Signal processing method and signal processing equipment based on DDMIMO radar and computer readable storage medium

Abstract

The invention relates to a signal processing method and signal processing equipment based on a DDMIMO radar and a computer readable storage medium. The signal processing method comprises signal modulation and demodulation. In the modulation stage, correlation between an outdoor driving scene target and a high-frequency Doppler position interval is determined firstly, and then DDMIMO radar TX channel non-equal-difference phase deviation parameters are designed; in the demodulation stage, 1D and two-dimensional FFT processing is sequentially carried out on echoes, the amplitude is calculated to obtain a distance and a Doppler two-dimensional matrix, Doppler-dimensional bottom noise is estimated through OS of distance points one by one, then distance-dimensional filtering is carried out, a CFAR threshold value is set based on the bottom noise, a peak value is screened to generate a mark map, corresponding Doppler points are accumulated according to phase parameters, and an accumulated value and a set threshold value are compared to determine an effective target point. According to the method, the multi-target Doppler dimension channel overlapping probability can be reduced, the small target detection capability is improved, and the target detection precision and reliability in a complex outdoor driving scene are ensured.

Description

Technical Field

[0001] This invention relates to the field of millimeter-wave radar signal processing technology, and in particular to a signal processing method, signal processing device, and computer-readable storage medium based on DDMIMO radar. Background Technology

[0002] As one of the core sensors in autonomous driving systems, automotive millimeter-wave radar transmits frequency-modulated continuous wave signals and uses the reflected echoes to estimate information such as the target's distance, speed, and angle. Its performance directly affects driving safety. DDMIMO radar, as one of the mainstream MIMO systems, applies phase modulation to different transmit antennas (TX channels) to separate the signals of each TX channel in the Doppler dimension. Compared with traditional time-division MIMO radar, it has higher transmit gain and anti-interference capabilities, and has received widespread attention in recent years.

[0003] In existing DDMIMO radar signal processing schemes, signal modulation mainly adopts an equally spaced phase offset design. For example, the phase offset of the six TX channels is set at equal intervals of 45°, and the corresponding Doppler dimension sub-bands are also equally spaced. The demodulation process is based on the "empty band separation" logic, which extracts the amplitude at a specific position in the Doppler dimension and accumulates iteratively, selecting the position corresponding to the maximum accumulated value as the target point. Figure 1 This diagram illustrates typical parameters for equally spaced phase modulation in the prior art, used to explain the phase offset configuration of six transmit channels (TX0–TX5) under different chirp waveforms. The “chirp number” in the diagram represents the chirp number within a transmit frame, including chirp0 to chirp255 (a total of 256 chirs, consistent with the number of waveforms in a vehicle-mounted radar transmit frame). The “FMCW wave” represents the linear frequency modulated continuous wave waveform corresponding to each chirp; the diagram shows a rising edge triangular wave, a typical transmit waveform for vehicle-mounted millimeter-wave radar. The “TX0–TX5 phase” represents the phase offset values ​​of the six transmit channels under the corresponding chirp. The characteristic of the phase configuration in the diagram is that the phase of channel TX0 is always 0°, while the phases of TX1–TX5 follow a 45° equally spaced increment pattern. For example, the phase of TX1 starts from 0° at chirp0, increasing by 45° at each chirp, reaching 315° at chirp255; the phases of the remaining TX channels also shift at equally spaced steps of 45°. Figure 2 yes Figure 1This is a schematic diagram of the Doppler dimension signal amplitude distribution curve of a single stationary target under equally spaced phase modulation. As shown in the figure, the X-axis represents the Doppler dimension index (range 0-300), corresponding to the Doppler dimension position of the radar echo after two-dimensional FFT processing, and is directly related to the target velocity. The Y-axis represents the Doppler dimension signal amplitude value, reflecting the energy intensity of the target echo. The peak values ​​at X positions (Doppler index) 1, 33, 65, 97, 129, and 161 in the figure correspond to... Figure 1 The Doppler response positions of the six transmit channels (TX0 to TX5) in the equally spaced phase modulation are determined, and the formation of these peaks and subsequent demodulation adopt the existing technology of using empty bands for TX demodulation. Specifically, in this scheme, the phase interval of the 6 TX channels is 45°. Since the entire Doppler dimension corresponds to a 360° phase range, the 360° phase can be divided into 8 sub-bands of 45° each. Six of these sub-bands correspond to the 6 TX channels, and the remaining 2 are blank sub-bands. During demodulation, the amplitude at the Doppler index N (N ranges from 1 to 32) is extracted first. Then, the amplitudes at Doppler_(N+32), Doppler_(N+64), Doppler_(N+96), Doppler_(N+128), Doppler_(N+160), Doppler_(N+192), and Doppler_(N+224) are extracted sequentially according to a pattern. Subsequently, 6 amplitudes are selected sequentially from each of these 8 amplitudes and accumulated (if the selection exceeds the index range, a loop is performed). Finally, 8 accumulated values ​​are obtained. By extracting the maximum value among these 8 accumulated values ​​and recording the starting Doppler position corresponding to the maximum value, demodulation can be completed.

[0004] However, existing technologies have the following significant drawbacks: In multi-target scenarios, equal-interval phase modulation results in a uniform distribution of each TX channel in the Doppler dimension, which can easily lead to a high degree of overlap between the TX channels of different targets (such as the TX0 channel of one target overlapping with the TX1 to TX5 channels of another target), causing demodulation errors and angle phase interference, resulting in angle estimation deviation. The energy radiation of each TX channel of a radar is affected by the antenna pattern, propagation path attenuation, and hardware deviations, resulting in inherent differences at different angles. When facing small targets, their echo energy is inherently weak, and the echo signal of some TX channels (sidelobe direction) is close to the noise level. Relying solely on the demodulation logic of "maximum accumulated value" can easily lead to misjudging the accumulated value of clutter as the target, resulting in the missed detection of small targets. The selection method of "maximum accumulated value" during demodulation will include the occasional high amplitude of clutter in the statistics, resulting in an overestimation of the noise floor in the subsequent CFAR (constant false alarm rate) processing. Small target signals are overwhelmed by the noise floor, reducing detection sensitivity.

[0005] To address the aforementioned issues, there is an urgent need for a DDMIMO radar signal processing method adapted to outdoor driving scenarios, which can optimize both modulation and demodulation to solve technical challenges such as channel overlap, misjudgment of small targets, and background noise overwhelmance. Summary of the Invention

[0006] To address the aforementioned problems in existing technologies, this invention proposes a signal processing method, signal processing device, and computer-readable storage medium based on DDMIMO radar. This method can reduce the probability of overlap in the Doppler dimension channels of multiple targets, improve the detection capability of small targets, and ensure the accuracy and reliability of target detection in complex outdoor driving scenarios.

[0007] Specifically, the present invention proposes a signal processing method based on DDMIMO radar, the signal processing method including a signal modulation method and a signal demodulation method; The signal modulation method includes the following steps: The TX channels of the DDMIMO radar are controlled to perform signal modulation using non-uniform phase offset parameters; The signal demodulation method includes the following steps: The radar echo signal is processed sequentially by one-dimensional FFT and two-dimensional FFT, and the amplitude value is calculated to obtain the range and Doppler two-dimensional signal matrix. Based on the two-dimensional signal matrix, the noise floor level in the Doppler dimension is estimated point by point. The noise floor estimation results at each distance point are then filtered in the distance dimension to obtain the noise floor value. A CFAR threshold is set based on the noise floor value, and peak target points are filtered out based on the CFAR threshold to generate a peak marker map. Based on the non-uniform phase offset parameters of each TX channel, corresponding Doppler points are selected at each position in the peak marker spectrum for accumulation to generate an accumulation spectrum; The accumulated value at each position in the accumulated map is compared with a set threshold to determine the valid target point; The non-equal arithmetic phase offset parameter is obtained in the following manner: Determine the high-frequency Doppler location range of the target in outdoor driving scenarios; Determine the non-uniform phase offset parameters of the TX channel of the DDMIMO radar.

[0008] According to an embodiment of the present invention, determining the Doppler position interval in obtaining the non-equal arithmetic phase offset parameters includes the following steps: Determine the basic parameters of the DDMIMO radar, including carrier frequency, wavelength, pulse repetition period, number of chirps per transmitted frame, and number of doppler-dimensional FFT points; Determine the velocity resolution, and based on the velocity resolution, determine the actual velocity range for different target types, and associate the actual velocity range with the Doppler position interval; Determining the non-uniform phase offset parameter in the signal modulation method includes the following steps: Set constraints; Based on the aforementioned constraints, non-uniform phase offset parameters for each of the TX channels are determined by combining the Doppler position intervals where the target frequently occurs.

[0009] According to one embodiment of the present invention, the constraint conditions include: The Doppler positions of each TX channel are non-arithmetically distributed within the target high-frequency range, and the difference between the Doppler indices of any two TX channels is greater than a set threshold. The phase offset parameters must meet the accuracy requirements of the hardware phase shifter of the DDMIMO radar, and the total phase range must cover 360°.

[0010] According to one embodiment of the present invention, the DDMIMO radar includes six TX channels TX0 to TX5, and the phase offset parameters corresponding to each TX channel are 0°, 50.625°, 123.75°, 219.375°, 281.25° and 320.625°, respectively.

[0011] According to one embodiment of the present invention, the OS method is used to estimate the noise floor level of the Doppler dimension.

[0012] According to one embodiment of the present invention, the distance dimension filtering process adopts a moving average filter, and the filter window size is 3 to 5 consecutive distance points.

[0013] According to one embodiment of the present invention, the signal-to-noise ratio of the CFAR threshold is 3.5dB, and peak filtering is performed using a cell-averaged constant false alarm rate detector.

[0014] According to one embodiment of the present invention, the set threshold is 4, and when the accumulated value is greater than 4, it is determined to be a valid target point; the Doppler position interval of each TX channel satisfies the formula Idop_m=Idop_total+ Nfft_dop / Q, where Idop_total is the target real Doppler index. is the phase coefficient of the TX channel, Nfft_dop is the number of FFT points in the Doppler dimension, and Q is the number of equal parts in the Doppler dimension.

[0015] The present invention also provides a signal processing device based on DDMIMO radar, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the signal processing method described in any of the preceding claims.

[0016] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the signal processing method as described in any of the preceding claims.

[0017] This invention provides a signal processing method, signal processing device, and computer-readable storage medium based on DDMIMO radar. Through the cooperation of non-equal phase modulation design and optimized demodulation process, it can reduce the probability of overlap of multiple target Doppler dimension channels, improve the detection capability of small targets, and ensure the accuracy and reliability of target detection in complex outdoor driving scenarios.

[0018] It should be understood that the above general description and the following detailed description of the invention are exemplary and illustrative, and are intended to provide further explanation of the invention as described in the claims. Attached Figure Description

[0019] The accompanying drawings are included to provide further explanation of the invention; they are incorporated into and constitute a part of this application. The drawings illustrate embodiments of the invention and, together with this specification, serve to explain the principles of the invention. In the drawings: Figure 1 A schematic diagram of typical parameters for equally spaced phase modulation in the prior art is shown.

[0020] Figure 2 yes Figure 1 A schematic diagram of the Doppler dimension signal amplitude distribution curve of a single stationary target under equal-interval phase modulation.

[0021] Figure 3 A flowchart of a signal processing method based on DDMIMO radar according to an embodiment of the present invention is shown.

[0022] Figure 4 A schematic diagram of non-uniform phase offset parameters according to an embodiment of the present invention is shown.

[0023] Figure 5 yes Figure 4 A schematic diagram of the Doppler dimension signal amplitude distribution curves of each TX channel under non-equal phase modulation.

[0024] Figure 6 A system block diagram of a signal processing device based on DDMIMO radar according to an embodiment of the present invention is shown. Detailed Implementation

[0025] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0026] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit this application or its application or use. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0027] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0028] Unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps described in these embodiments do not limit the scope of this application. It should also be understood that, for ease of description, the dimensions of the various parts shown in the drawings are not drawn to actual scale. Techniques, methods, and devices known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and devices should be considered part of the specification. In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values. It should be noted that similar reference numerals and letters in the following drawings denote similar items; therefore, once an item is defined in one drawing, it need not be further discussed in subsequent drawings.

[0029] In the description of this application, it should be understood that the orientation or positional relationship indicated by directional terms such as "front, back, up, down, left, right", "horizontal, vertical, horizontal" and "top, bottom" is usually based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of describing this application and simplifying the description. Unless otherwise stated, these directional terms do not indicate or imply that the device or element referred to must have a specific orientation or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on the scope of protection of this application; the directional terms "inner" and "outer" refer to the inner and outer contours relative to the outline of each component itself.

[0030] Furthermore, it should be noted that the use of terms such as "first" and "second" to define components is merely for the purpose of distinguishing the corresponding components. Unless otherwise stated, these terms have no special meaning and therefore should not be construed as limiting the scope of protection of this application. In addition, although the terminology used in this application is selected from commonly known and used terms, some terms mentioned in this application's specification may have been chosen by the applicant according to his or her judgment, and their detailed meanings are explained in the relevant sections of this description. Moreover, this application should be understood not only through the actual terms used, but also through the meaning implied by each term.

[0031] Figure 3 A flowchart of a signal processing method based on DDMIMO radar according to an embodiment of the present invention is shown. As shown in the figure, a signal processing method based on DDMIMO radar mainly addresses the shortcomings of traditional DDMIMO radar signal processing in vehicle-mounted scenarios, such as overlapping Doppler dimension channels for multiple targets and the ease with which small targets are submerged by background noise. It achieves accurate target detection through the coordinated use of signal modulation and demodulation. This signal processing method includes two parts: a signal modulation method and a signal demodulation method.

[0032] The signal modulation method includes the following steps: The TX channels of the DDMIMO radar are controlled using non-uniform phase offset parameters for signal modulation. These non-uniform phase offset parameters are obtained as follows: First, it is necessary to determine the Doppler position range where the target frequently appears in outdoor driving scenarios. This is because the core detection targets of vehicle radar, such as stationary obstacles, pedestrians, non-motorized vehicles, and vehicles under different road conditions, have clear speed range constraints. The speed resolution of the radar can map this speed range to the corresponding Doppler index range. Based on this, phase design can make the modulation parameters more suitable for the target distribution in the actual scenario and avoid redundant design of invalid areas. Next, the non-uniform phase offset parameters of the TX channels of the DDMIMO radar need to be determined. The design of the non-uniform phase offset parameters must ensure that each TX channel is discretely distributed within the high-frequency Doppler range of the target, thereby reducing the probability of channel overlap in the Doppler dimension of multiple targets from the root.

[0033] Signal demodulation methods include the following steps: First, the radar echo signal is processed sequentially using one-dimensional FFT and two-dimensional FFT, and the amplitude value is calculated to obtain the range and Doppler two-dimensional signal matrix. This step converts the time-domain echo signal into two-dimensional domain data that can directly reflect the target's range and velocity information, providing a foundation for subsequent processing. Subsequently, the noise floor level in the Doppler dimension is estimated point-by-point based on the two-dimensional signal matrix, and the noise floor estimation results are filtered in the distance dimension to obtain a stable noise floor value. Based on the aforementioned stable noise floor value, a CFAR threshold can be set, and peak target points can be selected based on this threshold to generate a peak marker spectrum. This step filters out most of the background noise through constant false alarm rate (CFAR) detection, retaining only high-confidence potential target points and avoiding invalid clutter from participating in subsequent calculations. Then, based on the non-uniform phase offset parameters of each TX channel, corresponding Doppler points are selected at each position in the peak marker spectrum and accumulated to generate an accumulation spectrum. The non-uniform phase offset corresponds to a specific Doppler position interval pattern. Accumulating according to this pattern can effectively correlate the target response of each TX channel and enhance the correlation of the target signal. Finally, the accumulated values ​​at each position in the accumulated spectrum are compared with the set threshold to determine the valid target points. This verification step can achieve consistency determination of the response of multiple TX channels, weaken the signal fluctuations caused by the angular energy difference of a single TX channel, and replace the traditional coarse judgment based on the single accumulated maximum value, thereby improving the detection accuracy of small targets and avoiding clutter misjudgment.

[0034] In some examples, determining the Doppler position range is crucial for establishing a correlation logic between "radar parameters - speed range - Doppler position" based on the actual needs of the vehicle scenario, ensuring that the phase design adapts to the target distribution characteristics.

[0035] First, the fundamental parameters of the DDMIMO radar need to be determined. These parameters include carrier frequency, wavelength, pulse repetition period, number of chirps per transmission frame, and the number of Doppler-dimensional FFT points. The preferred carrier frequency is 77GHz (the mainstream frequency band for vehicle-mounted 4D millimeter-wave radar). Wavelength, pulse repetition period, and number of chirps are prerequisites for calculating velocity resolution, while the number of Doppler-dimensional FFT points directly determines the range of the Doppler index. These fundamental parameters collectively form the basis for subsequent interval mapping calculations; the absence of any one parameter will lead to an estimation error in the Doppler position interval. Based on this, the radar's velocity resolution capability is determined through a preset velocity resolution formula. This velocity resolution represents the smallest target velocity difference that the radar can distinguish and is crucial for connecting the actual target velocity with the Doppler position range. Combining the actual velocity ranges of different core target types in outdoor driving scenarios (such as the low-speed range of stationary obstacles and pedestrians, and the medium-to-high-speed range of vehicles ahead on urban roads and highways), the velocity ranges of each target type are converted into corresponding Doppler index ranges using velocity resolution, ultimately locking down the Doppler position ranges where targets frequently appear. This design logic allows subsequent phase parameters to avoid redundant design in invalid regions, focusing on the Doppler range of the real target set, thereby improving the scenario adaptability of the modulation scheme. Determining non-uniform phase offset parameters in signal modulation methods includes the following steps: First, clear constraints must be set, including distributed constraints and hardware constraints; Based on the above constraints and the determined high-frequency Doppler position range of the target, non-uniform phase offset parameters for each TX channel are formulated. This design ensures the discrete distribution of each TX channel in the target concentration area, reducing the probability of multiple targets coinciding at the source, while also ensuring the feasibility of the parameters through hardware constraints. At the same time, it binds the phase offset to the target distribution characteristics, enabling the signal response of each TX channel to accurately correlate with the real target during subsequent demodulation.

[0036] In some examples, the constraints set in the signal modulation stage need to take into account both the performance requirements of the scenario and the feasibility of hardware engineering, so that the two can work together to ensure the effectiveness and feasibility of the non-equal arithmetic phase offset parameters.

[0037] The first constraint is that the Doppler positions of each TX channel are non-arithmically distributed within the target high-frequency range, and the difference in Doppler indices between any two TX channels is greater than a set threshold. This constraint is designed to address the core shortcomings of traditional equal-interval phase modulation. In traditional schemes, equal-interval phase modulation results in a uniform arrangement of Doppler positions for each TX channel. In multi-target scenarios in automotive environments, the TX channels of different targets are highly likely to overlap in the Doppler dimension, leading to demodulation errors and angular phase interference. The non-arithmically distributed design allows the Doppler responses of each TX channel to be discrete within the high-frequency range of the target set. Furthermore, by limiting the "index difference to greater than a set threshold," the spacing between Doppler positions of channels is further increased, fundamentally reducing the probability of channel overlap under multi-target conditions and ensuring that the responses of different TX channels to the same target can be clearly distinguished. The second constraint is that the phase shift parameters must meet the accuracy requirements of the DDMIMO radar's hardware phase shifter, and the total phase range must cover 360°. The accuracy of the hardware phase shifter typically needs to be ≤0.1°, which determines the practical feasibility of the phase shift. If the parameters exceed the phase shifter's accuracy range, the actual phase transmitted by the radar will deviate, causing the Doppler-dimensional channel separation effect to deviate from the design expectation. The total phase range must cover 360° because the phase response in the Doppler dimension is periodic; 360° corresponds to a complete Doppler cycle. Only by covering 360° of phase can the full-cycle characteristics of the Doppler dimension be matched, ensuring that all vehicle targets within all speed ranges (from stationary to high speed) are completely covered in the Doppler domain, avoiding speed folding or target response omissions.

[0038] These two constraints work together to solve the problem of multi-target overlap in existing technologies through non-arithmetic distribution, and to ensure that the parameters can be accurately realized in actual radar systems through hardware and phase period constraints, thus taking into account both performance optimization and engineering feasibility.

[0039] Figure 4A schematic diagram of non-equal arithmetic phase offset parameters according to an embodiment of the present invention is shown. As shown in the figure, in some examples, the phase offset configuration of the six TX channels TX0 to TX5 of the DDMIMO radar under different chirs within a single transmission frame is shown. "Chirp number" represents the chirp number within a transmission frame, including chirp0 to chirp255, totaling 256 chirs, consistent with the number of waveforms in the transmission frame of an automotive radar. "FMCW wave" represents the linear frequency modulated continuous wave waveform corresponding to each chirp. In the figure, it is a rising edge triangular wave, which is a typical transmission waveform of an automotive millimeter-wave radar. The phases of TX0 to TX5 represent the phase offset values ​​of the six transmission channels under the corresponding chirs. Among them, TX0 serves as the reference channel, and its phase is always kept at 0°. TX1 to TX5 adopt a non-equal arithmetic phase offset design, with phases of 50.625°, 123.75°, 219.375°, 281.25°, and 320.625°, respectively.

[0040] These parameters are obtained using the method described above, as explained below: The first step is to determine the Doppler position range where targets frequently appear in outdoor driving scenarios. First, the basic parameters of the DDMIMO radar are determined: including carrier frequency, wavelength λ, pulse repetition period T, the number of chirps (linear frequency modulated pulses) K in each transmitted frame, and the number of Doppler-dimensional FFT points Nfft_dop. Specifically, the preferred carrier frequency is 77 GHz, corresponding to a wavelength λ ≈ 3.897 mm; the preferred pulse repetition period T is 12 μs; the preferred number of chirps K in each transmitted frame is 256; and the preferred number of Doppler-dimensional FFT points Nfft_dop is 256 (index range 0–255). Next, the velocity resolution is calculated using the formula ΔV = λ / (2TK), which represents the minimum target velocity difference that the radar can distinguish. Based on the above optimized parameters, the velocity resolution ΔV ≈ 0.063 m / s, meaning that each Doppler index corresponds to a velocity change of approximately 0.063 m / s. Then, the speed range is correlated with the Doppler position interval. The core targets in outdoor driving scenarios include stationary obstacles, pedestrians, non-motorized vehicles, vehicles ahead on urban roads, and vehicles ahead on highways. Their actual speed range has clear constraints, which can be converted into the corresponding Doppler index interval by combining the speed resolution: Stationary obstacles / pedestrians: speed range 0–1.4 m / s (0–5 km / h), corresponding to Doppler index range 0–22; Non-motorized vehicles: speed range 1.4~4.2m / s (5~15km / h), corresponding to Doppler index range 22~67; For vehicles ahead on urban roads: speed range 4.2–16.7 m / s (15–60 km / h), corresponding to Doppler index interval 67–265, which is 67–9 after periodic extension; High-speed preceding vehicle: speed range 16.7~33.3m / s (60~120km / h), corresponding to Doppler index interval 9~17 (after periodic extension).

[0041] The Doppler position ranges where the target frequently appears were finally determined to be index 0-67 (corresponding to 0-4.2 m / s) and index 9-17 (corresponding to 16.7-33.3 m / s).

[0042] The second step is to set constraints, including distribution constraints: the Doppler positions of each TX channel are non-arithmically distributed within the target's high-frequency Doppler range, and the Doppler index difference between any two TX channels is ≥5 to avoid multiple target channels overlapping; hardware constraints: the phase offset parameter must meet the accuracy of the radar hardware phase shifter (≤0.1°), and the total phase range must cover 360° to ensure full coverage of the Doppler domain.

[0043] Based on the above constraints and the target high-frequency Doppler position range, non-uniform phase offset parameters were designed for the six TX channels (TX0 to TX5). The phase offset values ​​for each TX channel are 0°, 50.625°, 123.75°, 219.375°, 281.25° and 320.625°, respectively.

[0044] The non-uniformity of this parameter set is designed to address the shortcomings of traditional equal-interval phase modulation. Traditional schemes often use 45° equal-interval offsets, resulting in a uniform distribution of Doppler positions across all TX channels, leading to a high risk of overlap in multi-target scenarios. This parameter design ensures that the Doppler positions of each TX channel are discretely distributed within the Doppler range where the target frequently occurs. The Doppler index difference between any two channels meets a set threshold, fundamentally reducing the probability of channel overlap in multi-target scenarios. Simultaneously, this parameter set is adapted to the target speed distribution in outdoor driving scenarios, enabling clear separation of the TX channel responses to key targets such as pedestrians, non-motorized vehicles, and vehicles ahead in the Doppler domain. This provides a reliable signal foundation for subsequent demodulation stages, allowing for the cumulative verification and selection of effective target points and mitigating the impact of angular energy differences between TX channels.

[0045] In some examples, the OS (Ordered Statistics) method is used during the demodulation stage to estimate the noise floor level in the Doppler dimension. This choice is made to avoid the shortcomings of the traditional averaging method for noise floor estimation. The traditional averaging method is easily affected by target signals or clutter spikes, leading to an overestimation of the noise floor and thus compressing the survival space of small target signals. The OS method, on the other hand, selects the energy value at the median position after sorting the Doppler dimension signal energy at each distance point, which can effectively eliminate the influence of abnormal peaks and reflect the true background noise level. This provides a reliable benchmark for subsequent threshold setting and target selection, which is also a key prerequisite for ensuring the sensitivity of small target detection.

[0046] In some examples, the distance dimension filtering employs a moving average filter, with the filter window size set to 3–5 consecutive distance points. The moving average filter is used because the noise floor estimated by the OS method at a single distance point may exhibit local fluctuations. Furthermore, in vehicular scenarios, the background noise at adjacent distance points is highly correlated. Smoothing the noise floor at adjacent distance points further stabilizes the noise floor value, avoiding threshold distortion caused by local noise anomalies. Choosing 3–5 consecutive distance points as the window represents a balance between noise floor stability and signal response timeliness. If the window is too small, the filtering smoothing effect is insufficient, failing to effectively suppress fluctuations; if the window is too large, it may excessively blur the noise variation characteristics of the distance dimension, even affecting the noise floor estimation accuracy of small targets at close range. This window range ensures that while stabilizing the noise floor, the effective target distance dimension information is not lost.

[0047] The aforementioned demodulation process, including OS noise floor estimation, distance-dimensional filtering, peak filtering, and accumulation verification, along with the corresponding signal characteristics, can be obtained through... Figure 5 It is presented intuitively. Figure 5 yes Figure 4This diagram illustrates the Doppler dimension signal amplitude distribution curves of each TX channel under non-equal-arc phase modulation. As shown, the X-axis represents the Doppler dimension index, ranging from 0 to 300, directly related to the target velocity. The Y-axis represents the Doppler dimension signal amplitude value, reflecting the energy intensity of the target echo. The signal curves in the figure correspond to the Doppler domain response of a single stationary target after non-equal-arc phase modulation, and its peak distribution is more discrete compared to existing equal-interval phase modulation schemes. In the demodulation stage, the noise floor of the Doppler dimension at each distance point is first estimated using the OS method. After distance dimension filtering, a stable noise floor value is obtained. Then, peak target points are filtered using Doppler dimension CFAR processing (peak points are marked as 1, and non-peak points as 0) to generate a peak marker map. Finally, based on the Doppler position interval pattern corresponding to the phase of each TX channel, six Doppler points are selected at each position in the marker map and accumulated. The accumulated value is used to determine the effective target point.

[0048] In some examples, the signal-to-noise ratio (SNR) of the CFAR threshold is set to 3.5 dB, and a cell-averaged constant false alarm rate (CFAR) detector is used for peak filtering. The advantage of the cell-averaged CFAR detector is its ability to adapt to different noise environments. It dynamically adjusts the threshold by statistically analyzing the noise levels of reference cells around the target cell, adapting to the complex and varied background noise in vehicular scenarios (such as road clutter and environmental interference). Setting the SNR to 3.5 dB is a targeted design based on the aforementioned accurate noise floor estimation results. Traditional solutions require higher SNR thresholds due to overly high noise floor estimations, which can easily lead to the filtering of small target signals. This solution obtains the true noise floor through the OS method and distance-dimensional filtering. The low SNR threshold of 3.5 dB can effectively filter out most background noise while maximizing the retention of weak signals from small targets, avoiding missed detection of small targets and achieving a balance between noise suppression and target preservation.

[0049] In some examples, the threshold for determining a valid target point is set to 4; a point is considered a valid target point when the accumulated value is greater than 4. Simultaneously, the Doppler position intervals of each TX channel follow the formula Idop_m = Idop_total + ... Nfft_dop / Q, where Idop_total is the target real Doppler index. Here, Nfft_dop is the phase coefficient of the TX channel, and Q is the number of FFT points in the Doppler dimension. This formula establishes a precise correlation between the TX channel phase coefficients and the Doppler position, ensuring that the Doppler response points corresponding to the same real target for each TX channel can be accurately located based on the non-uniform phase offset parameter. This provides a clear positional basis for multi-channel signal accumulation, avoiding accumulation failures caused by Doppler position matching deviations. The threshold of 4 is set based on the multi-channel consistency verification logic formed by the configuration of 6 TX channels. Only when more than 2 / 3 of the TX channels detect peak signals (accumulated value > 4) is it considered a valid target point. This design effectively weakens the signal fluctuations caused by angular energy differences in a single TX channel, eliminates spurious peaks triggered by occasional clutter, ensures that the final output target point is a reliable response of the real target, and improves the accuracy of target detection.

[0050] Example: Taking the morning rush hour scenario on urban roads as the application scenario (including stationary obstacles, pedestrians traveling at 0-3 km / h, non-motorized vehicles traveling at 10-15 km / h, and urban vehicles traveling at 30-50 km / h), the target detection performance of the signal processing method provided by this invention is verified. The specific implementation process is as follows: I. Hardware and Waveform Parameter Configuration The DDMIMO 6TX+8RX millimeter-wave radar is used, with a transmission frame period of 50ms. Each transmission frame contains 256 chirs. When each chirp is transmitted, 6 transmission channels (TX0 to TX5) are activated simultaneously. The phase shift of each TX channel changes according to a fixed pattern under each chirp. The core waveform parameters are as follows: Center frequency: 76.5GHz; bandwidth: 750MHz; sampling rate: 40MHz; number of sampling points: 1024; Chirp period: 30.6 μs; number of chirp cycles: 256; The phase changes of TX0 to TX5 are 0°, 50.625°, 123.75°, 219.375°, 281.25°, and 320.625° respectively.

[0051] II. Association Rules between TX Channel and Doppler Position

[0052] Assuming the Doppler position of TX0 corresponding to the target point is N (the number of Doppler-dimensional FFT points is 256, and the index range is 0 to 255), the modulo rules for the Doppler positions of the other TX channels are as follows: TX1: (N+36)%256; TX2: (N+88)%256; TX3: (N+156)%256; TX4: (N+200)%256; TX5: (N+228)%256; This rule is based on the Doppler domain mapping result of non-equal phase offset parameters, ensuring that the Doppler positions of each TX channel are discretely distributed within the target high-frequency range.

[0053] III. Signal Processing Flow

[0054] Echo reception and preprocessing: The target reflected echo is received synchronously through 8 RX channels. 1024 points are collected at a sampling rate of 40MHz to obtain the time-domain echo signal of each RX channel. Two-dimensional signal matrix generation: Perform one-dimensional FFT (range dimension) and two-dimensional FFT (Doppler dimension) processing on the time-domain signal in sequence, and calculate the signal amplitude value to obtain the range-Doppler two-dimensional signal matrix; Noise floor estimation and filtering: The OS method is used to estimate the noise floor in the Doppler direction of each distance cell in the two-dimensional signal matrix; the noise floor estimation results are then filtered by moving average in the distance dimension (window size is 4 consecutive distance points) to obtain a stable noise floor value. Peak marker spectrum generation: Based on the filtered noise floor value, a CFAR threshold (signal-to-noise ratio 3.5dB) is set. The signal points under each distance cell are compared with the threshold. Peak points are selected and marked as 1, and non-peak points are marked as 0, generating a two-dimensional peak marker spectrum. Accumulation verification and target output: In the peak marker map, the marker values ​​corresponding to each TX channel are selected and accumulated according to the Doppler position rules; it is determined whether the accumulated value is greater than 4. If it is greater than 4, it is determined to be a valid target point. Finally, a two-dimensional target point map containing target distance and velocity information is output.

[0055] IV. Implementation Results

[0056] In the scenario of morning rush hour on urban roads, the detection performance of this embodiment is as follows: Multi-target Doppler channel overlap rate: less than 5%; Detection rate for small targets (pedestrians with RCS ≤ 0.1㎡): 92%; Background noise filtering rate: 90%; This embodiment verifies that the non-equal phase modulation and optimized demodulation process provided by the present invention can effectively solve the problems of multi-target overlap and small target omission in traditional solutions, and adapt to the detection needs of complex vehicle scenarios.

[0057] Figure 6A system block diagram of a signal processing device based on DDMIMO radar according to an embodiment of the present invention is shown. As shown, the signal processing device 600 may include an internal communication bus 601, a processor 602, a read-only memory (ROM) 603, a random access memory (RAM) 604, and a communication port 605. When applied to a personal computer, the signal processing device 600 may also include a hard disk 606. The internal communication bus 601 enables data communication between the components of the signal processing device 600. The processor 602 can perform judgments and issue prompts. In some embodiments, the processor 602 may consist of one or more processors. The communication port 605 enables data communication between the signal processing device 600 and external devices. In some embodiments, the signal processing device 600 can send and receive information and data from a network through the communication port 605. The signal processing device 600 may also include different types of program storage units and data storage units, such as a hard disk 606, a read-only memory (ROM) 603, and a random access memory (RAM) 604, capable of storing various data files used for computer processing and / or communication, as well as possible program instructions executed by the processor 602. The processor 602 executes these instructions to implement the main part of the method. The results processed by the processor 602 are transmitted to the user equipment via the communication port 605 and displayed on the user interface.

[0058] The above-described signal processing method can be implemented as a computer program, stored in the hard disk 606, and loaded into the processor 602 for execution to implement the signal processing method of this application.

[0059] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any of the aforementioned signal processing methods.

[0060] The specific implementation methods and technical effects of the signal processing device and computer-readable storage medium based on DDMIMO radar can be found in the above-described embodiment of the signal processing method based on DDMIMO radar provided by the present invention, and will not be repeated here.

[0061] The present invention provides a signal processing method, signal processing device, and computer-readable storage medium based on DDMIMO radar, which have the following beneficial effects: 1. The non-uniform phase modulation design adapts to the target Doppler distribution characteristics in outdoor driving scenarios, making each TX channel non-uniformly discrete in the high-frequency range of the target, which greatly reduces the probability of multiple target channels overlapping, avoids demodulation errors and angle phase interference, and improves the accuracy of angle estimation. 2. The OS method is used to estimate the noise floor, combined with distance-dimensional moving average filtering, which accurately reflects the background noise level, avoids the influence of TX channel energy differences and clutter on the noise floor estimation, reduces the signal-to-noise ratio of the CFAR threshold to 3.5dB, and effectively filters out most noise points; 3. The cumulative verification logic based on peak marker spectrum replaces the traditional coarse judgment of "maximum cumulative value". Even if the echo energy of small targets is low in some TX channels, as long as peak signals are detected in most channels (cumulative value > 4), it can still be accurately identified, solving the problem of small targets being submerged by background noise or clutter.

[0062] Those skilled in the art will further appreciate that the various illustrative logic blocks, modules, circuits, and algorithm steps described in conjunction with the embodiments disclosed herein can be implemented as electronic hardware, computer software, or a combination of both. To clearly illustrate this interchangeability between hardware and software, the various illustrative components, blocks, modules, circuits, and steps are described above in a generalized manner in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the specific application and the design constraints imposed on the overall system. Those skilled in the art may implement the described functionality in different ways for each specific application, but such implementation decisions should not be construed as departing from the scope of the invention.

[0063] The various illustrative logic modules and circuits described in conjunction with the embodiments disclosed herein may be implemented or performed using a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. The general-purpose processor may be a microprocessor, but in alternatives, it may be any conventional processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors cooperating with a DSP core, or any other such configuration.

[0064] The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of both. The software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to a processor such that the processor can read and write information to / from the storage medium. In an alternative, the storage medium may be integrated into the processor. The processor and storage medium may reside in an ASIC. The ASIC may reside in a user terminal. In an alternative, the processor and storage medium may reside as discrete components in the user terminal.

[0065] In one or more exemplary embodiments, the described functionality may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software as a computer program product, the functionality may be stored or transmitted as one or more instructions or code on or through a computer-readable medium. A computer-readable medium includes both computer storage media and communication media, encompassing any medium that facilitates the transfer of a computer program from one location to another. A storage medium may be any available medium accessible to a computer. By way of example and not limitation, such a computer-readable medium may include RAM, ROM, EEPROM, CD-ROM or other optical disc storage, disk storage or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and is accessible to a computer. Any connection is also legitimately referred to as a computer-readable medium. For example, if the software is transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, twisted pair, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of a medium. As used in this article, disk and disc include compact discs (CDs), laser discs, optical discs, digital multi-purpose discs (DVDs), floppy disks, and Blu-ray discs. Disks typically reproduce data magnetically, while discs reproduce data optically using lasers. Combinations of these should also be included within the scope of computer-readable media.

[0066] It will be apparent to those skilled in the art that various modifications and variations can be made to the exemplary embodiments described above without departing from the spirit and scope of the invention. Therefore, it is intended that this invention cover modifications and variations falling within the scope of the appended claims and their equivalents.

Claims

1. A signal processing method based on DDMIMO radar, the signal processing method comprising a signal modulation method and a signal demodulation method; The signal modulation method includes the following steps: The TX channels of the DDMIMO radar are controlled to perform signal modulation using non-uniform phase offset parameters; The signal demodulation method includes the following steps: The radar echo signal is processed sequentially by one-dimensional FFT and two-dimensional FFT, and the amplitude value is calculated to obtain the range and Doppler two-dimensional signal matrix. Based on the two-dimensional signal matrix, the noise floor level in the Doppler dimension is estimated point by point. The noise floor estimation results at each distance point are then filtered in the distance dimension to obtain the noise floor value. A CFAR threshold is set based on the noise floor value, and peak target points are filtered out based on the CFAR threshold to generate a peak marker map. Based on the non-uniform phase offset parameters of each TX channel, corresponding Doppler points are selected at each position in the peak marker spectrum for accumulation to generate an accumulation spectrum; The accumulated value at each position in the accumulated map is compared with a set threshold to determine the valid target point; The non-equal arithmetic phase offset parameter is obtained in the following manner: Determine the high-frequency Doppler location range of the target in outdoor driving scenarios; Determine the non-uniform phase offset parameters of the TX channel of the DDMIMO radar.

2. The signal processing method as described in claim 1, characterized in that, Determining the Doppler position interval in obtaining the non-arithmetic phase offset parameters includes the following steps: Determine the basic parameters of the DDMIMO radar, including carrier frequency, wavelength, pulse repetition period, number of chirps per transmitted frame, and number of doppler-dimensional FFT points; Determine the velocity resolution, and based on the velocity resolution, determine the actual velocity range for different target types, and associate the actual velocity range with the Doppler position interval; Determining the non-uniform phase offset parameter in the signal modulation method includes the following steps: Set constraints; Based on the aforementioned constraints, non-uniform phase offset parameters for each of the TX channels are determined by combining the Doppler position intervals where the target frequently occurs.

3. The signal processing method as described in claim 2, characterized in that, The constraints include: The Doppler positions of each TX channel are non-arithmetically distributed within the target high-frequency range, and the difference between the Doppler indices of any two TX channels is greater than a set threshold. The phase offset parameters must meet the accuracy requirements of the hardware phase shifter of the DDMIMO radar, and the total phase range must cover 360°.

4. The signal processing method as described in claim 3, characterized in that, The DDMIMO radar includes six TX channels, TX0 to TX5, with phase offset parameters corresponding to each TX channel as follows: 0°, 50.625°, 123.75°, 219.375°, 281.25° and 320.625°.

5. The signal processing method as described in claim 4, characterized in that, The OS method is used to estimate the noise floor level of the Doppler dimension.

6. The signal processing method as described in claim 5, characterized in that, The distance dimension filtering process uses a moving average filter with a filter window size of 3 to 5 consecutive distance points.

7. The signal processing method as described in claim 5, characterized in that, The signal-to-noise ratio of the CFAR threshold is 3.5dB, and peak filtering is performed using a unit-averaged constant false alarm rate detector.

8. The signal processing method as described in claim 5, characterized in that, The threshold value is set to 4, and a point is considered a valid target point when the accumulated value is greater than 4. The Doppler position interval pattern of each TX channel satisfies the formula Idop_m = Idop_total + Nfft_dop / Q, where Idop_total is the target real Doppler index. is the phase coefficient of the TX channel, Nfft_dop is the number of FFT points in the Doppler dimension, and Q is the number of equal parts in the Doppler dimension.

9. A signal processing device based on DDMIMO radar, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the signal processing method as claimed in any one of claims 1-8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the signal processing method as described in any one of claims 1-8.

Images