JP7727399B2 - Radar device, object detection method and program - Google Patents

Description

This disclosure relates to a radar device, an object detection method, and a program.

Radar devices are used, for example, in vehicles to detect the relative position of targets as seen from the vehicle. In such applications (e.g., detecting areas where a vehicle can be parked), it is preferable to detect the positions of multiple targets separately from one another, but there are cases where a signal corresponding to one target drowns out signals corresponding to other targets.

Patent Document 1 describes a radar device that changes the frequency of a pulse radar and detects target signals that are buried in side lobes in the process of processing multiple measurement results related to the target in a composite band.

Japanese Patent Application Laid-Open No. 2005-315738

However, the radar device described in Patent Document 1 does not separate targets when they are close to each other in the distance direction. Whether or not multiple targets can be distinguished as individual targets when they are located close to each other in the distance direction depends on the distance resolution of the radar device, so distance resolution can be improved by expanding the bandwidth of the transmitted signal. However, if the bandwidth is widened, the radar device needs to be configured to support the wideband, and transmitting a wide-bandwidth signal may affect other radar devices.

The objective of this invention is to provide a radar device, object detection method, and program that can achieve angular separation even with a small aperture and distance separation even with a narrow bandwidth.

The means for solving the problems of the present invention are as follows.
(1) A radar device comprising: a signal transmitting/receiving unit that receives signals reflected by one or more targets using a plurality of antennas that are actually or virtually located at different positions; and a separation processing unit that separates the angles of the targets based on the results of a calculation performed using the respective received signals received by the plurality of antennas and coefficients of an autoregressive model.
(2) The radar device according to (1), further comprising: an acquisition unit that acquires first amplitude information for each of the received signals with respect to a distance from the plurality of antennas; an extraction unit that extracts, from the first amplitude information for each of the received signals acquired by the acquisition unit, the first amplitude information corresponding to a predetermined distance from the plurality of antennas as information of an antenna channel signal for each of the antennas; a calculation unit that calculates coefficients of the autoregressive model based on the information of the antenna channel signal; and an expansion unit that estimates and expands the information of the antenna channel signal in an antenna direction based on the information of the antenna channel signal and the coefficients of the autoregressive model, wherein the separation processing unit processes the information of the expanded antenna channel signal expanded by the expansion unit to separate the first amplitude information for each target.
(3) The radar device of (1), further comprising: an acquisition unit that acquires first amplitude information with respect to a distance from the plurality of antennas for each of the received signals; an extraction unit that extracts, from the first amplitude information for each of the received signals acquired by the acquisition unit, the first amplitude information corresponding to a predetermined distance from the plurality of antennas as information of an antenna channel signal for each of the antennas; a first calculation unit that calculates coefficients of the autoregressive model based on the first amplitude information corresponding to the predetermined distance from the antenna; a second calculation unit that calculates phase information and second amplitude information corresponding to target signals reflected from the one or more targets based on the coefficients of the autoregressive model; and a position estimation unit that reconstructs target signals using the separation processing unit based on the phase information and the second amplitude information of the target signals calculated by the second calculation unit, and estimates positions of the one or more targets based on the reconstructed target signals.
(4) An object detection method comprising: a signal transmission/reception step of receiving signals reflected by one or more targets using a plurality of antennas arranged at different positions, either actually or virtually; and a separation processing step of separating the angles of the targets based on the results of calculations performed using the respective received signals received by the plurality of antennas and the coefficients of an autoregressive model.
(5) A program for causing a computer to execute a signal transmission/reception process for receiving signals reflected by one or more targets using a plurality of antennas arranged at different locations, either actually or virtually, and a separation processing process for separating the angles of the targets based on the results of calculations performed using the respective received signals received by the plurality of antennas and the coefficients of an autoregressive model.

This disclosure enables angular separation even with a small aperture, and distance separation even with a narrow bandwidth.

FIG. 1 is a block diagram showing the configuration of a radar device according to a first embodiment. FIG. 2 is a diagram illustrating an example of the configuration of an antenna of a signal transmitting/receiving unit. FIG. 3 is a diagram illustrating the antenna direction. FIG. 4 is a flowchart showing the procedure of signal processing by the radar device according to the first embodiment. FIG. 5 is a flowchart showing the procedure of signal processing by the radar device according to the first embodiment. FIG. 6 is a diagram illustrating the configuration of the signal processing unit related to the flowcharts shown in FIGS. FIG. 7 is a diagram showing the results (range angle map) detected by the peak detector. FIG. 8 is a block diagram showing the configuration of a radar device according to the second embodiment. FIG. 9 is a flowchart showing the procedure of signal processing by the radar device according to the second embodiment. FIG. 10 is a flowchart showing the procedure of signal processing by the radar device according to the second embodiment. FIG. 11 is a diagram illustrating the configuration of the signal processing unit related to the flowcharts shown in FIGS. FIG. 12 is a diagram showing the results of the distance evaluation process when the number of targets is assumed to be three. FIG. 13 is a diagram showing the results of the angle evaluation process when the number of targets is assumed to be three. FIG. 14 is a flowchart showing the procedure of the object detection method. FIG. 15 is a diagram showing the configuration of a computer. FIG. 16 is a diagram showing another configuration of a computer.

Next, we will explain an embodiment of the present invention.

(Regarding the first embodiment)
1 is a block diagram showing the configuration of a radar device 100 according to the first embodiment. The radar device 100 according to the first embodiment includes a signal transmitting/receiving unit 11 that receives a reflected wave from a target when a transmitted signal is reflected by the target, a signal processing unit 13A that processes the signal received by the signal transmitting/receiving unit 11, and a data processing unit 14. The radar device 100 is a device that radiates radio waves and receives the reflected wave from the target, and detects the presence of the target, its distance, relative speed, angle, etc. from the propagation time.

The signal transmitter/receiver unit 11 receives signals reflected by one or more targets using multiple antennas located at different real (physically) or virtually different positions.

Here, the configuration and operation of the signal transmitter/receiver 11 will be described. The signal transmitter/receiver 11 is composed of one or more transmitting antennas and one or more receiving antennas. There are several possible configurations for the transmitting and receiving antennas. In this embodiment, a configuration with multiple transmitting antennas and multiple receiving antennas is described as an example, but this configuration is not limited to this, and the signal transmitter/receiver 11 may also be composed of one transmitting antenna and multiple receiving antennas. Furthermore, the signal transmitter/receiver 11 employs the FCM (Fast-Chirp Modulation) method, but may also employ the FMCW (Frequency Modulated Continuous Wave) method or a pulse method. The FCM method continuously transmits a transmitted signal while rapidly sweeping it in units of several to several tens of microseconds, and measures distance, speed, and angle by analyzing the frequency difference with the reflected wave. The FMCW system uses frequency modulation (FM) as its modulation method, and measures distance, speed, and angle from the frequency difference (beat frequency) between the transmitted wave and the reflected wave. The pulse method emits radio waves intermittently, measuring distance by measuring the round-trip time of the pulse, and speed and angle by analyzing the frequency.

Furthermore, the frequency bands used by the radar device 100 are, for example, the quasi-millimeter wave band (24.05 to 24.25 GHz, 24 GHz band) and the millimeter wave band (76 to 81 GHz), but are not limited to these frequency bands.

The signal processing unit 13A includes a separation processing unit 13A-5 and a position estimation unit 13A-6, and supplies the results of signal processing to the data processing unit 14.

The separation processing unit 13A-5 separates the target angles based on the results of a calculation performed using each of the received signals received by the multiple antennas (signal transmission/reception unit 11) and the coefficients of an autoregressive model. In other words, the separation processing unit 13A-5 separates the multiple received signals into target signals reflected from each target based on the results of a calculation performed using each of the received signals and the coefficients of an autoregressive model.

The position estimation unit 13A-6 estimates the position of each target based on the target signals separated by the separation processing unit 13A-5.

The data processing unit 14 performs processes such as clustering and tracking on the digital data supplied from the signal processing unit 13A (information relating to the positions of each target estimated by the position estimation unit 13A-6) to detect the target. Information relating to the detected target is supplied to a higher-level device (not shown) (e.g., an ECU (Electric Control Unit)). The data processing unit 14 may also be provided in the higher-level device.

In this way, the radar device 100 separates the multiple received signals into target signals reflected from each target using the separation processing unit 13A-5, estimates the position of each target based on each target signal using the position estimation unit 13A-6, and performs object detection processing using the data processing unit 14 based on this estimation result.

Here, we will explain the case where the signal transmitter/receiver unit 11 is configured as a MIMO (multiple-input and multiple-output) system. Figure 2 shows two transmitting antennas Tx1 and Tx2 and four receiving antennas Rx1, Rx2, Rx3, and Rx4. The radar device 100 can virtually form receiving antennas (in the example shown in Figure 2, Rx5, Rx6, Rx7, and Rx8 are virtual receiving antennas) without changing the physical size, realizing an aperture A2 formed by the receiving antennas Rx1, Rx2, Rx3, and Rx4 and the virtual receiving antennas Rx5, Rx6, Rx7, and Rx8. This aperture A2 is wider than the aperture A1 formed by the receiving antennas Rx1, Rx2, Rx3, and Rx4. Note that the arrangement of the receiving and transmitting antennas is not limited to that shown in Figure 2. For example, the signal transmission/reception unit 11 may be configured with one transmitting antenna and eight receiving antennas (Tx1-Rx8). In this configuration, signals are received by sequentially switching between the eight receiving antennas. For example, the signal transmission/reception unit 11 may be configured with eight transmitting antennas and one receiving antenna (Tx8-Rx1). In this configuration, signals are transmitted by sequentially switching between the eight transmitting antennas.

Here, the procedure for transmitting and receiving signals by the signal transmitter/receiver unit 11 will be explained. The signal transmitter/receiver unit 11 outputs a transmission signal from transmitting antenna Tx1, and performs processing (first transmission/reception processing) in which the transmission signal is reflected by a target and received as a received signal by receiving antennas Rx1, Rx2, Rx3, and Rx4. Thereafter, the signal transmitter/receiver unit 11 outputs a transmission signal from transmitting antenna Tx2, and performs processing (second transmission/reception processing) in which the transmission signal is reflected by a target and received as a received signal by receiving antennas Rx1, Rx2, Rx3, and Rx4. The signal transmitter/receiver unit 11 repeatedly performs the first transmission/reception processing and the second transmission/reception processing.

In the radar device 100, the transmission signal transmitted from the transmitting antenna Tx1 is reflected by a target, and this reflected signal is received by the receiving antenna Rx1 and processed as a signal on antenna channel ch1. In the radar device 100, the transmission signal transmitted from the transmitting antenna Tx1 is reflected by a target, and this reflected signal is received by the receiving antenna Rx2 and processed as a signal on antenna channel ch2. In the radar device 100, the transmission signal transmitted from the transmitting antenna Tx1 is reflected by a target, and this reflected signal is received by the receiving antenna Rx3 and processed as a signal on antenna channel ch3. In the radar device 100, the transmission signal transmitted from the transmitting antenna Tx1 is reflected by a target, and this reflected signal is received by the receiving antenna Rx4 and processed as a signal on antenna channel ch4. In the radar device 100, the transmission signal transmitted from the transmitting antenna Tx2 is reflected by a target, and this reflected signal is received by the receiving antenna Rx1 and processed as a signal on antenna channel ch5. In the radar device 100, the transmission signal transmitted from the transmitting antenna Tx2 is reflected by a target, and this reflected signal is received by the receiving antenna Rx2 and processed as a signal on antenna channel ch6. In the radar device 100, the transmission signal transmitted from the transmitting antenna Tx2 is reflected by a target, and this reflected signal is received by the receiving antenna Rx3 and processed as a signal on antenna channel ch7. In the radar device 100, the transmission signal transmitted from the transmitting antenna Tx2 is reflected by a target, and this reflected signal is received by the receiving antenna Rx4 and processed as a signal on antenna channel ch81. Note that the antenna channel ch numbers described above are merely examples, and may be changed as appropriate depending on the combination of receiving and transmitting antennas.

Therefore, when the radar device 100 is configured with MIMO, it can improve target separation performance in the range direction even when the aperture of the physically arranged receiving antenna is small and the bandwidth is narrow.

Here, the detailed configuration of the signal processing unit 13A will be described. In addition to the separation processing unit 13A-5 and position estimation unit 13A-6 described above, the signal processing unit 13A also includes an acquisition unit 13A-1, an extraction unit 13A-2, a calculation unit 13A-3, and an expansion unit 13A-4.

The acquisition unit 13A-1 acquires first amplitude information for each received signal (antenna) relative to the distance from multiple antennas (signal transmission/reception unit 11). For example, if the signal transmission/reception unit 11 is configured with two transmitting antennas Tx1 and Tx2 and four receiving antennas Rx1, Rx2, Rx3, and Rx4 (configured in MIMO), the acquisition unit 13A-1 acquires first amplitude information from each antenna channel signal.

The first amplitude information _I1 relative to the distance can be expressed in complex form by equation (1). In the FCM method employed in this embodiment, the first amplitude information can be obtained by Fourier transforming the received signal.
I ₁ =A ₀ e ^iΨ ...(1)

_A0 indicates the amplitude (real number), and Ψ indicates the phase of the first amplitude information. Such first amplitude information includes information on the reflected signal from the target distance. The acquisition unit 13A-1 acquires the first amplitude information from the signal received by each receiving antenna (each antenna channel signal). Note that the amplitude of each received signal with respect to time (including the case of complex amplitude) can be used as the first amplitude information with respect to distance.

The extraction unit 13A-2, as will be described in detail below, extracts first amplitude information corresponding to a predetermined distance from each of the multiple antennas (signal transmission/reception unit 11) from the first amplitude information for each received signal acquired by the acquisition unit 13A-1 as antenna channel signal information (hereinafter, sometimes simply referred to as "antenna channel signal") for each antenna. This antenna channel signal information is extracted for each antenna channel signal. As described above, when using the amplitude of each received signal with respect to time as first amplitude information with respect to distance, the first amplitude information of a received signal received at a predetermined time after transmitting a transmitted signal can be used as the first amplitude information corresponding to the predetermined distance.

The calculation unit 13A-3 calculates the coefficients of an autoregressive model (hereinafter sometimes referred to as "AR coefficients") based on information about the antenna channel signal.

The expansion unit 13A-4 performs computational processing using an autoregressive model based on the antenna channel signal information and the coefficients (AR coefficients) of the autoregressive model, estimating and expanding the antenna channel signal information to the antenna direction. Here, antenna direction will be explained with reference to Figure 3. The antenna channel signal information contains multiple pieces of information corresponding to the distance from the antenna. The antenna direction is the direction of the antenna channel expanded during processing by the expansion unit 13A-4, and is the direction indicated by X in Figure 3. Furthermore, S1 indicates information corresponding to the distance contained in the signal of antenna channel ch1 (antenna channel signal information). S2 indicates information corresponding to the distance contained in the signal of antenna channel ch2 (antenna channel signal information). S3 indicates information corresponding to the distance contained in the signal of antenna channel ch3 (antenna channel signal information). S4 indicates information corresponding to the distance contained in the signal of antenna channel ch4 (antenna channel signal information). S5 indicates information corresponding to the distance contained in the signal of antenna channel ch5 (antenna channel signal information). S6 indicates information corresponding to the distance contained in the signal of antenna channel ch6 (antenna channel signal information). S7 indicates information corresponding to distance contained in the signal of antenna channel ch7 (antenna channel signal information). S8 indicates information corresponding to distance contained in the signal of antenna channel ch8 (antenna channel signal information). Sn indicates information corresponding to distance contained in the signal of antenna channel chn (extended nth antenna channel) (antenna channel signal information).

The separation processing unit 13A-5 performs a Fourier transform on the information of the extended antenna channel signal extended by the extension unit 13A-4 (hereinafter, sometimes simply referred to as the "extended antenna channel signal"), thereby separating the first amplitude information for each target. Note that the separation processing unit 13A-5 may also perform a Fourier transform on the information of the antenna channel signal extracted by the extraction unit 13A-2 and the information of the extended antenna channel signal extended by the extension unit 13A-4, thereby separating the first amplitude information for each target.

Here, we will explain the processing performed by the radar device 100 according to the present invention.

A one-dimensional (1D) column vector extracted by the extraction unit 13A-2 from a memory (not shown) as a signal for each antenna channel (antenna channel signal) for the jth signal in the time series is shown in equation (2). In equation (2), N, which indicates the antenna channel, is a subscript, and in this embodiment, N=8, but is not limited to N=8. Each x _k (k=1 to N) in the column vector in equation (2) indicates first amplitude information at distance R _j obtained for each antenna channel signal.
(x ₁ ~ x _N ) _j ^T ...(2)

Here, the AR coefficient is a one-dimensional column vector, and an AR model is applied to this signal. Equation (3) shows the AR coefficient. Equation (4) shows an example of the AR model.
(a ₁ ~ a _p ) _j ^T ...(3)
x _n =a ₁ x _n-1 +a ₂ x _n-2 +...+a _p x _n-p ...(4)

The number of AR coefficients, p, represents the order of the AR model and corresponds to the number of targets that can be represented. In this embodiment, p = 3. The order may be determined in advance, or it may be determined using, for example, AIC (Akaike Information Criteria) or MDL (Minimum Description Length), as described on pages 139 and 148 of the literature (Kikuma Nobuyoshi, Adaptive Antenna Technology, First Edition, October 2003).

To calculate the AR coefficient, a matrix equation (equation (5)) is formulated.
b=XA...(5)

The matrix size is as follows:
b: (n-p) row, column 1 X: (n-p) row, p column A: p row, column 1

The calculation unit 13A-3 calculates the AR coefficient using a generalized inverse matrix of X. Here, a generalized inverse matrix other than the generalized inverse matrix of X may be used to calculate the AR coefficient. For example, when a conjugate transpose (denoted by H) is multiplied from the left, if X ^H X is regular, an inverse matrix exists, and the AR coefficient can be found by diagonalizing this. Here, X ^H b = X ^H Xa is called a normal equation. Therefore, a generalized inverse matrix of X ^H X may be used.

The generalized inverse matrix can be calculated from the singular value decomposition of X. Equation (6) shows the singular value decomposition. Equation (7) shows the generalized inverse matrix.
X=USV ^H ...(6)
X ⁺ =VS ^{+ UH} ...(7)

In this embodiment, the formula shown in equation (7) was described, but it is also possible to formulate the formula using a correlation matrix and use the linear prediction method shown in the above-mentioned literature. It is also possible to calculate the Yule-Walker equation and find the AR coefficient.

The expansion unit 13A-4 estimates information about the antenna channel signal using the AR model of equation (4) based on the AR coefficients calculated by the calculation unit 13A-3 and the antenna channel signal, and expands the signal. The number of signals to expand can be determined based on the angle between the targets to be separated and the size of the antenna array (aperture, spacing, etc.). The expanded antenna channel signal is referred to as the expanded antenna channel signal.

The separation processing unit 13A-5 performs FFT (Fast Fourier Transform) processing on the extended antenna channel signal to calculate an extended angular spectrum signal. The position estimation unit 13A-6 performs peak detection processing on the extended angular spectrum signal calculated by the separation processing unit 13A-5.

The radar device 100 separates targets and more accurately determines the distance from the radar device 100 to the targets by performing FFT processing on the extended antenna channel signal using the separation processing unit 13A-5 to calculate an extended angular spectrum signal, and then detecting peaks based on the extended angular spectrum signal using the position estimation unit 13A-6. Furthermore, by performing the above-described processing on all antenna channels that make up the signal transmission/reception unit 11, the radar device 100 can also calculate (estimate) the azimuth angle of each target separately.

When targets are stationary relative to the radar device 100, it is possible to estimate the position of each target based on the expanded angular spectrum signal calculated by the separation processing unit 13A-5. However, performing peak detection using extreme value detection processing in the time series direction as described above is more preferable, as it allows targets to be separated and their positions to be estimated even when the targets are moving relative to the radar device 100.

Next, the specific operation of the radar device 100 (particularly the signal processing unit 13A) will be described. Figures 4 and 5 are flowcharts showing the signal processing procedure performed by the radar device 100. Figure 6 is a diagram used to explain the configuration of the signal processing unit 13A related to the flowcharts shown in Figures 4 and 5.

The following describes the antenna channel signal acquisition process ST1, which is a process prior to signal processing by the signal processing unit 13A, and which acquires an antenna channel signal; the AR extension processing process ST2, which is performed by the signal processing unit 13A and which performs extension processing using AR; and the target signal separation processing process ST3, which performs target signal separation processing. First, the antenna channel signal acquisition process ST1 will be described in detail.

In step ST11, the signal transmission/reception unit 11 sets the number of the first antenna channel. In this process, the signal transmission/reception unit 11 sets i=0. "i=0" is, for example, antenna channel ch1.

In step ST12, the signal transmitter/receiver 11 starts radar measurement. The signal transmitter/receiver 11 transmits a signal (transmission signal).

In step ST13, the signal transmitter/receiver unit 11 receives the reflected signal reflected by the object as a received signal, and processes this received signal to obtain (generate) a time series signal.

In step ST14, the signal transmission/reception unit 11 stores the time series signals acquired in step ST13 in memory (not shown) for each antenna channel.

In step ST15, the signal transmission/reception unit 11 increments the antenna channel number "i". The incrementing of the antenna channel number "i" is expressed as "i = i + 1".

In step ST16, the signal transmission/reception unit 11 determines whether the antenna channel number "i" updated in step ST15 has reached a predetermined number, "N-1." N indicates the number of antenna channels, and is "8" in this embodiment, but is not limited to "8." If "i = N-1" (Yes), proceed to step ST17; if "i = N-1" is not true (No), return to step ST12.

In step ST17, the signal transmitter/receiver unit 11 performs FFT processing on the time series signals for each antenna channel stored in memory (distance FFT processing). By performing FFT processing on the time series signals, the time series signals are converted into distance signals.

In step ST18, the signal transmission/reception unit 11 completes the acquisition (generation) of distance signals for each antenna channel (hereinafter sometimes referred to as "antenna channel signals"). All antenna channel signals are stored in memory (not shown).

In this way, in the antenna channel signal acquisition process ST1, a transmission signal is sent and the reflected signal is received as a received signal to acquire (generate) all antenna channel signals, which are then stored in memory (not shown).

Next, we will explain the details of the AR extension processing step ST2.

In step ST21, the sequence extraction unit 15 sets the time series number "j" of the first antenna channel signal. In this step, the sequence extraction unit 15 sets j = 0. The sequence extraction unit 15 corresponds to the acquisition unit 13A-1 and extraction unit 13A-2 described above.

In step ST22, the sequence extraction unit 15 retrieves the set jth antenna channel signal from memory. The sequence extraction unit 15 transmits the retrieved antenna channel signal to the AR coefficient calculation unit 16 and the signal expansion unit 19. The AR coefficient calculation unit 16 corresponds to the calculation unit 13A-3 described above. The signal expansion unit 19 corresponds to the expansion unit 13A-4 described above.

In step ST23, the AR coefficient calculation unit 16 generates a matrix equation from the antenna channel signals transmitted from the sequence extraction unit 15. The matrix equation corresponds to equation (7) described above. The AR coefficient calculation unit 16 calculates the AR coefficients from the generated equation based on the AR order (number p of AR coefficients) stored in the memory unit 17. In this embodiment, p = 3 is used for explanation, but this is not limited to p = 3.

In step ST24, the AR coefficient calculation unit 16 stores the AR coefficient calculated in step ST23 in the AR coefficient storage unit 18. The AR coefficient storage unit 18 may be the same as the memory unit 17.

In step ST25, the signal expansion unit 19 calculates an expanded signal (hereinafter sometimes referred to as the "expanded antenna channel signal") that has been estimated and expanded in the antenna direction using an AR model based on the AR coefficients stored in the AR coefficient storage unit 18 and the antenna channel signals extracted by the sequence extraction unit 15. This expanded signal is a signal that has been expanded by estimating the phase change for each antenna channel that depends on the target distance. The signal expansion unit 19 corresponds to the expansion unit 13A-4 described above.

In step ST26, the signal extension unit 19 stores the extension signals in memory in the order of the extended antenna channels.

In step ST27, the sequence extraction unit 15 increments the antenna channel signal number "j". The incrementing of the antenna channel signal number "j" is expressed as "j = j + 1".

In step ST28, the sequence extraction unit 15 determines whether the antenna channel signal number "j" updated in step ST27 has reached a predetermined number, "M-1." M indicates the number of samples of the antenna channel signal. If "j = M-1" (Yes), proceed to step ST29; if "j = M-1" is not true (No), return to step ST22.

In step ST29, the signal processing unit 13A completes the acquisition (generation) of the extended signal (extended antenna channel signal) for each extended antenna channel. This extended antenna channel signal is stored in memory.

In this way, in the AR extension processing step ST2, all antenna channel signals acquired in the antenna channel signal acquisition step ST1 are AR extended to acquire extended antenna channel signals.

Next, we will explain the details of the target signal separation processing step ST3.

In step ST31, the FFT calculation unit 20 performs FFT processing on the extended antenna channel signal to calculate an angular spectrum signal. The FFT calculation unit 20 performs FFT processing on the extended antenna signal for each distance stored in memory in the antenna direction to calculate an angular spectrum signal for each distance. The FFT calculation unit 20 corresponds to the separation processing unit 13A-5 described above.

In step ST32, the peak detection unit 21 detects peaks from the absolute values of the angular spectrum signal for each distance. The peak detection unit 21 corresponds to the position estimation unit 13A-6 described above. Figure 7 is a diagram showing the results (range angle map) detected by the peak detection unit 21. Figure 7 shows that two targets (target T11, target T12) are positioned at different angles and positions. The shading of the targets indicates the magnitude of the amplitude (peak height). Figure 7 shows that target T11 has a higher amplitude than target T12. Note that if peaks can be detected from the angular spectrum signal calculated by the FFT calculation unit 20, step ST32 may be skipped.

In step ST33, the sequence extraction unit 22 extracts the extended antenna channel signal corresponding to the angle at which the peak was detected for an arbitrary distance range. As shown in FIG. 7, the arbitrary distance range is the distance range R1 at the angle at which target T11 was detected, and an arbitrary distance range R2 at the angle at which target T11 was detected. Note that the arbitrary distance may be any distance range including 0 m, or may be a distance range such as 50 cm to 20 m. The extended antenna channel signal extracted in step ST33 is a signal distinguished as a single target. The sequence extraction unit 22 corresponds to the position estimation unit 13A-6 described above.

In step ST34, memory 23 stores the extended antenna channel signals extracted in step ST33 as signals distinguished as a single target. Note that memory 23 may be the same as storage unit 17.

In step ST35, the sequence extraction unit 22 determines whether or not it has extracted extended antenna channel signals corresponding to all of the peaks detected in step ST32. If it has extracted extended antenna channel signals corresponding to all of the peaks (Yes), it proceeds to step ST36; if it has not extracted extended antenna channel signals corresponding to all of the peaks (No), it returns to step ST33.

In step ST36, the signal processing unit 13A completes the acquisition (generation) of signals identified as single targets. All signals identified as single targets are stored in the memory 23.

In this way, in the AR extension processing step ST2, the radar device 100 applies AR processing to all antenna channel signals acquired in the antenna channel signal acquisition step ST1, estimating phase changes for each antenna channel that depend on the target distance to generate extended extended antenna channel signals. In the target signal separation processing step ST3, the radar device 100 performs FFT processing on these extended antenna channel signals to separate targets in the distance direction. Therefore, even with a small aperture, the radar device 100 can angularly separate targets that are close in the distance direction, and can distance-separate targets even with a narrow bandwidth (improving distance accuracy).

(Regarding the second embodiment)
Next, an embodiment will be described in which target-specific signals are reconstructed without performing the processing of expanding frequency-specific signal information using an AR model.

Figure 8 is a block diagram showing the configuration of a radar device 300 according to the second embodiment. Note that the same components as those in the radar device 100 according to the first embodiment are assigned the same numbers, and detailed descriptions will be omitted.

The radar device 300 comprises a signal transmitter/receiver 11, a signal processor 13B, and a data processor 14. The signal transmitter/receiver 11 receives signals reflected by one or more targets using multiple antennas located at different real (physically) or virtually different positions.

The signal processing unit 13B performs predetermined signal processing on the digital data stored in memory (not shown) and supplies the results of the signal processing to the data processing unit 14. Specifically, the signal processing unit 13B includes a separation processing unit 13B-5 and a position estimation unit 13B-6.

The separation processing unit 13B-5 separates the multiple received signals into target signals reflected from each target based on the results of calculations performed using the multiple received signals received by the signal transmission/reception unit 11 and the coefficients of the autoregressive model.

The position estimation unit 13B-6 estimates the position of each target based on the target signals separated by the separation processing unit 13B-5.

The data processing unit 14 performs processes such as clustering and tracking on the digital data supplied from the signal processing unit 13B (information relating to the positions of each target estimated by the position estimation unit 13B-6) to detect the target. Information relating to the detected target is supplied to a higher-level device (e.g., an ECU) (not shown). The data processing unit 14 may also be configured to be included in the higher-level device.

Here, the detailed configuration of the signal processing unit 13B will be described. In addition to the separation processing unit 13B-5 and position estimation unit 13B-6 described above, the signal processing unit 13B includes an acquisition unit 13B-1, an extraction unit 13B-2, a first calculation unit 13B-3, and a second calculation unit 13B-4.

The acquisition unit 13B-1 acquires first amplitude information for each received signal (antenna channel) relative to the distance from multiple antennas (signal transmission/reception unit 11). The acquisition unit 13B-1 has the same configuration as the acquisition unit 13A-1 of the radar device 100 according to the first embodiment.

The first amplitude information relative to distance includes amplitude information for the real component and amplitude information for the imaginary component. The acquisition unit 13B-1 acquires either or both of the amplitude information relative to distance and the phase information relative to distance for each antenna channel.

The extraction unit 13B-2 extracts, from the first amplitude information for each received signal (antenna channel) acquired by the acquisition unit 13B-1, first amplitude information corresponding to a predetermined distance from multiple antennas as antenna channel signal information for each antenna. The extraction unit 13B-2 has a configuration similar to that of the extraction unit 13A-2 of the radar device 100 according to the first embodiment.

The first calculation unit 13B-3 calculates the coefficients of the autoregressive model (AR coefficients) based on the first amplitude information corresponding to a predetermined distance from the antenna.

The second calculation unit 13B-4 calculates phase information and second amplitude information corresponding to target signals reflected from one or more targets based on the coefficients (AR coefficients) of the autoregressive model. In this embodiment, multiple targets are assumed, and the second calculation unit 13B-4 calculates phase information and second amplitude information corresponding to each target signal based on the AR coefficients. The second amplitude information refers to the amplitude information corresponding to the target signal.

The separation processing unit 13B-5 reconstructs each target signal based on the phase information and second amplitude information of each target signal calculated by the second calculation unit 13B-4.

The position estimation unit 13B-6 reconstructs the target signal using the separation processing unit 13B-6 based on the phase information and second amplitude information of the target signal calculated by the second calculation unit 13B-4, and estimates the position of one or more targets based on this reconstructed target signal. In this embodiment, multiple targets are assumed, and the position estimation unit 13B-6 estimates the position of each target based on the respective target signals reconstructed by the separation processing unit 13B-5.

In this way, the radar device 300 can reconstruct target-specific signals without performing expansion processing of frequency-specific signal information using an AR model.

Next, the specific operation of the radar device 300 (particularly the signal processing unit 13B) will be described. Figures 9 and 10 are flowcharts showing the signal processing procedure by the radar device 300. Figure 11 is a diagram used to explain the configuration of the signal processing unit 13B related to the flowcharts shown in Figures 9 and 10. Note that steps that are the same as those included in the signal processing procedure by the radar device 100 according to the first embodiment described using Figure 4 are assigned the same step numbers, and detailed explanations will be omitted.

The following describes the antenna channel signal acquisition process ST4, which acquires antenna channel signals and is a process that precedes signal processing by the signal processing unit 13B; the target signal separation processing process ST5 using AR, which separates and processes target signals and is performed by the signal processing unit 13B; and the distance, angle, and amplitude processing process ST7.

The antenna channel signal acquisition process ST4 (steps ST41 to ST48) is the same as the antenna channel signal acquisition process ST1 (steps ST11 to ST18) in the signal processing procedure by the radar device 100 according to the first embodiment. Therefore, in the antenna channel signal acquisition process ST4, a transmission signal is transmitted, and the reflected signal is received as a received signal to acquire (generate) all antenna channel signals, and these signals are stored in memory (not shown).

Next, we will explain the details of the target signal separation processing step ST5 using AR.

In step ST51, the sequence extraction unit 15 sets the time series number "j" of the first antenna channel signal. In this process, the sequence extraction unit 15 sets j = 0. The sequence extraction unit 15 corresponds to the acquisition unit 13B-1 and extraction unit 13B-2 described above.

In step ST52, the sequence extraction unit 15 extracts the jth antenna channel signal of the set time series from the memory. The sequence extraction unit 15 transmits the extracted antenna channel signal to the AR coefficient calculation unit 16. The AR coefficient calculation unit 16 corresponds to the first calculation unit 13B-3 described above.

In step ST53, the AR coefficient calculation unit 16 generates a matrix equation from the antenna channel signals transmitted from the sequence extraction unit 15. The matrix equation corresponds to equation (5) described above. The AR coefficient calculation unit 16 calculates the AR coefficients from the generated equation based on the AR order (number p of AR coefficients) stored in the memory unit 17. In this embodiment, p = 3 is used for explanation, but this is not limited to p = 3.

In step ST54, the AR coefficient calculation unit 16 stores the AR coefficient calculated in step ST53 in the AR coefficient storage unit 18. The AR coefficient storage unit 18 may be the same as the memory unit 17.

In step ST55, the characteristic polynomial calculation unit 31 generates a characteristic polynomial based on the AR coefficients stored in the AR coefficient storage unit 18 and calculates the roots. By calculating the roots, the characteristic polynomial calculation unit 31 can obtain phase information (phase terms) of the target-specific signal (target signal). The characteristic polynomial calculation unit 31 corresponds to the second calculation unit 13B-4 described above.

In step ST56, the characteristic polynomial calculation unit 31 stores the phase information of the target-specific signals in chronological order in the memory 32.

In step ST57, the target-specific amplitude calculation unit 33 generates a matrix equation for calculating the phase information of the target-specific signal and the second amplitude information (amplitude term) of the target-specific signal from the antenna channel signal. Note that, hereinafter, the second amplitude information (amplitude term) of the target-specific signal may be referred to as the "target-specific amplitude signal." The target-specific amplitude calculation unit 33 corresponds to the second calculation unit 13B-4 described above.

In step ST58, the target-specific amplitude calculation unit 33 calculates second amplitude information (target-specific amplitude signal) of the target-specific signal using the least squares method from the equation generated in step ST57.

In step ST59, the target-specific amplitude calculation unit 33 stores the second amplitude information of the target-specific signal (target-specific amplitude signal) calculated in step ST58 in the memory 34 in chronological order.

In step ST60, the sequence extraction unit 15 increments the number "j" in the time series of the target-specific amplitude signal. The incrementing of the number "j" in the time series of the target-specific amplitude signal is expressed as "j = j + 1."

In step ST61, the sequence extraction unit 15 determines whether the number (j) on the time series updated in step ST60 has reached a predetermined number, "M-1." M indicates the number of samples on the time series. If "j = M-1" (Yes), proceed to step ST62; if "j = M-1" is not true (No), return to step ST52.

In step ST62, the signal processing unit 13B completes acquisition of the target-specific signal (target signal). This target signal is stored in memory 34.

Next, we will explain the specific calculation processing steps of the AR target signal separation processing step ST5.

The characteristic polynomial generated in step ST55 is a polynomial having an AR coefficient for the time series "j" as shown in equation (8).
s(v)=a ₁ v ¹ + a ₂ v ² +...+a _p v ^p -1...(8)

The characteristic polynomial described above is a polynomial for finding the eigenvalues of the associated matrix shown in equation (9) from the AR coefficients. The eigenvalues are found by diagonalizing this. These eigenvalues become the roots λ _m (m is an integer from 1 to p) of p characteristic polynomials.

Also, a vertical vector in which the amplitudes for each target are arranged is placed as shown in equation (10).
(B ₁ ~B _p ) _j ^T ...(10)

A matrix is created in which the powers of the roots λ _m (m is an integer from 1 to p) of p characteristic polynomials are arranged, and the equation shown in equation (11) is considered.

By solving this equation, the target-specific amplitude B _m can be estimated. A signal with the target separated can be reconstructed from the roots λ _m of the p characteristic polynomials and the target-specific amplitude B _m . When solving this equation, for example, the least squares method can be used.

Next, we will explain the distance, angle, and amplitude processing step ST7.

In step ST71, the distance/angle/amplitude evaluation processing unit 36 sets the number "k" of the first target. In this process, the distance/angle/amplitude evaluation processing unit 36 sets the number of the first target to k = 0. The distance/angle/amplitude evaluation processing unit 36 corresponds to the separation processing unit 13B-5 and position estimation unit 13B-6 described above.

In step ST72, the distance/angle/amplitude evaluation processing unit 36 retrieves the second amplitude information (target-specific amplitude signal) of the target-specific signal from the memory 34 in chronological order and performs distance evaluation processing. Figure 12 shows the results of distance evaluation processing when the number of targets is assumed to be three (targets T21, T22, and T23). Note that while the waveforms of each target in Figure 12 are shown as smooth curves, this is merely an example and is not limited to smooth curves.

In step ST73, the distance/angle/amplitude evaluation processing unit 36 calculates the angular spectrum from the phase information (phase term) of the target-specific signal at the corresponding distance and performs angle evaluation processing. Figure 13 shows the results of angle evaluation processing when the number of targets is assumed to be three (targets T21, T22, and T23).

In step ST74, the distance/angle/amplitude evaluation processing unit 36 stores the amplitude, distance, and angle of the target in memory (not shown). Specifically, the distance/angle/amplitude evaluation processing unit 36 stores the amplitude and distance of the target obtained as a result of the distance evaluation processing in step ST72, and the angle of the target obtained as a result of the angle evaluation processing in step ST73 in memory.

In step ST75, the distance/angle/amplitude evaluation processing unit 36 increments the target number "k". The incrementing of the target number "k" is expressed as "k = k + 1".

In step ST76, the distance/angle/amplitude evaluation processing unit 36 determines whether the target number "k" updated in step ST75 has reached a predetermined number, "P-1." P indicates the number of targets. If "k = P-1" (Yes), proceed to step ST77; if "k = P-1" is not true (No), return to step ST72.

In step ST77, the distance/angle/amplitude evaluation processing unit 36 completes profile acquisition for all targets.

In this way, in the AR target signal separation processing step ST5, the radar device 300 generates a characteristic polynomial from the AR coefficients, finds the roots, calculates phase information of the target-specific signal, calculates second amplitude information of the target-specific signal using the phase information of the target-specific signal and the measurement data through a least-squares method, performs distance evaluation processing on the second amplitude information of the target-specific signal (target-specific amplitude signal) in a time series, and calculates an angular spectrum from the phase information of the target-specific signal and performs angle evaluation processing, thereby calculating the amplitude, distance, and angle of the target. Therefore, the radar device 300 can angularly separate targets close in the distance direction even with a small aperture, and can distance-separate targets even with a narrow bandwidth (improving distance accuracy). Note that the memories 32 and 34 may not be included. In this configuration, the characteristic polynomial calculation unit 31 transmits the phase information of the target-specific signal to the target-specific amplitude calculation unit 33 in chronological order. In addition, the target-specific amplitude calculation unit 33 transmits second amplitude information of the target-specific signal (target-specific amplitude signal) and phase information of the target-specific signal (phase term) to the distance/angle/amplitude evaluation processing unit 36.

(About the object detection method)
Next, an object detection method for the radar device 100 that improves target separation performance in the distance direction even in a narrow band will be described. Fig. 14 is a flowchart showing the steps of the object detection method that improves target separation performance in the distance direction even in a narrow band.

In step ST101, the signal transmission/reception unit 11 receives signals reflected by one or more targets using multiple antennas located at different real or virtual positions (signal transmission/reception process).

In step ST102, separation processing units 13A-5 and 13B-5 separate the target angles based on the results of calculations performed using the signals received by the multiple antennas and the coefficients of the autoregressive model (separation processing step).

With this configuration, the object detection method can separate multiple received signals into target signals reflected from each target using the separation processing step. Therefore, the object detection method can angularly separate targets that are close in the distance direction even with a small aperture, and can distance-separate targets even with a narrow bandwidth (improving distance accuracy).

(About the program)
The program for improving target separation performance in the distance direction even in a narrow band mainly comprises the following steps, and is executed by the computer 500 (hardware).

Step 1 (signal transmitting/receiving step): A step of receiving signals reflected by one or more targets by a plurality of antennas arranged at different positions, either actually or virtually.
Step 2 (separation processing step): A step of separating the angles of the target based on the results of calculations performed using the signals received by the multiple antennas and the coefficients of the autoregressive model.

The configuration and operation of computer 500 will now be described using FIG. 15. FIG. 15 is a diagram showing the configuration of computer 500. As shown in FIG. 15, computer 500 is configured by connecting a processor 501, memory 502, storage 503, input/output I/F 504, and communication I/F 505 on bus A. The functions and/or methods described in this disclosure are realized by the cooperation of these components.

Memory 502 is composed of RAM (Random Access Memory). RAM is composed of volatile or non-volatile memory.

Storage 503 is composed of ROM (Read Only Memory). ROM is composed of non-volatile memory, and is realized, for example, by an HDD (Hard Disc Drive), SSD (Solid State Drive), or Flash Memory. Storage 503 stores various programs, such as the programs realized in steps 1 and 2 described above.

A signal processing circuit 600 is connected to the input/output I/F 504. One or more transmitting antennas 601 and one or more receiving antennas 602 are connected to the signal processing circuit 600. There are multiple possible configurations of transmitting antennas and receiving antennas. The signal processing circuit 600, transmitting antenna 601, and receiving antenna 602 correspond to the signal transmitting/receiving unit 11 described above.

The processor 501 controls the overall operation of the computer 500. The processor 501 is an arithmetic unit that loads an operating system and various programs that realize various functions from the storage 503 into the memory 502 and executes the instructions contained in the loaded programs.

Specifically, when processor 501 receives a user operation, it reads a program (e.g., a program related to the present invention) stored in storage 503, loads the read program into memory 502, and executes the program. Furthermore, by processor 501 executing the processing program, various functions such as signal processing unit 13A and data processing unit 14 are realized.

The configuration of the processor 501 will now be described. The processor 501 may be realized, for example, by a CPU (Central Processing Unit), an MPU (Micro Processing Unit), a GPU (Graphics Processing Unit), various other computing devices, or a combination of these.

Furthermore, to realize the functions and/or methods described in this disclosure, some or all of the functions of the processor 501, memory 502, storage 503, etc. may be configured as dedicated hardware, such as a computer (hereinafter referred to as a processing circuit) 700. Figure 16 is a diagram showing the configuration of the processing circuit 700. The processing circuit 700 may be, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a combination thereof. A signal processing circuit 600 is connected to the processing circuit 700. One or more transmitting antennas 601 and one or more receiving antennas 602 are connected to the signal processing circuit 600.

Furthermore, while the processor 501 has been described as a single component, it is not limited to this and may be configured as a collection of multiple physically separate processors. In this specification, a program described as being executed by the processor 501 or instructions included in the program may be executed by a single processor 501, or may be executed in a distributed manner by multiple processors. Furthermore, a program executed by the processor 501 or instructions included in the program may be executed by multiple virtual processors.

The communication I/F 505 is an interface that conforms to a specific communication standard (e.g., CAN (Controller Area Network)) and communicates with an external higher-level device (e.g., an ECU) via wired or wireless communication.

In this way, the program according to this embodiment is executed on the computer 500, 700, and utilizes the signal processing circuit 600, the transmitting antenna 601, and the receiving antenna 602 to receive signals reflected by one or more targets in step 1 (signal transmission/reception step) using multiple antennas located at different real or virtual locations, and then performs processing to separate the target angles in step 2 (separation processing step) based on the results of a calculation using the received signals received by the multiple antennas and the coefficients of an autoregressive model. Therefore, by executing the program according to this embodiment on the computer 500, 700, and utilizing the signal processing circuit 600, the transmitting antenna 601, and the receiving antenna 602, it is possible to separate targets close in the distance direction by angle even with a small aperture, and to separate targets by distance even with a narrow bandwidth (improving distance accuracy).

The above describes in detail some of the embodiments of the present application based on the drawings, but these are merely examples, and the present invention can be implemented in other forms that incorporate various modifications and improvements based on the knowledge of those skilled in the art, including the aspects described in the disclosure of the present invention.

11 Signal transmitting/receiving unit 13A, 13B Signal processing unit 13A-1, 13B-1 Acquisition unit 13A-2, 13B-2 Extraction unit 13A-3 Calculation unit 13A-4 Expansion unit 13A-5, 13B-5 Separation processing unit 13A-6, 13B-6 Position estimation unit 13B-3 First calculation unit 13B-4 Second calculation unit 14 Data processing unit 15, 22 Sequence extraction unit 16 AR coefficient calculation unit 17 Memory unit 18 AR coefficient storage unit 19 Signal expansion unit 20 FFT calculation unit 21 Peak detection unit 23, 32, 34 Memory 31 Characteristic polynomial calculation unit 33 Target-specific amplitude calculation unit 36 Distance/angle/amplitude evaluation processing unit 100, 300 Radar device

Claims (5)

a signal transmitting/receiving unit that receives signals reflected by one or more targets using a plurality of antennas that are actually or virtually located at different positions;
an acquisition unit that acquires first amplitude information with respect to a distance from the plurality of antennas for each of the received signals;
an extracting unit that extracts, from the first amplitude information for each of the received signals acquired by the acquiring unit, the first amplitude information corresponding to a plurality of predetermined distances from the plurality of antennas as information of an antenna channel signal;
a calculation unit that represents first amplitude information obtained from the plurality of antennas by an autoregressive model of a predetermined order based on the information of the antenna channel signals extracted by the extraction unit, and calculates coefficients of the autoregressive model using a generalized inverse matrix or an autocorrelation function;
a separation processing unit that separates the angles of the targets based on a calculation result performed using each of the received signals received by the plurality of antennas and the coefficients of the autoregressive model calculated by the calculation unit . an extension unit that estimates and extends the information of the antenna channel signal in an antenna direction based on the information of the antenna channel signal and a coefficient of the autoregressive model,
The radar device according to claim 1 , wherein the separation processing unit performs a Fourier transform on information of the extended antenna channel signal extended by the extension unit, thereby separating the first amplitude information for each target. a second calculation unit that calculates phase information and second amplitude information corresponding to the target signals reflected from the one or more targets based on the coefficients of the autoregressive model;
2. The radar device according to claim 1, further comprising: a position estimation unit that reconstructs a target signal by the separation processing unit based on the phase information and the second amplitude information of the target signal calculated by the second calculation unit, and estimates positions of the one or more targets based on the reconstructed target signal. a signal transmitting and receiving step of receiving signals reflected by one or more targets by a plurality of antennas disposed at different positions, either actually or virtually;
an acquisition step of acquiring first amplitude information with respect to a distance from the plurality of antennas for each of the received signals;
an extraction step of extracting, as information on antenna channel signals, the first amplitude information corresponding to a plurality of predetermined distances from the plurality of antennas from the first amplitude information for each of the received signals acquired in the acquisition step;
a calculation step of expressing first amplitude information obtained from the plurality of antennas by an autoregressive model of a predetermined order based on the information of the antenna channel signals extracted in the extraction step, and calculating coefficients of the autoregressive model using a generalized inverse matrix or an autocorrelation function;
a separation processing step of separating the angles of the target based on a calculation result performed using each of the received signals received by the plurality of antennas and the coefficients of the autoregressive model calculated in the calculation step . On the computer,
a signal transmitting and receiving step of receiving signals reflected by one or more targets by a plurality of antennas disposed at different positions, either actually or virtually;
an acquisition step of acquiring first amplitude information with respect to a distance from the plurality of antennas for each of the received signals;
an extraction step of extracting, as information on antenna channel signals, the first amplitude information corresponding to a plurality of predetermined distances from the plurality of antennas from the first amplitude information for each of the received signals acquired in the acquisition step;
a calculation step of expressing first amplitude information obtained from the plurality of antennas by an autoregressive model of a predetermined order based on the information of the antenna channel signals extracted in the extraction step, and calculating coefficients of the autoregressive model using a generalized inverse matrix or an autocorrelation function;
a separation processing step of separating the angles of the target based on the calculation results performed using the signals received by the plurality of antennas and the coefficients of the autoregressive model calculated in the calculation step .