WO2021189206A1 - 雷达信号处理方法和雷达信号处理装置 - Google Patents

雷达信号处理方法和雷达信号处理装置 Download PDF

Info

Publication number
WO2021189206A1
WO2021189206A1 PCT/CN2020/080735 CN2020080735W WO2021189206A1 WO 2021189206 A1 WO2021189206 A1 WO 2021189206A1 CN 2020080735 W CN2020080735 W CN 2020080735W WO 2021189206 A1 WO2021189206 A1 WO 2021189206A1
Authority
WO
WIPO (PCT)
Prior art keywords
target
prior information
prediction value
signals
signal
Prior art date
Application number
PCT/CN2020/080735
Other languages
English (en)
French (fr)
Inventor
李斌
陶艺文
魏子平
李德建
李晓波
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Priority to CN202080004851.7A priority Critical patent/CN112689773B/zh
Priority to PCT/CN2020/080735 priority patent/WO2021189206A1/zh
Publication of WO2021189206A1 publication Critical patent/WO2021189206A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/02—Systems using reflection of radio waves, e.g. primary radar systems; Analogous systems
    • G01S13/50—Systems of measurement based on relative movement of target
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/66—Radar-tracking systems; Analogous systems
    • G01S13/72—Radar-tracking systems; Analogous systems for two-dimensional tracking, e.g. combination of angle and range tracking, track-while-scan radar

Definitions

  • This application relates to the field of radar signal processing, and more specifically, to a radar signal processing method and a radar signal processing device.
  • Vehicle-mounted millimeter-wave radar is one of the indispensable sensors in the field of autonomous driving because of its all-weather detection capability.
  • Vehicle-mounted millimeter wave radars usually use multiple input multiple output (MIMO) technology to obtain a large array aperture when the number of antennas is limited, thereby obtaining high angular resolution.
  • MIMO multiple input multiple output
  • Time division multiplexing (TDM)-MIMO transmission mode has the advantages of simple hardware implementation and low mutual coupling effect, and has become an important research direction of vehicle-mounted millimeter wave radar.
  • frequency modulated continuous wave FMWC is a commonly used transmission waveform in the TDM-MIMO transmission mode.
  • the TDM-MIMO radar of FMWC multiple virtual receiving antennas are formed between each transmitting (TX) antenna and each receiving (receive, RX) antenna.
  • TX transmitting
  • RX receiving
  • the range doppler map RD
  • the existing RD map combining method is mainly based on blind coherent combining, which has high complexity.
  • RD map combining uses non-coherent combining. Therefore, RD map merging can actually only achieve "semi-coherent merging".
  • the existing RD map merging method has high complexity, and the performance of target detection using the merged RD map is poor.
  • the present application provides a radar signal processing method and a radar signal processing device, which can improve the target detection performance of the RD map, and improve the detection signal-to-noise ratio and detection accuracy.
  • a radar signal processing method including: obtaining multiple received RX signals, and performing range-dimensional spectrum analysis and Doppler-dimensional spectrum analysis on each RX signal of the multiple RX signals to obtain the range Doppler RD pattern; according to the prior information of target tracking, the RD pattern of each RX signal is coherently processed, and the RD pattern after coherent processing of multiple RX signals is coherently superimposed.
  • the prior information of target tracking is the multi-channel The priori information outputted by the RX signal after the target tracking processing; the target detection is performed according to the RD pattern after the coherent superposition of multiple RX signals.
  • multiple RX signals can correspond to multiple virtual RX signals of MIMO radar, which can be all virtual RX signals or part of virtual RX signals; multiple RX signals can also correspond to multiple RX signals of SIMO radar, which can be All RX signals can also be partial RX signals. Selecting some RX signals from all RX signals and combining them can make the RD map of the selected partial RX signals more accurately achieve coherent combination, and secondly, it can reduce a certain amount of calculation and improve processing efficiency.
  • the radar signal processing method of the first aspect based on the target tracking prior information, performs coherent processing on the RD map of multiple RX signals and then performs coherent superposition, which can increase the combined gain of RX signals, that is, coherent processing and coherent superposition can make superimposed reception
  • the signal phase conforms to the coherence condition as much as possible, so that the strength of the received signal after superposition is significantly enhanced, which can improve the target detection performance of the RD map, and improve the detection signal-to-noise ratio and detection accuracy.
  • performing coherent processing on the RD pattern of each RX signal according to the target tracking prior information may include: according to the target tracking prior information, in the current frame of each RX signal At least one target is predicted on the RD graph, and the RD cell set of each target in at least one target is obtained; the RD cell set for each target is tracked according to the target in the prior information.
  • the prediction value of the radial velocity or the angle of the frame is phase-compensated for the RD cell set in the RD diagram of each RX signal.
  • the target RD cell set is based on the target tracking prior information, and the phase compensation of the RD map is also based on the target tracking prior information, which can effectively increase the combination gain and improve the target detection performance.
  • the RD cell set in the RD graph of each RX signal is phased according to the radial velocity prediction value or the angle prediction value of each target in the target tracking prior information in the current frame. Compensation may include: performing phase compensation on the RD cell set in the RD graph of each RX signal according to the radial velocity prediction value and/or angle prediction value of each target in the current frame in the target tracking prior information.
  • the target tracking prior information at least one target is predicted on the RD graph in the current frame of each RX signal, and the RD unit of each target in the at least one target is obtained.
  • the set may include: locating the RD cell set of each target on the RD map according to the distance prediction value and the radial velocity prediction value of each target in the current frame in the target tracking prior information.
  • the RD cell set of the target is located based on the distance prediction value and the radial velocity prediction value.
  • the RD cell set corresponding to the same target can be found on multiple RD maps, and for TDM -The different TX time of each TX antenna in the MIMO radar can compensate for the phase change of each RX signal caused by the target speed, which is beneficial to complete the coherent combination of the RD map between the RX channels, thereby improving the combination gain.
  • the distance prediction value and the radial velocity prediction value output by the target tracking are more accurate than the distance value and speed value of the RD map.
  • the RD cell set is also more accurate, which makes the gain of the coherent combination of the RD map higher.
  • the distance expansion factor ⁇ and the velocity expansion ⁇ are used to Select the possible range of target T on the RD map: distance range [(1- ⁇ )r k ,(1+ ⁇ )r k ], speed range [(1- ⁇ )v k ,(1+ ⁇ )v k ].
  • using the expansion factor to obtain the distance range and speed range is beneficial to more accurately and efficiently locate the RD area where the target is located.
  • the target tracking prior information at least one target is predicted on the RD graph in the current frame of each RX signal, and the RD unit of each target in the at least one target is obtained.
  • the set may include: locating the RD cell set of each target on the RD map according to the contour prediction value of each target in the current frame in the target tracking prior information.
  • the contour prediction value of the target in the current frame includes prediction information such as the position, orientation, and speed of the target, the RD cell set of the target is located based on the contour prediction value, and the target can be accurately mapped to the current frame.
  • the RD cell area is located, and the phase change of the RX signal caused by the speed of the target can be compensated according to the radial velocity corresponding to the RD cell set, which is beneficial to complete the coherent combination of the RD map between the RX channels, thereby improving the combination gain.
  • the contour prediction value of the target may include prediction information such as the position, orientation, and speed of the target.
  • the RD cell in the RD map of each RX signal is calculated according to the radial velocity prediction value or the angle prediction value of each target in the target tracking prior information in the current frame.
  • Collecting phase compensation may include: performing speed-dependent phase compensation on the RD unit set in the RD diagram of each RX signal according to the predicted value of the radial velocity of each target in the current frame in the target tracking prior information; Or, according to the angle prediction value of each target in the target tracking prior information in the current frame, perform angle-dependent phase compensation on the RD unit set in the RD diagram of each RX signal.
  • the coherent signal is enhanced based on the angle of the predicted target (that is, angle-dependent phase compensation) so that the RD map between more RX channels can be coherently combined, and the combined gain can be improved, thereby Improve the target detection performance of the RD map; and/or compensate for the phase change of the RX signal caused by the target's speed (that is, speed-dependent phase compensation) for the different TX time of each TX antenna appearing in the TDM-MIMO radar, so as to make more
  • the RD map between multiple RX channels can complete coherent combination, which can increase the combination gain, thereby improving the target detection performance of the RD map, such as detection signal-to-noise ratio and detection accuracy.
  • phase compensation of this possible implementation may include: performing a set of RD units in the RD diagram of each RX signal according to the prediction value of the radial velocity of each target in the target tracking prior information in the current frame. Speed-dependent phase compensation; and/or perform angle-dependent phase compensation on the RD unit set in the RD diagram of each RX signal according to the angle prediction value of each target in the target tracking prior information in the current frame.
  • the coherent superposition of the RD map after the coherent processing of the multiple RX signals may include: the RD map after the coherent processing of the multiple RX signals corresponds to the same one
  • the RD cell set of the target is added with complex values. This possible implementation manner performs complex value addition on the set of RD cells corresponding to the same target in the RD map after coherent processing to complete coherent combination, which can improve the combination gain.
  • the target tracking prior information output by the CHT algorithm may include the distance prediction value, velocity prediction value, angle prediction value, and convexity of each target in the current frame of at least one target.
  • Package forecast value The convex hull gate used in this possible implementation mode can make real-time adaptive adjustment of the target, and realize efficient target clustering and target association without manual parameter input.
  • the target association is performed by means of merging.
  • This possible implementation method uses mean merging for target association, compared to the global nearest neighbor (GNN) algorithm, strongest neighbor (SN) algorithm, and probabilistic data association that are commonly used in target association. association, PDA) algorithm, etc., the association performance has been greatly improved.
  • GNN global nearest neighbor
  • SN strongest neighbor
  • PDA probabilistic data association
  • a radar signal processing device including: an acquisition unit for acquiring multiple received RX signals, and performing range-dimensional spectrum analysis and Doppler-dimensional analysis on each of the multiple RX signals. Spectrum analysis to obtain the range Doppler RD map; the processing unit is used to perform coherent processing on the RD map of each RX signal obtained by the acquisition unit according to the prior information of target tracking, and the RD after coherent processing of the multiple RX signals The image is coherently superimposed.
  • the target tracking prior information is the prior information output after the target tracking processing of the multiple RX signals; the detection unit is used to perform the target based on the RD map after the coherent superposition of the multiple RX signals obtained by the processing unit Detection.
  • the processing unit is specifically configured to: predict at least one target on the RD map in the current frame of each RX signal according to the target tracking prior information, to obtain at least one target
  • the RD unit in the ensemble performs phase compensation.
  • the processing unit can be specifically used to: according to the target tracking a priori information, the radial velocity prediction value and/or the angle prediction value of each target in the current frame, for each channel of RX signal
  • the RD cells in the RD diagram are assembled for phase compensation.
  • the processing unit is specifically configured to: locate the distance prediction value and the radial velocity prediction value of each target in the current frame according to the target tracking prior information The set of RD units for each target.
  • the processing unit is specifically configured to locate the RD unit of each target on the RD map according to the contour prediction value of each target in the target tracking prior information. gather.
  • the processing unit is specifically configured to: according to the prediction value of the radial velocity of each target in the current frame in the target tracking a priori information, to determine the value in the RD diagram of each RX signal RD unit set for speed-dependent phase compensation; or based on the angle prediction value of each target in the target tracking prior information in the current frame, the RD unit set in the RD diagram of each RX signal is set for angle-dependent phase compensate.
  • the processing unit can be specifically used to: according to the target tracking a priori information of the radial velocity prediction value of each target in the current frame, for the RD unit in the RD diagram of each RX signal Set, perform speed-dependent phase compensation; and/or perform an angle-dependent phase based on the RD unit set in the RD diagram of each RX signal according to the angle prediction value of each target in the target tracking prior information compensate.
  • the processing unit is specifically configured to: perform complex value addition on the set of RD units corresponding to the same target in the RD graph after the coherent processing of the multiple RX signals.
  • the device further includes a tracking unit, and the target tracking prior information is calculated by the tracking unit through a convex hull tracking algorithm.
  • the target tracking prior information includes a distance prediction value, a velocity prediction value, an angle prediction value, and a convex hull prediction value of each target in at least one target in the current frame.
  • the device further includes a clustering association unit, configured to: obtain the predicted convex hull of the first target in the current frame or the target's predicted value from the convex hull prediction value in the target tracking prior information Predict the contour, cluster the points in the predicted convex hull or prediction wheel to obtain one or more cluster points; associate one or more cluster points that fall into the predicted convex hull or predicted contour with the first target .
  • a clustering association unit configured to: obtain the predicted convex hull of the first target in the current frame or the target's predicted value from the convex hull prediction value in the target tracking prior information Predict the contour, cluster the points in the predicted convex hull or prediction wheel to obtain one or more cluster points; associate one or more cluster points that fall into the predicted convex hull or predicted contour with the first target .
  • the signal processing device is applied to a vehicle-mounted radar.
  • a radar signal processing device including a processor, which is coupled with a memory, and can be used to execute the method in the first aspect or any one of the possible implementation manners of the first aspect.
  • the signal processing device further includes a memory.
  • the signal processing device further includes a communication interface, and the processor is coupled with the communication interface.
  • the signal processing apparatus is a network device.
  • the communication interface may be a transceiver, or an input/output interface.
  • the signal processing device is a chip or a chip system.
  • the communication interface may be an input/output interface, interface circuit, output circuit, input circuit, pin or related circuit on the chip or chip system.
  • the processor can also be embodied as a processing circuit or a logic circuit.
  • a processor including: an input circuit, an output circuit, and a processing circuit.
  • the processing circuit is configured to receive signals through the input circuit and transmit signals through the output circuit, so that the first aspect and the method in any one of the possible implementation manners of the first aspect are implemented.
  • the above-mentioned processor may be a chip
  • the input circuit may be an input pin
  • the output circuit may be an output pin
  • the processing circuit may be a transistor, a gate circuit, a flip-flop, and various logic circuits.
  • the input signal received by the input circuit may be received and input by, for example, but not limited to, a receiver
  • the signal output by the output circuit may be, for example, but not limited to, output to the transmitter and transmitted by the transmitter
  • the input circuit and output The circuit can be the same circuit, which is used as an input circuit and an output circuit at different times.
  • a processing device including a processor and a memory.
  • the processor is configured to read instructions stored in the memory to execute the first aspect and the method in any one of the possible implementation manners of the first aspect.
  • processors there are one or more processors and one or more memories.
  • the memory may be integrated with the processor, or the memory and the processor may be provided separately.
  • ROM Read only memory
  • the memory can be non-transitory (non-transitory) memory, for example, only set in different On the chip.
  • sending instruction information may be a process of outputting instruction information from the processor
  • receiving capability information may be a process of receiving input capability information by the processor.
  • the processed output data may be output to the transmitter, and the input data received by the processor may come from the receiver.
  • the transmitter and receiver can be collectively referred to as a transceiver.
  • the processor in the above fifth aspect may be a chip, and the processor may be implemented by hardware or software.
  • the processor When implemented by hardware, the processor may be a logic circuit, an integrated circuit, etc.; when implemented by software
  • the processor may be a general-purpose processor, which is implemented by reading software codes stored in the memory.
  • the memory may be integrated in the processor, may be located outside the processor, and exist independently.
  • a computer program product includes: a computer program (also called code, or instruction), which when the computer program is executed, causes the computer to execute the first aspect and any one of the first aspect.
  • a computer program also called code, or instruction
  • a computer-readable storage medium stores a computer program (also called code, or instruction) when it runs on a computer, so that the computer executes the first aspect and the first aspect described above.
  • a computer program also called code, or instruction
  • the method in any possible implementation.
  • the second aspect to the seventh aspect and the corresponding possible implementation manners can be applied to the vehicle-mounted radar.
  • a radar including a receiver and a processor, the receiver is configured to receive multiple received RX signals, and the processor is configured to perform the first aspect and any one of the possible possibilities of the first aspect according to the multiple RX signals. The method in the implementation mode.
  • V2X vehicle to everything
  • LTE-V vehicle-to-vehicle
  • V2X vehicle-to-everything
  • Figure 1 is a schematic diagram of radar signal processing.
  • Fig. 2 is a schematic diagram of radar signal processing according to an embodiment provided by the present application.
  • FIG. 3 is a schematic flowchart of a radar signal processing method according to an embodiment of the present application.
  • Fig. 4 is a schematic diagram of a TDM-MIMO frame according to an embodiment provided by the present application.
  • FIG. 5 is a schematic diagram of the distance prediction of the target T in the current frame according to an embodiment provided by the present application.
  • Fig. 6 is a schematic diagram of target distribution according to an embodiment provided by the present application.
  • FIG. 7 is a comparison diagram of the coherent combination detection result of the embodiment provided by the present application and the existing incoherent combination detection result.
  • FIG. 8 is a schematic diagram of the result of target clustering and target association according to an embodiment provided by the present application.
  • FIG. 9 is a schematic diagram of the result of target tracking according to an embodiment provided by the present application.
  • Fig. 10 is a schematic block diagram of a radar signal processing device according to an embodiment of the present application.
  • FIG. 11 is a schematic block diagram of a radar signal processing device according to another embodiment provided by the present application.
  • the embodiments of the present application are not only applicable to MIMO radars, but also applicable to single input multiple output (SIMO) radars.
  • MIMO radar it has multiple TX antennas and multiple RX antennas (for example, M TX antennas and N RX antennas, M and N are positive integers), which can form multiple (M ⁇ N) virtual RX channels, Obtain multiple virtual RX signals.
  • SIMO radar with one TX antenna and multiple RX antennas, multiple RX channels can be formed to obtain multiple RX signals.
  • virtual RX channels or RX channels are collectively referred to as receiving channels (RX channels), and virtual RX signals or RX signals are collectively referred to as RX signals.
  • the superposition of multiple RD maps corresponding to the multiple RX signals received by the TDM-MIMO radar is also called "RD map combination", and is usually “semi-coherent combination”.
  • the existing RD map merges into blind coherent merge, which belongs to coherent merge, that is, when the angle information of the target is unknown, the multiple RX signals are phase compensated at each angle. That is, it is combined once at each angle, and the superposition/combination method of this signal has a high computational complexity.
  • the RD map combination used can only be incoherent combination.
  • the existing RD map merging fails to effectively use the prior information of the target's position, distance, angle, speed, contour and other information output by target tracking, and the calculation complexity is high, the merging gain is small, and the merged RD map is used for processing.
  • the target detection performance is poor, that is, the detection signal-to-noise ratio is low, and the detection accuracy is low.
  • radar signal processing can also include target clustering, target association, and target tracking based on received signals.
  • target clustering fails to use prior information such as the position, distance, angle, speed, and contour of the target that has been output by target tracking, which leads to a single cluster when the target is clustered into multiple clusters
  • the target is split into multiple targets, multiple targets are clustered into one target, or the target is partially occluded, etc., resulting in poor target clustering effect and low efficiency.
  • the clustering radius cannot be applied to the coexistence of large targets (such as buses, etc.) and small targets (such as people).
  • FIG. 1 is a schematic diagram of radar signal processing. As shown in Figure 1, first obtain the signal data. After acquiring the data, perform corresponding processing on the data, for example, perform a two-dimensional FFT (2D-FFT) to obtain an RD map. In the existing solution, the aforementioned "semi-coherent merging" is performed based on the RD map, and then the target detection is performed based on the merged RD map. After target detection, target clustering and target association and target tracking can be performed in sequence.
  • 2D-FFT two-dimensional FFT
  • target clustering, target association, and target tracking are performed by separate modules; in other solutions, target clustering and target association can be performed by the same module, or target association and target tracking It may be executed by the same module, or target clustering, target association, and target tracking may be executed by the same module; this embodiment of the present application does not limit this.
  • FIG. 2 is a schematic diagram of radar signal processing according to an embodiment provided by the present application.
  • the embodiment of the present application can perform target detection based on the output of target tracking, for example, target tracking prior information.
  • the embodiment of the present application may also perform target clustering based on the target tracking prior information output by the target tracking. That is, the corrected target's position, distance, angle, speed, contour and other information obtained by target tracking are used to provide reference information for signal processing.
  • FIG. 3 is a schematic flowchart of a radar signal processing method 300 according to an embodiment of the present application. As shown in FIG. 3, the method 300 may include the following steps.
  • S310 Obtain multiple RX signals, and perform range-dimensional spectrum analysis and Doppler-dimensional spectrum analysis on each RX signal of the multiple RX signals to obtain a range Doppler map RD map.
  • S320 Perform coherent processing on the RD map of each RX signal according to the prior information of target tracking, and perform coherent superposition on the RD map after the coherent processing of multiple RX signals.
  • the prior information of target tracking is to target multiple RX signals. Priori information output after tracking processing.
  • S330 Perform target detection according to the RD map after coherent superposition of multiple RX signals.
  • the radar signal processing method of the embodiment of the present application performs coherent processing on the RD map of multiple RX signals based on target tracking prior information and then performs coherent superposition, which can increase the combined gain of RX signals, that is, coherent processing and coherent superposition can make the superposition
  • coherent processing and coherent superposition can make the superposition
  • the phase of the received signal meets the coherence condition as much as possible, so that the strength of the received signal after superposition is significantly enhanced, which can improve the target detection performance of the RD map, and improve the detection signal-to-noise ratio and detection accuracy.
  • the multiple RX signals obtained by S310 may be M ⁇ N virtual RX signals corresponding to M ⁇ N virtual RX channels. In some other embodiments of the present application, the multiple RX signals obtained by S310 may also be part of the virtual RX signals among the M ⁇ N virtual RX signals.
  • the RD map participating in the coherent superposition can be selected according to the quality of the virtual RX signal received by the virtual RX channel and/or the correlation between the virtual RX channels. For example, multiple virtual RX signals with relatively high signal quality among M ⁇ N RX signals may be selected; RX signals corresponding to multiple virtual RX channels with high correlation may also be selected.
  • the embodiment of the present application does not limit how to select and obtain multiple RX signals.
  • each TX antenna uses TDM-MIMO to transmit FMWC signals.
  • TX antenna 1 to TX antenna 12 are represented as TX1-TX12, and TX1-TX12 transmit chirp pulse signals (chirp1, chirp2,..., chirp12) in turn, forming a TDM-MIMO transmission cycle .
  • Fig. 4 is a schematic diagram of a TDM-MIMO frame according to an embodiment provided by the present application.
  • a frame includes one or more TDM and MIMO transmission cycles.
  • the RD map can be obtained in S310 through the following methods.
  • Each RX channel performs a two-dimensional FFT (2D-FFT) on the received RX signal of a frame (for example, the current frame), that is, two-bit spectrum analysis, where the RX signal may be an FMWC signal, including multiple chirp pulse signals.
  • 2D-FFT can be performed by distance-dimensional FFT (also known as distance-dimensional spectrum analysis, fast FFT, ranging FFT, fast-FFT) and Doppler FFT (also known as Doppler FFT, slow FFT, speed measurement FFT, namely slow-FFT).
  • RDM i perform 2D-FFT on the RX signal to obtain the RD map of the RX channel, denoted as RDM i , where i is the complete set or subset of the set ⁇ 1,2,3,...,M ⁇ N ⁇ , that is, calculate M ⁇ N RD map of all or part of the RX channels in each RX channel.
  • S320 performs coherent processing on the RD map of each RX signal according to the target tracking prior information, which may include: according to the target tracking prior information, on the RDmap in the current frame of each RX signal Predict at least one target to obtain the RD cell set of each target in at least one target; for each target RD cell set, follow the target tracking prior information in the radial direction of each target in the current frame Speed prediction value or angle prediction value, phase compensation is performed on the RD cell set in the RD diagram of each RX signal.
  • the RD cell set may include one or more RD cells.
  • the RD cell set for determining the target is based on the target tracking prior information, and the phase compensation of the RD map is also based on the target tracking prior information, which can effectively increase the combination gain and improve the target detection performance.
  • performing phase compensation on the RD cell set in the RD diagram of each RX signal may include: according to the target Tracking the radial velocity prediction value and/or angle prediction value of each target in the prior information in the current frame, and performing phase compensation on the RD cell set in the RD graph of each RX signal.
  • At least one target is predicted on the RD map in the current frame of each RX signal to obtain the RD cell set of each target in the at least one target, which may include : According to the distance prediction value and radial velocity prediction value of each target in the current frame in the target tracking prior information, locate the RD cell set of each target on the RD map. In this embodiment, the RD cell set of the target is located based on the predicted value of the distance and the predicted value of the radial velocity.
  • the RD cell set corresponding to the same target can be found on multiple RD maps, and for TDM-MIMO
  • the different TX time of each TX antenna in the radar can compensate for the phase change of each RX signal caused by the target speed, which is beneficial to complete the coherent combination of the RD map between the RX channels, thereby improving the combination gain.
  • the RD maps of all M ⁇ N RX channels are selected for merging.
  • the distance prediction value r k is the unambiguous distance value of the target T estimated by the k-th RX channel;
  • the radial velocity prediction value v k is the unambiguous speed value of the target T estimated by the k-th RX channel.
  • the corresponding target tracking module will estimate the target T for the current frame based on the target tracking prior information of the target T and the target motion law The distance prediction value r k and the radial velocity prediction value v k .
  • the target tracking prior information may include the state information (or the predicted value in the current frame) of the target T estimated after the previous frame tracking and filtering, including one or more of the following: a distance predicted value or a spatial coordinate system X coordinate, Y coordinate, Z coordinate; velocity prediction value, including radial velocity prediction value, velocity vector prediction value, etc.; angle prediction value, including azimuth angle prediction value and pitch angle prediction value; and if the target tracking algorithm uses convexity which will be described in detail below
  • the convex hull prediction value can be obtained in the packet tracking algorithm.
  • the target tracking algorithm (such as the extended Kalman filter algorithm) is used to predict the distance and radial velocity of the target T at the current moment according to the movement law of the target.
  • a priori model based on the target movement is used to predict the target position in one step.
  • a priori model of target motion can be a constant velocity (CV) model, a constant turn rate and velocity (CTRV) model, a constant turn rate and acceleration (Constant Turn Rate and Acceleration, CTRA) model.
  • CV constant velocity
  • CTRV constant turn rate and velocity
  • CTRA Constant Turn Rate and Acceleration
  • the target tracking algorithm is based on the CV model, and the predicted value data of the target T in the current frame can be obtained according to the predicted value data of the target T in the previous frame.
  • FIG. 5 is a schematic diagram of the distance prediction of the target T in the current frame according to an embodiment provided by the present application.
  • T frame is the time length of one frame.
  • Other embodiments of the present application may also use other algorithms to calculate the distance prediction value r k and the radial velocity prediction value v k , which is not limited in the embodiment of the present application.
  • the velocity value of the target T displayed on the RD map (calculated from the Doppler frequency) is inaccurate. This is because TDM-MIMO causes the velocity measurement range to become smaller, so the RD map is introduced The speed value is blurred, and the distance value has a similar situation.
  • the distance prediction value and the radial velocity prediction value output by the target tracking are more accurate than the distance value and speed value of the RD map.
  • the RD cell set is also more accurate, which makes the gain of the coherent combination of the RD map higher.
  • At least one target is predicted on the RD map in the current frame of each RX signal to obtain the RD cell set of each target in the at least one target.
  • a contour prediction value can be output for the point cloud corresponding to the target T, so as to obtain the set of RD cells corresponding to the point cloud corresponding to the contour prediction value in the current frame, which is marked as ⁇ cell k ⁇ .
  • the contour prediction value can be output by the tracking algorithm in the form of convex hull calculation, or it can be estimated according to the method of target recognition and classification, but the center, orientation and other parameters of the contour prediction value are output by the tracking algorithm. Since the contour prediction value of the target in the current frame includes prediction information such as the position, orientation, and speed of the target, locating the target RD cell set based on the contour prediction value can accurately locate the target in the RD cell area corresponding to the current frame. It is also possible to compensate the phase change of the RX signal caused by the speed of the target according to the radial velocity corresponding to the RD cell set, which is beneficial to complete the coherent combination of the RD map between the RX channels, thereby increasing the combination gain.
  • the RD cell set in the RD map of each RX signal is performed according to the radial velocity prediction value and/or the angle prediction value of each target in the current frame in the target tracking prior information.
  • the phase compensation may include: according to the prediction value of the radial velocity of each target in the current frame in the target tracking prior information, perform velocity-dependent phase compensation on the RD cell set in the RD map of each RX signal; or The angle prediction value of each target in the current frame in the target tracking prior information, and the RD cell set in the RD map of each RX signal is subjected to angle-dependent phase compensation.
  • the enhancement of coherent signals (that is, angle-dependent phase compensation) based on the predicted angle of the target can enable the RD map between more RX channels to be coherently combined, and the combined gain can be increased, thereby increasing RD map's target detection performance.
  • the phase change of the RX signal caused by the speed of the target can be compensated (that is, the speed-dependent phase compensation), so that the RD map between more RX channels can be Completing the coherent combination can increase the combination gain, thereby improving the target detection performance of the RD map, such as the detection signal-to-noise ratio and detection accuracy.
  • phase compensation of this possible implementation may include: performing a set of RD units in the RD diagram of each RX signal according to the prediction value of the radial velocity of each target in the target tracking prior information in the current frame. Speed-dependent phase compensation; and/or perform angle-dependent phase compensation on the RD unit set in the RD diagram of each RX signal according to the angle prediction value of each target in the target tracking prior information in the current frame.
  • FIG. 6 is a schematic diagram of the target distribution of an embodiment provided in the present application.
  • the RD map includes a distance dimension and a speed dimension.
  • the RD map includes target 1, target 2, and target 3.
  • v k and ⁇ k respectively represent the radial velocity v k and the angle ⁇ k of the target T predicted by the target tracking in the previous frame, and are parameters used for phase compensation of the RD cell of the target.
  • Each RD cell set ⁇ cell k ⁇ performs phase compensation according to the following formula.
  • d R is the distance between adjacent RX antennas.
  • ⁇ t is the receiving delay of the RXn antenna relative to the reference virtual RX antenna, and the delay is equal to the transmission delay of the TX antenna corresponding to the RXn antenna of the TDM-MIMO radar relative to the reference TX antenna.
  • the repetition period of a chirp pulse signal is T PRF
  • the transmission order of the actual TX antenna corresponding to the virtual RXn antenna is the tth (referring to the order of the TX antenna corresponding to the virtual RXn antenna is the first)
  • ⁇ t (t -1) T PRF
  • Y received (n) is the received signal after an RD cell on the RD map of the nth RX channel has completed angle phase compensation and speed phase compensation.
  • is the amplitude signal of Y received (n), It is the initial phase of Y received (n).
  • the coherent superposition of the RD map after the coherent processing of the multiple RX signals in S320 may include: the RD map after the coherent processing of the multiple RX signals corresponds to the same target
  • the RD cell set performs complex value addition.
  • the multiple RD maps after the speed-related and angle-related phase compensation are coherently combined, also called coherent superposition, to obtain the signal to be detected corresponding to the target T
  • coherent superposition also called coherent superposition
  • the multiple RD maps after the speed-related and angle-related phase compensation are coherently combined, also called coherent superposition, to obtain the signal to be detected corresponding to the target T
  • CFAR constant false alarm ratio
  • FIG. 7 is a comparison diagram of the coherent combination detection result of the embodiment provided by the present application and the existing incoherent combination detection result.
  • the graph in the upper left graph and the contour graph in the upper right graph of FIG. 7 are the coherent merge detection results of the embodiment of the application.
  • the graph in the lower left graph and the contour graph in the lower right graph of FIG. 7 are the existing incoherent merge detection results. Comparing the received signal strength of the target in the circle, the received signal strength of the coherent combined detection is significantly stronger than that of the incoherent combined detection, that is, the signal-to-noise ratio is improved, which makes the target easier to be detected.
  • the method 300 can be regarded as a target detection process based on a priori information of target tracking.
  • radar signal processing includes not only target detection, but also target clustering, target association, and target tracking based on received signals.
  • the priori information of target tracking in each embodiment of the present application is obtained by target tracking.
  • Most existing target tracking algorithms use circular wave gates, which have low tracking efficiency, and the results of target clustering and target association based on circular wave gate algorithms are not ideal.
  • the priori information of target tracking in each embodiment of the present application may be calculated through convex hull tracking (CHT) algorithm. Compared with the traditional circular wave gate, the convex wave gate is more matched with the shape of the target, which can greatly improve the correlation performance.
  • CHT convex hull tracking
  • the target tracking prior information output by the CHT algorithm may include the distance prediction value, velocity prediction value, angle prediction value, and convex hull prediction value of each target in at least one target in the current frame.
  • the embodiment of the present application can realize dynamic tracking of multiple targets based on the target tracking prior information combined with the target movement law or target movement model.
  • Target tracking can include predicting the position of the target in the current frame; and real-time tracking of multiple dynamic targets in multiple consecutive frames, including target addition (for example, adding a new observation value to the target), target deletion (for example, Delete targets that cannot be associated with observations), target clustering, target association and target tracking, etc.
  • the filter update algorithm in the CHT algorithm may use extended Kalman filter or cubature Kalman filter (CKF).
  • the idea of the CHT algorithm is to first determine the large field of view (that is, determine a convex hull for a target), and then merge multiple clusters that fall into the convex hull according to the prediction result of the convex hull to overcome the inability to optimize the field of view and resolution at the same time Difficult problem.
  • the method 300 may further include: obtaining the predicted convex hull of the first target in the current frame or the predicted contour of the target through the convex hull predicted value in the prior information of target tracking, and predicting the convex hull or the predicted contour of the target in the current frame.
  • the points in the wheel are clustered to obtain one or more cluster points; one or more cluster points falling in the predicted convex hull or predicted contour are associated with the first target.
  • the embodiment of the present application can effectively solve the problem of discrimination such as occlusion by introducing target tracking prior information during target clustering.
  • the convex hull gate used in the embodiment of the present application can perform real-time adaptive adjustment of the target, and realize efficient target clustering and target association without manual parameter input.
  • the existing target association usually uses, for example, the global nearest neighbor (GNN) algorithm, the strongest neighbor (SN) algorithm, the probabilistic data association (PDA) algorithm, etc.
  • GNN global nearest neighbor
  • SN strongest neighbor
  • PDA probabilistic data association
  • Some embodiments of the present application adopt a mean value combination method to perform target association.
  • the specific steps of the CHT algorithm may include: a) Initialization: convex hull initialization. b) Prediction: Predict the target convex hull of the current frame based on the estimated velocity prediction value of the target of the current frame in the previous frame and the size of the initialized convex hull. c) Target association: For the points that fall into the convex hull, the association with a single target is completed based on CKF. Specifically, the convex hull calculated based on the target point cloud can be predicted, and the cluster points that fall into the convex hull can be associated by means of merging, instead of using traditional association methods (for example, GNN algorithm, SN Algorithm, PDA algorithm, etc.) for target association. d) Update: Update the convex hull of the target based on the result of target association.
  • FIG. 8 is a schematic diagram of the result of target clustering and target association according to an embodiment provided by the present application.
  • the largest frame in Figure 8 is a truck
  • the medium-sized frame is an electric vehicle
  • the small frame is a pedestrian.
  • FIG. 9 is a schematic diagram of the result of target tracking according to an embodiment provided by the present application. It can be seen from FIG. 8 and FIG. 9 that the signal processing method of the embodiment of the present application can achieve accurate clustering, correlation and tracking of large targets (such as trucks) and small targets (such as electric vehicles and pedestrians).
  • FIG. 10 is a schematic block diagram of a radar signal processing device 1000 according to an embodiment of the present application.
  • the signal processing device 1000 includes: an acquisition unit 1010, configured to acquire multiple received RX signals, and perform range-dimensional spectrum analysis and Doppler-dimensional spectrum analysis on each of the multiple RX signals.
  • the processing unit 1020 is used to perform coherent processing on the RD map of each RX signal obtained by the acquisition unit 1010 according to the target tracking prior information, and the RD after the coherent processing of the multiple RX signals
  • the graph is coherently superimposed, and the target tracking prior information is the prior information output after target tracking processing is performed on the multiple RX signals;
  • the detection unit 1030 is used to coherently superimpose the RD map of the multiple RX signals obtained by the processing unit 1020, Perform target detection.
  • the processing unit 1020 may be specifically configured to: according to the target tracking prior information, the processing unit 1020 predicts at least one target on the RD diagram in the current frame of each RX signal, and obtains each of the at least one target.
  • the RD unit set of the target; the processing unit 1020 aims at the RD unit set of each target, and calculates the RD of each RX signal according to the prediction value of the radial velocity or the angle of each target in the current frame in the target tracking prior information
  • the RD unit assembly in the figure performs phase compensation.
  • processing unit 1020 may be specifically configured to: according to the target tracking a priori information, the radial velocity prediction value and/or the angle prediction value of each target in the current frame, for the RD cell in the RD diagram of each RX signal Assemble for phase compensation.
  • the processing unit 1020 may be specifically used to locate the RD of each target on the RD map according to the predicted value of the distance and the predicted value of the radial velocity of each target in the current frame in the target tracking prior information. Unit collection.
  • the processing unit 1020 may be specifically configured to locate the RD unit set of each target on the RD map according to the contour prediction value of each target in the target tracking prior information in the current frame.
  • the processing unit 1020 may be specifically configured to: perform a set of RD units in the RD diagram of each RX signal according to the predicted value of the radial velocity of each target in the target tracking prior information in the current frame. Speed-dependent phase compensation; or the processing unit 1020 performs angle-dependent phase compensation on the RD unit set in the RD diagram of each RX signal according to the angle prediction value of each target in the target tracking prior information in the current frame. It should be understood that the processing unit 1020 may be specifically used to: perform speed-dependent velocity-dependent on the RD unit set in the RD diagram of each RX signal according to the predicted value of the radial velocity of each target in the current frame in the target tracking prior information. Phase compensation; and/or perform angle-dependent phase compensation on the RD unit set in the RD diagram of each RX signal according to the angle prediction value of each target in the current frame in the target tracking prior information.
  • the processing unit 1020 may be specifically configured to: perform complex value addition on the set of RD units corresponding to the same target in the RD graph after the coherent processing of the multiple RX signals.
  • the device 1000 further includes a tracking unit 1040, and the target tracking prior information is calculated by the tracking unit 1040 through a convex hull tracking algorithm.
  • the target tracking prior information includes a distance prediction value, a velocity prediction value, an angle prediction value, and a convex hull prediction value of each target in at least one target in the current frame.
  • the device further includes a clustering association unit 1050, which is used to obtain the predicted convex hull of the first target in the current frame or the predicted contour of the target from the convex hull predicted value in the target tracking prior information, and compare the predicted convex hull of the target
  • the points in the package or prediction wheel are clustered to obtain one or more cluster points; one or more cluster points falling in the predicted convex hull or predicted contour are associated with the first target.
  • the signal processing device 1000 is applied to a vehicle-mounted radar.
  • each unit in the signal processing device 1000 can be used to implement the corresponding operations in the above method embodiments.
  • FIG. 11 is a schematic block diagram of a signal processing apparatus 1100 according to another embodiment provided by the present application.
  • the signal processing apparatus 1100 includes: a processor 1110, the processor 1110 is coupled with a memory, the memory is used to store computer programs or instructions or and/or data, and the processor 1110 is used to execute computer programs or instructions stored in the memory. And/or data, so that the method in the above method embodiment is executed.
  • the signal processing apparatus 1100 may include a memory 1120 for performing the foregoing operations.
  • the signal processing device 1110 may further include a communication interface, and the processor is coupled with the communication interface.
  • the signal processing apparatus 1100 is a network device.
  • the communication interface may be a transceiver, or an input/output interface.
  • the signal processing device 1100 is a chip or a chip system.
  • the communication interface may be an input/output interface, interface circuit, output circuit, input circuit, pin, or related circuit on the chip or chip system.
  • the processor can also be embodied as a processing circuit or a logic circuit.
  • each module in the signal processing device 1100 may be used to implement the corresponding operations in the above method embodiment, and may correspond to each unit in the above processing device 1000.
  • the application also provides a processor, including: an input circuit, an output circuit, and a processing circuit.
  • the processing circuit is used to receive the signal through the input circuit and transmit the signal through the output circuit, so that the method in the above method embodiment is realized.
  • the above-mentioned processor may be a chip
  • the input circuit may be an input pin
  • the output circuit may be an output pin
  • the processing circuit may be a transistor, a gate circuit, a flip-flop, and various logic circuits.
  • the input signal received by the input circuit may be received and input by, for example, but not limited to, a receiver
  • the signal output by the output circuit may be, for example, but not limited to, output to the transmitter and transmitted by the transmitter
  • the input circuit and output The circuit can be the same circuit, which is used as an input circuit and an output circuit at different times.
  • sending instruction information may be a process of outputting instruction information from the processor
  • receiving capability information may be a process of receiving input capability information by the processor.
  • the processed output data may be output to the transmitter, and the input data received by the processor may come from the receiver.
  • the transmitter and receiver can be collectively referred to as a transceiver.
  • the above-mentioned processor may be a chip, and the processor may be realized by hardware or software.
  • the processor may be a logic circuit, an integrated circuit, etc.; when realized by software, the processing
  • the processor may be a general-purpose processor, which is implemented by reading software codes stored in the memory.
  • the memory may be integrated in the processor, may be located outside the processor, and exist independently.
  • the processor mentioned in the embodiments of the present application may include a central processing unit (central processing pnit, CPU), a network processor (network processor, NP), or a combination of a CPU and NP.
  • the processor may further include a hardware chip.
  • the above-mentioned hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD) or a combination thereof.
  • the above-mentioned PLD may be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof.
  • the memory mentioned in the embodiments of the present application may be a volatile memory (volatile memory) or a non-volatile memory (non-volatile memory), or may include both volatile and non-volatile memory.
  • the non-volatile memory can be read-only memory (ROM), programmable read-only memory (programmable ROM, PROM), erasable programmable read-only memory (erasable PROM, EPROM), and electrically available Erase programmable read-only memory (electrically EPROM, EEPROM), flash memory (flash memory), hard disk (HDD) or solid-state drive (SSD).
  • the volatile memory may be random access memory (RAM), which is used as an external cache.
  • RAM random access memory
  • static random access memory static random access memory
  • dynamic RAM dynamic RAM
  • DRAM dynamic random access memory
  • synchronous dynamic random access memory synchronous DRAM, SDRAM
  • double data rate synchronous dynamic random access memory double data rate SDRAM, DDRSDRAM
  • enhanced synchronous dynamic random access memory enhanced SDRAM, ESDRAM
  • synchronous connection dynamic random access memory serial DRAM, SLDRAM
  • direct rambus RAM direct rambus RAM, DR RAM
  • the processor is a general-purpose processor, DSP, ASIC, FPGA or other programmable logic device, discrete gate or transistor logic device, or discrete hardware component
  • the memory storage module
  • the present application also provides a computer program product.
  • the computer program product includes a computer program (also called code or instruction), which when the computer program is executed, causes the computer to execute the method in the above method embodiment.
  • the computer-readable storage medium stores a computer program (also called code, or instruction) when it runs on a computer, so that the computer executes the method in the above method embodiment. .
  • the present application also provides a radar, including a receiver and a processor, the receiver is configured to receive multiple received RX signals, and the processor is configured to execute the method in the foregoing method embodiment according to the multiple RX signals.
  • V2X vehicle to everything
  • LTE-vehicle LTE-V
  • V2V vehicle to vehicle
  • Internet of Vehicles Internet of Vehicles
  • MTC highway semi-automatic lane toll collection
  • LTE-machine to machine LTE-machine to machine
  • LTE-M machine-to-machine communication systems
  • IoT Internet of Things
  • the devices provided by the embodiments of the present application may be implemented in whole or in part by software, hardware, firmware, or any combination thereof.
  • software When implemented by software, it can be implemented in the form of a computer program product in whole or in part.
  • the computer program product includes one or more computer instructions.
  • the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.
  • the computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from a website, computer, server, or data center.
  • the computer-readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server or a data center integrated with one or more available media.
  • the usable medium may be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a high-density digital video disc (digital video disc, DVD)), or a semiconductor medium (for example, SSD).
  • the size of the sequence numbers of the above-mentioned processes does not mean the order of execution, and the execution order of the processes should be determined by their functions and internal logic, and should not correspond to the implementation process of the embodiments of the present application. Constitute any limitation.
  • the disclosed system, device, and method can be implemented in other ways.
  • the device embodiments described above are merely illustrative, for example, the division of the units is only a logical function division, and there may be other divisions in actual implementation, for example, multiple units or components may be combined or It can be integrated into another system, or some features can be ignored or not implemented.
  • the displayed or discussed mutual coupling or direct coupling or communication connection may be indirect coupling or communication connection through some interfaces, devices or units, and may be in electrical, mechanical or other forms.
  • the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the objectives of the solutions of the embodiments.
  • the functional units in the various embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units may be integrated into one unit.

Landscapes

  • Engineering & Computer Science (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

一种雷达信号处理方法和装置，方法包括：获取多路RX信号，对多路RX信号中的每一路RX信号进行距离维谱分析和多普勒维谱分析，得到RD图（S310）；根据目标跟踪先验信息，对每一路RX信号的RD图进行相干处理，对多路RX信号的相干处理后的RD图进行相干叠加（S320）；根据多路RX信号相干叠加后的RD图，进行目标检测（S330）；方法基于目标跟踪先验信息对多路RX信号的RD map进行相干处理进而进行相干叠加，可以提高RX信号的合并增益，即相干处理和相干叠加能够使得叠加的接收信号相位尽可能符合相干条件，使得叠加后接收信号的强度明显增强，从而能够改善RD map的目标检测性能，提高检测信噪比和检测精度。

Description

雷达信号处理方法和雷达信号处理装置 技术领域
本申请涉及雷达信号处理领域，并且更具体地，涉及一种雷达信号处理方法和雷达信号处理装置。
背景技术
车载毫米波雷达因为具有全天候的探测能力，在自动驾驶领域是不可或缺的传感器之一。车载毫米波雷达通常使用多输入多输出(multiple input multiple output，MIMO)技术，以在天线数有限的情况下获得大的阵列孔径，从而获得高角度分辨率。时分多路复用(time division multiplexing，TDM)-MIMO发射模式具有硬件实现简单、互耦效应低等优点，成为车载毫米波雷达的一个重要研究方向。其中，调频连续波(frequency modulated continuous waveform，FMWC)是TDM-MIMO发射模式的一种常用的发射波形。
对于FMWC的TDM-MIMO雷达，各个发送(transmit，TX)天线与各个接收(receive，RX)天线之间形成多个虚拟接收天线。对于每个虚拟接收天线对应的虚拟接收通道，在针对虚拟RX信号应用测距快速傅里叶变换(fast fourier transform，FFT)和测速FFT后，可以得到距离多谱勒图(range doppler map，RD map)。对多个虚拟接收通道的RD map进行合并，之后可以根据合并后的RD map进行目标检测。对于同一TX对应的多路虚拟RX信号，现有的RD map合并方式主要基于盲相干合并，复杂度高。对于不同TX对应的多路虚拟RX信号，RD map合并采用的则是非相干合并。因而，RD map合并实际上只能做到“半相干合并”。现有的RD map合并方式复杂度高，并且使用该合并后的RD map进行目标检测性能差。
发明内容
本申请提供一种雷达信号处理方法和雷达信号处理装置，能够改善RD map的目标检测性能，提高检测信噪比和检测精度。
第一方面，提供了一种雷达信号处理方法，包括：获取多路接收RX信号，对多路RX信号中的每一路RX信号进行距离维的谱分析和多普勒维的谱分析，得到距离多普勒RD图；根据目标跟踪先验信息，对每一路RX信号的RD图进行相干处理，对多路RX信号的相干处理后的RD图进行相干叠加，目标跟踪先验信息为对多路RX信号进行目标跟踪处理后输出的先验信息；根据多路RX信号相干叠加后的RD图，进行目标检测。
应理解，多路RX信号可以对应MIMO雷达的多路虚拟RX信号，可以是所有虚拟RX信号，也可以是部分虚拟RX信号；多路RX信号也可以对应SIMO雷达的多路RX信号，可以是所有RX信号，也可以是部分RX信号。从所有RX信号中选取出部分RX信号进行合并，一来可以使得选出的部分RX信号的RD map更准确地做到相干合并，二来可以减小一定的计算量，提高处理效率。
第一方面的雷达信号处理方法，基于目标跟踪先验信息对多路RX信号的RD map进行相干处理进而进行相干叠加，可以提高RX信号的合并增益，即相干处理和相干叠加能够使得叠加的接收信号相位尽可能符合相干条件，使得叠加后接收信号的强度明显增强，从而能够改善RD map的目标检测性能，提高检测信噪比和检测精度。
在第一方面的一种可能的实现方式中，根据目标跟踪先验信息，对每一路RX信号的RD图进行相干处理，可以包括：根据目标跟踪先验信息，在每一路RX信号当前帧中的RD图上对至少一个目标进行预测，得到至少一个目标中每个目标的RD单元(RD cell)集合；针对每个目标的RD cell集合，根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值或角度预测值，对每一路RX信号的RD图中的RD cell集合进行相位补偿。本可能的实现方式中确定目标的RD cell集合基于目标跟踪先验信息，对RD map进行相位补偿也基于目标跟踪先验信息，可以有效提高合并增益，改善目标检测性能。应理解，本可能得实现方式中，根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值或角度预测值，对每一路RX信号的RD图中的RD cell集合进行相位补偿，可以包括：根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值和/或角度预测值，对每一路RX信号的RD图中的RD cell集合进行相位补偿。
在第一方面的一种可能的实现方式中，根据目标跟踪先验信息，在每一路RX信号当前帧中的RD图上对至少一个目标进行预测，得到至少一个目标中每个目标的RD单元集合，可以包括：根据目标跟踪先验信息中的每个目标在当前帧的距离预测值和径向速度预测值，在RD map上定位出每个目标的RD cell集合。在本可能的实现方式中，基于距离预测值和径向速度预测值定位目标的RD cell集合，对于多路RX信号可以在多个RD map上找到同一个目标对应的RD cell集合，并且对于TDM-MIMO雷达中的各TX天线的不同TX时间，可以补偿由目标速度导致的各路RX信号的相位变化，有利于完成RX通道之间的RD map的相干合并，从而提高合并增益。相对于使用RD map的距离值和速度值来定位目标的RD cell集合，使用目标跟踪输出的距离预测值和径向速度预测值比RD map的距离值和速度值更准确，定位出的目标的RD cell集合也更精确，使得RD map的相干合并的增益更高。
在第一方面的一种可能的实现方式中，在RD map上获得目标T在当前帧的距离预测值r k和径向速度预测值v k后，利用距离扩展因子α和速度扩展β，在RD map上分别选取目标T可能出现的范围：距离范围[(1-α)r k,(1+α)r k]，速度范围[(1-β)v k,(1+β)v k]。在本可能的实现方式中，利用扩展因子求得距离范围和速度范围，有利于更准确高效的定位出目标所在的RD区域。
在第一方面的一种可能的实现方式中，根据目标跟踪先验信息，在每一路RX信号当前帧中的RD图上对至少一个目标进行预测，得到至少一个目标中每个目标的RD单元集合，可以包括：根据目标跟踪先验信息中的每个目标在当前帧的轮廓预测值，在RD map上定位出每个目标的RD cell集合。在本可能的实现方式中，由于当前帧中目标的轮廓预测值包括有目标的位置、朝向、速度等预测信息，基于轮廓预测值定位目标的RD cell集合，可以准确地将目标在当前帧对应的RD cell区域定位出来，同时也可以根据RD cell集合对应的径向速度补偿目标的速度导致的RX信号的相位变化，有利于完成RX通道之间的RD map的相干合并，从而提高合并增益。应理解，目标的轮廓预测值可以包括目标 的位置、朝向、速度等预测信息。
在第一方面的一种可能的实现方式中，根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值或角度预测值，对每一路RX信号的RD map中的RD cell集合进行相位补偿，可以包括：根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值，对每一路RX信号的RD图中的RD单元集合，进行速度依赖的相位补偿；或根据目标跟踪先验信息中的每个目标在当前帧的角度预测值，对每一路RX信号的RD图中的RD单元集合，进行角度依赖的相位补偿。在本可能的实现方式中，基于预测的目标的角度进行相干信号的增强(即角度依赖的相位补偿)从而使更多的RX通道之间的RD map可以完成相干合并，能够提高合并增益，从而提高RD map的目标检测性能；和/或对于TDM-MIMO雷达中出现的各TX天线的不同TX时间，补偿目标的速度导致的RX信号的相位变化(即速度依赖的相位补偿)，从而使更多的RX通道之间的RD map可以完成相干合并，能够提高合并增益，从而提高RD map的目标检测性能，例如检测信噪比和检测精确度。应理解，本可能的实现方式相位补偿，可以包括：根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值，对每一路RX信号的RD图中的RD单元集合，进行速度依赖的相位补偿；和/或根据目标跟踪先验信息中的每个目标在当前帧的角度预测值，对每一路RX信号的RD图中的RD单元集合，进行角度依赖的相位补偿。
在第一方面的一种可能的实现方式中，对多路RX信号的相干处理后的RD map进行相干叠加，可以包括：对多路RX信号的相干处理后的RD map中的对应于同一个目标的RD cell集合进行复数值相加。本可能的实现方式对相干处理后的RD map中的对应于同一个目标的RD cell集合进行复数值相加完成相干合并，能够提高合并增益。
在第一方面的一种可能的实现方式中，CHT算法的输出的目标跟踪先验信息，可以包括至少一个目标中每个目标在当前帧的距离预测值、速度预测值、角度预测值和凸包预测值。本可能的实现方式使用的凸包波门可以对目标进行实时自适应调整，在无需手动参数输入的情况下实现高效的目标聚类与目标关联。
在第一方面的一种可能的实现方式中，采用均值合并的方式进行目标关联。本可能的实现方式使用均值合并的方式进行目标关联，相比目标关联通常使用的全局最近邻(global nearest neighbor，GNN)算法、最强邻(strongest neighbor，SN)算法、概率数据关联(probabilistic data association，PDA)算法等，关联性能有很大提高。
第二方面，提供了一种雷达信号处理装置，包括：获取单元，用于获取多路接收RX信号，对多路RX信号中的每一路RX信号进行距离维的谱分析和多普勒维的谱分析，得到距离多普勒RD图；处理单元，用于根据目标跟踪先验信息，对获取单元得到的每一路RX信号的RD图进行相干处理，对多路RX信号的相干处理后的RD图进行相干叠加，目标跟踪先验信息为对多路RX信号进行目标跟踪处理后输出的先验信息；检测单元，用于根据处理单元得到的多路RX信号相干叠加后的RD图，进行目标检测。
在第二方面的一种可能的实现方式中，处理单元具体用于：根据目标跟踪先验信息，在每一路RX信号当前帧中的RD图上对至少一个目标进行预测，得到至少一个目标中每个目标的RD单元集合；针对每个目标的RD单元集合，根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值或角度预测值，对每一路RX信号的RD图中的RD单元集合进行相位补偿。应理解，本可能得实现方式中，处理单元具体可以用于：根据目标跟 踪先验信息中的每个目标在当前帧的径向速度预测值和/或角度预测值，对每一路RX信号的RD图中的RD cell集合进行相位补偿。
在第二方面的一种可能的实现方式中，处理单元具体用于：根据目标跟踪先验信息中的每个目标在当前帧的距离预测值和径向速度预测值，在RD图上定位出每个目标的RD单元集合。
在第二方面的一种可能的实现方式中，处理单元具体用于：根据目标跟踪先验信息中的每个目标在当前帧的轮廓预测值，在RD图上定位出每个目标的RD单元集合。
在第二方面的一种可能的实现方式中，处理单元具体用于：根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值，对每一路RX信号的RD图中的RD单元集合，进行速度依赖的相位补偿；或根据目标跟踪先验信息中的每个目标在当前帧的角度预测值，对每一路RX信号的RD图中的RD单元集合，进行角度依赖的相位补偿。应理解，本可能得实现方式中，处理单元具体可以用于：根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值，对每一路RX信号的RD图中的RD单元集合，进行速度依赖的相位补偿；和/或根据目标跟踪先验信息中的每个目标在当前帧的角度预测值，对每一路RX信号的RD图中的RD单元集合，进行角度依赖的相位补偿。
在第二方面的一种可能的实现方式中，处理单元具体用于：对多路RX信号的相干处理后的RD图中的对应于同一个目标的RD单元集合进行复数值相加。
在第二方面的一种可能的实现方式中，装置还包括跟踪单元，目标跟踪先验信息是跟踪单元通过凸包跟踪算法计算得到的。
在第二方面的一种可能的实现方式中，目标跟踪先验信息包括至少一个目标中每个目标在当前帧的距离预测值、速度预测值、角度预测值和凸包预测值。
在第二方面的一种可能的实现方式中，装置还包括聚类关联单元，用于：通过目标跟踪先验信息中的凸包预测值得到当前帧中第一目标的预测凸包或目标的预测轮廓，对预测凸包或预测轮郭中的点进行聚类，得到一个或多个聚类点；将落入预测凸包或预测轮廓的一个或多个聚类点与第一目标进行关联。
在第二方面的一种可能的实现方式中，该信号处理装置应用于车载雷达中。
第三方面，提供了一种雷达信号处理装置，包括处理器，处理器与存储器耦合，可用于执行第一方面或第一方面中任一种可能实现方式中的方法。可选地，该信号处理装置还包括存储器。可选地，该信号处理装置还包括通信接口，处理器与通信接口耦合。
在第三方面的一种可能的实现方式中，该信号处理装置为网络设备。当该信号处理装置为网络设备时，通信接口可以是收发器，或，输入/输出接口。
在第三方面的一种可能的实现方式中，该信号处理装置为芯片或芯片系统。当该信号处理装置为芯片或芯片系统时，通信接口可以是该芯片或芯片系统上的输入/输出接口、接口电路、输出电路、输入电路、管脚或相关电路等。处理器也可以体现为处理电路或逻辑电路。
第四方面，提供了一种处理器，包括：输入电路、输出电路和处理电路。处理电路用于通过输入电路接收信号，并通过输出电路发射信号，使得第一方面以及第一方面中任一种可能的实现方式中的方法被实现。
在具体实现过程中，上述处理器可以为芯片，输入电路可以为输入管脚，输出电路可 以为输出管脚，处理电路可以为晶体管、门电路、触发器和各种逻辑电路等。输入电路所接收的输入的信号可以是由例如但不限于接收器接收并输入的，输出电路所输出的信号可以是例如但不限于输出给发射器并由发射器发射的，且输入电路和输出电路可以是同一电路，该电路在不同的时刻分别用作输入电路和输出电路。
第五方面，提供了一种处理装置，包括处理器和存储器。该处理器用于读取存储器中存储的指令，以执行第一方面以及第一方面中任一种可能的实现方式中的方法。
在第五方面的一种可能的实现方式中，处理器为一个或多个，存储器为一个或多个。
在第五方面的一种可能的实现方式中，存储器可以与处理器集成在一起，或者存储器与处理器分离设置。
读存储器(read only memory，ROM)，其可以与处理器集成在同一块芯片上，也可以在具体实现过程中，存储器可以为非瞬时性(non-transitory)存储器，例如只分别设置在不同的芯片上。
应理解，相关的数据交互过程例如发送指示信息可以为从处理器输出指示信息的过程，接收能力信息可以为处理器接收输入能力信息的过程。具体地，处理输出的数据可以输出给发射器，处理器接收的输入数据可以来自接收器。其中，发射器和接收器可以统称为收发器。
上述第五方面中的处理器可以是一个芯片，该处理器可以通过硬件来实现也可以通过软件来实现，当通过硬件实现时，该处理器可以是逻辑电路、集成电路等；当通过软件来实现时，该处理器可以是一个通用处理器，通过读取存储器中存储的软件代码来实现，该存储器可以集成在处理器中，可以位于该处理器之外，独立存在。
第六方面，提供了一种计算机程序产品，计算机程序产品包括：计算机程序(也可以称为代码，或指令)，当计算机程序被运行时，使得计算机执行上述第一方面以及第一方面任一种可能的实现方式中的方法。
第七方面，提供了一种计算机可读存储介质，计算机可读存储介质存储有计算机程序(也可以称为代码，或指令)当其在计算机上运行时，使得计算机执行上述第一方面以及第一方面任一种可能的实现方式中的方法。
第二方面至第七方面及相应的可能的实现方式可以应用于车载雷达中。
第八方面，提供了一种雷达，包括接收器和处理器，接收器用于接收多路接收RX信号，处理器用于根据多路RX信号，执行上述第一方面以及第一方面任一种可能的实现方式中的方法。
第一方面至第八方面及相应的可能的实现方式可以应用于通信系统，例如车联万物(vehicle to everything，V2X)通信系统、车路通信(LTE-vehicle，LTE-V)系统、车对车(vehicle to vehicle，V2V)通信系统、车联网、公路半自动车道收费(manual toll collection，MTC)系统、LTE机器对机器(LTE-machine to machine，LTE-M)通信系统、机器对机器(machine to machine，M2M)通信系统、物联网(internet of things，IoT)等中。
附图说明
图1是雷达信号处理的示意图。
图2是本申请提供的一个实施例的雷达信号处理的示意图。
图3是本申请提供的一个实施例的雷达信号处理方法的示意性流程图。
图4是本申请提供的一个实施例的TDM-MIMO的帧的示意图。
图5是本申请提供的一个实施例的目标T在当前帧的距离预测示意图。
图6是本申请提供的一个实施例的目标分布的示意图。
图7是本申请提供的实施例的相干合并检测结果和现有的非相干合并检测结果得对比图。
图8是本申请提供的一个实施例的目标聚类和目标关联的结果示意图。
图9是本申请提供的一个实施例的目标跟踪的结果示意图。
图10是本申请提供的一个实施例的雷达信号处理装置的示意性框图。
图11是本申请提供的另一个实施例的雷达信号处理装置的示意性框图。
具体实施方式
下面将结合附图，对本申请中的技术方案进行描述。
应理解，本申请实施例不仅适用于MIMO雷达，还适用于单输入多输出(single input multiple output，SIMO)雷达。对于MIMO雷达，其具有多个TX天线和多个RX天线(例如，M个TX天线和N个RX天线，M和N为正整数)，可以形成多路(M×N路)虚拟RX通道，获得多路虚拟RX信号。对于SIMO雷达，具有一个TX天线和多个RX天线，可以形成多路RX通道，获得多路RX信号。本申请将虚拟RX通道或RX通道统称为接收通道(RX通道)，将虚拟RX信号或RX信号统称为RX信号。
现有的技术中，TDM-MIMO雷达接收的多路RX信号对应的多个RD map的叠加，又称为“RD map的合并”，通常为“半相干合并”。对于同一TX对应的多路RX信号，现有的RD map合并为盲相干合并，属于相干合并，即在未知目标的角度信息时，将多路RX信号在每个角度上都做一次相位补偿，也就是在每个角度上合并一次，这种信号的叠加/合并方式计算复杂度很高。另外，对于不同TX对应的多路RX信号，由于RD map上不使用速度信息，采用的RD map合并只能是非相干合并。而且，现有的RD map合并未能有效利用目标跟踪输出的目标的位置、距离、角度、速度、轮廓等先验信息，计算复杂度高，合并增益小，并且使用该合并后的RD map进行目标检测性能差，即检测信噪比低、检测精度低。
雷达信号处理除包括目标检测外，还可以包括基于接收信号进行的目标聚类、目标关联和目标跟踪等处理。其中，现有的目标聚类未能利用目标跟踪已经输出的目标的位置、距离、角度、速度、轮廓等先验信息，这导致在目标聚类成多个类(cluster)时，会出现单个目标分裂成多个目标、多个目标聚类成一个目标、或出现目标被部分遮挡等问题，使得目标聚类效果差、效率低。例如，不同大小的目标聚类半径要求不一致时，造成聚类半径无法同时适用于大目标(如公交车等)、小目标(如人等)共存的情况。
综上，MIMO雷达信号处理一般都是开环的，不将目标跟踪的计算结果反馈回来影响当前的信号处理。图1是雷达信号处理的示意图。如图1所示，首先获取信号的数据。在获取数据后，对数据进行相应的处理，例如进行二维FFT(2D-FFT)得到RD map。现有的方案中，基于RD map进行前文的“半相干合并”，之后基于合并后的RD map进行目标检测。目标检测后，可以依次进行目标聚类和目标关联以及目标跟踪。应理解，在一些方 案中，目标聚类、目标关联和目标跟踪分别是单独的模块执行的；在另一些方案中，目标聚类和目标关联可以由同一个模块执行，或者目标关联和目标跟踪可以由同一个模块执行，或者目标聚类、目标关联和目标跟踪可以由同一个模块执行；本申请实施例对此不作限定。
本申请根据目标跟踪输出的目标的位置、距离、角度、速度、轮廓等目标跟踪先验信息，实现闭环的雷达信号处理，能够改善前文描述的问题。图2是本申请提供的一个实施例的雷达信号处理的示意图。如图2所示，与图1的方案相比，本申请的实施例可以基于目标跟踪的输出，例如目标跟踪先验信息，来进行目标检测。本申请的实施例还可以基于目标跟踪输出的目标跟踪先验信息，来进行目标聚类。即，利用目标跟踪得到的修正的目标的位置、距离、角度、速度、轮廓等信息，为信号处理提供参考信息。
本申请的实施例提供了一种信号处理方法。图3是本申请提供的一个实施例的雷达信号处理方法300的示意性流程图。如图3所示，该方法300可以包括以下步骤。
S310，获取多路RX信号，对多路RX信号中的每一路RX信号进行距离维的谱分析和多普勒维的谱分析，得到距离多普勒图RD map。
S320，根据目标跟踪先验信息，对每一路RX信号的RD map进行相干处理，对多路RX信号的相干处理后的RD map进行相干叠加，目标跟踪先验信息为对多路RX信号进行目标跟踪处理后输出的先验信息。
S330，根据多路RX信号相干叠加后的RD map，进行目标检测。
本申请实施例的雷达信号处理方法，基于目标跟踪先验信息对多路RX信号的RD map进行相干处理进而进行相干叠加，可以提高RX信号的合并增益，即相干处理和相干叠加能够使得叠加的接收信号相位尽可能符合相干条件，使得叠加后接收信号的强度明显增强，从而能够改善RD map的目标检测性能，提高检测信噪比和检测精度。
以MIMO雷达为例，在本申请的一些实施例中，S310获取的多路RX信号可以是M×N个虚拟RX通道对应的M×N个虚拟RX信号。在本申请的另一些实施例中，S310获取的多路RX信号也可以是M×N个虚拟RX信号中的部分虚拟RX信号。可以根据虚拟RX通道接收的虚拟RX信号的质量，和/或虚拟RX通道之间的相关性选取参与相干叠加的RD map。例如，可以选取M×N个RX信号中信号质量相对较高的多路虚拟RX信号；也可以选取相关性高的多个虚拟RX通道对应的RX信号。对于SIMO雷达，可以选取物理距离较近的多个RX通道对应的RX信号，尤其是在天线阵列尺寸较大的情况下。从所有RX信号中选取出部分RX信号进行合并，一来可以使得选出的部分RX信号的RD map更准确地做到相干合并，二来可以减小一定的计算量，提高处理效率。本申请的实施例对多路RX信号如何选取和获得不作限定。
TDM-MIMO雷达中，每个TX天线采用TDM-MIMO的方式发送FMWC信号。例如，当M＝12时，TX天线1至TX天线12表示为TX1-TX12，TX1-TX12轮流发送啁啾(chirp)脉冲信号(chirp1，chirp2，…，chirp12)，形成一个TDM-MIMO发送周期。图4是本申请提供的一个实施例的TDM-MIMO的帧的示意图。一个帧包括一个或多个TDM MIMO发送周期。
S310中得到RD map可以通过以下方法。每一个RX通道对接收的一帧(例如当前帧)RX信号进行二维FFT(2D-FFT)，即二位谱分析，其中，该RX信号可以为FMWC信 号，包含多个chirp脉冲信号。2D-FFT可以是分别进行距离维FFT(也称为距离维谱分析、快速FFT、测距FFT，即fast-FFT)和多谱勒维FFT(也称为多谱勒维谱分析、慢速FFT、测速FFT，即slow-FFT)。对RX信号进行2D-FFT可以获得该RX通道的RD map，记为RDM i，其中，i是集合{1,2,3,…,M×N}的全集或子集，即计算M×N个RX通道中全部或部分RX通道的RD map。
在本申请的一些实施例中，S320根据目标跟踪先验信息，对每一路RX信号的RD图进行相干处理，可以包括：根据目标跟踪先验信息，在每一路RX信号当前帧中的RDmap上对至少一个目标进行预测，得到至少一个目标中每个目标的RD单元(RD cell)集合；针对每个目标的RD cell集合，根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值或角度预测值，对每一路RX信号的RD图中的RD cell集合进行相位补偿。在这些实施例中，RD cell集合中可以包括一个或多个RD cell。确定目标的RD cell集合基于目标跟踪先验信息，对RD map进行相位补偿也基于目标跟踪先验信息，可以有效提高合并增益，改善目标检测性能。
应理解，根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值或角度预测值，对每一路RX信号的RD图中的RD cell集合进行相位补偿，可以包括：根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值和/或角度预测值，对每一路RX信号的RD图中的RD cell集合进行相位补偿。
可选地，作为一个实施例，根据目标跟踪先验信息，在每一路RX信号当前帧中的RD map上对至少一个目标进行预测，得到至少一个目标中每个目标的RD cell集合，可以包括：根据目标跟踪先验信息中的每个目标在当前帧的距离预测值和径向速度预测值，在RD map上定位出每个目标的RD cell集合。在该实施例中，基于距离预测值和径向速度预测值定位目标的RD cell集合，对于多路RX信号可以在多个RD map上找到同一个目标对应的RD cell集合，并且对于TDM-MIMO雷达中的各TX天线的不同TX时间，可以补偿由目标速度导致的各路RX信号的相位变化，有利于完成RX通道之间的RD map的相干合并，从而提高合并增益。
在一个具体的例子中，假设选取全部M×N个RX通道的RD map进行合并。可以利用当前帧之前的帧(例如上一帧)目标跟踪对目标(例如，目标T，可以为一辆车)估计出的针对当前帧的距离预测值r k，k＝1,2,…,M×N、径向速度预测值v k，k＝1,2,…,M×N，实现各个RX通道的RD map上目标T所在RD区域(即RD cell集合)的选取。其中，距离预测值r k为第k个RX通道估计出的目标T的无模糊的距离值；径向速度预测值v k为第k个RX通道估计出的目标T的无模糊的速度值。
在该例子中，针对当前帧中目标T所在RD区域的选取，上一帧结束时，目标跟踪的相应模块会根据目标T的目标跟踪先验信息和目标运动规律，估计得到目标T针对当前帧的距离预测值r k和径向速度预测值v k。
目标跟踪先验信息可以包括上一帧跟踪滤波后估计出的目标T的状态信息(或在当前帧的预测值)，包括以下的一种或多种：距离预测值或空间坐标系X坐标、Y坐标、Z坐标；速度预测值，包括径向速度预测值、速度矢量预测值等；角度预测值，包括方位角预测值和俯仰角预测值；以及如果目标跟踪算法采用下文将详细描述的凸包跟踪算法时可以 得到凸包预测值等。
利用目标跟踪算法(例如扩展卡尔曼滤波算法)并根据目标移动规律，对目标T在当前时刻的距离、径向速度进行预测。根据目标移动规律表示基于目标运动的先验模型对目标位置进行一步预测。例如，目标运动的先验模型可以是恒速度(constant velocity，CV)模型、恒转向率与速度(constant turn rate and velocity，CTRV)模型、恒转向率模型与加速度(Constant Turn Rate and Acceleration，CTRA)模型中的一种或多种，本申请实施例对此不做限定。
在一个具体的例子中，利用目标跟踪算法基于CV模型，根据上一帧时目标T的预测值数据可以得到当前帧时目标T的预测值数据。图5是本申请提供的一个实施例的目标T在当前帧的距离预测示意图。如图5所示，基于余弦定理，根据上一帧时目标T的距离预测值r k′、速度矢量预测值v r,k′(矢量参数，非径向速度)、速度矢量预测值v r,k′的方向与雷达径向(雷达与目标T之间连线)之间的夹角θ d′，计算出目标T在当前帧的距离预测值为
Figure PCTCN2020080735-appb-000001
其中T frame为一帧的时间长度。本申请其他实施例也可以通过其他算法计算距离预测值r k和径向速度预测值v k，本申请实施例对此不做限定。
在每个RX通道的RD map上获得目标T在当前帧的距离预测值r k，k＝1,2,…,M×N和径向速度预测值v k，k＝1,2,…,M×N后，利用距离扩展因子α和速度扩展β，在RD map上分别选取目标T可能出现的范围：距离范围[(1-α)r k,(1+α)r k]，速度范围[(1-β)v k,(1+β)v k]。利用扩展因子求得距离范围和速度范围，有利于更准确高效的定位出目标所在的RD区域。在此范围内，在各个RDM i上定位出一个或多个RD单元(cell)，将上述一个或多个RD cell的集合标记为{cell k}。
应理解，在该实施例中，RD map上显示的目标T的速度值(由多谱勒频率计算得到)是不准确的，这是因为TDM-MIMO导致测速范围变小，从而引入RD map上速度值的模糊，距离值也有类似的情况。相对于使用RD map的距离值和速度值来定位目标的RD cell集合，使用目标跟踪输出的距离预测值和径向速度预测值比RD map的距离值和速度值更准确，定位出的目标的RD cell集合也更精确，使得RD map的相干合并的增益更高。
可选地，作为另一个实施例，根据目标跟踪先验信息，在每一路RX信号当前帧中的RD map上对至少一个目标进行预测，得到至少一个目标中每个目标的RD cell集合，可以包括：根据目标跟踪先验信息中的每个目标在当前帧的轮廓预测值，在RD map上定位出每个目标的RD cell集合。具体而言，可以在目标跟踪时对于目标T对应的点云输出一个轮廓预测值，从而获得该轮廓预测值对应的点云在当前帧对应的RD cell的集合，标记为{cell k}。该轮廓预测值可以采用凸包计算的方式由跟踪算法输出，也可根据目标识别分类的方式估计得到，但该轮廓预测值的中心、朝向等参量由跟踪算法输出。由于当前帧中目标的轮廓预测值包括有目标的位置、朝向、速度等预测信息，基于轮廓预测值定位目标的RD cell集合，可以准确地将目标在当前帧对应的RD cell区域定位出来，同时也可以根据 RD cell集合对应的径向速度补偿目标的速度导致的RX信号的相位变化，有利于完成RX通道之间的RD map的相干合并，从而提高合并增益。
在本申请的一些实施例中，根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值和/或角度预测值，对每一路RX信号的RD map中的RD cell集合进行相位补偿，可以包括：根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值，对每一路RX信号的RD map中的RD cell集合，进行速度依赖的相位补偿；或根据目标跟踪先验信息中的每个目标在当前帧的角度预测值，对每一路RX信号的RD map中的RD cell集合，进行角度依赖的相位补偿。在这些实施例中，基于预测的目标的角度进行相干信号的增强(即角度依赖的相位补偿)，可以使更多的RX通道之间的RD map可以完成相干合并，能够提高合并增益，从而提高RD map的目标检测性能。对于TDM-MIMO雷达中出现的各TX天线的不同TX时间，可以补偿目标的速度导致的RX信号的相位变化(即速度依赖的相位补偿)，从而使更多的RX通道之间的RD map可以完成相干合并，能够提高合并增益，从而提高RD map的目标检测性能，例如检测信噪比和检测精确度。
应理解，本可能的实现方式相位补偿，可以包括：根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值，对每一路RX信号的RD图中的RD单元集合，进行速度依赖的相位补偿；和/或根据目标跟踪先验信息中的每个目标在当前帧的角度预测值，对每一路RX信号的RD图中的RD单元集合，进行角度依赖的相位补偿。
各个目标对应的RD cell位于RD map上的各个区域，图6是本申请提供的一个实施例的目标分布的示意图。如图6所示，RD map包括距离维度和速度维度。RD map上包括目标1、目标2和目标3。v k和θ k分别表示上一帧时目标跟踪所预测出的目标T的径向速度v k和角度θ k，是用于对目标的RD cell进行相位补偿时用到的参量。其中，指θ k水平角(azimuth angle)和/或俯仰角(elevation angle)。
在一个具体的实施例中，基于当前帧之前的帧(例如上一帧)目标跟踪对目标T估计出的针对当前帧的径向速度预测值v k和角度预测值θ k，对识别出的各RD cell集合{cell k}按如下公式进行相位补偿。
Figure PCTCN2020080735-appb-000002
该例子中，做出速度依赖的相位补偿，也做出角度依赖的相位补偿。其中，第n个RX通道对应的RX天线的角度补偿相位为
Figure PCTCN2020080735-appb-000003
速度补偿相位为
Figure PCTCN2020080735-appb-000004
n为虚拟RX天线的序号，即n＝1,2,...,M×N。d R是相邻RX天线的间距。Δt是RXn天线相对参考虚拟RX天线的接收时延，该时延等于由TDM-MIMO雷达的RXn天线对应的TX天线相对于参考TX天线的发送时延。例如，一个chirp脉冲信号的重复周期为T PRF，虚拟RXn天线对应的实际TX天线的发送顺序为第t个(参考虚拟RXn天线对应的TX天线的顺序为第1个)， 则Δt＝(t-1)T PRF。Y received(n)是第n个RX通道的RD map上一个RD cell做完角度相位补偿、速度相位补偿后的接收信号。|Y target(n)|是Y received(n)的幅度信号，
Figure PCTCN2020080735-appb-000005
是Y received(n)的初始相位。
在本申请的一些实施例中，S320中对多路RX信号的相干处理后的RD map进行相干叠加，可以包括：对多路RX信号的相干处理后的RD map中的对应于同一个目标的RD cell集合进行复数值相加。
在一个具体的例子中，对完成速度相关和角度相关的相位补偿后的多个RD map进行相干合并，也称为相干叠加，得到目标T对应的待检测的信号
Figure PCTCN2020080735-appb-000006
对目标跟踪输出的所有目标都执行上述相位补偿操作，获得相干叠加后的RD mapRDM intergrated,针对RDM intergrated进行目标检测，例如恒虚警率(constant false alarm ratio，CFAR)检测。
图7是本申请提供的实施例的相干合并检测结果和现有的非相干合并检测结果得对比图。图7的左上图的曲线图和右上图的等高图为本申请的实施例的相干合并检测结果。图7的左下图的曲线图和右下图的等高图为现有的非相干合并检测结果。对比圆圈中的目标的接收信号强度，相干合并检测较非相干合并检测的接收信号强度明显增强，即信噪比提高，这使得目标更容易被检测出来。
应理解，可以认为方法300是基于目标跟踪先验信息的目标检测过程。
前文提到雷达信号处理除包括目标检测外，还可以包括基于接收信号进行的目标聚类、目标关联和目标跟踪等处理。本申请各实施例的目标跟踪先验信息是目标跟踪得到的。现有的目标跟踪的算法大多采用圆形波门，跟踪效率低，并且基于圆形波门的算法下的目标聚类和目标关联的结果也并不理想。本申请各实施例的目标跟踪先验信息可以是通过凸包跟踪(convex hull tracking，CHT)算法计算得到的。相较于传统的圆形波门，凸包波门与目标的外形更为匹配，因而可以极大提升关联性能。
下面详细说明本申请实施例的基于目标跟踪先验信息的CHT算法。CHT算法的输出的目标跟踪先验信息，可以包括至少一个目标中每个目标在当前帧的距离预测值、速度预测值、角度预测值和凸包预测值。
本申请的实施例可以根据目标跟踪先验信息结合目标移动规律或目标移动模型，实现对多个目标的动态跟踪。目标跟踪可以包括对当前帧的目标的位置进行预测；以及在多个连续帧中，对多个动态目标实时跟踪，包括目标添加(例如，添加新的观测值为目标)、目标删除(例如，删除不能与观测值关联的目标)、目标聚类、目标关联与目标跟踪等。其中，CHT算法中的滤波更新算法可以采用扩展卡尔曼滤波或容积卡尔曼滤波(cubature Kalman filter，CKF)。CHT算法的思想是先确定大视野(即针对一个目标确定一个凸包)，再根据凸包的预测结果，将落入凸包的多个聚类进行合并，克服视野和分辨率无法同时最优的难题。
在本申请的一些实施例中，方法300还可以包括：通过目标跟踪先验信息中的凸包预测值得到当前帧中第一目标的预测凸包或目标的预测轮廓，对预测凸包或预测轮郭中的点 进行聚类，得到一个或多个聚类点；将落入预测凸包或预测轮廓的一个或多个聚类点与第一目标进行关联。本申请的实施例通过在目标聚类时引入目标跟踪先验信息，可以有效解决遮挡等分辨问题。本申请的实施例使用的凸包波门可以对目标进行实时自适应调整，在无需手动参数输入的情况下实现高效的目标聚类与目标关联。
现有的目标关联通常使用例如全局最近邻(global nearest neighbor，GNN)算法、最强邻(strongest neighbor，SN)算法、概率数据关联(probabilistic data association，PDA)算法等，这些算法下的关联性能有待提高。本申请的一些实施例采用均值合并的方式进行目标关联。
在一个具体的例子中，CHT算法的具体步骤可以包括：a)起始：凸包(convex hull)初始化。b)预测：基于上一帧估计出的针对当前帧的目标的速度预测值和初始化的凸包的尺寸，对当前帧的目标凸包进行预测。c)目标关联：对落入凸包内的点，基于CKF完成与单个目标的关联。具体而言，可以对基于目标点云计算的凸包进行预测，并对落入该凸包的聚类点以均值合并的方式进行关联，而非以传统的关联方法(例如，GNN算法、SN算法、PDA算法等)进行目标关联。d)更新：基于目标关联的结果对目标的凸包进行更新。
图8是本申请提供的一个实施例的目标聚类和目标关联的结果示意图。图8中最大的框为卡车、中等大小的框为电动车、小框为行人。图9是本申请提供的一个实施例的目标跟踪的结果示意图。由图8和图9可以看出，本申请实施例的信号处理方法可以实现大目标(例如卡车)与小目标(例如电动车和行人)的准确聚类、关联与跟踪。
应理解，本申请各实施例的信号处理方法可以应用于车载雷达的信号处理中。
前文对本申请实施例的方法进行了详细说明，下面对本申请实施例的装置进行说明。
图10是本申请提供的一个实施例的雷达信号处理装置1000的示意性框图。如图10所示，信号处理装置1000包括：获取单元1010，用于获取多路接收RX信号，对多路RX信号中的每一路RX信号进行距离维的谱分析和多普勒维的谱分析，得到距离多普勒RD图；处理单元1020，用于根据目标跟踪先验信息，对获取单元1010得到的每一路RX信号的RD图进行相干处理，对多路RX信号的相干处理后的RD图进行相干叠加，目标跟踪先验信息为对多路RX信号进行目标跟踪处理后输出的先验信息；检测单元1030，用于根据处理单元1020得到的多路RX信号相干叠加后的RD图，进行目标检测。
在一些实施例中，处理单元1020具体可以用于：处理单元1020根据目标跟踪先验信息，在每一路RX信号当前帧中的RD图上对至少一个目标进行预测，得到至少一个目标中每个目标的RD单元集合；处理单元1020针对每个目标的RD单元集合，根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值或角度预测值，对每一路RX信号的RD图中的RD单元集合进行相位补偿。应理解，处理单元1020具体可以用于：根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值和/或角度预测值，对每一路RX信号的RD图中的RD cell集合进行相位补偿。
在一些实施例中，处理单元1020具体可以用于：根据目标跟踪先验信息中的每个目标在当前帧的距离预测值和径向速度预测值，在RD图上定位出每个目标的RD单元集合。
在一些实施例中，处理单元1020具体可以用于：根据目标跟踪先验信息中的每个目标在当前帧的轮廓预测值，在RD图上定位出每个目标的RD单元集合。
在一些实施例中，处理单元1020具体可以用于：根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值，对每一路RX信号的RD图中的RD单元集合，进行速度依赖的相位补偿；或处理单元1020根据目标跟踪先验信息中的每个目标在当前帧的角度预测值，对每一路RX信号的RD图中的RD单元集合，进行角度依赖的相位补偿。应理解，处理单元1020具体可以用于：根据目标跟踪先验信息中的每个目标在当前帧的径向速度预测值，对每一路RX信号的RD图中的RD单元集合，进行速度依赖的相位补偿；和/或根据目标跟踪先验信息中的每个目标在当前帧的角度预测值，对每一路RX信号的RD图中的RD单元集合，进行角度依赖的相位补偿。
在一些实施例中，处理单元1020具体可以用于：对多路RX信号的相干处理后的RD图中的对应于同一个目标的RD单元集合进行复数值相加。
在一些实施例中，装置1000还包括跟踪单元1040，目标跟踪先验信息是跟踪单元1040通过凸包跟踪算法计算得到的。
在一些实施例中，目标跟踪先验信息包括至少一个目标中每个目标在当前帧的距离预测值、速度预测值、角度预测值和凸包预测值。
在一些实施例中，装置还包括聚类关联单元1050，用于：通过目标跟踪先验信息中的凸包预测值得到当前帧中第一目标的预测凸包或目标的预测轮廓，对预测凸包或预测轮郭中的点进行聚类，得到一个或多个聚类点；将落入预测凸包或预测轮廓的一个或多个聚类点与第一目标进行关联。
在一些实施例中，该信号处理装置1000应用于车载雷达中。
应理解，该信号处理装置1000中的各单元可以用于实现上文方法实施例中相应的操作。
图11是本申请提供的另一个实施例的信号处理装置1100的示意性框图。如图11所示，信号处理装置1100包括：处理器1110，处理器1110与存储器耦合，存储器用于存储计算机程序或指令或者和/或数据，处理器1110用于执行存储器存储的计算机程序或指令和/或者数据，使得上文方法实施例中的方法被执行。
可选地，该信号处理装置1100可以包括存储器1120，用于执行上述操作。
可选地，该信号处理装置1110还可以包括通信接口，处理器与通信接口耦合。
在一些实施例中，该信号处理装置1100为网络设备。当该信号处理装置1100为网络设备时，通信接口可以是收发器，或，输入/输出接口。
在一些实施例中，该信号处理装置1100为芯片或芯片系统。当该信号处理装置1100为芯片或芯片系统时，通信接口可以是该芯片或芯片系统上的输入/输出接口、接口电路、输出电路、输入电路、管脚或相关电路等。处理器也可以体现为处理电路或逻辑电路。
应理解，该信号处理装置1100中的各模块可以用于实现上文方法实施例中相应的操作，并可以对应于上文的处理装置1000中的各单元。
本申请还提供了一种处理器，包括：输入电路、输出电路和处理电路。处理电路用于通过输入电路接收信号，并通过输出电路发射信号，使得上文方法实施例中的方法被实现。
在具体实现过程中，上述处理器可以为芯片，输入电路可以为输入管脚，输出电路可以为输出管脚，处理电路可以为晶体管、门电路、触发器和各种逻辑电路等。输入电路所接收的输入的信号可以是由例如但不限于接收器接收并输入的，输出电路所输出的信号可 以是例如但不限于输出给发射器并由发射器发射的，且输入电路和输出电路可以是同一电路，该电路在不同的时刻分别用作输入电路和输出电路。
应理解，相关的数据交互过程例如发送指示信息可以为从处理器输出指示信息的过程，接收能力信息可以为处理器接收输入能力信息的过程。具体地，处理输出的数据可以输出给发射器，处理器接收的输入数据可以来自接收器。其中，发射器和接收器可以统称为收发器。
上述处理器可以是一个芯片，该处理器可以通过硬件来实现也可以通过软件来实现，当通过硬件实现时，该处理器可以是逻辑电路、集成电路等；当通过软件来实现时，该处理器可以是一个通用处理器，通过读取存储器中存储的软件代码来实现，该存储器可以集成在处理器中，可以位于该处理器之外，独立存在。
应理解，本申请实施例中提及的处理器可以包括中央处理器(central processing pnit，CPU)，网络处理器(network processor，NP)或者CPU和NP的组合。处理器还可以进一步包括硬件芯片。上述硬件芯片可以是专用集成电路(application-specific integrated circuit，ASIC)，可编程逻辑器件(programmable logic device，PLD)或其组合。上述PLD可以是复杂可编程逻辑器件(complex programmable logic device，CPLD)，现场可编程逻辑门阵列(field-programmable gate array，FPGA)，通用阵列逻辑(generic array logic，GAL)或其任意组合。
还应理解，本申请实施例中提及的存储器可以是易失性存储器(volatile memory)或非易失性存储器(non-volatile memory)，或可包括易失性和非易失性存储器两者。其中，非易失性存储器可以是只读存储器(read-only memory，ROM)、可编程只读存储器(programmable ROM，PROM)、可擦除可编程只读存储器(erasable PROM，EPROM)、电可擦除可编程只读存储器(electrically EPROM，EEPROM)、快闪存储器(flash memory)、硬盘(hard disk drive，HDD)或固态硬盘(solid-state drive，SSD)。易失性存储器可以是随机存取存储器(random access memory，RAM)，其用作外部高速缓存。通过示例性但不是限制性说明，许多形式的RAM可用，例如静态随机存取存储器(static RAM，SRAM)、动态随机存取存储器(dynamic RAM，DRAM)、同步动态随机存取存储器(synchronous DRAM，SDRAM)、双倍数据速率同步动态随机存取存储器(double data rate SDRAM，DDRSDRAM)、增强型同步动态随机存取存储器(enhanced SDRAM，ESDRAM)、同步连接动态随机存取存储器(synchlink DRAM，SLDRAM)和直接内存总线随机存取存储器(direct rambus RAM，DR RAM)。
需要说明的是，当处理器为通用处理器、DSP、ASIC、FPGA或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件时，存储器(存储模块)集成在处理器中。
应注意，本文描述的存储器旨在包括但不限于这些和任意其它适合类型的存储器。
本申请还提供了一种计算机程序产品，计算机程序产品包括：计算机程序(也可以称为代码，或指令)，当计算机程序被运行时，使得计算机执行上述方法实施例中的方法。
本申请还提供了一种计算机可读存储介质，计算机可读存储介质存储有计算机程序(也可以称为代码，或指令)当其在计算机上运行时，使得计算机执行上述方法实施例中的方法。
本申请还提供了一种雷达，包括接收器和处理器，接收器用于接收多路接收RX信号， 处理器用于根据多路RX信号，执行上述方法实施例中的方法。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的系统、装置和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
应理解，本申请的方法、装置、设备、处理器、计算机程序产品、计算机可读存储介质或芯片等可以应用于车载雷达中。
应理解，本申请的方法、装置、设备、处理器、计算机程序产品、计算机可读存储介质或芯片等可以应用于通信系统，例如车联万物(vehicle to everything，V2X)通信系统、车路通信(LTE-vehicle，LTE-V)系统、车对车(vehicle to vehicle，V2V)通信系统、车联网、公路半自动车道收费(manual toll collection，MTC)系统、LTE机器对机器(LTE-machine to machine，LTE-M)通信系统、机器对机器(machine to machine，M2M)通信系统、物联网(internet of things，IoT)等中。
本申请实施例提供给的设备，可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当使用软件实现时，可以全部或部分地以计算机程序产品的形式实现。所述计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行所述计算机指令时，全部或部分地产生按照本申请实施例所述的流程或功能。所述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。所述计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一个计算机可读存储介质传输，例如，所述计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线(例如同轴电缆、光纤、数字用户线(digital subscriber line，DSL))或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。所述计算机可读存储介质可以是计算机能够存取的任何可用介质或者是包含一个或多个可用介质集成的服务器、数据中心等数据存储设备。所述可用介质可以是磁性介质(例如，软盘、硬盘、磁带)、光介质(例如，高密度数字视频光盘(digital video disc，DVD))、或者半导体介质(例如，SSD)等。
应理解，本文中涉及的第一、第二以及各种数字编号仅为描述方便进行的区分，并不用来限制本申请的范围。
应理解，本文中术语“和/或”，仅仅是一种描述关联对象的关联关系，表示可以存在三种关系，例如，A和/或B，可以表示：单独存在A，同时存在A和B，单独存在B这三种情况。另外，本文中字符“/”，一般表示前后关联对象是一种“或”的关系。
应理解，在本申请实施例中，上述各过程的序号的大小并不意味着执行顺序的先后，各过程的执行顺序应以其功能和内在逻辑确定，而不应对本申请实施例的实施过程构成任何限定。
本领域普通技术人员可以意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请的范围。
在本申请所提供的几个实施例中，应该理解到，所揭露的系统、装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如，所述单元的 划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。
另外，在本申请各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。
以上所述，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应所述以权利要求的保护范围为准。

Claims (22)

  1. 一种雷达信号处理方法，其特征在于，包括：
    获取多路接收RX信号，对所述多路RX信号中的每一路RX信号进行距离维的谱分析和多普勒维的谱分析，得到距离多普勒RD图；
    根据目标跟踪先验信息，对所述每一路RX信号的RD图进行相干处理，对所述多路RX信号的相干处理后的RD图进行相干叠加，所述目标跟踪先验信息为对所述多路RX信号进行目标跟踪处理后输出的先验信息；
    根据所述多路RX信号相干叠加后的RD图，进行目标检测。
  2. 根据权利要求1所述的方法，其特征在于，所述根据目标跟踪先验信息，对所述每一路RX信号的RD图进行相干处理，包括：
    根据所述目标跟踪先验信息，在所述每一路RX信号当前帧中的RD图上对至少一个目标进行预测，得到所述至少一个目标中每个目标的RD单元集合；
    针对所述每个目标的RD单元集合，根据所述目标跟踪先验信息中的所述每个目标在所述当前帧的径向速度预测值或角度预测值，对所述每一路RX信号的RD图中的所述RD单元集合进行相位补偿。
  3. 根据权利要求2所述的方法，其特征在于，所述根据所述目标跟踪先验信息，在所述每一路RX信号当前帧中的RD图上对至少一个目标进行预测，得到所述至少一个目标中每个目标的RD单元集合，包括：
    根据所述目标跟踪先验信息中的所述每个目标在所述当前帧的距离预测值和径向速度预测值，在所述RD图上定位出所述每个目标的RD单元集合。
  4. 根据权利要求2所述的方法，其特征在于，所述根据所述目标跟踪先验信息，在所述每一路RX信号当前帧中的RD图上对至少一个目标进行预测，得到所述至少一个目标中每个目标的RD单元集合，包括：
    根据所述目标跟踪先验信息中的所述每个目标在所述当前帧的轮廓预测值，在所述RD图上定位出所述每个目标的RD单元集合。
  5. 根据权利要求2至4中任一项所述的方法，其特征在于，所述根据所述目标跟踪先验信息中的所述每个目标在所述当前帧的径向速度预测值或角度预测值，对所述每一路RX信号的RD图中的所述RD单元集合进行相位补偿，包括：
    根据所述目标跟踪先验信息中的所述每个目标在所述当前帧的径向速度预测值，对所述每一路RX信号的RD图中的所述RD单元集合，进行速度依赖的相位补偿；或
    根据所述目标跟踪先验信息中的所述每个目标在所述当前帧的角度预测值，对所述每一路RX信号的RD图中的所述RD单元集合，进行角度依赖的相位补偿。
  6. 根据权利要求1至5中任一项所述的方法，其特征在于，所述对所述多路RX信号的相干处理后的RD图进行相干叠加，包括：
    对所述多路RX信号的相干处理后的RD图中的对应于同一个目标的RD单元集合，按RD单元对应的复数值进行相加。
  7. 根据权利要求1至6中任一项所述的方法，其特征在于，所述目标跟踪先验信息 是通过凸包跟踪算法计算得到的。
  8. 根据权利要求7所述的方法，其特征在于，所述目标跟踪先验信息包括所述至少一个目标中每个目标在当前帧的距离预测值、速度预测值、角度预测值和凸包预测值。
  9. 根据权利要求7或8所述的方法，其特征在于，所述方法还包括：
    通过所述目标跟踪先验信息中的凸包预测值得到当前帧中第一目标的预测凸包或目标的预测轮廓，对所述预测凸包或预测轮廓中的点进行聚类，得到一个或多个聚类点；
    将落入所述预测凸包或所述预测轮廓的所述一个或多个聚类点与所述第一目标进行关联。
  10. 一种雷达信号处理装置，其特征在于，包括：
    获取单元，用于获取多路接收RX信号，对所述多路RX信号中的每一路RX信号进行距离维的谱分析和多普勒维的谱分析，得到距离多普勒RD图；
    处理单元，用于根据目标跟踪先验信息，对所述获取单元得到的所述每一路RX信号的RD图进行相干处理，对所述多路RX信号的相干处理后的RD图进行相干叠加，所述目标跟踪先验信息为对所述多路RX信号进行目标跟踪处理后输出的先验信息；
    检测单元，用于根据所述处理单元得到的所述多路RX信号相干叠加后的RD图，进行目标检测。
  11. 根据权利要求10所述的装置，其特征在于，所述处理单元具体用于：
    根据所述目标跟踪先验信息，在所述每一路RX信号当前帧中的RD图上对至少一个目标进行预测，得到所述至少一个目标中每个目标的RD单元集合；
    针对所述每个目标的RD单元集合，根据所述目标跟踪先验信息中的所述每个目标在所述当前帧的径向速度预测值或角度预测值，对所述每一路RX信号的RD图中的所述RD单元集合进行相位补偿。
  12. 根据权利要求11所述的装置，其特征在于，所述处理单元具体用于：
    根据所述目标跟踪先验信息中的所述每个目标在所述当前帧的距离预测值和径向速度预测值，在所述RD图上定位出所述每个目标的RD单元集合。
  13. 根据权利要求11所述的装置，其特征在于，所述处理具体用于：
    根据所述目标跟踪先验信息中的所述每个目标在所述当前帧的轮廓预测值，在所述RD图上定位出所述每个目标的RD单元集合。
  14. 根据权利要求11至13中任一项所述的装置，其特征在于，所述处理单元具体用于：
    根据所述目标跟踪先验信息中的所述每个目标在所述当前帧的径向速度预测值，对所述每一路RX信号的RD图中的所述RD单元集合，进行速度依赖的相位补偿；或
    根据所述目标跟踪先验信息中的所述每个目标在所述当前帧的角度预测值，对所述每一路RX信号的RD图中的所述RD单元集合，进行角度依赖的相位补偿。
  15. 根据权利要求10至14中任一项所述的装置，其特征在于，所述处理单元具体用于：
    对所述多路RX信号的相干处理后的RD图中的对应于同一个目标的RD单元集合进行复数值相加。
  16. 根据权利要求10至15中任一项所述的装置，其特征在于，所述装置还包括跟踪 单元，所述目标跟踪先验信息是所述跟踪单元通过凸包跟踪算法计算得到的。
  17. 根据权利要求16所述的装置，其特征在于，所述目标跟踪先验信息包括所述至少一个目标中每个目标在当前帧的距离预测值、速度预测值、角度预测值和凸包预测值。
  18. 根据权利要求16或17所述的装置，其特征在于，所述装置还包括聚类关联单元，用于：
    通过所述目标跟踪先验信息中的凸包预测值得到当前帧中第一目标的预测凸包或目标的预测轮廓，对所述预测凸包或预测轮郭中的点进行聚类，得到一个或多个聚类点；
    将落入所述预测凸包或所述预测轮廓的所述一个或多个聚类点与所述第一目标进行关联。
  19. 根据权利要求10至18中任一项所述的装置，其特征在于，所述装置应用于车载雷达中。
  20. 一种信号处理装置，其特征在于，包括处理器和存储器，所述处理器与所述存储器耦合，所述存储器用于存储计算机程序或指令，所述处理器用于执行所述存储器中的所述计算机程序或指令，使得权利要求1至9中任一项所述的方法被执行。
  21. 一种计算机可读存储介质，其特征在于，存储有计算机程序或指令，所述计算机程序或指令用于实现权利要求1至9中任一项所述的方法。
  22. 一种雷达，其特征在于，包括接收器和处理器，所述接收器用于接收多路接收RX信号，所述处理器用于根据所述多路RX信号，执行如权利要求1至9中任一项所述的方法。
PCT/CN2020/080735 2020-03-23 2020-03-23 雷达信号处理方法和雷达信号处理装置 WO2021189206A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN202080004851.7A CN112689773B (zh) 2020-03-23 2020-03-23 雷达信号处理方法和雷达信号处理装置
PCT/CN2020/080735 WO2021189206A1 (zh) 2020-03-23 2020-03-23 雷达信号处理方法和雷达信号处理装置

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2020/080735 WO2021189206A1 (zh) 2020-03-23 2020-03-23 雷达信号处理方法和雷达信号处理装置

Publications (1)

Publication Number Publication Date
WO2021189206A1 true WO2021189206A1 (zh) 2021-09-30

Family

ID=75457710

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/080735 WO2021189206A1 (zh) 2020-03-23 2020-03-23 雷达信号处理方法和雷达信号处理装置

Country Status (2)

Country Link
CN (1) CN112689773B (zh)
WO (1) WO2021189206A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114999087A (zh) * 2022-05-24 2022-09-02 深圳康佳电子科技有限公司 一种保护隐私的监控方法、装置、介质及终端

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102243303A (zh) * 2011-04-13 2011-11-16 电子科技大学 一种基于呼吸特征的静止人体穿墙定位方法
US8305261B2 (en) * 2010-04-02 2012-11-06 Raytheon Company Adaptive mainlobe clutter method for range-Doppler maps
CN103454624A (zh) * 2013-09-22 2013-12-18 河海大学 基于降维稀疏重构空时谱的直接数据域动目标检测方法
CN106054138A (zh) * 2016-07-29 2016-10-26 西安电子科技大学 一种ddma波形的参差多普勒频率偏移选择方法
CN106707247A (zh) * 2017-03-24 2017-05-24 武汉大学 一种基于紧凑天线阵的高频海洋雷达目标检测方法
CN107102318A (zh) * 2017-05-16 2017-08-29 武汉大学 一种数字音频广播外辐射源雷达目标探测系统与方法

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105334507B (zh) * 2015-11-18 2017-11-03 西安电子科技大学 基于极化多特征的对海面漂浮雷达目标的检测方法
CN106199548B (zh) * 2016-06-30 2019-01-11 西安电子科技大学 基于四极化通道融合的海面漂浮微弱雷达目标的检测方法
US20180024239A1 (en) * 2017-09-25 2018-01-25 GM Global Technology Operations LLC Systems and methods for radar localization in autonomous vehicles

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8305261B2 (en) * 2010-04-02 2012-11-06 Raytheon Company Adaptive mainlobe clutter method for range-Doppler maps
CN102243303A (zh) * 2011-04-13 2011-11-16 电子科技大学 一种基于呼吸特征的静止人体穿墙定位方法
CN103454624A (zh) * 2013-09-22 2013-12-18 河海大学 基于降维稀疏重构空时谱的直接数据域动目标检测方法
CN106054138A (zh) * 2016-07-29 2016-10-26 西安电子科技大学 一种ddma波形的参差多普勒频率偏移选择方法
CN106707247A (zh) * 2017-03-24 2017-05-24 武汉大学 一种基于紧凑天线阵的高频海洋雷达目标检测方法
CN107102318A (zh) * 2017-05-16 2017-08-29 武汉大学 一种数字音频广播外辐射源雷达目标探测系统与方法

Also Published As

Publication number Publication date
CN112689773B (zh) 2022-03-29
CN112689773A (zh) 2021-04-20

Similar Documents

Publication Publication Date Title
US11927668B2 (en) Radar deep learning
US11885872B2 (en) System and method for camera radar fusion
Brodeski et al. Deep radar detector
CN111679266B (zh) 汽车毫米波雷达稀疏阵列栅瓣虚假目标识别方法及系统
CN112098990A (zh) 车载高分辨毫米波雷达对于中高速车辆的检测与跟踪方法
US11899132B2 (en) Super-resolution enhancement techniques for radar
CN113536850B (zh) 基于77g毫米波雷达的目标物体大小测试方法和装置
WO2021129581A1 (zh) 一种信号处理方法及装置
WO2021189206A1 (zh) 雷达信号处理方法和雷达信号处理装置
US20230139751A1 (en) Clustering in automotive imaging
US20210323560A1 (en) Vehicle speed calculation method, system, device, and storage medium
WO2023124780A1 (zh) 点云数据增强方法、装置、计算机设备、系统及存储介质
Kim et al. Deep-learning based multi-object detection and tracking using range-angle map in automotive radar systems
WO2022226948A1 (zh) 一种目标的特征提取方法及装置
CN111522010B (zh) 汽车防撞雷达信号处理方法和系统
CN114859337A (zh) 数据处理方法、装置、电子设备、计算机存储介质
WO2021196165A1 (zh) 频率分析方法、装置及雷达
Hayashi et al. In corporation of Super-resolution Doppler Analysis and Compressed Sensing Filter for UWB Human Body Imaging Radar
CN116027288A (zh) 生成数据的方法、装置、电子设备及存储介质
Ren et al. Research and Implementation of 77GHz Automotive Radar Target Detection Technology
TWI834772B (zh) 雷達深度學習
JP7222952B2 (ja) 電子機器、電子機器の制御方法、及びプログラム
CN116008944B (zh) 一种毫米波fmcw雷达空间维信源数判断方法和装置
CN116828394B (zh) 追踪通信方法、追踪通信装置、电子设备以及存储介质
WO2020097903A1 (zh) 测角方法以及雷达设备

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20927072

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20927072

Country of ref document: EP

Kind code of ref document: A1