an analog-to-digital converter (ADC) 27 connected in series. Each mixer 25 receives the received signal R from the connected receiving antenna 12. x and receiving the transmitted signal T generated by the oscillator 23 x and the received signal R x With the transmitted signal T x Composed of a beat signal S beat Each low-pass filter 26 receives the difference frequency signal S from the connected mixer 25. beat High-frequency filtering is performed. Each ADC 27 receives the filtered difference frequency signal from the connected low-pass filter 26, digitally samples it to obtain a digital data point, and outputs it to the digital signal processor 3.

[0037] See also Figure 3 and Figure 4 Since the transmitting antenna 11 transmits a sequential number of transmission signals (T... x (That is, chirps), therefore the receiving antenna 12 receives a series of consecutive received signals R. x Therefore, the digital data received by the digital signal processor 3 is also a series of consecutive data points. The digital signal processor 3 performs a first FFT (also known as Range FFT) on the multiple data points from each channel to obtain location information related to distance. Each set of transmitted signals T... x The corresponding received signal R x The spectrum obtained after Range FFT processing is as follows Figure 4 As shown. Where the beat frequency f b (beat frequency, or difference frequency) formula (1) is as follows.

[0038]

[0039] Where R represents distance, c represents the speed of light, Δf represents bandwidth, and T C This represents the chirps period.

[0040] The digital signal processor 3 performs a second FFT (also known as Doppler FFT) on multiple consecutively generated spectrograms to obtain velocity information, which generates a two-dimensional Fast Fourier Transform (2D-FFT) matrix as shown in Figure 5(a), with distance on the vertical axis and velocity on the horizontal axis. The digital signal processor 3 can further calculate the orientation of the target 9 from multiple digital data from multiple channels, generating angle information. Therefore, the digital signal processor 3 can construct a radar data cube as shown in Figure 5(b) from the digital data originating from all receiving antennas 12. The radar data cube in Figure 5(b) includes range, velocity, and angle information, with the right side being the angle-range matrix. In this embodiment, the digital signal processor 3 uses at least one of MATLAB, Simulink, and the Phased Array System Toolbox to construct the radar data cube, but is not limited to these tools. Using the radar data cube, the digital signal processor 3 reads out the target object's location within a target grid in the range-angle matrix using peak detection. The coordinates of this target grid are used as the approximate location information, including range and angle values. In other words, the digital signal processor 3 can obtain the approximate location information of the target object 9, including range and angle values. In this embodiment, the digital signal processor 3 controls the transmit segment of the radio frequency module 2 to adjust the transmit signal T. x This enhances the gain and increases the reflection intensity of the target object 9 based on the angle value in the approximate location information of the target object 9.

[0041] Since the angular resolution of a typical 2T4R antenna is approximately 14 degrees, the horizontal axis of the angle-range matrix diagrams in Figures 5(c) and 5(d) is represented in 14 degrees. The following example illustrates this using the radar system 100 of the vehicle detecting two targets A and B at different distances. The angle-range matrix diagram in Figure 5(c) illustrates the actual distances and angles of targets A and B: 25 meters at a distance of 15 degrees and 55 meters at an angle of 27 degrees, respectively. However, since both targets A and B fall within the 14-28 degree range in the angle-range matrix diagram, due to this limited angular resolution, only the results for target A (distance 25 meters, angle 21 degrees) and target B (distance 55 meters, angle 21 degrees) can be obtained from Figure 5(d).

[0042] See also Figure 6The digital signal processor 3 of the radar system 100 for vehicles of the present invention performs further analysis on the angle-range matrix diagram shown in FIG5(d).

[0043] In step S1, the digital signal processor 3 reads the approximate location information of the target object from the angle-distance matrix. Taking Figure 5(d) as an example, the data for the grid where target object A is located (hereinafter referred to as target grid A) is read as a distance of 25 meters and an angle of 21 degrees; this is the approximate location information of target object A. The data for the grid where target object B is located (hereinafter referred to as target grid B) is read as a distance of 55 meters and an angle of 21 degrees; this is the approximate location information of target object B.

[0044] In step S2, the digital signal processor 3 analyzes whether the distance value in the approximate location information exceeds a preset distance threshold. In this embodiment, since 2D-FFT is sufficient for angular resolution within 10 meters, a preset distance threshold of 10 meters is used. However, this invention is not limited to this, and the threshold set internally in the actual product may vary depending on the requirements or application environment. If this step determines that the location information is correct, then step S3 is continued. If not, no further processing is required, meaning the approximate location information is sufficient. Therefore, the digital signal processor 3 sends the information back to the control device of the vehicle (not shown) for subsequent analysis and processing, such as assisted driving.

[0045] In step S3, the digital signal processor 3 performs two-dimensional multiple signal classification (2D-MUSIC) on the digital data to determine the precise distance and angle values ​​of the target object. Taking the target grids A and B in Figure 5(d) as examples, since the distance is greater than 10 meters, both grid data are processed using 2D-MUSIC.

[0046] In this embodiment, the digital signal processor 3 uses the `phased.MUSICEstimator2D` tool in Matlab to perform 2D-MUSIC calculations to estimate the direction of arrival (DoA) for the received digital data. The process involves establishing the convariance matrix of the digital data, calculating the autoconvariance matrix of the eigenvectors of the convariance matrix, and so on. Finally, it calculates the probability of the target object being located in multiple detailed grids within the target grids A and B, selects the detailed grid with the highest probability, and outputs it as the precise distance and angle values ​​of the target object. Since the `phased.MUSICEstimator2D` tool is prior art and not a key technical feature of this invention, the calculation details are not elaborated here.

[0047] Since the accuracy of the MUSIC algorithm is related to the number of digital samples taken by the ADC 27 of the RF module 2, but not to the number of virtual antennas, and the number of digital samples taken by the ADC 27 is sufficient to obtain accurate angle values, this invention uses the 2D-MUSIC algorithm to accurately estimate the angle and distance of the target grid, which can compensate for the low resolution of 2D-FFT in terms of angle. It is worth noting that although the 2D-MUSIC algorithm itself has a large computational load, the 2D-MUSIC calculation is limited to the angle and distance range of target grids A and B to find the highest probability of finding the accurate angle and distance of the target object, significantly reducing the computational load compared to using the 2D-MUSIC algorithm for all grids.

[0048] It should be noted that this embodiment uses a 2D-MUSIC algorithm to obtain both precise distance and angle values ​​simultaneously. However, only the precise angle value is needed to compensate for the shortcomings of 2D-FFT. Nevertheless, the advantage of obtaining both distance and angle values ​​simultaneously is that they can be directly compared and supplemented with the 2D-FFT calculation results. Taking the aforementioned targets A and B as examples, if the 1D-MUSIC algorithm can obtain precise angle values ​​of 15 degrees and 27 degrees, when integrating them with the 2D-FFT calculation results, it is impossible to determine whether it is "25 meters, 15 degrees angle" or "55 meters, 15 degrees angle", or "25 meters, 27 degrees angle" or "55 meters, 27 degrees angle". Further multi-baseline measurement methods, such as interferometry, must be used to find the correspondence with the distance. When the 2D-MUSIC algorithm is used, the precise distance and angle values ​​can be directly calculated and matched with the approximate location information to determine that target A is 25 meters away and 15 degrees at an angle, and target B is 55 meters away and 27 degrees at an angle.

[0049] In step S4, the digital signal processor 3 outputs the precise distance and angle values ​​of the targets A and B to the vehicle's control device for subsequent analysis, such as assisted driving. In other embodiments, after resolving the precise angle value, the digital signal processor 3 can control the radio frequency module 2 to adjust the transmission signal T. x This changes the method from enhancing the gain based on the angle value in the approximate location information to enhancing the gain based on the precise angle value.

[0050] Furthermore, in this embodiment, the radar system 100 also transmits the transmitted signal T by adjusting the transmitting antenna 11. x The phase difference creates a beamforming effect, i.e., beamforming technology is employed. The digital signal processor 3 can, after obtaining an approximate angle value indicating the target's position, or after obtaining a precise angle value, control the transmitter segment of the radio frequency module 2 to adjust the transmitted signal T. x The gain is increased by, for example, 2 to 3 dB based on this angle value to improve the reflection intensity of the target object. In this way, the transmitted signal T emitted by the transmitting antenna 11... x The horizontal angular range is reduced, the energy is more concentrated, and noise reflected from objects outside the lane is also reduced.

[0051] In summary, the radar system 100 for vehicles of the present invention first uses 2D-FFT to estimate the distance and angle of the target object 9, and then performs 2D-MUSIC analysis on the grid in the angle-distance matrix diagram of the target object 9 whose distance exceeds a preset threshold. In this way, high-resolution angle values ​​can be obtained for distant objects without upgrading antenna hardware specifications or significantly increasing the amount of computation.

[0052] The above description is merely an embodiment of the present invention and should not be construed as limiting the scope of the present invention. Any simple equivalent changes and modifications made in accordance with the claims and description of the present invention shall still fall within the scope of the present invention.

Claims

1. A radar system for a vehicle, suitable for detecting targets, the radar system for a vehicle comprising: An antenna module, comprising at least one transmitting antenna and multiple receiving antennas; The radio frequency module includes a transmitting segment connected to the transmitting antenna and a receiving segment connected to the receiving antenna. The transmitting segment is used to generate a signal T. x The transmitting wave is transmitted through the transmitting antenna, and the receiving segment receives the reflected wave generated by the target object reflecting the transmitting wave from the receiving antenna. The reflected wave carries the receiving signal R. x The receiving segment will receive each of the received signals R x With the transmitted signal T x They are respectively composed of difference frequency signals; as well as A signal processor, connected to the radio frequency module, performs two-dimensional fast Fourier transform (2D-FFT) processing on the difference frequency signal to obtain approximate location information of the target object, including distance values. It determines whether the distance value is greater than a preset distance threshold. If so, it performs two-dimensional multi-signal classification (2D-MUSIC) calculations based on the approximate location information to determine the precise distance and angle values ​​of the target object. The receiving section of the radio frequency module includes several channels, the same number as the receiving antenna, each connected to the receiving antenna. Each channel receives the received signal R from the connected receiving antenna. x The signal processor processes the difference frequency signal and digital sampling to generate multiple digital data. It performs two FFT processes on the multiple digital data from the channel and establishes a radar data cube including range, velocity, and angle information. The signal processor reads the approximate location information of the target object from the radar data cube through peak detection, including range and angle values. It reads that the target object is located within a target grid in the range-angle matrix of the radar data cube, using the coordinates of the target grid as the approximate location information. The signal processor performs 2D-MUSIC calculations to obtain the range correlation matrix and angle correlation matrix within the target grid area, and calculates the probability of the target object being located in multiple detailed grids within the target grid. It then selects the detailed grid with the highest probability of occurrence and reads it as the precise range and angle values ​​of the target object.

2. The radar system for vehicles according to claim 1, characterized in that: The signal processor controls the transmit segment of the RF module to adjust the transmit signal T. x This enhances the gain based on either the angle value or the precise angle value in the approximate location information of the target object, thereby increasing the reflection intensity of the target object.

3. A target detection method, implemented by a radar system, the radar system comprising an antenna module having at least one transmitting antenna and multiple receiving antennas, and a radio frequency module having a transmitting section connected to the transmitting antennas and a receiving section connected to the receiving antennas, characterized in that: The radar system also includes a signal processor connected to the radio frequency module; the target detection method includes: The transmitter segment of this RF module generates a signal with respect to the transmitted signal T. x The transmitted wave is transmitted through the transmitting antenna; The receiving section of the radio frequency module receives the reflected wave generated by the target object reflecting the transmitted wave from the receiving antenna. This reflected wave carries the received signal R. x and each of the received signals R x With the transmitted signal T x They respectively form the difference frequency signal; and The signal processor The difference frequency signal is processed by a two-dimensional fast Fourier transform (2D-FFT) to obtain approximate location information about the target object, including distance values. Determine whether the distance value is greater than a preset distance threshold, and If so, then based on the approximate location information as the calculation range, a two-dimensional multi-signal classification (2D-MUSIC) calculation is performed to determine the precise distance and precise angle values ​​of the target object; wherein, the receiving segment of the radio frequency module includes several channels, the same number as the receiving antenna, each connected to the receiving antenna, wherein, The steps performed by the receiving section of the radio frequency module include: each channel receiving the received signal R from the connected receiving antenna. x The difference frequency signal is obtained through processing, and multiple digital data are generated through digital sampling; and The steps by which the signal processor obtains approximate location information about the target object include: Perform two FFT processes on multiple digital data from the channel; Establish a radar data cube that includes distance, velocity, and angle information; The target object is read from the radar data cube by peak detection and is located within a target grid in the range-angle matrix of the radar data cube. The coordinates of the target grid are used as the approximate location information, including the range and angle values. The feature is that the signal processor performs 2D-MUSIC calculation based on the approximate location information as the calculation range. This is implemented using the phased.MUSICEstimator2D tool in Matlab. The probability of the target object being located in multiple detailed grids within the target grid is calculated, and the detailed grid with the highest probability of occurrence is selected and output as the precise range and angle values ​​of the target object.

4. The target detection method according to claim 3, characterized in that: The target detection method also includes: the signal processor controlling the transmitter segment of the radio frequency module to adjust the transmission signal T. x This enhances the gain based on either the angle value or the precise angle value in the approximate location information of the target object, thereby increasing the reflection intensity of the target object.

Images