TWI697688B - Frequency modulated continuous wave processing device - Google Patents

Abstract

The invention discloses a frequency modulation continuous wave processing device, which comprises a frequency modulated continuous wave radar and a frequency modulated continuous wave processor. The frequency modulated continuous wave processor controls one of the antennas of the frequency modulated continuous wave radar to transmit a plurality of chirp signals to the target in a coherent processing time interval, and the target object reflects the frequency sweep signals to form echo signals. The frequency modulated continuous wave processor receives and demodulates the echo signals through the antenna of the frequency modulated continuous wave radar, thereby generating distance frequency domain signals, speed frequency domain signals or azimuth frequency domain signals respectively. The frequency modulation continuous wave processor uses the wave peaks of the frequency domain signals, the speed frequency domain signals or the frequency domain signals to establish a plurality of time varying models, and thereby calculating the distance, the speed or the azimuth of the target relative to the frequency modulated continuous wave radar and reducing the complexity of data fitting.

Description

Frequency modulated continuous wave (FMCW) processing device

The present invention relates to a processing device, and more particularly to a frequency modulated continuous wave (FMCW) processing device.

Frequency modulated continuous wave (FMCW) radar is a very mature technology. The development of radar technology by various companies mainly focuses on the processing of radar data, which mainly includes radar target detection, target parameter estimation, target recognition, and more back-end radar data application. The parameters include distance, speed and azimuth. In terms of target parameter acquisition, the azimuth angle is generally more difficult. The well-known high-resolution techniques include the Multiple Signal Classification (MUSIC) algorithm and the estimation of signal parameter based on the rotation invariant technology. Via rotational invariant techniques, ESPRIT) algorithm, minimum absolute value convergence and selection operator (LASSO) algorithm, and various algorithms.

An important application of FM continuous wave radar is automotive radar. The characteristics of the acquired data are complex, poor signal-to-noise ratio (SNR), and rapid changes. The above-mentioned technology is difficult to achieve high performance, low computational complexity and high real-time in such an environment. In addition, in a multi-user environment, mutual interference between radars will be a serious problem. Conventional anti-interference techniques include staggering the use of frequency bands, staggering the transmission time, frequency hopping, and adding spread spectrum coding. Staggering the use of frequency bands or transmission time does not require much change to the radar itself, but it is necessary to coordinate users in advance. The addition of frequency hopping mechanism or spread spectrum coding will greatly increase the complexity of radar signal processing.

Therefore, the present invention proposes a FM continuous wave processing device for the above-mentioned problems. Set up to solve the problems caused by conventional knowledge.

The main purpose of the present invention is to provide a frequency modulated continuous wave (FMCW) processing device, which fits the parameters of multiple time-varying models to effectively improve the signal-to-noise ratio and accurately calculate the distance of the target relative to the frequency modulated continuous wave radar , Speed or azimuth angle, and effectively reduce the complexity of data fitting, and improve the real-time data processing.

Another object of the present invention is to provide a FM continuous wave processing device, which uses a grouping algorithm to classify data into a plurality of groups, performs subsequent processing on each group, and then combines the results of each group processing Combine them to effectively reduce the interference caused by the position estimation of complex data and improve the real-time data processing.

Another object of the present invention is to provide a FM continuous wave processing device, which reduces the time and probability of the frequency sweep signal appearing in the coherent processing time interval, so as to reduce the probability of multi-user interference.

To achieve the above objective, the present invention provides a frequency modulated continuous wave (FMCW) processing device, which includes a frequency modulated continuous wave radar and a frequency modulated continuous wave processor. The FM continuous wave processor is electrically connected to the FM continuous wave radar. The FM continuous wave radar has multiple antennas, and there are multiple targets in front of all the antennas. The FM continuous wave processor is electrically connected to the FM continuous wave radar, and the FM continuous wave processor controls one of the antennas of the FM continuous wave radar to transmit complex numbers to all targets in a coherent processing time interval. A chirp signal, all the targets reflect all the chirp signals to form a plurality of echo signals. The FM continuous wave processor receives all the echo signals through all the antennas of the FM continuous wave radar, and demodulates all the echo signals, and for all the echo signals in the fast time dimension and slow time dimension ) Or space dimension to perform Fourier transform to generate multiple distance frequency domain signals, multiple speed frequency domain signals or multiple azimuth frequency domain signals, distance frequency domain signals, speed The frequency domain signal or the azimuth frequency domain signal is a wave signal with multiple crests, and the number of all crests is the same as the number of all targets. The FM continuous wave processor uses all the peaks of all distance frequency domain signals, all speed frequency domain signals or all azimuth frequency domain signals to create multiple time-varying models. The number of all time-varying models is the same as the number of all peaks. FM continuous wave processing The device calculates the distance, speed or azimuth angle of all targets relative to the FM continuous wave radar based on all time-varying models.

In an embodiment of the present invention, the FM continuous wave processor uses a clustering algorithm to cluster the same distance frequency domain signal to form a plurality of sub-distance frequency domain signals, and all the sub-distance frequency domain signals correspond to a plurality of For different distance ranges, the FM continuous wave processor distinguishes all targets into different groups according to all distance ranges, and performs Fourier transform on all echo signals in the slow time dimension or space dimension according to this different group, to generate all speeds respectively Frequency domain signal or all azimuth frequency domain signal.

In an embodiment of the present invention, the FM continuous wave processor calculates the beat frequency of each sweep signal and its corresponding echo signal corresponding to the same target, and based on the beat frequency, the dynamic programming algorithm ( dynamic programming algorithm) Get the distance corresponding to the peak of all distance frequency domain signals with all distance ranges.

In an embodiment of the present invention, the FM continuous wave processor uses a grouping algorithm to group the same speed frequency domain signal to form a plurality of sub-speed frequency domain signals, and all the sub-speed frequency domain signals correspond to a plurality of different speed ranges, respectively. The FM continuous wave processor distinguishes all targets into different groups according to all speed ranges, and performs Fourier transforms on all echo signals in the fast time dimension or space dimension according to the different groups to generate all distance frequency domain signals or All azimuth frequency domain signals.

In one embodiment of the present invention, the FM continuous wave processor calculates the beat frequency of each sweep signal and its corresponding echo signal corresponding to the same target, and obtains it according to the beat frequency, dynamic programming algorithm and all speed ranges The speed corresponding to the peak of all speed frequency domain signals.

In an embodiment of the present invention, the FM continuous wave processor uses a grouping algorithm to compare the same The one-sided frequency domain signals are grouped to form a plurality of sub-azimuth frequency domain signals. All the sub-azimuth frequency domain signals correspond to a plurality of different azimuth ranges. The FM continuous wave processor distinguishes all targets into different groups according to all the azimuth ranges, and According to the different groups, Fourier transform is performed on all the echo signals in the fast time dimension or the slow time dimension to generate all distance frequency domain signals or all speed frequency domain signals respectively.

In an embodiment of the present invention, the FM continuous wave processor calculates the beat frequency of each sweep signal and its corresponding echo signal corresponding to the same target, and obtains it according to the beat frequency, dynamic programming algorithm and all azimuth ranges The azimuth angle corresponding to the peak of all azimuth frequency domain signals.

In one embodiment of the present invention, the clustering algorithm is k-means algorithm or Density-based Spatial Clustering of Applications with Noise (DBSCAN).

In an embodiment of the present invention, the FM continuous wave processor uses a Constant False-Alarm Rate (CFAR) to detect all peaks.

In an embodiment of the present invention, the FM continuous wave processor calculates the beat frequency of each sweep signal and its corresponding echo signal corresponding to the same target, and performs a fast time dimension on all echo signals according to the beat frequency Perform Fourier transform to generate all range frequency domain signals.

In an embodiment of the present invention, the FM continuous wave processor calculates the beat frequency of each sweep signal and its corresponding echo signal corresponding to the same target, and calculates the difference between the beat frequencies corresponding to the same target, And according to the difference, all echo signals are Fourier transformed in the slow time dimension to generate all speed frequency domain signals.

In an embodiment of the present invention, the FM continuous wave processor obtains the time point of the echo signal corresponding to the same target object received by all antennas, calculates the time difference at this time point, and uses the time difference to calculate the time difference between all echo signals. Fourier transformation is performed in the spatial dimension to generate all azimuth frequency domain signals.

In an embodiment of the present invention, the time-varying model is represented by an n-th degree polynomial, and n is greater than or equal to 1.

In an embodiment of the present invention, the FM continuous wave processor stores a plurality of azimuth frequency response signals corresponding to different azimuth angles, and compares all azimuth frequency response signals according to the minimum absolute value convergence and selection operator (LASSO) algorithm With all azimuth frequency domain signals, to obtain the azimuth angles corresponding to the peaks of all azimuth frequency domain signals.

In an embodiment of the present invention, the FM continuous wave processor stores a plurality of speed frequency response signals corresponding to different speeds, and compares all speed frequency response signals with the minimum absolute value convergence and selection operator (LASSO) algorithm All speed frequency domain signals to obtain the speed corresponding to the peak of all speed frequency domain signals.

In one embodiment of the present invention, the FM continuous wave processor stores a plurality of distance frequency response signals corresponding to different distances, and compares all distance frequency response signals with the minimum absolute value convergence and selection operator (LASSO) algorithm All distance frequency domain signals to obtain the distance corresponding to the peak of all distance frequency domain signals.

In an embodiment of the present invention, the corresponding time intervals of all the frequency sweep signals in the coherent processing time interval are staggered with each other, and the corresponding time intervals are randomly distributed. The coherent processing time interval is 1-10 milliseconds (ms), and all corresponding times The interval accounts for less than or equal to 20% of the coherent processing time interval and greater than 0.

In order to enable your reviewers to have a better understanding and understanding of the structural features of the present invention and the effects achieved, I would like to provide a better embodiment diagram and detailed descriptions. The description is as follows:

10: FM continuous wave radar

12: FM continuous wave processor

14: Antenna

16: target

16’: Target

16": target

Figure 1 is a circuit block diagram of an embodiment of the frequency modulated continuous wave (FMCW) processing device of the present invention.

Figure 2 is a waveform diagram of an embodiment of the frequency sweep signal of the present invention.

Figure 3 is a waveform diagram of the distance frequency domain signal, speed frequency domain signal, or azimuth frequency domain signal of the present invention.

Figure 4 is a waveform diagram of another embodiment of the frequency sweep signal of the present invention.

The following description of "one embodiment" or "an embodiment" refers to at least one specific element, structure, or feature related to the embodiment. Therefore, multiple descriptions of "one embodiment" or "an embodiment" appearing in various places below are not directed to the same embodiment. Furthermore, specific components, structures, and features in one or more embodiments can be combined in an appropriate manner.

Please refer to FIG. 1, FIG. 2 and FIG. 3 below to introduce the frequency modulated continuous wave (FMCW) processing device of the present invention, which includes a frequency modulated continuous wave radar 10 and a frequency modulated continuous wave processor 12. The FM continuous wave processor 12 is electrically connected to the FM continuous wave radar 10. The frequency modulated continuous wave radar 10 can be used as a vehicle radar, a surveillance radar, or a synthetic aperture radar (SAR). The FM continuous wave radar 10 has a plurality of antennas 14. There are a plurality of targets 16, 16' and 16" in front of all the antennas 14. In this embodiment, three targets 16, 16' and 16" are taken as an example. The FM continuous wave processor 12 is electrically connected to the FM continuous wave radar 10, and the FM continuous wave processor 12 controls one of the antennas 14 of the FM continuous wave radar 10 in a coherent processing time interval T0-T1 (coherent processing time interval) Send multiple chirp signals to all targets 16, 16' and 16” in sequence. As shown in Figure 2, a diagonal solid line represents a sweep signal. All targets 16, 16' and 16 "Reflects all sweep signals to form multiple echo signals, which are represented by diagonal dashed lines. The FM continuous wave processor 12 receives all the echo signals through all the antennas 14 of the FM continuous wave radar 10, and demodulates all the echo signals, and for all the echo signals in the fast time dimension and the slow time dimension ( slow time dimension) or space dimension (Fourier transform) to generate a plurality of distance frequency domain signals, a plurality of speed frequency domain signals or a plurality of azimuth frequency domain signals, each A distance frequency domain signal, each speed frequency domain signal or each azimuth frequency domain signal is a wave signal with multiple peaks A, B, and C, each distance frequency domain signal, each speed frequency domain signal or each azimuth The number of peaks A, B, and C of the frequency domain signal is the same as the number of all targets 16, and peaks A, B, and C correspond to targets 16, 16', and 16”, respectively. Since the wave signal is interfered by noise, Therefore, not all wave crests can represent targets 16, 16' and 16". Only the more obvious crests A, B and C can represent targets 16, 16' and 16". Therefore, as shown in Figure 2, The FM continuous wave processor 12 uses a constant false alarm rate (Constant False-Alarm Rate, CFAR) to detect peaks A, B and C. The FM continuous wave processor 12 uses all distance frequency domain signals, all speed frequency domain signals or Establish multiple time varying models for all peaks A, B, and C of all azimuth frequency domain signals, and the number of all time varying models is the same as the number of all peaks A, B, and C. A time varying model represents a target Motion behavior of 16, 16' or 16”, because the motion behavior of target object 16, 16' or 16” can be expressed by polynomial, so the time-varying model is expressed by n-th degree polynomial, n is greater than or Equal to 1. The FM continuous wave processor 12 calculates the distance, speed or azimuth angle of all targets 16, 16' and 16" relative to the FM continuous wave radar 10 according to all time-varying models.

In one embodiment of the present invention, the FM continuous wave processor 12 calculates the beat frequency Fb of each sweep signal and its corresponding echo signal corresponding to the same target 16, 16' or 16", and According to the beat frequency Fb, all echo signals are Fourier transformed in the fast time dimension to generate all distance frequency domain signals. The FM continuous wave processor 12 uses clustering algorithm to group the same distance frequency domain signals to A plurality of sub-distance frequency domain signals are formed, and all the sub-distance frequency domain signals correspond to a plurality of different distance ranges. The FM continuous wave processor 12 distinguishes all targets 16, 16' and 16” into different groups according to all distance ranges, and This different group performs Fourier transform on all echo signals in the slow time dimension or the space dimension to generate all speed frequency domain signals or all azimuth frequency domain signals respectively. In an embodiment of the present invention, the clustering algorithm can be a k-means algorithm or a density-based spatial clustering application with noise (DBSCAN, Density-based Spatial Clustering of Applications with Noise). Taking Figure 2 as an example, the same distance frequency domain signal is divided into two sub-distance frequency domain signals. One distance frequency domain signal corresponds to one distance range, and the other distance frequency domain signal corresponds to another distance range. Two different distance ranges correspond to different targets 16, 16' and 16". One distance range corresponds to one target 16 and the other distance range corresponds to two targets 16' and 16". Therefore, the FM continuous wave processor 12 can first perform Fourier transform on all the echo signals of the target 16 in the slow time dimension or the space dimension to generate all speed frequency domain signals or all azimuth frequency domain signals, respectively. Then, the FM continuous wave processor 12 performs Fourier transform on all the echo signals of the targets 16' and 16" in the slow time dimension or the space dimension to generate all speed frequency domain signals or all azimuth frequency domain signals, respectively. Use the clustering algorithm to divide the data into multiple groups, and perform subsequent processing on each group, and then merge the processing results of each group to effectively reduce the interference caused by the position estimation of the complex data, and Improve the real-time data processing.

In an embodiment of the present invention, the FM continuous wave processor 12 calculates the beat frequency Fb of each sweep signal and its corresponding echo signal corresponding to the same target 16, 16' or 16", and calculates the same target The difference of the beat frequency Fb corresponding to the object 16, 16' or 16", and according to this difference, all echo signals are Fourier transformed in the slow time dimension to generate all velocity frequency domain signals. The FM continuous wave processor 12 uses a grouping algorithm to group the same speed frequency domain signal to form a plurality of sub-speed frequency domain signals, and all the sub-speed frequency domain signals correspond to a plurality of different speed ranges. The FM continuous wave processor 12 distinguishes all targets 16, 16', and 16" into different groups according to all speed ranges, and performs Fourier transform on all echo signals in the fast time dimension or the space dimension according to the different groups. Generate all distance frequency domain signals or all azimuth frequency domain signals respectively. In one embodiment of the present invention, the clustering algorithm can be k-means algorithm or a noise density based spatial clustering application method (DBSCAN). , Density-based Spatial Clustering of Applications with Noise). Taking Figure 2 as an example, the same speed frequency domain signal is divided into two sub-speed frequency domain signals, one of which corresponds to a speed range, and the other speed frequency domain signal Correspond to another speed range. Two different speed ranges correspond to different targets 16, 16' and 16", one of the speed ranges corresponds to one target 16 and the other speed range corresponds to two targets 16' and 16". Therefore, the FM continuous wave processor 12 can first perform Fourier transform on all the echo signals of the target 16 in the fast time dimension or the space dimension to generate all the range frequency domain signals or all azimuth frequency domain signals respectively. Then, the FM continuous wave processor 12 performs Fourier transform on all the echo signals of the targets 16' and 16" in the fast time dimension or the space dimension to respectively generate all distance frequency domain signals or all azimuth frequency domain signals. The present invention Use the clustering algorithm to divide the data into multiple groups, and perform subsequent processing on each group, and then merge the processing results of each group to effectively reduce the interference caused by the position estimation of the complex data, and Improve the real-time data processing.

In an embodiment of the present invention, the FM continuous wave processor 12 obtains the time points of the echo signals corresponding to the same target object 16, 16' or 16" received by all the antennas 14, and calculates the time difference of all time points, And use all time differences to perform Fourier transform on all echo signals in the spatial dimension to generate all azimuth frequency domain signals. The FM continuous wave processor 12 uses grouping algorithm to group the same azimuth frequency domain signals to form a plurality of sub-azimuth frequencies All sub-azimuth frequency-domain signals correspond to a plurality of different azimuth ranges. The FM continuous wave processor 12 distinguishes all targets 16, 16', and 16" into different groups according to all azimuth ranges, and according to the different groups. Perform Fourier transform on all echo signals in fast time dimension or slow time dimension to generate all distance frequency domain signals or all velocity frequency domain signals respectively. In an embodiment of the present invention, the clustering algorithm may be a k-means algorithm or a Density-based Spatial Clustering of Applications with Noise (DBSCAN). Taking Figure 2 as an example, the same azimuth frequency domain signal is divided into two sub-azimuth frequency domain signals, where one azimuth frequency domain signal corresponds to one azimuth range, and the other azimuth frequency domain signal corresponds to another azimuth range. Two different azimuth ranges correspond to different targets 16, 16' and 16". One azimuth range corresponds to one target 16, and the other azimuth range corresponds to two targets 16' and 16". Therefore, the FM continuous wave processor 12 can first perform Fourier transform on all the echo signals of the target 16 in the fast time dimension or the slow time dimension to generate all distance frequency domain signals or all speed frequency domain signals, respectively. then, The FM continuous wave processor 12 then performs Fourier transform on all the echo signals of the targets 16' and 16" in the fast time dimension or the slow time dimension to respectively generate all distance frequency domain signals or all speed frequency domain signals. The present invention utilizes The grouping algorithm divides the data into multiple groups, performs subsequent processing on each group, and then combines the processing results of each group to effectively improve the timeliness of data processing.

，是方位角，d是天線之間的間距，f _c 是中心頻率，c則是光速。若具波峰A的距離頻域訊號與此距離頻率響應訊號相同時，代表此距離頻域訊號之波峰A所對應的距離為1公里。依此類推，調頻連續波處理器12取得所有距離頻域訊號之波峰A、B與C所對應的距離。 The following describes the method of the present invention to obtain the distances, speeds and azimuths corresponding to the peaks A, B and C. If the beat frequency Fb of each sweep signal and its corresponding echo signal corresponding to the same target 16, 16' or 16" is not obtained, the FM continuous wave processor 12 calculates the same target 16, 16' or 16 "Corresponding to the beat frequency Fb of each sweep signal and its corresponding echo signal. If the beat frequency Fb of each sweep signal and its corresponding echo signal corresponding to the same target 16, 16' or 16" has been obtained, the FM continuous wave processor 12 can directly use the same target 16, 16' or 16" corresponds to the beat frequency Fb of each sweep signal and its corresponding echo signal. The beat frequency Fb of each sweep signal corresponding to the same target 16, 16' or 16" and its corresponding echo signal may have a specific relationship, for example, a linear relationship, and the FM continuous wave processor 12 is based on the same target 16, 16' or 16" corresponds to the beat frequency Fb of each sweep signal and its corresponding echo signal, dynamic programming algorithm and all different distance ranges to obtain all the peaks A, B and C of the frequency domain signal at all distances Corresponding distance. In addition to the above methods, the FM continuous wave processor 12 can pre-store a plurality of distance frequency response signals corresponding to different distances, and compare all distance frequency response signals and distances according to the minimum absolute value convergence and selection operator (LASSO) algorithm Frequency domain signal to obtain the distance corresponding to the peak of the frequency domain signal. Assuming that the distance frequency response signal corresponding to 1 km is

, Is the azimuth angle, d is the distance between antennas, f _c is the center frequency, and c is the speed of light. If the distance frequency domain signal with peak A is the same as the distance frequency response signal, it means that the distance corresponding to the peak A of the distance frequency domain signal is 1 km. By analogy, the FM continuous wave processor 12 obtains all distances corresponding to the peaks A, B, and C of the frequency domain signal.

The FM continuous wave processor 12 substitutes the distances corresponding to all peaks A into the n-degree polynomial to obtain the time-varying model and its parameters of the target 16 in distance, and substitutes the distances corresponding to all peaks B into The time-varying model and parameters of the target 16' in distance can be obtained from the nth degree polynomial, and the distance corresponding to all wave crests C can be substituted into the nth degree polynomial to get the time-varying model and its parameters of the target 16" in distance. The FM continuous wave processor 12 calculates the relative distances of all targets 16, 16' and 16" to the FM continuous wave radar 10 according to all time-varying models. The invention fits the parameters of a plurality of time-varying models to effectively improve the signal-to-noise ratio, and accurately calculates the distance of the target 16, 16' and 16" relative FM continuous wave radar 10, and effectively reduces the complexity of data fitting , And improve the real-time data processing.

，是方位角，d是天線之間的間距，f _c 是中心頻率，c則是光速。若具波峰A的速度頻域訊號與此速度頻率響應訊號相同時，代表此速度頻域訊號之波峰A所對應的速度為1公里/分。依此類推，調頻連續波處理器12取得所有速度頻域訊號之波峰A、B與C所對應的速度。 If the beat frequency Fb of each sweep signal and its corresponding echo signal corresponding to the same target 16, 16' or 16" is not obtained, the FM continuous wave processor 12 calculates the same target 16, 16' or 16 "Corresponding to the beat frequency Fb of each sweep signal and its corresponding echo signal. If the beat frequency Fb of each sweep signal and its corresponding echo signal corresponding to the same target 16, 16' or 16" has been obtained, the FM continuous wave processor 12 can directly use the same target 16, 16' or 16" corresponds to the beat frequency Fb of each sweep signal and its corresponding echo signal. The beat frequency Fb of each sweep signal corresponding to the same target 16, 16' or 16" and its corresponding echo signal may have a specific relationship, for example, a linear relationship, and the FM continuous wave processor 12 is based on the same target 16, 16' or 16” corresponds to the beat frequency Fb of each sweep signal and its corresponding echo signal, dynamic programming algorithm and all different speed ranges to obtain the speeds corresponding to the peaks A, B and C of all speed frequency domain signals. In addition to the above methods, the FM continuous wave processor 12 can pre-store a plurality of speed frequency response signals corresponding to different speeds, and compare all speed frequency response signals and speeds according to the minimum absolute value convergence and selection operator (LASSO) algorithm Frequency domain signal to obtain the speed corresponding to the peak of the speed frequency domain signal. Assuming that the speed frequency response signal corresponding to 1 km/min is

, Is the azimuth angle, d is the distance between antennas, f _c is the center frequency, and c is the speed of light. If the speed frequency domain signal with peak A is the same as the speed frequency response signal, it means that the speed corresponding to the peak A of the speed frequency domain signal is 1 km/min. By analogy, the FM continuous wave processor 12 obtains the speeds corresponding to the peaks A, B, and C of all speed frequency domain signals.

The FM continuous wave processor 12 substitutes the speeds corresponding to all peaks A into the nth degree polynomial The time-varying model and its parameters of the target 16's speed can be obtained. Substituting the speeds corresponding to all wave crests B into the n-degree polynomial can obtain the time-varying model and its parameters of the target 16' in speed. Substituting the speed into the n-degree polynomial can obtain the time-varying model and its parameters of the target 16" speed. The FM continuous wave processor 12 calculates all targets 16, 16' and 16" relative FM continuous according to all the time-varying models Wave radar 10 speed. The invention fits the parameters of multiple time-varying models to effectively improve the signal-to-noise ratio, and accurately calculates the speed of the target 16, 16' and 16" relative frequency modulated continuous wave radar 10, and effectively reduces the complexity of data fitting , And improve the real-time data processing.

，是方位角，d是天線之間的間距，f _c 是中心頻率，c則是光速。若具波峰A的方位頻域訊號與此方位頻率響應訊號相同時，代表此方位頻域訊號之波峰A所對應的方位為15度。依此類推，調頻連續波處理器12取得所有方位頻域訊號之波峰A、B與C所對應的方位角。 If the beat frequency Fb of each sweep signal and its corresponding echo signal corresponding to the same target 16, 16' or 16" is not obtained, the FM continuous wave processor 12 calculates the same target 16, 16' or 16 "Corresponding to the beat frequency Fb of each sweep signal and its corresponding echo signal. If the beat frequency Fb of each sweep signal and its corresponding echo signal corresponding to the same target 16, 16' or 16" has been obtained, the FM continuous wave processor 12 can directly use the same target 16, 16' or 16" corresponds to the beat frequency Fb of each sweep signal and its corresponding echo signal. The beat frequency Fb of each sweep signal corresponding to the same target 16, 16' or 16" and its corresponding echo signal may have a specific relationship, for example, a linear relationship, and the FM continuous wave processor 12 is based on the same target 16, 16' or 16" corresponds to the beat frequency Fb of each sweep signal and its corresponding echo signal, dynamic programming algorithm and all different azimuth ranges to obtain the azimuths corresponding to the peaks A, B and C of all azimuth frequency domain signals. In addition to the above methods, the FM continuous wave processor 12 can prestore a plurality of azimuth frequency response signals corresponding to different azimuths, and compare all azimuth frequency response signals and azimuths according to the minimum absolute value convergence and selection operator (LASSO) algorithm Frequency domain signal to obtain the azimuth angle corresponding to the peak of the azimuth frequency domain signal. Assuming that the azimuth frequency response signal corresponding to 15 degrees is

, Is the azimuth angle, d is the distance between antennas, f _c is the center frequency, and c is the speed of light. If the azimuth frequency domain signal with peak A is the same as the azimuth frequency response signal, it means that the azimuth corresponding to the peak A of the azimuth frequency domain signal is 15 degrees. By analogy, the FM continuous wave processor 12 obtains the azimuth angles corresponding to the peaks A, B, and C of all azimuth frequency domain signals.

The FM continuous wave processor 12 substitutes the azimuth angles corresponding to all crests A into the nth degree polynomial to obtain the time-varying model of the target 16 in azimuth and its parameters, and substitutes the azimuth angles corresponding to all crests B into the nth degree polynomial to obtain The time-varying model and its parameters of the target 16' in azimuth. Substituting the azimuth angles corresponding to all wave crests C into the n-degree polynomial can obtain the time-varying model and its parameters of the target 16" in the azimuth. FM continuous wave processor 12 Calculate the azimuth angles of all targets 16, 16' and 16" relative to the FM CW radar 10 based on all time-varying models. The invention fits the parameters of a plurality of time-varying models to effectively improve the signal-to-noise ratio, and accurately calculates the azimuth angle of the target 16, 16' and 16" relative frequency modulated continuous wave radar 10, and effectively reduces the complexity of data fitting And improve the real-time data processing.

In an embodiment of the present invention, as shown in Figure 4, an oblique solid line represents a sweep signal, and the echo signal is represented by an oblique dashed line. The corresponding time of all sweep signals in the coherent processing time interval T0-T1 The intervals are staggered to avoid overlapping, and all corresponding time intervals are randomly distributed. The coherent processing time interval is 1-10 milliseconds (ms). All corresponding time intervals account for less than or equal to 20% of the coherent processing time interval T0-T1, and greater than 0 . The present invention reduces the time and probability of the frequency sweep signal appearing in the coherent processing time interval T0-T1 to reduce subsequent processing interference, while reducing the probability of multi-user interference, and improving the accuracy of estimated distance, speed or azimuth.

In summary, the present invention fits the parameters of multiple time-varying models to effectively improve the signal-to-noise ratio, and accurately calculates the distance, speed or azimuth of the target relative to the FM continuous wave radar, reducing the complexity of data fitting , And improve the real-time data processing.

The above is only a preferred embodiment of the present invention, and is not used to limit the scope of implementation of the present invention. Therefore, all the shapes, structures, features and spirits described in the scope of the patent application of the present invention are equally changed and modified. , Should be included in the scope of patent application of the present invention.

10: FM continuous wave radar

12: FM continuous wave processor

14: Antenna

16: target

16’: Target

16": target

Claims (17)

A frequency modulated continuous wave (FMCW) processing device, which includes: a frequency modulated continuous wave radar, which has a plurality of antennas with a plurality of targets in front of the antennas; and a frequency modulated continuous wave processor electrically connected to the frequency modulated continuous wave Wave radar, the FM continuous wave processor controls one of the antennas of the FM continuous wave radar to sequentially transmit a plurality of frequency sweep signals to the targets in a coherent processing time interval ( chirp signals), the targets reflect the sweep signals to form a plurality of echo signals, the FM continuous wave processor receives the echo signals through the antennas of the FM continuous wave radar, and demodulates the Some echo signals, and Fourier transform is performed on these echo signals in fast time dimension, slow time dimension or space dimension to generate plural The distance frequency domain signal, a plurality of speed frequency domain signals or a plurality of azimuth frequency domain signals, the distance frequency domain signal, the speed frequency domain signal or the azimuth frequency domain signal is a wave signal with a plurality of peaks, and the number of these peaks As the number of the targets is the same, the FM continuous wave processor uses the distance frequency domain signals, the speed frequency domain signals or the azimuth frequency domain signals to establish a plurality of time-varying models. The number of variable models is the same as the number of the wave crests, and the FM continuous wave processor calculates the distance, speed or azimuth angle of the targets relative to the FM continuous wave radar according to the time-varying models. The FM continuous wave processing device according to claim 1, wherein the FM continuous wave processor uses a clustering algorithm to cluster the same distance frequency domain signal to form a plurality of sub-distance frequency domain signals. The distance frequency domain signal corresponds to a plurality of different distance ranges, and the FM continuous wave processor distinguishes according to these distance ranges The targets are in different groups, and the echo signals are Fourier transformed in the slow time dimension or the space dimension according to the different groups to generate the velocity frequency domain signals or the azimuth frequency domains respectively Signal. The FM continuous wave processing device according to claim 2, wherein the FM continuous wave processor calculates the beat frequency of each sweep signal corresponding to the same target object and the corresponding echo signal, According to the beat frequency, dynamic programming algorithm and the distance ranges, the distances corresponding to the peaks of the distance frequency domain signals are obtained. The FM continuous wave processing device according to claim 1, wherein the FM continuous wave processor uses a clustering algorithm to cluster the same velocity frequency domain signal to form a plurality of sub-velocity frequency domain signals. The speed frequency domain signals correspond to a plurality of different speed ranges. The FM continuous wave processor distinguishes the targets into different groups according to the speed ranges, and according to the different groups, the echo signals are in the fast time Fourier transform is performed on the dimension or the spatial dimension to generate the distance frequency domain signals or the azimuth frequency domain signals respectively. The FM continuous wave processing device according to claim 4, wherein the FM continuous wave processor calculates the beat frequency of each sweep signal corresponding to the same target object and its corresponding echo signal, And according to the beat frequency, dynamic programming algorithm and the speed ranges, the speeds corresponding to the peaks of the speed frequency domain signals are obtained. The FM continuous wave processing device according to claim 1, wherein the FM continuous wave processor uses a clustering algorithm to group the same azimuth frequency domain signal to form a plurality of sub-azimuth frequency domain signals. The azimuth frequency domain signals respectively correspond to a plurality of different azimuth ranges, and the FM continuous wave processor distinguishes according to these azimuth ranges The targets are in different groups, and the echo signals are Fourier transformed in the fast time dimension or the slow time dimension according to the different groups to generate the distance frequency domain signals or the speed frequency signals respectively Domain signal. The FM continuous wave processing device according to claim 6, wherein the FM continuous wave processor calculates the beat frequency of each sweep signal corresponding to the same target object and its corresponding echo signal, The azimuth angles corresponding to the peaks of the azimuth frequency domain signals are obtained according to the beat frequency, dynamic programming algorithm and the azimuth ranges. The FM continuous wave processing device according to claim 2, 4, or 6, wherein the grouping algorithm is a k-means algorithm or a density-based spatial grouping method with noise (DBSCAN, Density-based Spatial) Clustering of Applications with Noise). The FM continuous wave processing device according to claim 1, wherein the FM continuous wave processor uses a Constant False-Alarm Rate (CFAR) to detect the peaks. The FM continuous wave processing device according to claim 1, wherein the FM continuous wave processor calculates the beat frequency of each sweep signal corresponding to the same target object and its corresponding echo signal, According to the beat frequency, Fourier transform is performed on the echo signals in the fast time dimension to generate the distance frequency domain signals. The FM continuous wave processing device according to claim 1, wherein the FM continuous wave processor calculates the beat frequency of each sweep signal corresponding to the same target object and its corresponding echo signal, And calculate the difference of the beat frequency corresponding to the same target, and perform Fourier transform on the slow time dimension on the echo signals according to the difference to generate the speed frequency domain signals. The FM continuous wave processing device according to claim 1, wherein the FM continuous wave processor is Obtain the time point of the echo signal corresponding to the same target object received by the antennas, calculate the time difference of the time point, and use the time difference to Fourier transform the echo signals in the spatial dimension to Generate these azimuth frequency domain signals. The FM continuous wave processing device according to claim 1, wherein the time-varying model is represented by an n-th degree polynomial, and n is greater than or equal to 1. The FM continuous wave processing device according to claim 1, wherein the FM continuous wave processor stores a plurality of azimuth frequency response signals corresponding to different azimuth angles, and converges according to the minimum absolute value and compares the selection operator (LASSO) algorithm The azimuth frequency response signals and the azimuth frequency domain signals are used to obtain the azimuth angles corresponding to the peaks of the azimuth frequency domain signals. The FM continuous wave processing device according to claim 1, wherein the FM continuous wave processor stores a plurality of speed frequency response signals corresponding to different speeds, and compares them according to the minimum absolute value convergence and selection operator (LASSO) algorithm The speed frequency response signals and the speed frequency domain signals are used to obtain the speed corresponding to the peak of the speed frequency domain signals. The FM continuous wave processing device according to claim 1, wherein the FM continuous wave processor stores a plurality of distance frequency response signals corresponding to different distances, and compares them according to the minimum absolute value convergence and selection operator (LASSO) algorithm The distance frequency response signals and the distance frequency domain signals are used to obtain the distances corresponding to the peaks of the distance frequency domain signals. The frequency modulated continuous wave processing device according to claim 1, wherein the corresponding time intervals of the frequency sweep signals in the coherent processing time interval are staggered, and the corresponding time intervals are randomly distributed, and the coherent processing time interval is 1-10 In milliseconds (ms), the corresponding time interval is less than or equal to 20% of the coherent processing time interval, and is greater than 0.

Images