Idle sampling window based millimeter wave radar noise compensation circuit and method

Abstract

This invention relates to the field of electromagnetic compatibility and adaptive background noise compensation technology for millimeter-wave radar receiving links, and particularly to a millimeter-wave radar noise compensation circuit and method based on idle sampling windows. The method involves acquiring radar frame timing information, target detection information, and interference source timing information of internal periodic electromagnetic interference sources; establishing corresponding idle sampling window groups at the same intra-frame position in multiple consecutive radar frames to collect background noise samples; determining the amplitude and phase parameters, repetition confidence, and timing correlation of candidate background noise components; identifying the background noise component to be compensated when a threshold condition is met; selecting the disturbed node based on injection path calibration data and injecting an amplitude-reverse compensation signal; and collecting residual noise in subsequent corresponding idle sampling windows and correcting the compensation control parameters.

Description

Technical Field

[0001] This invention relates to the field of electromagnetic compatibility and adaptive background noise compensation technology for millimeter-wave radar receiving links, and particularly to a millimeter-wave radar noise compensation circuit and method based on an idle sampling window. Background Technology

[0002] Millimeter-wave radar typically detects the range, velocity, and azimuth of a target by transmitting a frequency-modulated signal and receiving the echo signal reflected from the target. The receiver link of a millimeter-wave radar usually includes a receiving antenna, RF front-end, mixer unit, intermediate frequency signal processing unit, analog-to-digital converter (ADC) unit, and digital signal processing unit. During the operation of a millimeter-wave radar, in addition to the target echo, interference from transmission path leakage, internal electromagnetic coupling, power conversion noise, digital clock coupling, communication interface disturbances, and external electromagnetic interference can all enter the receiver link, causing issues such as increased noise floor, analog signal amplitude fluctuations, or occupancy of the ADC input dynamic range.

[0003] Existing millimeter-wave radars typically employ methods such as filtering, shielding, grounding, background data estimation, or digital crosstalk compensation to suppress interference in the receiving link. For example, patent application CN116804739A discloses an adaptive transmitter-receiver crosstalk cancellation method for radar systems. This method selects a dataset from radar data output by the radar sensor and performs target detection on the selected dataset. When no target is detected in the dataset, it updates the crosstalk compensation factor between the transmission and receiving paths based on the dataset and compensates the radar data based on the crosstalk compensation factor. When a target is detected in the dataset, the existing crosstalk compensation factor is maintained.

[0004] The aforementioned existing technologies primarily address transmitter-receiver crosstalk formed by the coupling of the transmission path to the receiving path in millimeter-wave radar, using the radar digital data output from the receiving link as the object of crosstalk estimation and compensation. However, for periodic or quasi-periodic background noise generated by power conversion units, digital clock units, communication interfaces, or other internal electromagnetic coupling paths, the formation path, frequency composition, amplitude and phase states, and intra-frame occurrence positions of this periodic or quasi-periodic background noise differ from transmitter-receiver crosstalk, making it difficult to directly characterize using a crosstalk model between the transmission and receiving paths. Furthermore, compensation based on the digital data output from the receiving link typically occurs after analog-to-digital conversion, resulting in a timing lag in the compensation process compared to the background noise entering the analog receiving link. When background noise has already caused compression in the intermediate frequency amplification unit, occupancy of the dynamic range of the analog-to-digital conversion input, or clipping of the sampling signal, subsequent digital compensation, while reducing the corresponding interference components in the digital data, struggles to recover the target echo information lost during the analog receiving link and analog-to-digital conversion process.

[0005] During continuous operation of millimeter-wave radar, changes in power conversion status, digital clock activity status, communication interface operation status, installation structure, and grounding status can alter the frequency, amplitude, phase, or intra-frame location of background noise. If the sampling data used for background noise identification and compensation effect evaluation includes valid target echoes, or if the sampling data before and after compensation correspond to different intra-frame locations and different interference source operating states, the changes between the sampling data may simultaneously include changes in target echoes, scene changes, and background noise, easily leading to inaccurate background noise identification results or incorrect updates to compensation control parameters.

[0006] Furthermore, existing digital compensation is typically performed after the received signal has undergone analog-to-digital conversion. When background noise has already caused compression of the analog receiver link, occupancy of the dynamic range of the analog-to-digital conversion input, or clipping of the sampled signal, subsequent digital compensation struggles to recover the target echo information lost before the analog-to-digital conversion. Therefore, relying solely on radar data after analog-to-digital conversion for crosstalk estimation and compensation is insufficient to accommodate background noise generated by internal periodic electromagnetic interference sources, whose amplitude and phase states vary with operating conditions.

[0007] Therefore, how to reduce the interference of effective target echoes and changes in radar frame operation status on background noise identification and compensation effect evaluation during the normal operation of millimeter-wave radar, and how to implement reliable adaptive compensation for background noise related to periodic electromagnetic interference sources inside the millimeter-wave radar and having cross-frame repeating amplitude and phase characteristics, have become technical problems that need to be solved. Summary of the Invention

[0008] To overcome the aforementioned technical deficiencies, the present invention aims to provide a millimeter-wave radar noise compensation circuit and method based on idle sampling windows. The present invention establishes corresponding idle sampling window groups within consecutive radar frames that share the same intra-frame sampling position identifier and have no effective targets. It identifies the background noise components to be compensated by combining the timing of internal periodic electromagnetic interference sources, repetition confidence, and timing correlation. Based on injection path calibration data, it selects the disturbed nodes and implements amplitude-reverse compensation. Finally, it uses the residual noise and repetition confidence from subsequent corresponding idle sampling windows to feedback and correct the compensation control parameters.

[0009] This invention discloses a millimeter-wave radar noise compensation method based on an idle sampling window, comprising:

[0010] The radar frame timing information, target detection information, and interference source timing information corresponding to the internal periodic electromagnetic interference source of the millimeter-wave radar are acquired. The internal periodic electromagnetic interference source includes at least one of the following: power conversion unit, digital clock unit, and communication interface.

[0011] Based on radar frame timing information and target detection information, corresponding idle sampling window groups with the same intra-frame sampling position identifier are established in multiple consecutive radar frames, and background noise samples are obtained in the corresponding idle sampling window groups.

[0012] The frequency, amplitude, phase, and repetition confidence of candidate background noise components are determined based on the background noise samples, and the temporal correlation is determined based on the frequency correspondence and intra-frame time correspondence between the candidate background noise components and the temporal information of the interference source.

[0013] When the repetition confidence reaches the repetition confidence threshold and the temporal correlation reaches the temporal correlation threshold, the candidate background noise component is determined as the background noise component to be compensated. Based on the background noise component to be compensated and the injection path calibration data corresponding to multiple candidate disturbed nodes, the disturbed node is selected from multiple candidate disturbed nodes, and the compensation control parameters are determined.

[0014] A compensation signal is generated based on the compensation control parameters. The compensation signal is injected into the disturbed node through the compensation injection unit so that the phase difference between the phase of the compensation signal after it reaches the disturbed node and the phase of the background noise component to be compensated at the disturbed node is not greater than the phase deviation threshold of 180°.

[0015] In subsequent radar frames, residual noise samples are acquired within the corresponding idle sampling windows that have intra-frame sampling position identifiers. When the target detection information indicates that the corresponding idle sampling window is in a state without effective targets and the temporal correlation degree corresponding to the corresponding idle sampling window reaches the temporal correlation degree threshold, the compensation control parameters are maintained or corrected based on the background noise samples and residual noise samples, and the upper limit of the compensation amplitude and the parameter update step size are determined based on the repetition confidence.

[0016] In one implementation, establishing a corresponding idle sampling window group includes:

[0017] Based on the radar frame timing information, at least one of the following is determined as a candidate sampling interval: the transmission silence interval, the sampling interval corresponding to the range gate where no effective target output is formed, and the sampling interval marked as a non-target interval by the target detection information;

[0018] Configure an intra-frame sampling position identifier for each radar frame candidate sampling interval, consisting of at least one of a range gate identifier and an intra-frame time position;

[0019] Candidate sampling intervals in multiple consecutive radar frames that have the same intra-frame sampling position identifier, target detection information indicating that there is no effective target, and intra-frame position deviation of the corresponding noise spectrum peak is not greater than the position deviation threshold are assigned to the same corresponding idle sampling window group.

[0020] When the target detection information indicates that any candidate sampling interval is in a valid target state, a target echo leakage state, or a strong reflection clutter state, the corresponding candidate sampling interval is removed from the corresponding idle sampling window group.

[0021] In one implementation, determining the repetition confidence and temporal correlation includes:

[0022] Background noise samples are time-aligned across radar frames according to the intra-frame sampling position identifier. After time alignment, DC component removal, outlier sampling point removal, and complex spectrum transformation are performed on the background noise samples.

[0023] Based on the results of the complex spectrum transformation, determine the cross-frame frequency deviation, amplitude dispersion, cross-frame phase deviation, and intra-frame position deviation of each candidate background noise component in multiple consecutive radar frames;

[0024] The repetition reliability is determined based on cross-frame frequency deviation, amplitude dispersion, cross-frame phase deviation, and intra-frame position deviation.

[0025] Determine the interference source frequency deviation between the frequency of the candidate background noise component and the fundamental frequency, harmonic frequency, or sub-frequency of the internal periodic electromagnetic interference source, and determine the interference time deviation between the intra-frame time position of the candidate background noise component and the interference action time period represented by the interference source timing information.

[0026] The temporal correlation degree is determined based on the frequency deviation of the interference source and the time deviation of the interference period.

[0027] In one implementation, determining the upper limit of the compensation magnitude and the parameter update step size based on the repetition confidence level includes:

[0028] The range of repeatability confidence is divided into a stop compensation interval, a parameter hold interval, and a parameter update interval;

[0029] When the repetition confidence level is in the stop compensation range, stop injecting compensation signals into the disturbed node;

[0030] When the repeatability confidence is within the parameter holding range, maintain the previous set of stable compensation control parameters and reduce the upper limit of the compensation amplitude according to the preset amplitude attenuation amount;

[0031] The parameter update interval is divided into multiple consecutive confidence sub-intervals. When the repeat confidence is within the parameter update interval, the upper limit of the compensation range and the parameter update step size are determined according to the confidence sub-interval to which the repeat confidence belongs.

[0032] The range of intra-frame position deviation is divided into multiple position deviation intervals, and a compensation time extension is configured for each position deviation interval. The compensation time is determined according to the position deviation interval to which the intra-frame position deviation of the background noise component to be compensated belongs and the corresponding compensation time extension, so that the compensation time is located outside the adjacent effective target echo occupied interval.

[0033] In one implementation, obtaining injection path calibration data and selecting disturbed nodes includes:

[0034] During the compensation calibration phase, calibration signals at multiple calibration frequencies are injected through multiple candidate disturbed nodes, including intermediate frequency bias nodes, analog-to-digital conversion reference nodes, and grounding reference sensitive nodes.

[0035] Obtain calibration sampling data for each candidate disturbed node before and after the injection of calibration signal, and determine the path amplitude transfer, path phase offset, linear injection range, additional noise of compensation branch, and upper limit of safe compensation amplitude for each candidate disturbed node based on the calibration sampling data.

[0036] The path amplitude transfer amount, path phase offset, linear injection range, additional noise amount of compensation branch, and upper limit of safe compensation amplitude are associated with the candidate disturbed node identifier and calibration frequency to form injection path calibration data;

[0037] Based on the amplitude of the background noise component to be compensated and the path amplitude propagation amount corresponding to each candidate disturbed node, the candidate compensation amplitude corresponding to each candidate disturbed node is determined respectively.

[0038] Based on the amplitude and phase of the background noise component to be compensated, the candidate compensation amplitude corresponding to each candidate disturbed node, the path amplitude transfer amount, and the path phase offset, the estimated residual noise amount corresponding to each candidate disturbed node is determined respectively.

[0039] From the candidate disturbed nodes whose deviation between the calibrated frequency and the frequency of the background noise component to be compensated is not greater than the frequency matching threshold, whose candidate compensation amplitude is within the linear injection range and the upper limit of the safe compensation amplitude, and whose additional noise amount of the compensation branch is not greater than the additional noise threshold, the candidate disturbed node with the smallest estimated residual noise amount is selected as the disturbed node.

[0040] In one implementation, determining and executing compensation control parameters includes:

[0041] Determine the compensation frequency control amount based on the frequency of the background noise component to be compensated;

[0042] The compensation amplitude control amount is determined based on the amplitude of the background noise component to be compensated and the path amplitude propagation amount corresponding to the disturbed node, and the compensation amplitude control amount is limited to the linear injection range, the safe compensation amplitude upper limit, and the compensation amplitude upper limit determined based on the repetition confidence.

[0043] The compensation phase control quantity is determined based on the phase of the background noise component to be compensated, the 180° phase reversal, and the path phase offset corresponding to the disturbed node.

[0044] A reference compensation signal is generated based on the compensation frequency control quantity. The reference compensation signal is then subjected to amplitude adjustment, phase adjustment, impedance isolation, DC isolation, amplitude limiting and injection control in sequence through a controllable amplitude adjustment unit, a controllable phase adjustment unit, a buffer isolation unit, a DC isolation unit, a limiting protection unit and a compensation injection switch to form a compensation signal. The buffer isolation unit, the DC isolation unit, the limiting protection unit and the compensation injection switch constitute the compensation injection unit.

[0045] The compensation period is determined based on the intra-frame sampling position identifier, and the compensation injection switch is controlled to inject the compensation signal into the disturbed node during the compensation period.

[0046] In one implementation, correcting the compensation control parameters includes:

[0047] Under the compensation amplitude for phase search below the upper limit of the compensation amplitude, the compensation signal is injected sequentially according to multiple candidate compensation phase control quantities, and residual noise samples are obtained in the corresponding idle sampling windows with the same intra-frame sampling position identifier in subsequent radar frames.

[0048] The residual noise energy is determined based on the residual noise sample corresponding to each candidate compensation phase control quantity, and a phase-residual noise energy sequence is generated according to each candidate compensation phase control quantity and the corresponding residual noise energy.

[0049] If there is a local minimum of residual noise energy in the phase-residual noise energy sequence, and the deviation of the candidate compensation phase control quantity corresponding to the local minimum in multiple consecutive radar frames is not greater than the phase stability threshold, the candidate compensation phase control quantity corresponding to the local minimum is determined as the compensation phase control quantity to be used.

[0050] Under the waiting compensation phase control quantity, the compensation signal is injected sequentially according to multiple candidate compensation amplitude control quantities, and the candidate compensation amplitude control quantity with the smallest corresponding residual noise energy is determined as the waiting compensation amplitude control quantity.

[0051] The direction of subsequent parameter updates is determined based on the difference in residual noise energy before and after using the standby compensation phase control quantity and the standby compensation amplitude control quantity, and the compensation control parameters are corrected according to the parameter update step size.

[0052] In one embodiment, the method further includes:

[0053] When the background noise samples before compensation and the residual noise samples after compensation used for comparison come from corresponding idle sampling windows with the same intra-frame sampling position identifier, and the target detection information of the two corresponding idle sampling windows both indicate that there are no effective targets, and the temporal correlation degree of the two corresponding idle sampling windows both reach the temporal correlation degree threshold, the residual noise samples after compensation are used to correct the compensation control parameters.

[0054] If the target detection information of any corresponding idle sampling window indicates the existence of a valid target, or if the temporal correlation degree corresponding to any corresponding idle sampling window is lower than the temporal correlation degree threshold, discard the corresponding compensated residual noise sample and maintain the previous set of stable compensation control parameters.

[0055] If there is no local minimum in the phase-residual noise energy sequence, or if the deviation of the candidate compensation phase control quantity corresponding to the local minimum is greater than the phase stability threshold in multiple consecutive radar frames, the background noise component to be compensated is marked as an uncompensable noise component, and the compensation injection switch is switched to the off state.

[0056] If the residual noise energy after the compensation control parameter update is greater than the residual noise energy before the compensation control parameter update, restore the compensation control parameters before the update and reduce the parameter update step size.

[0057] In one implementation, when there are multiple background noise components to be compensated that satisfy both the repetition confidence threshold and the temporal correlation threshold, the method further includes:

[0058] Based on the repetition confidence, amplitude, temporal correlation with the corresponding internal periodic electromagnetic interference source, intra-frame time position, and injection path calibration data corresponding to multiple candidate disturbed nodes of each background noise component to be compensated, the compensation priority of each background noise component to be compensated and the corresponding disturbed node are determined.

[0059] When multiple background noise components to be compensated correspond to the same disturbed node and the corresponding compensation periods overlap, a combined compensation signal containing multiple corresponding frequency components is generated, and the compensation amplitude control amount and compensation phase control amount of each corresponding frequency component are determined respectively.

[0060] When multiple background noise components to be compensated correspond to different disturbed nodes, each background noise component to be compensated is assigned to a compensation channel connected to the corresponding disturbed node.

[0061] When the number of available compensation channels is less than the number of background noise components to be compensated, multiple background noise components to be compensated are compensated in a time-division manner according to the compensation priority. After each switching of background noise components to be compensated, the corresponding idle sampling window determined by the intra-frame sampling position identifier corresponding to the switched background noise component to be compensated is used to re-determine the corresponding repetition confidence and temporal correlation.

[0062] The present invention also provides a millimeter-wave radar noise compensation circuit based on an idle sampling window, including a millimeter-wave radar receiving link, a radar frame timing interface, a target detection interface, an interference source timing interface, an idle sampling window matching unit, a noise acquisition unit, a background noise parameter estimation unit, an injection path calibration and storage unit, an interference node selection unit, a control unit, a compensation signal generation unit, an amplitude and phase adjustment unit, and a compensation injection unit.

[0063] The millimeter-wave radar receiving link is equipped with a monitoring node and multiple candidate interference nodes, including an intermediate frequency bias node, an analog-to-digital conversion reference node, and a ground reference sensitive node.

[0064] The radar frame timing interface is used to acquire radar frame timing information, the target detection interface is used to acquire target detection information, and the interference source timing interface is used to acquire interference source timing information corresponding to the internal periodic electromagnetic interference source.

[0065] The radar frame timing interface, target detection interface, and interference source timing interface are respectively connected to the idle sampling window matching unit. The idle sampling window matching unit is respectively connected to the sampling control terminal and the control unit of the noise acquisition unit. The idle sampling window matching unit is used to establish corresponding idle sampling window groups with the same intra-frame sampling position identifier in multiple consecutive radar frames according to the radar frame timing information and target detection information, and output the target status and intra-frame sampling position identifier of the corresponding idle sampling window group to the control unit.

[0066] The input of the noise acquisition unit is connected to the monitoring node, and the output of the noise acquisition unit is connected to the background noise parameter estimation unit. The background noise parameter estimation unit is connected to the idle sampling window matching unit, the interference source timing interface, and the control unit, respectively. The noise acquisition unit is used to acquire background noise samples within the corresponding idle sampling window group. The background noise parameter estimation unit is used to generate the frequency, amplitude, phase, and repetition confidence of candidate background noise components based on the background noise samples, and to determine the timing correlation degree based on the frequency correspondence and intra-frame time correspondence between the candidate background noise components and the timing information of the interference source.

[0067] The injection path calibration storage unit stores the path amplitude transfer amount, path phase offset, linear injection range, additional noise amount of compensation branch, and upper limit of safe compensation amplitude corresponding to each candidate disturbed node. The disturbed node selection unit is connected to the background noise parameter estimation unit, the injection path calibration storage unit, and the control unit, respectively, and is used to determine one candidate disturbed node as the disturbed node from multiple candidate disturbed nodes based on the candidate background noise components and the injection path calibration data in the injection path calibration storage unit.

[0068] The output of the compensation signal generation unit is connected in sequence to the amplitude and phase adjustment unit and the compensation injection unit. The compensation injection unit is connected to multiple candidate disturbed nodes through multiple selectable compensation injection paths. The amplitude and phase adjustment unit includes a controllable amplitude adjustment unit and a controllable phase adjustment unit. The compensation injection unit includes a buffer isolation unit, a DC isolation unit, a limiting protection unit, and a compensation injection switch corresponding to the multiple selectable compensation injection paths.

[0069] The control unit is connected to the disturbed node selection unit, the compensation signal generation unit, the controllable amplitude adjustment unit, the controllable phase adjustment unit, and the compensation injection switch, respectively. It is used to determine the compensation control parameters based on the candidate background noise components and the injection path calibration data corresponding to the disturbed node when the repetition confidence reaches the repetition confidence threshold and the temporal correlation reaches the temporal correlation threshold. It controls the corresponding compensation injection path to be turned on based on the disturbed node selection result output by the disturbed node selection unit, and determines the upper limit of the compensation amplitude and the parameter update step size based on the repetition confidence.

[0070] The noise acquisition unit is also used to acquire residual noise samples from the corresponding idle sampling window with intra-frame sampling position identifier in subsequent radar frames after compensation is performed. The background noise parameter estimation unit is also used to determine the noise energy before compensation based on the background noise samples and the residual noise energy based on the residual noise samples. The control unit is also used to maintain or correct the compensation control parameters based on the noise energy before compensation and the residual noise energy when the target state indicates that the corresponding idle sampling window is in a state without effective targets and the temporal correlation degree reaches the temporal correlation degree threshold.

[0071] Compared with existing technologies, the above technical solution has the following advantages:

[0072] 1. Based on radar frame timing information and target detection information, this invention establishes corresponding idle sampling window groups with the same intra-frame sampling position identifier in multiple consecutive radar frames, so that the acquisition of background noise samples avoids effective target echoes, and the noise samples before and after compensation correspond to the same distance position and intra-frame working period, thereby reducing the interference of target echoes and changes in radar frame operating status on background noise identification and compensation effect evaluation.

[0073] 2. This invention determines the repetition confidence based on the frequency, amplitude, phase, and intra-frame position of candidate background noise components, and determines the temporal correlation by combining the frequency correspondence and intra-frame time correspondence between candidate background noise components and internal periodic electromagnetic interference sources. This enables the identification of periodic or quasi-periodic background noise with cross-frame repetition characteristics from random noise, sporadic noise, and scene change signals, reducing the possibility of false activation of compensation branches.

[0074] 3. This invention calibrates the path amplitude transfer, path phase offset, linear injection range, additional noise of the compensation branch, and upper limit of the safe compensation amplitude of multiple candidate disturbed nodes. Based on the background noise component to be compensated, the disturbed node that matches the coupling path of the background noise can be selected. This can avoid injecting the compensation signal into an unsuitable circuit node and improve the amplitude matching degree between the compensation signal and the background noise component.

[0075] 4. This invention generates a compensation signal based on the background noise component to be compensated and the injection path calibration data, and injects the compensation signal into the disturbed node through an amplitude and phase adjustment unit and a compensation injection unit. This allows the compensation signal to form an inverse amplitude and phase relationship with the background noise component to be compensated after reaching the disturbed node. This can reduce the amplitude of the background noise in the millimeter-wave radar receiving link before the background noise enters subsequent digital processing, and reduce the occupation of the linear range and dynamic range of the receiving link by the background noise.

[0076] 5. The present invention acquires residual noise samples in the corresponding idle sampling windows with the same intra-frame sampling position identifier in subsequent radar frames, and corrects the compensation control parameters only when there is no effective target state and the timing correlation conditions remain effective, so that the residual noise feedback has consistent sampling conditions, which can reduce the error update of compensation parameters caused by target appearance, change of interference source state or change of sampling position.

[0077] 6. This invention controls the start / stop state of the compensation signal, the upper limit of the compensation amplitude, and the parameter update step size based on the repetition reliability. It also employs phase search, amplitude search, parameter hold, parameter backoff, and uncompensable noise exit mechanisms to match the compensation intensity and parameter adjustment speed with the repetition stability of the background noise, thereby improving the convergence stability and operational safety of the closed-loop compensation process.

[0078] 7. This invention can target multiple background noise components to be compensated, and adopt multi-frequency combination compensation, parallel compensation of multiple compensation channels, or time-division compensation according to the disturbed node and compensation period. After the compensation object is switched, the repetition reliability and temporal correlation are reconfirmed, thereby adapting to application scenarios where there are multiple periodic electromagnetic interference sources or multiple noise frequency components inside the millimeter-wave radar. Attached Figure Description

[0079] Figure 1 This is a schematic diagram of the overall structure of the millimeter-wave radar noise compensation circuit based on the idle sampling window in an embodiment of the present invention;

[0080] Figure 2 This is a schematic diagram of the process for determining the idle sampling window, estimating the background noise parameters, and controlling the amplitude and phase compensation closed loop in an embodiment of the present invention.

[0081] Figure 3 This is a schematic diagram comparing the target echo component, the pre-compensation sampled signal, and the post-compensation sampled signal in an embodiment of the present invention. Figure 3 (a) is a schematic diagram of the time-domain waveform of the target echo component in an embodiment of the present invention. Figure 3 (b) is a schematic diagram of the time-domain waveform of the sampled signal before background noise compensation in an embodiment of the present invention. Figure 3 (c) is a schematic diagram of the time-domain waveform of the sampled signal after background noise compensation in an embodiment of the present invention. Detailed Implementation

[0082] The following provides a detailed description of a millimeter-wave radar noise compensation method based on an idle sampling window. This embodiment is used to illustrate the specific implementation process of the millimeter-wave radar noise compensation method and does not constitute a limitation on the millimeter-wave radar operating frequency band, radar frame structure, sampling frequency, device model, location of disturbed nodes, compensation frequency range, or compensation parameter values. While maintaining the concepts of idle sampling window background noise perception, internal periodic electromagnetic interference source correlation identification, amplitude and phase compensation of disturbed nodes, and residual noise closed-loop feedback, those skilled in the art can adjust the corresponding parameters and hardware structure according to the millimeter-wave radar's operating mode, receiver link structure, type of internal periodic electromagnetic interference source, and product electromagnetic compatibility requirements.

[0083] The millimeter-wave radar described in this embodiment refers to a radar device that uses millimeter-wave frequency electromagnetic waves to transmit and receive target echoes, and obtains target range, velocity, or azimuth information based on the target echoes. The millimeter-wave radar can operate in a frequency-modulated continuous wave mode, or it can employ other operating modes with periodic radar frame timing and analog-to-digital conversion sampling timing. The millimeter-wave radar can be applied to scenarios such as vehicle target detection, industrial ranging, security monitoring, environmental perception of unmanned equipment, or indoor human status detection.

[0084] The millimeter-wave radar receiving link described in this embodiment refers to the signal transmission path formed by the target echo entering the millimeter-wave radar from the receiving antenna, and sequentially passing through the RF receiving front-end, mixing unit, intermediate frequency amplification unit, filtering unit, analog-to-digital conversion unit, and digital data interface. Depending on the specific structure of the millimeter-wave radar chip, the millimeter-wave radar receiving link may include at least one of a low-noise amplifier, a quadrature mixer, a variable gain amplifier, an anti-aliasing filter, and an input buffer.

[0085] The background noise described in this embodiment refers to periodic or quasi-periodic noise, excluding target echoes and unpredictable random thermal noise, that enters the millimeter-wave radar receiving link through spatial coupling paths, conducted coupling paths, common impedance paths, or reference potential paths, and exhibits repetitive frequency, amplitude, phase, or intra-frame position characteristics in multiple consecutive radar frames. This background noise can be generated by power conversion units, digital clock units, communication interfaces, digital processing units, wiring harnesses, or grounding structures. For periodic background noise formed by external spatial electromagnetic coupling, it is only included in the determination process of the background noise component to be compensated when the corresponding synchronization trigger signal, periodic identifier, or equivalent interference source timing information is available; external background noise for which corresponding timing information cannot be obtained is only monitored and does not drive the timing correlation compensation process in this embodiment.

[0086] The internal periodic electromagnetic interference source described in this embodiment refers to a functional unit located inside the millimeter-wave radar, or sharing power supply, clock, communication, or grounding structures with the millimeter-wave radar, and periodically generating electromagnetic disturbances according to a determined or measurable operating sequence. The internal periodic electromagnetic interference source includes at least one of a power conversion unit, a digital clock unit, and a communication interface. The power conversion unit can be a switching power supply, a DC-DC converter, or a power management chip; the digital clock unit can be a master control clock, a data conversion clock, or a serial bus clock; the communication interface can be a serial peripheral interface, an integrated circuit bus interface, a controller area network interface, an Ethernet interface, or a universal asynchronous transceiver interface.

[0087] See Figure 1The millimeter-wave radar system used in this embodiment includes a millimeter-wave radar receiving link, a radar frame timing interface, a target detection interface, an interference source timing interface, an idle sampling window matching unit, a noise acquisition unit, a background noise parameter estimation unit, an injection path calibration and storage unit, a disturbed node selection unit, a control unit, a compensation signal generation unit, an amplitude and phase adjustment unit, and a compensation injection unit.

[0088] The millimeter-wave radar receiving link is equipped with monitoring nodes and multiple candidate disturbed nodes. The monitoring nodes are nodes used to acquire voltage signals or digital sampling data corresponding to the background noise in the millimeter-wave radar receiving link. These monitoring nodes can be intermediate frequency output nodes, analog-to-digital converter input nodes, analog-to-digital converter output data interfaces, bias monitoring nodes, or ground reference monitoring nodes.

[0089] The candidate disturbed nodes refer to nodes that, through circuit analysis or calibration, are confirmed to be potentially susceptible to background noise coupling and capable of receiving compensation signals. These candidate disturbed nodes include intermediate frequency (IF) bias nodes, analog-to-digital (ADC) conversion reference nodes, and ground reference sensitive nodes. IF bias nodes are nodes used to provide a DC operating point for IF amplifier circuits or input buffer circuits; ADC conversion reference nodes are nodes used to provide a reference voltage or potential for the ADC conversion process; and ground reference sensitive nodes are ground nodes whose potential changes can couple to the millimeter-wave radar receiving link and affect the IF signal or the ADC conversion result.

[0090] The disturbed node refers to the actual compensation injection node selected from multiple candidate disturbed nodes based on the injection path calibration data and the frequency, amplitude, and phase of the background noise component to be compensated. Candidate disturbed nodes represent the set of nodes available before compensation, while disturbed nodes represent the nodes that have been selected and used for actual injection of compensation signals during the compensation process corresponding to the current background noise component to be compensated.

[0091] Figure 1 The millimeter-wave radar receiving link, disturbed node, noise acquisition unit, background noise parameter estimation unit, control unit, compensation signal generation unit, amplitude and phase adjustment unit, and compensation injection unit shown constitute a closed-loop signal flow. The compensated residual noise can be acquired again by the noise acquisition unit in the corresponding idle sampling window of subsequent radar frames and sent to the background noise parameter estimation unit and control unit. Figure 1 This represents both the forward injection relationship of the compensation signal from the compensation signal generation unit to the disturbed node and the feedback relationship of the residual noise from the millimeter-wave radar receiving link back to the control unit.

[0092] In this embodiment, the millimeter-wave radar operates in a frequency-modulated continuous wave (FM) mode, with a frequency band of 77 GHz to 81 GHz. The single radar frame period is 40 ms, including a 32 ms effective transmission and target echo acquisition period, and an 8 ms interval between transmission silence and data processing. Each radar frame consists of 64 linear frequency modulated (LFM) pulses, and each LFM pulse includes 256 analog-to-digital (ADC) sampling points. These parameters are only used to illustrate one specific implementation and do not constitute a limitation on the radar frame period, the number of LFM pulses, or the number of ADC sampling points.

[0093] The actual sampled signal in the millimeter-wave radar receiver link consists of the target echo component, the background noise component, and the random noise component. The actual sampled signal is represented as follows:

[0094] (1)

[0095] In the formula, Indicates time The actual sampled signal, in volts; Indicates time The target echo component, in volts; Indicates time The background noise component, in volts; Indicates time The random noise component, in volts; The time is expressed in seconds, based on the radar frame synchronization time. The target echo component is the effective signal formed by the reflection of the millimeter-wave radar signal after it is transmitted by the target; the background noise component has a frequency, amplitude, phase, or position within a frame that can be estimated in multiple consecutive radar frames; the random noise component does not have a stable predictable amplitude-phase and time relationship, and therefore is not used as an active amplitude-phase compensation target.

[0096] The radar frame timing interface acquires radar frame timing information. This radar frame timing information refers to data describing the start and end times of each operating phase within a radar frame period for the millimeter-wave radar. It includes at least one of the following: radar frame start time, start time of each linear frequency modulation (LFM) pulse, transmit enable period, analog-to-digital conversion (ADC) sampling enable period, transmit silence period, signal processing period, and radar frame end time. The radar frame timing information can be obtained from the millimeter-wave radar chip's frame synchronization pin, LFM pulse synchronization pin, ADC sampling enable signal, or on-chip registers.

[0097] The target detection interface acquires target detection information. This target detection information refers to data indicating the presence of valid target echoes at different range gates, different Doppler units, or different intra-frame time intervals. It includes at least one of the following: range gate occupancy status, target detection flag, target confidence level, target echo amplitude, strong reflection clutter flag, and target echo leakage flag.

[0098] The term "valid target state" refers to the detection results within the corresponding sampling interval meeting the target output conditions of the millimeter-wave radar. The term "no valid target state" refers to the absence of detection results meeting the target output conditions within the corresponding sampling interval. "No valid target state" is not equivalent to a zero sampling data amplitude; rather, it indicates that the sampling data was not recognized by the target detection and processing unit as a valid target requiring output.

[0099] The target echo leakage state refers to the state where the effective target echo extends into adjacent sampling intervals due to filter sidelobes, range spectrum leakage, time window edge effects, or analog link transition response. The strong reflection clutter state refers to the state where the signal energy generated by reflections from fixed objects, nearby highly reflective objects, or installation structures exceeds a preset clutter threshold, but is not considered as the current target output. Although target echo leakage and strong reflection clutter may not form an effective target output, they may still affect background noise estimation and are therefore not used to establish corresponding idle sampling window groups.

[0100] The interference source timing interface acquires interference source timing information. This interference source timing information refers to the time data used to describe the switching actions, clock flips, data transmissions, current jumps, or other electromagnetic disturbances of an internal periodic electromagnetic interference source within a radar frame period. The interference source timing information can be obtained from the pulse width modulation synchronization signal of the power conversion unit, the clock gating signal of the digital clock unit, the data transmission enable signal of the communication interface, or the device runtime timing recorded in the control unit.

[0101] See Figure 2 Upon system power-up, radar frame synchronization is performed first. Radar frame synchronization refers to mapping target detection information, interference source timing information, and background noise sampling data to the same intra-frame time coordinate, using the start time of the radar frame as a common zero point. Through radar frame synchronization, it is possible to determine whether a certain background noise component appears at the same or similar intra-frame time position in different radar frames, and whether the background noise component matches the action period of the corresponding internal periodic electromagnetic interference source.

[0102] The control unit determines candidate sampling intervals based on radar frame timing information and target detection information. A candidate sampling interval refers to a sampling period or range gate interval that has not yet completed cross-frame consistency verification but preliminarily meets the background noise sampling conditions in the current radar frame. The candidate sampling interval can consist of at least one of the following: the transmission silence interval, the range gate sampling interval where no effective target output has been formed, and the sampling interval where the target detection information is marked as a non-target interval.

[0103] The transmission silence interval refers to the time interval during which the millimeter-wave radar transmission link does not output millimeter-wave transmission signals for target detection, or the transmission power is lower than a preset transmission threshold. The range gate corresponding sampling interval that does not form a valid target output refers to the sampling data interval corresponding to one or more range gates after range dimension processing, where the target detection information indicates that the range gate has not formed a valid target output. The non-target interval refers to the sampling interval explicitly marked by the target detection processing unit as not participating in the current valid target output based on the target mask, detection result, or tracking result.

[0104] Configure intra-frame sampling position identifiers for each candidate sampling interval. The intra-frame sampling position identifier is identification data used to distinguish the spatial and temporal positions of the candidate sampling interval in the radar frame, and can be composed of at least one of the following: range gate identifier, linear frequency modulated pulse sequence number, analog-to-digital conversion sampling point interval, and intra-frame temporal position.

[0105] In this embodiment, the intra-frame sampling position identifier is composed of a range gate identifier and an intra-frame time position. For example, for the 36th range gate located in the candidate sampling interval of 12.4ms to 12.8ms after the start of the radar frame, the intra-frame sampling position identifier can be recorded as "range gate 36 - intra-frame time 12.4ms to 12.8ms". The intra-frame sampling position identifier is used to ensure that the background noise samples before compensation and the residual noise samples after compensation originate from the same range position and the same millimeter-wave radar operating stage, thereby avoiding misinterpreting changes in the target scene or changes in the millimeter-wave radar operating state as compensation effects.

[0106] Before establishing the corresponding idle sampling window group, the noise acquisition unit first performs presampling on each candidate sampling interval. The background noise parameter estimation unit then determines the noise spectral peak and intra-frame position in the candidate sampling interval based on the presampling data. Specifically, each candidate sampling interval is divided into multiple overlapping time sub-windows, and complex spectral transformation or orthogonal accumulation at preset frequencies is performed on each time sub-window to obtain the narrowband energy of each time sub-window at the candidate background noise frequency. The center time of the time sub-window with the largest narrowband energy is determined as the intra-frame position of the noise spectral peak.

[0107] For the Each radar frame divides the candidate sampling interval into segments of length [length missing]. Each sampling point, with an adjacent time window interval of [number] times. The time sub-window of the sampling point, the first The complex spectral values ​​of each time window are represented as follows:

[0108] (2)

[0109] In the formula, Indicates the first The first radar frame The time window at the Complex spectral values ​​at each frequency unit; Indicates the radar frame number; Indicates the time sub-window sequence number; Indicates the frequency unit number; Indicates the first The presampled background noise values ​​of the corresponding sampling points in each radar frame are expressed in volts. This indicates the sampling point interval between adjacent time windows; Indicates the sampling point number within the time sub-window; Indicates the first The time sub-window function values ​​of each sampling point; Indicates the length of the time sub-window; Represents the imaginary unit; This represents the natural exponential function.

[0110] For the The candidate background noise component, at the _n_th ... The intra-frame position of the spectral peak in each radar frame is represented as follows:

[0111] (3)

[0112] In the formula, Indicates the first The candidate background noise components in the first The intra-frame position of the spectral peak in each radar frame, in seconds; Indicates the candidate background noise component number; Indicates the corresponding candidate sampling interval in the th order. The start time in each radar frame, in seconds; Indicates the first The sequence number of the narrowband energy maximum time sub-window corresponding to each candidate background noise component; This indicates the sampling point interval between adjacent time windows; Indicates the length of the time sub-window; This indicates the sampling frequency of the noise acquisition unit, measured in Hertz.

[0113] The time window number with the highest narrowband energy is represented as:

[0114] (4)

[0115] In the formula, This indicates the time window number that maximizes the narrowband energy. Indicates the sub-window number that participated in the search; Indicates the first The first radar frame The time window at the The complex spectral values ​​of each candidate background noise component at the corresponding frequency unit; Indicates the first The candidate background noise components in the first The corresponding frequency unit number in each radar frame; Represents the square of the modulus of a complex number; This represents the independent variable that makes the subsequent function reach its maximum value.

[0116] In other implementations, orthogonal demodulation can be performed on the candidate background noise frequencies to obtain the corresponding complex envelope, and the sampling time when the amplitude of the complex envelope reaches its maximum value can be determined as the intra-frame position of the noise spectrum peak.

[0117] The control unit searches for candidate sampling intervals with the same intra-frame sampling position identifier in multiple consecutive radar frames. When candidate sampling intervals in multiple consecutive radar frames have the same intra-frame sampling position identifier, the target detection information all indicate a state of no valid target, and the intra-frame position deviation of the corresponding noise spectrum peak is not greater than the position deviation threshold, the candidate sampling interval is assigned to the same corresponding idle sampling window group.

[0118] The corresponding idle sampling window refers to a sampling interval in different radar frames that has the same intra-frame sampling position identifier and satisfies the condition of no valid target. The corresponding idle sampling window group refers to a data set composed of multiple corresponding idle sampling windows in multiple consecutive radar frames. The corresponding idle sampling window group is not a window in a single radar frame, but a set of sampling windows at the same position established across multiple radar frames.

[0119] In this embodiment, a corresponding idle sampling window group is established using five consecutive radar frames, and the position deviation threshold is set to 20 μs. The number of consecutive radar frames can be set from 3 to 20 frames, depending on the target's movement speed, radar frame period, and background noise stability. The position deviation threshold can be determined based on radar frame synchronization jitter, interference source-triggered jitter, and time sub-window positioning error. In one calibration method, the intra-frame positions of the spectral peaks of at least 100 consecutive radar frames are recorded when the internal periodic electromagnetic interference source is in a stable operating state. The 95th percentile value of the corresponding position deviation statistical distribution, plus the radar frame synchronization error, is used as the position deviation threshold.

[0120] When any candidate sampling interval exhibits a valid target state, a target echo leakage state, or a strong reflection clutter state, the control unit removes the corresponding candidate sampling interval from the corresponding idle sampling window group. For candidate sampling intervals that do not generate valid target output in multiple consecutive radar frames, but whose spectral peak positions continuously change with the target's trajectory, these intervals are also not considered as corresponding idle sampling windows. This reduces the possibility of weak target echoes, target sidelobes, or motion clutter being incorrectly identified as background noise.

[0121] The noise acquisition unit can reuse the analog-to-digital conversion sampling resources in the millimeter-wave radar receiving link, or it can use an independent low-speed monitoring channel. Reusing analog-to-digital conversion sampling resources means directly extracting background noise samples from the data output by the original analog-to-digital conversion unit of the millimeter-wave radar according to the intra-frame sampling position identifier, without needing to set up an independent analog-to-digital converter. Using an independent low-speed monitoring channel means setting up an independent sampling circuit at the monitoring node to obtain monitoring data characterizing the background noise at a sampling frequency lower than the sampling frequency of the main receiving link of the millimeter-wave radar.

[0122] In the implementation of reusing analog-to-digital conversion sampling resources, the control unit enables data extraction according to the intra-frame sampling position identifier and reads background noise samples from the corresponding idle sampling window. In the implementation using an independent low-speed monitoring channel, the independent low-speed monitoring channel can be connected to an intermediate frequency monitoring node, an analog-to-digital conversion reference node, a ground reference sensitive node, or an offset monitoring node.

[0123] The background noise sample refers to the time-domain discrete sampling data obtained by the noise acquisition unit within the corresponding idle sampling window. For the first... The corresponding idle sampling window in each radar frame, the noise acquisition unit obtains a length of... Discrete background noise samples . Indicates the radar frame number; This indicates the sampling point number in the idle sampling window corresponding to the current radar frame; This indicates the total number of sampling points contained in the corresponding idle sampling window.

[0124] The background noise parameter estimation unit first removes the DC component from the discrete background noise samples. The background noise sample after removing the DC component is represented as follows:

[0125] (5)

[0126] In the formula, Indicates the first The first radar frame The background noise sample value after removing the DC component is in volts; Indicates the first The first radar frame One original background noise sample value, in volts; This indicates the total number of sampling points contained in the corresponding idle sampling window; This indicates the sequence number of the summation sampling point used when calculating the average value; This indicates the current sampling point number.

[0127] The background noise parameter estimation unit also performs outlier removal on the background noise samples. Outlier samples are those generated by transient pulses, analog-to-digital conversion glitches, or occasional communication flips, and deviate from the median by more than six times the absolute median difference. In one embodiment, the background noise parameter estimation unit calculates the median and absolute median difference of the background noise samples, identifies samples deviating from the median by more than six times the absolute median difference as outlier samples, and replaces them using the linear interpolation results of adjacent normal samples.

[0128] After removing the DC component and eliminating outlier sampling points, a window function is applied to the background noise samples, and a complex spectrum transform is performed.

[0129] (6)

[0130] In the formula, Indicates the first The first radar frame Complex spectral values ​​of each frequency unit; Indicates the radar frame number; Indicates the frequency unit number; Indicates the sampling point number; This indicates the total number of sampling points contained in the corresponding idle sampling window; This represents the background noise sample value after removing the DC component and replacing abnormal sampling points, in volts. Indicates the first The window function value corresponding to each sampling point; Represents the imaginary unit; This represents the natural exponential function. The window function can be a Hanning window, a Hamming window, or a Blackman window, used to reduce spectral leakage caused by finite-length sampling.

[0131] The background noise parameter estimation unit performs a peak search on the complex spectral values ​​and identifies peaks whose amplitudes reach the background noise threshold as candidate background noise components. These candidate background noise components refer to noise frequency components that have detectable peaks in the current corresponding idle sampling window group but have not yet been determined by repeatability confidence and temporal correlation.

[0132] For the The frequency of each of the candidate background noise components in the m-th radar frame is expressed as:

[0133] (7)

[0134] In the formula, Indicates the first The candidate background noise components in the first The frequency in each radar frame, measured in Hertz; Indicates the frequency unit number of the corresponding spectral peak; This indicates the sampling frequency of the noise acquisition unit, measured in Hertz. This indicates the total number of sampling points contained in the corresponding idle sampling window; Indicates the candidate background noise component number; Indicates the radar frame number.

[0135] No. The candidate background noise components in the first The amplitude in each radar frame is represented as follows:

[0136] (8)

[0137] In the formula, Indicates the first The candidate background noise components in the first Peak amplitude in each radar frame, in volts; Represents the complex spectral value of the corresponding spectral peak; Indicates the first The window function value corresponding to each sampling point; This indicates the total number of sampling points contained in the corresponding idle sampling window; This indicates the sequence number of the sampling point for summation.

[0138] No. The candidate background noise components in the first The phase in each radar frame is represented as follows:

[0139] (9)

[0140] In the formula, Indicates the first The candidate background noise components in the first Phase in a radar frame, in radians; Represents the complex spectral value of the corresponding spectral peak; This represents the argument function of a complex number.

[0141] In this embodiment, the phase formulas all use radians as the unit of calculation. When the phase is expressed in degrees in tables or text, it is calculated according to... The phase is converted in radians. The phase in different radar frames uses the radar frame synchronization time as a common zero-phase reference.

[0142] In other implementations, amplitude, frequency, and phase can also be obtained through phase-locked detection, quadrature demodulation, peak detection after narrowband filtering, or quadrature accumulation at preset frequency points. For microcontrollers with limited computing resources, quadrature accumulation can be performed only on the fundamental frequency, harmonic frequency, and sub-frequency of the internal periodic electromagnetic interference source; for field-programmable gate arrays or radar-specific processors, fast Fourier transform, interpolation spectrum analysis, or high-resolution spectrum estimation can be used.

[0143] For continuous The first radar frame Each candidate background noise component is used to calculate the cross-frame frequency deviation, amplitude dispersion, cross-frame phase deviation, and intra-frame position deviation.

[0144] No. The average frequency across frames for each candidate background noise component is expressed as:

[0145] (10)

[0146] In the formula, Indicates the first The cross-frame average frequency of each candidate background noise component, in Hertz; Indicates the number of radar frames used in the repeatability evaluation; Indicates the radar frame number; Indicates the first The candidate background noise components in the first The frequency in a radar frame is measured in Hertz.

[0147] No. The cross-frame frequency standard deviation of each candidate background noise component is expressed as:

[0148] (11)

[0149] In the formula, Indicates the first The cross-frame frequency standard deviation of each candidate background noise component, in Hertz; Indicates the number of radar frames used in the repeatability evaluation; Indicates the radar frame number; Indicates the first The candidate background noise components in the first Frequency in a radar frame; This indicates the average frequency across frames.

[0150] No. The average amplitude of each candidate background noise component across frames is expressed as:

[0151] (12)

[0152] In the formula, Indicates the first The average amplitude of each candidate background noise component across frames, in volts; Indicates the number of radar frames used in the repeatability evaluation; Indicates the radar frame number; Indicates the first The candidate background noise components in the first The amplitude in each radar frame, measured in volts.

[0153] No. The standard deviation of the amplitude of each candidate background noise component is expressed as:

[0154] (13)

[0155] In the formula, Indicates the first The amplitude standard deviation of each candidate background noise component, in volts; Indicates the number of radar frames used in the repeatability evaluation; Indicates the radar frame number; Indicates the first The candidate background noise components in the first The amplitude in each radar frame; This indicates the average amplitude across the corresponding frames.

[0156] Because phase is periodic, the phase deviation across frames is determined using circular statistical methods. The phase concentration of each candidate background noise component is expressed as:

[0157] (14)

[0158] In the formula, Indicates the first Phase concentration of each candidate background noise component; Indicates the number of radar frames used in the repeatability evaluation; Indicates the radar frame number; Indicates the first The candidate background noise components in the first Phase in a radar frame, in radians; Represents the imaginary unit; It represents the modulus of a complex number.

[0159] No. The cross-frame phase deviation of each candidate background noise component is expressed as:

[0160] (15)

[0161] In the formula, Indicates the first The cross-frame phase deviation of each candidate background noise component, in radians; Indicates the corresponding phase concentration; This represents the phase constant to prevent the logarithmic input from being zero. This indicates retrieving the maximum value from the set. This represents the natural logarithm function. In this embodiment, Pick The phase constant can be determined based on the minimum positive number representation range of the computing processor. Using circular statistical methods can avoid 359° and 1° being incorrectly identified as a phase difference of 358°.

[0162] No. The average position of each candidate background noise component across frames is represented as follows:

[0163] (16)

[0164] In the formula, Indicates the first The average position of each candidate background noise component across frames, in seconds; Indicates the number of radar frames used in the repeatability evaluation; Indicates the radar frame number; Indicates the first The candidate background noise components in the first The position of the spectral peak within a radar frame, in seconds.

[0165] No. The intra-frame location standard deviation of each candidate background noise component is expressed as:

[0166] (17)

[0167] In the formula, Indicates the first The intra-frame position standard deviation of each candidate background noise component, in seconds; Indicates the number of radar frames used in the repeatability evaluation; Indicates the radar frame number; Indicates the intra-frame position of the corresponding spectral peak; This indicates the average position across frames.

[0168] Normalized cross-frame frequency offset is expressed as:

[0169] (18)

[0170] In the formula, Indicates the first Normalized cross-frame frequency deviation of each candidate background noise component; This represents the standard deviation of frequency across frames, in Hertz. This indicates the allowable frequency deviation for repeatability evaluation, expressed in Hertz (Hz). This indicates taking the minimum value in the set.

[0171] The normalized amplitude dispersion is expressed as:

[0172] (19)

[0173] In the formula, Indicates the first Normalized amplitude dispersion of each candidate background noise component; Indicates the standard deviation of amplitude, in volts; This indicates the allowable magnitude deviation for repeatability evaluation, expressed in volts.

[0174] Normalized inter-frame phase deviation is expressed as:

[0175] (20)

[0176] In the formula, Indicates the first Normalized cross-frame phase deviation of candidate background noise components; Indicates the phase deviation across frames, in radians; This indicates that the repeatability evaluation allows for phase deviation, expressed in radians.

[0177] The normalized intra-frame positional offset is expressed as:

[0178] (twenty one)

[0179] In the formula, Indicates the first Normalized intra-frame positional offset of each candidate background noise component; This represents the standard deviation of the intra-frame position, in seconds. This indicates the allowable time and location deviation for repeatability evaluation, in seconds.

[0180] The allowable frequency deviation, allowable amplitude deviation, allowable phase deviation, and allowable time and position deviation for repeatability evaluation are determined based on the statistical results of multiple consecutive radar frames under stable interference conditions, noise acquisition frequency resolution, radar frame synchronization jitter, and analog-to-digital conversion quantization error. In one embodiment, the 95th percentile of the corresponding deviation statistical distribution under stable interference conditions is used as the initial allowable deviation, and it is corrected based on the number of false start compensations during factory calibration.

[0181] No. The repetition confidence of each candidate background noise component is expressed as:

[0182] (twenty two)

[0183] In the formula, Indicates the first The repeatability confidence of each candidate background noise component; This represents the normalized cross-frame frequency deviation weight; This represents the normalized cross-frame frequency deviation; Indicates the weight of the normalized magnitude dispersion; Indicates the dispersion of the normalized amplitude; This represents the normalized cross-frame phase deviation weight; This represents the normalized cross-frame phase deviation; This represents the normalized intra-frame positional offset weights; Represents the normalized intra-frame positional offset; each weight satisfies .

[0184] The weights for normalized cross-frame frequency deviation, normalized amplitude dispersion, normalized cross-frame phase deviation, and normalized intra-frame position deviation can be determined based on the ability of various deviations to distinguish between stable interference samples and random noise samples. Statistical analysis is performed on both stable interference samples and random noise samples. When a deviation has a high degree of distinguishability between the two types of samples, its corresponding weight is increased; when a deviation is significantly affected by sampling quantization errors or frame synchronization jitter, its corresponding weight is decreased.

[0185] In this embodiment, the weight for normalized cross-frame frequency deviation is 0.30, the weight for normalized amplitude dispersion is 0.15, the weight for normalized cross-frame phase deviation is 0.35, and the weight for normalized intra-frame position deviation is 0.20. When the normalized cross-frame frequency deviation is 0.08, the normalized amplitude dispersion is 0.12, the normalized cross-frame phase deviation is 0.06, and the normalized intra-frame position deviation is 0.10, the repetition confidence is:

[0186]

[0187] In the formula, Indicates the first The repetition confidence of each candidate background noise component; 0.30, 0.15, 0.35 and 0.20 represent the weights corresponding to the four types of normalized biases; 0.08, 0.12, 0.06 and 0.10 represent the input values ​​of normalized cross-frame frequency bias, normalized amplitude dispersion, normalized cross-frame phase bias and normalized intra-frame position bias, respectively.

[0188] The closer the repetition confidence is to 1, the more stable the frequency, amplitude, phase, and intra-frame occurrence position of the candidate background noise component; the closer the repetition confidence is to the lower limit of the value, the more likely the candidate background noise component is to belong to random noise, occasional noise, or scene change signal.

[0189] The background noise parameter estimation unit also determines the temporal correlation between candidate background noise components and internal periodic electromagnetic interference sources. The temporal correlation is used to represent the degree to which the frequency and intra-frame occurrence time of the candidate background noise components match the operating frequency and operating period of the corresponding internal periodic electromagnetic interference sources.

[0190] For the An internal periodic electromagnetic interference source forms a set of interference candidate frequencies based on its corresponding fundamental frequency, harmonic frequency, and sub-frequency. The background noise parameter estimation unit selects the interference candidate frequency with the smallest cross-frame average frequency deviation from the candidate background noise components from each set of interference candidate frequencies. The normalized interference source frequency deviation is expressed as:

[0191] (twenty three)

[0192] In the formula, Indicates the first Normalized interference source frequency deviation of each candidate background noise component; Indicates the serial number of the internal periodic electromagnetic interference source; Indicates the first A set of candidate frequencies for interference from an internal periodic electromagnetic interference source; This represents a candidate frequency in the set of interference candidate frequencies, expressed in Hertz; Indicates the first The cross-frame average frequency of each candidate background noise component, in Hertz; This indicates the allowable frequency deviation for timing correlation, in Hertz. and These represent the minimum frequency deviation among each internal periodic electromagnetic interference source and its corresponding candidate frequency set.

[0193] Let the operating time of the internal periodic electromagnetic interference source corresponding to the minimum frequency deviation be . The time distance from the average intra-frame position of the candidate background noise component to the action period is expressed as:

[0194] (twenty four)

[0195] In the formula, Indicates the first The time distance from the average cross-frame position of each candidate background noise component to the corresponding interference action period, in seconds; Indicates the start time of the corresponding interference action period, in seconds; Indicates the end time of the corresponding interference action period, in seconds; Indicates the first The average position across frames of each candidate background noise component, in seconds. When the average position across frames is within the interference action period, the time distance is 0; when the average position across frames is before or after the interference action period, the time distance is the difference between the average position across frames and the nearest time period boundary.

[0196] The normalized interference time period bias is expressed as:

[0197] (25)

[0198] In the formula, Indicates the first Normalized interference time period deviation of each candidate background noise component; This represents the time distance from the average position across frames to the corresponding interference action period, in seconds. This indicates the allowable time deviation for timing associations, in seconds.

[0199] No. The temporal correlation degree of each candidate background noise component is expressed as:

[0200] (26)

[0201] In the formula, Indicates the first Temporal correlation of candidate background noise components; This represents the normalized interference source frequency deviation weight; This indicates the normalized interference source frequency deviation; This represents the weight of the normalized interference period deviation; Represents the normalized interference time period deviation; the weights satisfy... .

[0202] The normalized interference source frequency deviation weight and the normalized interference time period deviation weight can be determined based on the frequency stability and jitter of the internal periodic electromagnetic interference source. When the fundamental frequency and harmonic frequency stability of the internal periodic electromagnetic interference source is higher than the stability of the action period, the normalized interference source frequency deviation weight is increased; when the action period of the internal periodic electromagnetic interference source has a more stable synchronization relationship with the radar frame timing, the normalized interference time period deviation weight is increased.

[0203] For example, the switching frequency of the power conversion unit is 400kHz, the first The cross-frame average frequency of the candidate background noise components is 399.6 kHz. These components appear 18 μs after the power conversion unit switches on, corresponding to a center time of 20 μs for the interference action. With the allowed frequency deviation for timing correlation set to 5 kHz and the allowed time deviation for timing correlation set to 50 μs, the normalized interference source frequency deviation is 0.08, and the normalized interference time deviation is 0.04. The weight of the normalized interference source frequency deviation is 0.60, and the weight of the normalized interference time deviation is 0.40. Therefore, the timing correlation degree is:

[0204]

[0205] In the formula, Indicates the first The temporal correlation degree of each candidate background noise component; 0.60 and 0.40 represent the normalized interference source frequency deviation weight and the normalized interference time period deviation weight, respectively; 0.08 and 0.04 represent the input values ​​of the normalized interference source frequency deviation and the normalized interference time period deviation, respectively.

[0206] In this embodiment, the repeatability confidence threshold is set to 0.82, and the temporal correlation threshold is set to 0.80. The repeatability confidence threshold and the temporal correlation threshold are determined based on the statistical results of samples under stable interference, random noise, and target echo leakage conditions, to ensure that the pass rate of stable interference samples reaches a preset proportion, and that the false pass rates of random noise and target echo leakage samples are lower than preset proportions. Specifically, stable interference samples, random noise samples, and target echo leakage samples can be collected separately, the repeatability confidence and temporal correlation of each sample can be calculated, and the corresponding thresholds can be determined based on the balance between the pass rate of stable interference samples and the false pass rate of unstable interference samples.

[0207] The control unit only determines the candidate background noise component as the background noise component to be compensated when the repetition confidence of the candidate background noise component reaches the repetition confidence threshold and the temporal correlation reaches the temporal correlation threshold. Figure 2 The decision boxes indicate whether the repetition confidence level and the temporal correlation level have reached the repetition confidence level threshold and the temporal correlation level threshold, respectively. If either condition is not met, compensation is not initiated and the process returns to the establishment of the corresponding idle sampling window group.

[0208] To ensure that the compensation action matches the reliability of the background noise estimation, this embodiment divides the repeatability confidence range into a stop compensation interval, a parameter hold interval, and a parameter update interval. When the repeatability confidence is less than 0.55, it is in the stop compensation interval, and the control unit stops injecting compensation signals to the disturbed node; when the repeatability confidence is greater than or equal to 0.55 and less than 0.82, it is in the parameter hold interval, and the control unit maintains the previous set of stable compensation control parameters and gradually reduces the upper limit of the compensation amplitude by 5% per parameter update cycle; when the repeatability confidence is greater than or equal to 0.82, it is in the parameter update interval.

[0209] The parameter update interval is further divided into multiple consecutive confidence sub-intervals. The compensation amplitude upper limit, amplitude parameter update step size and phase parameter update step size corresponding to each confidence sub-interval are shown in Table 1.

[0210] Table 1. Reliability Sub-intervals and Compensation Parameter Update Configuration Table

[0211]

[0212] The upper limit of the compensation amplitude is the operational control limit determined by the control unit within the safe compensation amplitude limit based on the current repetition confidence level; the amplitude parameter update step size is the numerical increment that allows changes to the compensation amplitude control quantity in a single compensation parameter correction; the phase parameter update step size is the angular increment that allows changes to the compensation phase control quantity in a single compensation parameter correction. The confidence sub-intervals and corresponding control parameters in Table 1 can be calibrated based on the linear range of the receiving link, the response speed of the compensation branch, the pass rate of stable interference samples, and the number of convergence frames for parameter search.

[0213] When the repetition confidence level is 0.92 and the upper limit of the security compensation amplitude corresponding to the disturbed node is 0.30V, the repetition confidence level belongs to... The confidence sub-interval is defined, with an upper limit of compensation amplitude of 0.195V, an amplitude parameter update step size of 5mV, and a phase parameter update step size of 2°.

[0214] The control unit also determines the compensation period based on the intra-frame position deviation. The compensation period is the intra-frame time interval during which the compensation injection switch is allowed to conduct and inject compensation signals into the disturbed node. When the intra-frame position deviation is no greater than 10 μs, the compensation period is extended by 20 μs on each side of the average occurrence position of the candidate background noise; when the intra-frame position deviation is greater than 10 μs but no greater than 30 μs, the compensation period is extended by 50 μs on each side of the average occurrence position of the candidate background noise; when the intra-frame position deviation is greater than 30 μs, the control unit does not perform gated compensation, or only maintains the previous set of stable compensation control parameters. The extended compensation period must not enter the adjacent valid target echo-occupied interval.

[0215] Before performing actual compensation, the system performs injection path calibration on multiple candidate disturbed nodes during product testing or installation. The injection path is the signal transmission path from the output of the compensation signal generation unit, through the amplitude and phase adjustment unit, the compensation injection unit, and the circuit connection structure, to the candidate disturbed node.

[0216] During the compensation calibration phase, the control unit sets the millimeter-wave radar to a state without effective targets, maintains stable radar frame timing and the operating state of internal periodic electromagnetic interference sources, and controls the compensation signal generation unit to generate calibration signals with known amplitude and known phase at multiple calibration frequencies. The calibration signals are sequentially injected into the corresponding candidate disturbed nodes through the amplitude and phase adjustment unit and the compensation injection unit, and the noise acquisition unit acquires the calibration sampling data before and after the calibration signal injection.

[0217] For the same candidate disturbed node and the same calibration frequency, the background noise parameter estimation unit extracts the complex spectral values ​​at the corresponding frequency from the calibration sampling data before and after the calibration signal injection. The difference between the complex spectral values ​​after and before the calibration signal injection is used to characterize the complex response formed at the monitoring node after the calibration signal passes through the compensation injection path. Based on the ratio between the complex response and the complex amplitude and phase of the calibration signal output by the compensation signal generation unit, the path amplitude transfer and path phase offset of the corresponding injection path are determined.

[0218] When determining the linear injection range, the control unit maintains the calibration frequency and calibration phase unchanged, and sequentially injects calibration signals according to multiple increasing calibration amplitudes, comparing the actual complex response at each calibration amplitude with the complex response predicted according to the path amplitude propagation. When the deviation of the actual response amplitude from the predicted response amplitude is not greater than a preset linear deviation threshold, and the deviation of the actual response phase from the predicted response phase is not greater than a preset linear phase deviation threshold, the corresponding calibration amplitude is included in the linear injection range. The upper limit of the linear injection range is the maximum calibration amplitude that continuously satisfies the linearity condition.

[0219] The additional noise level of the compensation branch is determined by comparing the noise energy evaluation values ​​when the compensation branch is open and when the compensation branch is on but the compensation signal amplitude is zero. The additional noise level of the compensation branch is obtained by subtracting the noise energy evaluation value when the compensation branch is open from the noise energy evaluation value when the compensation branch is on but the compensation signal amplitude is zero.

[0220] The upper limit of the safety compensation amplitude is determined based on the upper limit of the linear injection range, the allowable AC disturbance amplitude of the candidate disturbed node, the linear output range of the amplitude and phase adjustment unit, and the limiting threshold of the limiting protection unit. The upper limit of the safety compensation amplitude shall not exceed the upper limit of the linear injection range, and shall not cause the DC operating point, analog-to-digital conversion reference voltage, or ground reference potential of the candidate disturbed node to exceed the corresponding allowable range.

[0221] The path amplitude transfer amount, path phase offset amount, linear injection range, additional noise amount of compensation branch, and upper limit of safe compensation amplitude are respectively associated with the candidate disturbed node identifier and calibration frequency, and stored in the injection path calibration storage unit.

[0222] To ensure that the background noise amplitude and phase acquired by the monitoring nodes correspond to each candidate disturbed node, a background noise observation transfer coefficient is determined from the monitoring node to each candidate disturbed node during the injection path calibration process. This background noise observation transfer coefficient is used to convert the background noise amplitude and phase acquired at the monitoring node to the background noise amplitude and phase at the candidate disturbed node. This observation transfer relationship is a data conversion relationship based on the candidate disturbed node calibration data and does not change the compensation injection path or the disturbed node selection process.

[0223] In one implementation, the internal periodic electromagnetic interference source to be calibrated is kept in stable operation. Complex spectral values ​​of the monitoring node and the corresponding candidate disturbed node are acquired at the same radar frame position and the same interference frequency. The background noise observed amplitude transfer and background noise observed phase offset are determined based on the ratio between the complex spectral value of the candidate disturbed node and the complex spectral value of the monitoring node. When the monitoring node and the candidate disturbed node are the same node, the background noise observed amplitude transfer is set to 1, and the background noise observed phase offset is set to 0. When the monitoring node and the candidate disturbed node are different, amplitude and phase reduction are performed according to the corresponding background noise observed transfer coefficient.

[0224] No. The candidate disturbed nodes are at the _th ... The complex transfer coefficient of the compensation injection path at a given frequency is expressed as:

[0225] (27)

[0226] In the formula, Indicates the first The candidate disturbed nodes are at the _th ... Complex transfer coefficient of the compensation injection path at a given frequency; Indicates the sequence number of the candidate disturbed node; Indicates the calibration frequency sequence number; Indicates the magnitude of the transmission along the corresponding path; This represents the corresponding path phase offset, in radians. Represents the imaginary unit; This represents the natural exponential function.

[0227] Monitoring nodes up to the The background noise observation transfer coefficient of each candidate disturbed node is expressed as:

[0228] (28)

[0229] In the formula, Indicates the monitoring node up to the 1st Observed transmission coefficients of background noise for each candidate disturbed node; This indicates the observed amplitude transfer of background noise; This indicates the observed phase shift in background noise, expressed in radians. Indicates the sequence number of the candidate disturbed node; Indicates the calibration frequency sequence number; It represents the imaginary unit.

[0230] This embodiment was calibrated at three calibration frequencies: 200kHz, 400kHz, and 800kHz. A set of injection path calibration data is shown in Table 2.

[0231] Table 2. Calibration data for different candidate disturbed nodes

[0232]

[0233] When the frequency of the background noise component to be compensated lies between two adjacent calibration frequencies, linear interpolation is performed on the complex transfer coefficients corresponding to the two adjacent calibration frequencies. Taking the compensation of the complex transfer coefficient of the injection path as an example, the interpolation result is expressed as follows:

[0234] (29)

[0235] In the formula, Indicates the first The candidate disturbed nodes are at the background noise frequency to be compensated. Complex transfer coefficient of the compensation injection path; Indicates the first The candidate disturbed nodes are at the _th ... Complex transfer coefficient of the compensation injection path at a given frequency; Indicates the first The candidate disturbed nodes are at the _th ... Complex transfer coefficient of the compensation injection path at a given frequency; Indicates the first The frequency of the background noise component to be compensated is expressed in Hertz. and Indicates that it is located at The adjacent calibration frequencies on both sides, in Hertz; Indicates the sequence number of the candidate disturbed node; This indicates the calibration frequency number.

[0236] The background noise observation transfer coefficient is interpolated using the same method as in equation (29). After interpolation, the corresponding path amplitude transfer and path phase offset are obtained by the modulus and argument of the complex transfer coefficient, respectively. The frequency matching threshold is determined based on the calibration frequency interval, the background noise frequency estimation error, and the interpolation error, and the background noise frequency to be compensated is within the effective calibration frequency range.

[0237] Based on the background noise observation transmission coefficient, the background noise amplitude obtained at the monitoring node is converted to the next level. One candidate disturbed node:

[0238] (30)

[0239] In the formula, Indicates the first The background noise component to be compensated is in the first... The reduced amplitude at each candidate disturbed node, in volts; Indicates the monitoring node up to the 1st The candidate disturbed nodes are in the frequency Observe the amplitude transfer of background noise; Indicates the number obtained at the monitoring node The amplitude of each background noise component to be compensated, in volts; This indicates the frequency of the background noise to be compensated, in Hertz.

[0240] Background noise phase obtained at the monitoring node is converted to the th One candidate disturbed node:

[0241] (31)

[0242] In the formula, Indicates the first The background noise component to be compensated is in the first... The reduced phase at each candidate disturbed node, in radians; Indicates the number obtained at the monitoring node The phase of each background noise component to be compensated, in radians; Indicates the monitoring node up to the 1st The candidate disturbed nodes are in the frequency The phase shift is observed under background noise, and the unit is radians.

[0243] No. The candidate compensation magnitude for each candidate disturbed node is expressed as follows:

[0244] (32)

[0245] In the formula, Indicates that for the first The background noise component to be compensated and the first The candidate compensation range determined by each candidate disturbed node, in volts; The background noise component to be compensated is in the th... The reduced amplitude at each candidate disturbed node, in volts; Indicates the first The candidate disturbed nodes are in the frequency The path amplitude propagation amount.

[0246] No. The candidate compensation phase corresponding to each candidate disturbed node is represented as follows:

[0247] (33)

[0248] In the formula, Indicates that for the first The background noise component to be compensated and the first The candidate compensation phase determined by each candidate disturbed node, in radians; The background noise component to be compensated is in the th... The reduced phase at each candidate disturbed node, in radians; Indicates a 180° phase reversal; Indicates the first The candidate disturbed nodes are in the frequency The path phase offset, in radians.

[0249] No. The estimated residual noise corresponding to each candidate disturbed node is expressed as follows:

[0250] (34)

[0251] In the formula, Indicates the first The background noise component to be compensated is in the first... The estimated residual noise at each candidate disturbed node is expressed in volts squared. The background noise component to be compensated is in the th... The reduced amplitude at each candidate disturbed node, in volts; Indicates the corresponding converted phase, in radians; Indicates the magnitude of the transmission along the corresponding path; Indicates the candidate compensation range, in volts; Indicates the candidate compensation phase, in radians; This represents the corresponding path phase offset, in radians. Indicates the first The compensation branch corresponding to the candidate disturbed node is in the first... The amount of additional noise near the background noise frequency to be compensated, in volts squared; It represents the imaginary unit.

[0252] The estimated residual noise is a theoretical evaluation quantity obtained from the injection path calibration data, used for initial selection among multiple candidate disturbed nodes. The actual residuals caused by injection path calibration errors, device temperature drift, and changes in installation status are corrected through closed-loop feedback of residual noise in the corresponding idle sampling window.

[0253] The control unit selects the candidate disturbed node with the smallest estimated residual noise from among the candidate disturbed nodes that meet the following conditions: the calibrated frequency range covers the background noise frequency to be compensated; the candidate compensation amplitude is within the linear injection range; the candidate compensation amplitude does not exceed the upper limit of the safe compensation amplitude and the upper limit of the current compensation amplitude; and the additional noise of the compensation branch is not greater than the additional noise threshold.

[0254] The additional noise threshold is determined based on the baseline noise floor under the condition of the compensation branch being disconnected, the allowable increase in noise floor of the receiving link, and the signal-to-noise ratio required for weak target detection. The phase deviation threshold is determined based on the phase resolution of the controllable phase adjustment unit, the path phase repetition error, and the radar frame synchronization phase jitter. The phase stabilization threshold is determined based on the distribution of phase search results from multiple consecutive radar frames under stable interference conditions.

[0255] For example, the frequency of the background noise component to be compensated is 399.6 kHz, the peak amplitude at the monitoring node is 0.16 V, and the phase is 9°. Based on the calibration data of the analog-to-digital converter (ADC) reference node near 400 kHz in Table 2, the background noise amplitude converted to the ADC reference node is approximately 0.152 V, and the converted phase is approximately 17°. The corresponding candidate compensation amplitude is approximately 0.167 V, and the candidate compensation phase is approximately 180°. Since the candidate compensation amplitude is lower than the current compensation amplitude limit of 0.195 V, and the additional noise in the compensation branch corresponding to the ADC reference node is lower than that of other candidate disturbed nodes, the ADC reference node is selected as the disturbed node.

[0256] The control unit determines the compensation frequency control amount based on the frequency of the background noise component to be compensated, determines the compensation amplitude control amount based on the reduced amplitude and path amplitude propagation of the background noise component to be compensated at the disturbed node, and determines the compensation phase control amount based on the reduced phase, 180° phase reversal and path phase offset of the background noise component to be compensated at the disturbed node.

[0257] The compensation signal generation unit can employ a digital-to-analog converter, a numerically controlled oscillator, a phase-locked loop frequency divider output circuit, a quadrature local oscillator component generation circuit, or a programmable waveform generation circuit. The amplitude and phase adjustment unit includes a controllable amplitude adjustment unit and a controllable phase adjustment unit. The controllable amplitude adjustment unit can employ a digital step attenuator, a variable gain amplifier, a voltage-controlled attenuator, or a resistor network attenuator; the controllable phase adjustment unit can employ a digital phase shifter, a vector modulator, a quadrature weighting network, a delay line, or an analog phase adjustment network.

[0258] The control unit can control the controllable amplitude adjustment unit and the controllable phase adjustment unit via a serial peripheral interface, an integrated circuit bus interface, a general-purpose input / output interface, a digital-to-analog conversion control voltage, or a pulse width modulation filter voltage. The reference compensation signal sequentially passes through the controllable amplitude adjustment unit, the controllable phase adjustment unit, the buffer isolation unit, the DC isolation unit, the limiting protection unit, and the compensation injection switch to form the compensation signal. In other embodiments, the connection order of the controllable phase adjustment unit and the controllable amplitude adjustment unit can be interchanged.

[0259] The buffer isolation unit reduces the reverse impact of impedance changes at the disturbed node on the compensation signal generation unit and reduces the reverse coupling of signals in the millimeter-wave radar receiving link to the compensation signal generation unit. The DC isolation unit prevents the compensation signal from altering the DC bias of the disturbed node; it can be a coupling capacitor, isolation transformer, or differential coupling network. The amplitude limiting protection unit limits the amplitude of the compensation signal under power-on transients, abnormal control commands, or strong external interference. The compensation injection switch disconnects the compensation injection path when compensation stops, during power-on reset, or in abnormal protection states.

[0260] No. The individual background noise components to be compensated are represented as follows:

[0261] (35)

[0262] In the formula, Indicates time The first One background noise component to be compensated, in volts; This represents the amplitude of the background noise to be compensated obtained at the monitoring node, in volts. This indicates the frequency of the background noise to be compensated, in Hertz. Indicates the phase of the background noise to be compensated, in radians; This represents the time based on the radar frame synchronization time, in seconds. This indicates the sequence number of the background noise component to be compensated.

[0263] Regarding the first The compensation signal generated by each disturbed node is represented as follows:

[0264] (36)

[0265] In the formula, Indicates time For the first The background noise component to be compensated and the first The compensation signal generated by each disturbed node is in volts. This indicates the output amplitude corresponding to the compensation amplitude control quantity, in volts. This indicates the frequency corresponding to the compensation frequency control quantity, in Hertz (Hz). This indicates the phase corresponding to the compensation phase control quantity, in radians; Indicates time, in seconds.

[0266] The control unit ensures that the phase difference between the phase of the compensation signal after it reaches the disturbed node and the phase of the background noise component to be compensated at the disturbed node, relative to 180°, is no greater than the phase deviation threshold. In this embodiment, the phase deviation threshold is set to 12°. When the phase adjustment resolution and path phase stability meet the requirements, the phase deviation threshold can be set to 3° to 8°.

[0267] The compensation injection switch can be continuously turned on or gated according to the compensation period. When the background noise component to be compensated only appears at a local position in the radar frame period, the control unit presets the controllable amplitude adjustment unit and the controllable phase adjustment unit before the start of the compensation period, and turns on the compensation injection switch after the corresponding device reaches a stable state. The gating opening and closing process can be set with a soft start time and a soft close time to avoid additional spikes caused by sudden changes in the compensation signal.

[0268] After compensation is performed, the noise acquisition unit acquires residual noise samples within the corresponding idle sampling windows with the same intra-frame sampling position identifier in subsequent radar frames. The background noise samples before compensation and the residual noise samples after compensation used for comparison have the same range gate identifier and the same intra-frame time position identifier.

[0269] The temporal correlation state of the corresponding idle sampling window after compensation is determined as follows: When the residual noise spectral peak still reaches the spectral peak detection threshold, the temporal correlation degree is recalculated based on the residual noise frequency, the residual noise intra-frame position, and the timing information of the interference source; when the residual noise spectral peak is lower than the spectral peak detection threshold due to compensation, the disappearance of the spectral peak is not used as the condition for temporal discordation. Instead, the temporal correlation degree before compensation is maintained based on whether the action frequency, action period, and intra-frame sampling position identifier of the corresponding idle sampling window of the internal periodic electromagnetic interference source remain within the allowable deviation range. If the action frequency, action period, or position of the corresponding idle sampling window exceeds the allowable deviation range, the temporal correlation degree after compensation is set below the temporal correlation degree threshold, and the use of the corresponding residual noise sample to update the compensation control parameters is stopped.

[0270] The noise energy before compensation and the residual noise energy after compensation referred to in this paper are both noise energy evaluation quantities. These noise energy evaluation quantities are the average values ​​of the squares of the sampled voltages within the corresponding idle sampling window, expressed in volts squared. The noise energy before compensation is expressed as:

[0271] (37)

[0272] In the formula, This represents the noise energy evaluation quantity within the corresponding idle sampling window before compensation, in volts squared. This indicates the total number of sampling points contained in the corresponding idle sampling window; Indicates the sampling point number; This indicates the first sample in the corresponding idle sampling window before compensation. One input reference voltage sample value, in volts; This represents the square of the sampled voltage amplitude.

[0273] The residual noise energy after compensation is expressed as:

[0274] (38)

[0275] In the formula, This represents the residual noise energy evaluation quantity within the corresponding idle sampling window after compensation, in volts squared. This indicates the total number of sampling points contained in the corresponding idle sampling window; Indicates the sampling point number; This indicates the first sample in the corresponding idle sampling window after compensation. One input reference voltage sample value, in volts. and All values ​​are calculated from the analog-to-digital conversion code value based on the full-scale voltage and gain calibration coefficient of the analog-to-digital conversion.

[0276] The noise suppression rate is expressed as:

[0277] (39)

[0278] In the formula, Indicates the noise suppression rate; This represents the noise energy before compensation; This represents the residual noise energy after compensation. This represents the energy constant to prevent the denominator from being too small, and its unit is volt squared. In this embodiment, Pick The energy constant can be determined based on the square of the minimum quantization voltage for analog-to-digital conversion and the numerical precision of the processor.

[0279] When the noise energy before compensation is The residual noise energy after compensation is At that time, the noise suppression rate was:

[0280]

[0281] In the formula, 0.0248V² represents the noise energy before compensation; 0.0036V² represents the residual noise energy after compensation; and 0.000001V² represents the energy constant.

[0282] In this embodiment, the preset inhibition rate threshold is set to 0.80. Figure 2 The decision box in the diagram indicates whether the noise suppression rate has reached the preset suppression rate threshold. When the noise suppression rate reaches the preset suppression rate threshold, the control unit maintains the current compensation control parameters; when the noise suppression rate is lower than the preset suppression rate threshold, the control unit continues to fine-tune the compensation phase control amount and the compensation amplitude control amount. The preset suppression rate threshold is determined based on the residual noise allowed by the receiving link, the signal-to-noise ratio required for target detection, and the closed-loop convergence speed.

[0283] When correcting the compensation control parameters, the control unit first performs a phase search at a compensation amplitude lower than the upper limit of the compensation amplitude. In this embodiment, the compensation amplitude for phase search is set to 50% of the upper limit of the compensation amplitude, and multiple candidate compensation phase control quantities are set sequentially with an initial phase search interval of 30°.

[0284] In the aforementioned numerical calculations, the initial absolute compensation phase control value calculated based on the background noise phase of the disturbed node and the path phase offset was 180°. Therefore, the phase search data in Table 3 represents an absolute phase scan around the initial absolute compensation phase control value of 180°, rather than directly using 180° phase inversion without considering the path phase offset.

[0285] Table 3. Data table of phase search and amplitude search process

[0286]

[0287] The control unit verifies the local minimum of the phase-residual noise energy sequence in four consecutive radar frames. The local minimum phases obtained in the four radar frames are 178°, 181°, 179°, and 180°, respectively. Since the phase deviation is no greater than the phase stability threshold of 3°, 180° is determined as the phase control variable to be used for compensation. Subsequently, an amplitude search is performed while keeping the phase control variable to be used unchanged, and the value corresponding to the minimum residual noise energy of 0.17V is determined as the amplitude control variable to be used for compensation.

[0288] When the residual noise energy decreases after the compensation control parameters are updated, the control unit continues to fine-tune the parameter update step size according to the corresponding confidence sub-interval along the current parameter change direction; when the residual noise energy increases after the compensation control parameters are updated, the control unit restores the previous set of stable compensation control parameters before the update, and reduces the current amplitude parameter update step size and phase parameter update step size to 50% of the original step size.

[0289] The previous set of stable compensation control parameters refers to the compensation control parameters that, within at least one previous parameter holding period, enable the noise suppression rate to reach a preset suppression rate threshold without causing target detection anomalies. The control unit can store the previous set of stable compensation control parameters in non-volatile memory or the controller's internal parameter area.

[0290] When there is no local minimum in the phase-residual noise energy sequence, or when the deviation of the candidate compensation phase control quantity corresponding to the local minimum exceeds the phase stability threshold in multiple consecutive radar frames, the control unit marks the corresponding background noise component to be compensated as an uncompensable noise component and switches the compensation injection switch to the off state. When three consecutive parameter adjustments fail to reduce the residual noise energy, the control unit also records the corresponding noise event as an uncompensable event.

[0291] When a valid target state appears in any corresponding idle sampling window used for comparison, or when the temporal correlation of any corresponding idle sampling window before or after compensation is lower than the temporal correlation threshold, the control unit discards the corresponding residual noise sample after compensation and maintains the previous set of stable compensation control parameters. The control unit can also reduce the compensation amplitude in strong target detection mode, freeze phase adjustment when abnormal fluctuations occur in the target detection result, or switch to monitoring-only mode.

[0292] When there are multiple background noise components to be compensated that simultaneously meet the repetition confidence threshold and the temporal correlation threshold within the corresponding idle sampling window, the control unit determines the compensation priority based on the repetition confidence, amplitude, temporal correlation, intra-frame time position, and injection path calibration data of each background noise component to be compensated.

[0293] In one implementation, the components are first sorted by repetition confidence from high to low; if the repetition confidence is the same, they are then sorted by temporal correlation from high to low; if both repetition confidence and temporal correlation are the same, they are then sorted by background noise amplitude from high to low; if they are still the same, the background noise components with a longer overlap with the target detection sensitive period are compensated first. The target detection sensitive period is an intra-frame period determined by target detection information where weak target echoes may occur or the target detection threshold is low.

[0294] When multiple background noise components to be compensated correspond to the same disturbed node and their compensation periods overlap, the compensation signal generation unit generates a combined compensation signal containing multiple corresponding frequency components. The combined compensation signal without amplitude scaling is represented as:

[0295] (40)

[0296] In the formula, Indicates time For the first The unscaled combined compensation signal generated by the disturbed node is in volts. Represents a combination of multiple frequency components; Indicates allocation to the first A set of background noise components to be compensated, where there are multiple disturbed nodes and overlapping compensation periods. Indicates the sequence number of the background noise component to be compensated; Indicates the first The compensation amplitude control amount corresponding to each background noise component to be compensated, in volts; This indicates the corresponding compensation frequency, in Hertz (Hz). This indicates the corresponding compensation phase control amount, in radians; Indicates time, in seconds.

[0297] The peak amplitude of the combined compensated signal is expressed as:

[0298] (41)

[0299] In the formula, Indicates the first The peak amplitude of the unscaled combined compensated signal corresponding to each disturbed node, in volts; Indicates the first The combined compensation period corresponding to each disturbed node; Indicates the time within the combined compensation period; This indicates a combined compensation signal that has not undergone amplitude scaling; This indicates that the maximum value is taken during the combined compensation period.

[0300] The combined compensation scaling factor is expressed as:

[0301] (42)

[0302] In the formula, Indicates the first The combined compensation scaling factor corresponding to each disturbed node; Indicates the first The minimum value among the upper limit of the linear injection range, the upper limit of the safety compensation range, and the upper limit of the current compensation range for each disturbed node, in volts; This represents the peak amplitude of the unscaled combined compensated signal, in volts. This represents the amplitude constant to prevent the denominator from being zero, and its unit is volts. This indicates taking the minimum value in the set. Amplitude constant. It can be set to a positive number that is no greater than the minimum output resolution of the compensation signal generation unit.

[0303] The actual output combined compensation signal is expressed as:

[0304] (43)

[0305] In the formula, Indicates time For the first The actual combined compensation signal output by each disturbed node is expressed in volts. Indicates the combined compensation scaling factor; This represents the combined compensated signal without amplitude scaling, measured in volts.

[0306] Multiple frequency components have corresponding compensation amplitude control and compensation phase control, and after combined compensation scaling, the total peak amplitude of the combined compensation signal does not exceed the upper limit of the linear injection range of the compensation branch, the upper limit of the safe compensation amplitude, and the upper limit of the current compensation amplitude.

[0307] When multiple background noise components to be compensated correspond to different disturbed nodes, the control unit assigns each component to a compensation channel connected to the corresponding disturbed node. The compensation circuit can use a single-channel structure, or two or three parallel compensation channels. When the number of available compensation channels is less than the number of background noise components to be compensated, the control unit performs time-division compensation according to the compensation priority. After each switch of background noise components to be compensated, the control unit uses the corresponding idle sampling window determined by the intra-frame sampling position identifier corresponding to the switched background noise component to re-determine the repetition confidence and temporal correlation.

[0308] In a specific operational process, after the system is powered on, it first enters the baseline establishment phase. The baseline establishment phase includes confirmation of the corresponding idle sampling window, recording of the basic noise level, self-test of the compensation branch, and status check of the compensation injection switch.

[0309] The baseline noise floor is a reference energy obtained by statistically analyzing the noise energy within the corresponding idle sampling window under the condition that the compensation branch is closed and there is no effective target echo. The control unit also records the additional noise of the compensation branch when the compensation branch is open but not outputting a compensation signal, and confirms the linear injection range according to different pre-calibrated compensation amplitudes. When the additional noise of the compensation branch exceeds the additional noise threshold, or when the compensation injection switch is abnormal, the control unit keeps the compensation branch closed.

[0310] After the baseline is established, the system enters the normal monitoring phase. The control unit continuously compares the background noise sample corresponding to the current idle sampling window with the baseline noise floor. When the amplitude of the candidate background noise component exceeds the entry threshold for multiple consecutive radar frames, the repetition confidence reaches the repetition confidence threshold, and the temporal correlation reaches the temporal correlation threshold, the compensation control phase begins. When the amplitude of the background noise component to be compensated is below the exit threshold for multiple consecutive radar frames, the control unit stops compensation. The entry threshold is higher than the exit threshold to form a compensation start-stop loop and reduce frequent switching of the compensation injection switch. The entry and exit thresholds are determined based on the baseline noise floor, the signal-to-noise ratio required for target detection, and the allowed switching frequency of the compensation injection switch.

[0311] See Figure 3 , Figure 3 (a) shows the target echo component located at approximately 0.56 of the normalized time position; Figure 3 (b) shows the pre-compensated sampled signal after background noise is superimposed on the target echo; Figure 3 (c) shows the sampled signal after compensation. Figure 3 (c) shows that the residual background noise amplitude is reduced after compensation while the target echo position and basic shape are maintained. It does not mean that the compensation process changes the target echo itself. Figure 3 (a) to Figure 3 (c) The horizontal axis represents the normalization time, and the vertical axis represents the normalization magnitude.

[0312] To illustrate the calculation process of closed-loop compensation for the corresponding idle sampling window, a simulation model was established based on the aforementioned calibration parameters. The noise sampling frequency in the simulation was 2MHz, the corresponding idle sampling window contained 1024 sampling points, and the full-scale analog-to-digital converter input was ±1V. The internal periodic electromagnetic interference source was a 400kHz power conversion ripple, with a peak amplitude of 0.16V at the monitoring node. The root mean square amplitude of the random noise was 0.03V. The target distance was set to 8.5m. Each compensation method was simulated continuously for 200 radar frames. The simulation data is used to illustrate the parameter calculations and the relative relationships between different control methods and is not considered as actual measured data from the prototype.

[0313] Different compensation methods use the same radar frame structure, noise sampling frequency, internal periodic electromagnetic interference source, random noise intensity, target echo model, candidate disturbed nodes, injection path calibration data, initial compensation amplitude and phase, number of running frames and evaluation method, only changing the compensation activation method, compensation parameter update method or feedback sampling window matching method.

[0314] In the non-compensation mode, the compensation injection switch remains off, and the noise acquisition and target detection processes continue to run.

[0315] In the fixed amplitude-phase compensation method, the disturbed node, compensation amplitude control amount, and compensation phase control amount are determined by the corresponding idle sampling window that satisfies the repetition confidence threshold and the temporal correlation threshold in the first set of compensation cycles. The compensation amplitude control amount and compensation phase control amount remain unchanged in subsequent compensation cycles, and the amplitude and phase are not adjusted based on residual noise energy. The fixed amplitude-phase compensation method still loads fixed compensation control parameters according to the preset compensation cycle and records the number of times the fixed compensation control parameters are incorrectly loaded when the conditions of no valid target or temporal correlation are not met.

[0316] In the arbitrary idle sampling window feedback method, the disturbed node, initial compensation control parameters, parameter update step size and feedback algorithm are the same as the corresponding idle sampling window closed-loop compensation method. However, the residual noise samples after compensation only satisfy the condition of no effective target and are not required to have the same intra-frame sampling position identifier as the background noise samples before compensation.

[0317] In the corresponding idle sampling window closed-loop compensation method, the background noise samples before compensation and the residual noise samples after compensation are respectively from the corresponding idle sampling windows with the same intra-frame sampling position identifier, and the repetition confidence judgment, temporal correlation judgment, disturbed node selection, phase search, amplitude search and residual noise feedback correction are performed simultaneously.

[0318] Table 4 Simulation results for different compensation methods

[0319]

[0320] The noise floor in Table 4 is the average spectral amplitude calculated within the spectral range excluding the target echo, expressed in decibels relative to the full scale of the analog-to-digital conversion; the target echo signal-to-noise ratio is the ratio of the peak power of the target echo to the background noise power near the target; the number of parameter error updates is the number of times the residual noise energy increases after the compensation control parameter or the fixed compensation control parameter loading state is updated, and the sample state check determines that the increase is caused by a change in the state of the target echo or the interference source; the number of stable convergence frames is the number of radar frames experienced from the start of compensation until the noise suppression rate continuously reaches the preset suppression rate threshold; the change in the peak value of the target echo is the difference in the peak amplitude of the target echo before and after compensation.

[0321] In applications where the power conversion unit generates switching frequency ripple, the interference source timing interface acquires the pulse width modulation synchronization signal of the power conversion unit, the background noise parameter estimation unit determines the corresponding ripple frequency and intra-frame position, and the control unit initiates compensation after the repetition reliability and timing correlation reach the corresponding thresholds. When the power conversion unit experiences frequency drift due to load changes, the control unit re-determines the timing correlation and re-selects the calibration frequency range and compensation control parameters based on the updated background noise frequency to be compensated.

[0322] In application scenarios where periodic data transmission disturbances occur at the communication interface, the interference source timing interface acquires the data transmission enable signal of the communication interface. The control unit limits the compensation period to the data transmission period of the communication interface and the extended period determined based on the intra-frame position deviation. When the valid target echo enters the corresponding intra-frame sampling position, the control unit stops using the corresponding residual noise sample to update the compensation control parameters and maintains the previous set of stable compensation control parameters or stops compensation.

[0323] In application scenarios where the digital clock unit generates fundamental frequency and harmonic interference, the background noise parameter estimation unit matches the candidate background noise frequency with the fundamental frequency, harmonic frequency and division frequency of the digital clock unit; when there are multiple background noise components that meet the conditions to be compensated, the control unit adopts combined compensation, parallel compensation or time-division compensation according to the compensation priority.

[0324] The aforementioned method utilizes corresponding idle sampling windows with the same intra-frame sampling position identifier in multiple consecutive radar frames to obtain background noise samples. It determines the repetition confidence based on the cross-frame frequency, amplitude, phase, and intra-frame position of the background noise components, and determines the temporal correlation based on the correspondence between the background noise frequency and occurrence time and the internal periodic electromagnetic interference source. Compensation is only applied to the background noise components that meet the corresponding conditions. The control unit selects the disturbed node based on the multi-frequency calibration injection path calibration data, controls the compensation signal to form an inverse amplitude-phase relationship at the disturbed node, and uses the residual noise in the subsequent corresponding idle sampling windows to correct the compensation control parameters, thus forming a background noise identification, compensation execution, and feedback correction process.

[0325] The following is combined Figure 1 The following provides a further description of a millimeter-wave radar noise compensation circuit based on an idle sampling window. In this embodiment, the compensation circuit is used to execute the aforementioned millimeter-wave radar noise compensation method based on an idle sampling window, and the relevant data processing is performed according to the data sources, calculation relationships, and control conditions defined in the aforementioned method embodiments.

[0326] See Figure 1 The millimeter-wave radar noise compensation circuit based on the idle sampling window includes a millimeter-wave radar receiving link, a radar frame timing interface, a target detection interface, an interference source timing interface, an idle sampling window matching unit, a noise acquisition unit, a background noise parameter estimation unit, an injection path calibration and storage unit, a disturbed node selection unit, a control unit, a compensation signal generation unit, an amplitude and phase adjustment unit, and a compensation injection unit. These units form a closed-loop circuit structure encompassing background noise acquisition, background noise parameter estimation, disturbed node selection, compensation signal generation, compensation signal injection, and residual noise feedback.

[0327] The millimeter-wave radar receiving link is used to receive the superimposed signal of target echo and background noise. The millimeter-wave radar receiving link may include a receiving antenna, RF receiving front-end, mixing unit, intermediate frequency amplification unit, filtering unit, analog-to-digital conversion unit, and digital data interface. The compensation circuit is connected to the millimeter-wave radar receiving link in an additional manner, without changing the millimeter-wave radar's transmitted waveform, receiving mixing method, or the basic processing flow of the target detection algorithm.

[0328] The millimeter-wave radar receiver link is equipped with monitoring nodes and multiple candidate disturbed nodes. Monitoring nodes are used to acquire analog signals or digital sampled data that characterize the background noise of the millimeter-wave radar receiver link. Monitoring nodes can be located at the intermediate frequency output, analog-to-digital converter input, analog-to-digital converter output data interface, bias monitoring nodes, or ground reference monitoring nodes.

[0329] Multiple candidate disturbed nodes include intermediate frequency (IF) bias nodes, analog-to-digital (ADC) reference nodes, and ground reference sensitive nodes. IF bias nodes provide a DC operating point for the IF amplification unit or IF input buffer unit; ADC reference nodes provide a reference voltage or potential for the ADC process; and ground reference sensitive nodes are grounded nodes whose potential changes can couple to the millimeter-wave radar receiving link and affect the IF signal or ADC result. In other embodiments, candidate disturbed nodes may also include radio frequency (RF) input reference nodes, IF link reference nodes, ADC input buffer nodes, or power supply reference nodes.

[0330] The radar frame timing interface is used to acquire radar frame timing information. It can connect to the radar frame synchronization terminal, linear frequency modulation pulse synchronization terminal, transmit enable terminal, analog-to-digital conversion sampling enable terminal, or timing register interface of a millimeter-wave radar chip. The radar frame timing information includes at least one of the following: radar frame start time, transmit enable period, analog-to-digital conversion sampling period, transmit silence period, data processing period, and radar frame end time.

[0331] The target detection interface is used to acquire target detection information. It can connect to the digital signal processor, target detection processing unit, or on-chip controller of the millimeter-wave radar. Target detection information includes at least one of the following: range gate occupancy status, valid target indicator, target echo amplitude, target confidence level, target echo leakage indicator, and strong reflection clutter indicator.

[0332] The interference source timing interface is used to acquire the interference source timing information corresponding to the internal periodic electromagnetic interference source. The interference source timing interface can be connected to the pulse width modulation synchronization terminal of the power conversion unit, the clock gating terminal of the digital clock unit, the data transmission enable terminal of the communication interface, or the controller port used to record the operating status of the internal periodic electromagnetic interference source. The interference source timing information is used to indicate the operating frequency, switching action period, clock flipping period, or data transmission period of the internal periodic electromagnetic interference source.

[0333] The radar frame timing interface, target detection interface, and interference source timing interface are connected to the idle sampling window matching unit, respectively. The idle sampling window matching unit is connected to the sampling control terminal and control unit of the noise acquisition unit, respectively. Based on the radar frame timing information and target detection information, the idle sampling window matching unit establishes corresponding idle sampling window groups with the same intra-frame sampling position identifier in multiple consecutive radar frames, and outputs the target status and intra-frame sampling position identifier of the corresponding idle sampling window group to the control unit. The interference source timing information is mapped by the idle sampling window matching unit to the same intra-frame time reference as the radar frame timing information and sent to the background noise parameter estimation unit.

[0334] The idle sampling window matching unit can be implemented by timing comparison logic in a field-programmable gate array, timers and comparison programs in a microcontroller, timing state machines in a radar on-chip controller, or a combination of the above structures. The idle sampling window matching unit stores at least one of the following: range gate identifier, intra-frame time position, linear frequency modulated pulse sequence number, and analog-to-digital conversion sampling point interval, and uses the stored information as an intra-frame sampling position identifier.

[0335] When the idle sampling window matching unit confirms that the current sampling interval belongs to the corresponding idle sampling window, it outputs a sampling enable signal to the noise acquisition unit; when the target detection information indicates that there is a valid target, target echo leakage or strong reflection clutter in the corresponding sampling interval, it stops outputting the sampling enable signal or discards the corresponding sampling data.

[0336] The input of the noise acquisition unit is connected to the monitoring node, and the output of the noise acquisition unit is connected to the background noise parameter estimation unit. The noise acquisition unit is used to acquire background noise samples within the corresponding idle sampling window group.

[0337] In one implementation, the noise acquisition unit reuses the analog-to-digital converter (ADC) in the millimeter-wave radar receiving link, extracting ADC sampling data within the corresponding idle sampling window via a direct memory access controller or on-chip data bus. In another implementation, the noise acquisition unit employs an independent low-speed ADC, whose analog input is connected to an intermediate frequency monitoring node, an ADC reference node, an offset monitoring node, or a ground reference monitoring node.

[0338] The noise acquisition unit may also include a sample-and-hold circuit, an anti-aliasing filter circuit, a data buffer, and sampling trigger logic. The data buffer is used to store background noise samples from multiple consecutive radar frames, and the sampling trigger logic is used to start or stop background noise acquisition based on the sampling enable signal output by the idle sampling window matching unit.

[0339] The background noise parameter estimation unit is connected to the noise acquisition unit, the idle sampling window matching unit, the interference source timing interface, and the control unit, respectively. Based on the background noise samples within the corresponding idle sampling window group, the background noise parameter estimation unit determines the frequency, amplitude, phase, and repetition confidence of candidate background noise components, and determines the timing correlation degree based on the frequency correspondence and intra-frame time correspondence between the candidate background noise components and the interference source timing information.

[0340] The background noise parameter estimation unit can be implemented by a microcontroller, digital signal processor, field-programmable gate array, radar on-chip processor, or multiple processing resources of such processors. The background noise parameter estimation unit can perform DC component removal, outlier sampling point removal, window function processing, complex spectrum transformation, spectral peak search, cross-radar frame time alignment, cross-radar frame phase comparison, and intra-frame position extraction.

[0341] When the background noise parameter estimation unit is implemented using a field-programmable gate array, the complex spectrum transformation and narrowband energy calculation can be implemented by hardware multipliers, accumulators, and memory; when the background noise parameter estimation unit is implemented using a microcontroller or digital signal processor, it can be implemented by calling fast Fourier transform, quadrature demodulation, or narrowband filtering functions through program calls.

[0342] The injection path calibration storage unit is connected to the disturbed node selection unit. The injection path calibration storage unit stores the injection path calibration data corresponding to each candidate disturbed node. The injection path calibration data includes path amplitude propagation, path phase offset, linear injection range, additional noise from the compensation branch, and upper limit of the safe compensation amplitude.

[0343] The injection path calibration storage unit can be made of flash memory, electrically erasable programmable read-only memory, non-volatile random access memory, or the controller's internal parameter storage area. The injection path calibration data can be written during the product's factory calibration phase, or it can be recalibrated and updated after the millimeter-wave radar is installed.

[0344] In an implementation with multiple calibration frequencies, the injection path calibration storage unit stores the corresponding path amplitude transfer and path phase offset according to the candidate disturbed node identifier and calibration frequency, respectively. When the frequency of the background noise component to be compensated is between two adjacent calibration frequencies, the control unit or the disturbed node selection unit interpolates the injection path calibration data corresponding to the two adjacent calibration frequencies to obtain the injection path parameters at the current background noise frequency to be compensated.

[0345] The injection path calibration storage unit can also store background noise observation transmission data from the monitoring node to each candidate disturbed node, which is used to convert the background noise amplitude and phase obtained at the monitoring node into the background noise amplitude and phase at each candidate disturbed node. When the monitoring node and the candidate disturbed node coincide, the background noise observation amplitude transmission amount is 1, and the background noise observation phase offset amount is 0.

[0346] The disturbed node selection unit is connected to the background noise parameter estimation unit, the injection path calibration storage unit, and the control unit, respectively. The disturbed node selection unit determines one candidate disturbed node as the disturbed node from multiple candidate disturbed nodes based on the frequency, amplitude, and phase of the candidate background noise components and the injection path calibration data corresponding to each candidate disturbed node.

[0347] When performing the selection, the disturbed node selection unit first excludes candidate disturbed nodes whose calibrated frequency range does not cover the background noise frequency to be compensated, then excludes candidate disturbed nodes whose required compensation amplitude exceeds the linear injection range or the upper limit of the safe compensation amplitude, and excludes candidate disturbed nodes whose additional noise in the compensation branch exceeds the additional noise threshold; among the remaining candidate disturbed nodes, the candidate disturbed node with the smallest estimated residual noise is selected as the disturbed node.

[0348] The control terminal of the compensation signal generation unit is connected to the control unit. The compensation signal generation unit generates a reference compensation signal with the same frequency as the background noise component to be compensated, based on the compensation frequency control value. The compensation signal generation unit can employ a combination of a digitally controlled oscillator and a digital-to-analog converter, a phase-locked loop frequency divider output circuit, a quadrature local oscillator generation circuit, a direct digital frequency synthesizer, or a programmable waveform generator circuit.

[0349] The compensation signal generation unit can output a single-frequency reference compensation signal or a combined reference compensation signal containing multiple discrete frequency components. When there are multiple background noise components to be compensated that correspond to the same disturbed node and whose compensation periods overlap, the control unit determines the compensation frequency control amount, compensation amplitude control amount, and compensation phase control amount for each frequency component, and controls the compensation signal generation unit to generate the combined reference compensation signal.

[0350] The amplitude and phase adjustment unit is connected between the compensation signal generation unit and the compensation injection unit. The amplitude and phase adjustment unit includes a controllable amplitude adjustment unit and a controllable phase adjustment unit.

[0351] The controllable amplitude adjustment unit is used to adjust the amplitude of the reference compensation signal according to the compensation amplitude control value. The controllable amplitude adjustment unit can be a digital step attenuator, a variable gain amplifier, a voltage-controlled attenuator, or a resistor network attenuator. The controllable phase adjustment unit is used to adjust the phase of the reference compensation signal according to the compensation phase control value. The controllable phase adjustment unit can be a digital phase shifter, a vector modulator, a quadrature weighting network, a delay line, or an analog phase adjustment network.

[0352] In one implementation, the reference compensation signal first passes through a controllable amplitude adjustment unit and then through a controllable phase adjustment unit; in another implementation, the reference compensation signal first passes through a controllable phase adjustment unit and then through a controllable amplitude adjustment unit. Both connection sequences can achieve amplitude and phase adjustment of the compensation signal, and the appropriate connection method can be selected according to the linearity, insertion loss, and control interface of the device used.

[0353] The compensation injection unit is connected between the amplitude and phase adjustment unit and multiple candidate disturbed nodes. The compensation injection unit is connected to each candidate disturbed node via multiple selectable compensation injection paths. Each compensation injection path includes a buffer isolation unit, a DC isolation unit, a limiting protection unit, a compensation injection switch, and corresponding circuit connections. Multiple compensation injection paths can share a buffer isolation unit, a DC isolation unit, or a limiting protection unit, or they can each have their own dedicated unit.

[0354] The buffer isolation unit is used to reduce the reverse impact of impedance changes at disturbed nodes on the compensation signal generation unit and amplitude-phase adjustment unit, and to reduce the reverse coupling of the millimeter-wave radar receiving link signal to the compensation signal generation unit. The buffer isolation unit can be a voltage follower, emitter follower, source follower, isolation amplifier, or a buffer circuit with high input impedance and low output impedance.

[0355] The DC isolation unit is used to prevent the compensation signal from altering the DC bias of the disturbed node. The DC isolation unit can be a coupling capacitor, an isolation transformer, or a differential AC coupling network. The amplitude limiting protection unit is used to limit the amplitude of the compensation signal to within the upper limit of the safe compensation amplitude; it can be a diode limiting network, an active limiter, or a comparator-controlled gain limiting circuit.

[0356] Each compensation injection switch corresponds to a compensation injection path. Based on the disturbed node selection result output by the disturbed node selection unit, the control unit activates the compensation injection switch corresponding to the selected disturbed node and deactivates the compensation injection switches corresponding to other candidate disturbed nodes. The compensation injection switches can be analog switches, radio frequency switches, field-effect transistor switches, or relays.

[0357] The compensation injection unit injects a compensation signal into the disturbed node via small-signal coupling, ensuring that the compensation signal primarily affects the background noise component at the disturbed node without altering its normal DC operating point. The specific structure of the compensation injection unit can be determined based on the frequency range, impedance, voltage range, and allowable load of the disturbed node.

[0358] The control unit is connected to the idle sampling window matching unit, the background noise parameter estimation unit, the disturbed node selection unit, the compensation signal generation unit, the controllable amplitude adjustment unit, the controllable phase adjustment unit, and each compensation injection switch. The control unit can be a microcontroller, a digital signal processor, a field-programmable gate array, a radar on-chip controller, or a combination of the above controllers.

[0359] The control unit receives the frequency, amplitude, phase, repetition confidence, and temporal correlation of candidate background noise components, and allows compensation to be initiated when both the repetition confidence and temporal correlation reach a repetition confidence threshold and a temporal correlation threshold. Based on the disturbed nodes determined by the disturbed node selection unit and the corresponding injection path calibration data, the control unit determines the compensation frequency control amount, compensation amplitude control amount, compensation phase control amount, compensation injection switch status, and compensation period.

[0360] The control unit also determines the upper limit of the compensation amplitude and the parameter update step size based on the repetition confidence level. When the repetition confidence level does not meet the parameter update condition, the control unit stops compensation or maintains the previous set of stable compensation control parameters; when the repetition confidence level meets the parameter update condition, the control unit determines the upper limit of the compensation amplitude, the amplitude parameter update step size, and the phase parameter update step size according to the corresponding confidence level sub-interval.

[0361] The control unit can control the compensation signal generation unit, controllable amplitude adjustment unit, controllable phase adjustment unit, and compensation injection switch through a serial peripheral interface, integrated circuit bus interface, general-purpose input / output interface, digital-to-analog conversion control voltage, or pulse width modulation filter voltage.

[0362] The control unit ensures that the phase difference between the phase of the compensation signal after it reaches the disturbed node and the phase of the background noise component to be compensated at the disturbed node is within the anti-phase deviation threshold range, and limits the amplitude of the compensation signal to within the linear injection range, the upper limit of the safe compensation amplitude, and the upper limit of the current compensation amplitude.

[0363] The noise acquisition unit is also used to acquire residual noise samples from corresponding idle sampling windows with the same intra-frame sampling position identifier in subsequent radar frames after compensation is performed. The background noise parameter estimation unit determines the noise energy before compensation based on the background noise samples, determines the residual noise energy after compensation based on the residual noise samples, and sends the noise energy before compensation and the residual noise energy after compensation to the control unit.

[0364] The idle sampling window matching unit outputs the target state and intra-frame sampling position identifier of the corresponding idle sampling window to the control unit. The target state indicates that the corresponding idle sampling window is in a state without a valid target, and when the temporal correlation of the corresponding idle sampling window reaches the temporal correlation threshold, the control unit compares the noise energy before compensation with the residual noise energy after compensation. If the residual noise energy after compensation meets the preset suppression condition, the control unit maintains the current compensation control parameters; if the residual noise energy after compensation does not meet the preset suppression condition, the control unit adjusts the compensation amplitude control amount or the compensation phase control amount.

[0365] When performing the compensation parameter search, the control unit first controls the controllable phase adjustment unit to perform phase scanning at an amplitude below the upper limit of the compensation amplitude to determine the local minimum value of the residual noise energy, and then maintains the corresponding compensation phase control amount and controls the controllable amplitude adjustment unit to perform amplitude search.

[0366] When the residual noise energy increases after compensation compared to before the parameter update, the control unit restores the previous set of stable compensation control parameters and reduces the parameter update step size. When the phase scan cannot form a stable local minimum of residual noise energy, or when multiple consecutive parameter adjustments fail to reduce the residual noise energy, the control unit disconnects the compensation injection switch and marks the corresponding background noise component as an uncompensable noise component.

[0367] When the target status indicates that a valid target, target echo leakage, or strong reflected clutter has appeared in the corresponding idle sampling window, the control unit does not use the corresponding residual noise sample to correct the compensation control parameters. When the target detection result shows abnormal fluctuations, the compensation signal exceeds the upper limit of the safe compensation amplitude, the additional noise of the compensation branch exceeds the additional noise threshold, or the compensation injection switch is in an abnormal state, the control unit performs at least one of the following protective operations: reducing the compensation amplitude, freezing phase adjustment, switching to monitoring-only state, or disconnecting the compensation injection switch.

[0368] In one specific implementation, the control unit is implemented using a combination of a field-programmable gate array (FPGA) and a microcontroller. The FPGA is used to receive radar frame synchronization signals, control the noise acquisition unit, buffer background noise samples, and perform complex spectrum transformation. The microcontroller is used to calculate repetition confidence, timing correlation, and compensation control parameters, and controls a numerically controlled oscillator, a digital step attenuator, a digital phase shifter, and multiple analog switches through a serial peripheral interface. The injection path calibration storage unit uses non-volatile memory, the compensation injection switch uses analog switches, the DC isolation unit uses coupling capacitors, and the limiting protection unit uses a bidirectional limiting diode network.

[0369] In another implementation, the idle sampling window matching unit, background noise parameter estimation unit, disturbed node selection unit, and control unit are all implemented by an on-chip radar controller with corresponding computing capabilities, while the compensation signal generation unit is implemented using the programmable local oscillator or digital-to-analog converter resources within the radar chip. When the millimeter-wave radar chip does not have amplitude and phase adjustment resources, the controllable amplitude adjustment unit and the controllable phase adjustment unit are implemented using external numerically controlled analog devices.

[0370] The millimeter-wave radar noise compensation circuit based on idle sampling windows obtains corresponding input information through the radar frame timing interface, target detection interface, and interference source timing interface. It establishes corresponding idle sampling window groups through an idle sampling window matching unit, determines candidate background noise components through a noise acquisition unit and a background noise parameter estimation unit, determines disturbed nodes through an injection path calibration storage unit and a disturbed node selection unit, and injects compensation signals into the disturbed nodes through a compensation signal generation unit, an amplitude and phase adjustment unit, and a compensation injection unit. Finally, it uses the residual noise in the subsequent corresponding idle sampling windows to form a closed-loop feedback.

[0371] It should be noted that the embodiments of the present invention have better implementability and are not intended to limit the present invention in any way. Any person skilled in the art may use the above-disclosed technical content to change or modify it into equivalent effective embodiments. However, any modifications or equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. A method for noise compensation in millimeter-wave radar based on an idle sampling window, characterized in that, include: The radar frame timing information, target detection information, and interference source timing information of the internal periodic electromagnetic interference source are acquired. The internal periodic electromagnetic interference source is at least one of the power conversion unit, digital clock unit, and communication interface. Based on the radar frame timing information and the target detection information, corresponding idle sampling window groups with the same intra-frame sampling position identifier are established in multiple consecutive radar frames, and background noise samples are obtained. The frequency, amplitude, phase, and repetition confidence of candidate background noise components are determined based on the background noise samples, and the temporal correlation degree is determined based on the frequency correspondence and intra-frame time correspondence between the candidate background noise components and the temporal information of the interference source. When the repetition confidence and the temporal correlation reach the corresponding threshold, the candidate background noise component is determined as the background noise component to be compensated. The disturbed node is selected and the compensation control parameters are determined based on the background noise component to be compensated and the injection path calibration data of multiple candidate disturbed nodes. A compensation signal is generated based on the compensation control parameters, and the compensation signal is injected into the disturbed node through the compensation injection unit, so that the phase difference between the compensation signal arriving at the disturbed node and the background noise component to be compensated is not greater than the phase deviation threshold relative to 180°. In subsequent radar frames, residual noise samples are acquired within the corresponding idle sampling windows that have the same intra-frame sampling position identifier. When there is no effective target in the corresponding idle sampling window and the temporal correlation reaches the temporal correlation threshold, the compensation control parameters are maintained or corrected based on the background noise samples and the residual noise samples. The upper limit of the compensation amplitude and the parameter update step size are determined based on the repetition confidence.

2. The millimeter-wave radar noise compensation method based on an idle sampling window according to claim 1, characterized in that, Establishing the corresponding idle sampling window group includes: Based on the radar frame timing information, at least one of the following is determined as a candidate sampling interval: the transmission silence interval, the sampling interval corresponding to the range gate where no effective target output is formed, and the sampling interval marked as a non-target interval by the target detection information; Configure the candidate sampling interval in each radar frame with an intra-frame sampling position identifier consisting of at least one of a range gate identifier and an intra-frame time position; Candidate sampling intervals in multiple consecutive radar frames that have the same intra-frame sampling position identifier, whose target detection information all indicate that they are in a state without effective targets, and whose intra-frame position deviation of the corresponding noise spectrum peak is not greater than the position deviation threshold are classified into the same corresponding idle sampling window group. When the target detection information indicates that any candidate sampling interval is in a valid target state, a target echo leakage state, or a strong reflection clutter state, the corresponding candidate sampling interval is removed from the corresponding idle sampling window group.

3. The millimeter-wave radar noise compensation method based on an idle sampling window according to claim 2, characterized in that, Determining the repetition confidence and the temporal correlation includes: The background noise samples are time-aligned across radar frames according to the intra-frame sampling position identifier, and the time-aligned background noise samples are subjected to DC component removal, abnormal sampling point removal, and complex spectrum transformation. Based on the results of the complex spectrum transformation, determine the cross-frame frequency deviation, amplitude dispersion, cross-frame phase deviation, and intra-frame position deviation of each candidate background noise component in multiple consecutive radar frames; The repetition confidence level is determined based on the cross-frame frequency deviation, the amplitude dispersion, the cross-frame phase deviation, and the intra-frame position deviation. Determine the interference source frequency deviation between the frequency of the candidate background noise component and the fundamental frequency, harmonic frequency, or sub-frequency of the internal periodic electromagnetic interference source, and determine the interference time period deviation between the intra-frame time position of the candidate background noise component and the interference action time period represented by the interference source timing information. The timing correlation degree is determined based on the frequency deviation of the interference source and the time deviation of the interference period.

4. The millimeter-wave radar noise compensation method based on an idle sampling window according to claim 3, characterized in that, Determining the upper limit of the compensation magnitude and the parameter update step size based on the repeatability confidence includes: The range of the repeatability confidence is divided into a stop compensation interval, a parameter retention interval, and a parameter update interval; When the repetition confidence level is within the stop compensation range, the injection of the compensation signal into the disturbed node is stopped; When the repeatability confidence is within the parameter holding range, the previous set of stable compensation control parameters is maintained, and the upper limit of the compensation amplitude is reduced according to the preset amplitude attenuation amount; The parameter update interval is divided into multiple consecutive confidence sub-intervals. When the repeat confidence is within the parameter update interval, the upper limit of the compensation magnitude and the parameter update step size are determined according to the confidence sub-interval to which the repeat confidence belongs. The range of values ​​for the intra-frame position deviation is divided into multiple position deviation intervals, and a compensation time period extension is configured for each position deviation interval. The compensation time period is determined based on the position deviation interval to which the intra-frame position deviation of the background noise component to be compensated belongs and the corresponding compensation time period extension, so that the compensation time period is located outside the adjacent effective target echo occupied interval.

5. The millimeter-wave radar noise compensation method based on an idle sampling window according to claim 1, characterized in that, Obtaining the injection path calibration data and selecting the disturbed node includes: During the compensation calibration phase, calibration signals at multiple calibration frequencies are injected through the multiple candidate disturbed nodes, including intermediate frequency bias nodes, analog-to-digital conversion reference nodes, and grounding reference sensitive nodes. Obtain calibration sampling data of each candidate disturbed node before and after the calibration signal is injected, and determine the path amplitude transfer amount, path phase offset amount, linear injection range, additional noise amount of compensation branch and upper limit of safe compensation amplitude of the injection path corresponding to each candidate disturbed node based on the calibration sampling data. The path amplitude transfer amount, the path phase offset amount, the linear injection range, the additional noise amount of the compensation branch, and the upper limit of the safety compensation amplitude are associated with the candidate disturbed node identifier and the calibration frequency to form the injection path calibration data; Based on the amplitude of the background noise component to be compensated and the path amplitude propagation amount corresponding to each candidate disturbed node, the candidate compensation amplitude corresponding to each candidate disturbed node is determined respectively. Based on the amplitude and phase of the background noise component to be compensated, the candidate compensation amplitude corresponding to each candidate disturbed node, the path amplitude transmission amount, and the path phase offset, the estimated residual noise amount corresponding to each candidate disturbed node is determined respectively. From the candidate disturbed nodes whose deviation between the calibrated frequency and the frequency of the background noise component to be compensated is not greater than the frequency matching threshold, whose candidate compensation amplitude is within the linear injection range and the upper limit of the safe compensation amplitude, and whose additional noise amount of the compensation branch is not greater than the additional noise threshold, the candidate disturbed node with the smallest estimated residual noise amount is selected as the disturbed node.

6. The millimeter-wave radar noise compensation method based on an idle sampling window according to claim 5, characterized in that, Determining and executing the compensation control parameters includes: The compensation frequency control amount is determined based on the frequency of the background noise component to be compensated. The compensation amplitude control amount is determined based on the amplitude of the background noise component to be compensated and the path amplitude propagation amount corresponding to the disturbed node, and the compensation amplitude control amount is limited to the linear injection range, the safe compensation amplitude upper limit, and the compensation amplitude upper limit determined based on the repetition confidence. The compensation phase control amount is determined based on the phase of the background noise component to be compensated, the 180° phase inversion, and the path phase offset corresponding to the disturbed node; A reference compensation signal is generated according to the compensation frequency control quantity. The reference compensation signal is then subjected to amplitude adjustment, phase adjustment, impedance isolation, DC isolation, amplitude limiting and injection control in sequence through a controllable amplitude adjustment unit, a controllable phase adjustment unit, a buffer isolation unit, a DC isolation unit, a limiting protection unit and a compensation injection switch to form the compensation signal. The buffer isolation unit, the DC isolation unit, the limiting protection unit and the compensation injection switch constitute the compensation injection unit. The compensation period is determined based on the intra-frame sampling position identifier, and the compensation injection switch is controlled to inject the compensation signal into the disturbed node during the compensation period.

7. The millimeter-wave radar noise compensation method based on an idle sampling window according to claim 6, characterized in that, The correction of the compensation control parameters includes: Under a phase search compensation amplitude lower than the upper limit of the compensation amplitude, the compensation signal is injected sequentially according to multiple candidate compensation phase control quantities, and residual noise samples are obtained in the corresponding idle sampling windows with the same intra-frame sampling position identifier in subsequent radar frames. The residual noise energy is determined based on the residual noise sample corresponding to each candidate compensation phase control quantity, and a phase-residual noise energy sequence is generated according to each candidate compensation phase control quantity and the corresponding residual noise energy. If there is a local minimum of residual noise energy in the phase-residual noise energy sequence, and the deviation of the candidate compensation phase control quantity corresponding to the local minimum in multiple consecutive radar frames is not greater than the phase stability threshold, the candidate compensation phase control quantity corresponding to the local minimum is determined as the compensation phase control quantity to be used. Under the pending compensation phase control quantity, the compensation signal is injected sequentially according to multiple candidate compensation amplitude control quantities, and the candidate compensation amplitude control quantity with the smallest corresponding residual noise energy is determined as the pending compensation amplitude control quantity. The subsequent parameter update direction is determined based on the residual noise energy difference before and after using the pending compensation phase control amount and the pending compensation amplitude control amount, and the compensation control parameters are corrected according to the parameter update step size.

8. The millimeter-wave radar noise compensation method based on an idle sampling window according to claim 7, characterized in that, Also includes: When the background noise sample before compensation and the residual noise sample after compensation used for comparison are respectively from corresponding idle sampling windows with the same intra-frame sampling position identifier, and the target detection information of the two corresponding idle sampling windows both indicate that there is no effective target, and the temporal correlation degree corresponding to the two corresponding idle sampling windows both reach the temporal correlation degree threshold, the residual noise sample after compensation is used to correct the compensation control parameters. If the target detection information of any corresponding idle sampling window indicates the existence of a valid target, or if the temporal correlation degree corresponding to any corresponding idle sampling window is lower than the temporal correlation degree threshold, the corresponding compensated residual noise sample is discarded, and the previous set of stable compensation control parameters is maintained. If the local minimum does not exist in the phase-residual noise energy sequence, or if the deviation of the candidate compensation phase control quantity corresponding to the local minimum in multiple consecutive radar frames is greater than the phase stability threshold, the background noise component to be compensated is marked as an uncompensable noise component, and the compensation injection switch is switched to the off state. If the residual noise energy after the compensation control parameter update is greater than the residual noise energy before the compensation control parameter update, the compensation control parameters before the update are restored, and the parameter update step size is reduced.

9. The millimeter-wave radar noise compensation method based on an idle sampling window according to claim 4, characterized in that, When there are multiple background noise components to be compensated that satisfy the repetition confidence threshold and the temporal correlation threshold, the method further includes: Based on the repetition confidence, amplitude, temporal correlation with the corresponding internal periodic electromagnetic interference source, intra-frame time position, and injection path calibration data corresponding to the multiple candidate disturbed nodes of each background noise component to be compensated, the compensation priority of each background noise component to be compensated and the corresponding disturbed node are determined. When multiple background noise components to be compensated correspond to the same disturbed node and the corresponding compensation periods overlap, a combined compensation signal containing multiple corresponding frequency components is generated, and the compensation amplitude control amount and compensation phase control amount of each corresponding frequency component are determined respectively. When multiple background noise components to be compensated correspond to different disturbed nodes, each background noise component to be compensated is assigned to a compensation channel connected to the corresponding disturbed node. When the number of available compensation channels is less than the number of background noise components to be compensated, multiple background noise components to be compensated are compensated in a time-division manner according to the compensation priority. After each switching of background noise components to be compensated, the corresponding idle sampling window determined by the intra-frame sampling position identifier corresponding to the switched background noise component to be compensated is used to re-determine the corresponding repetition confidence and temporal correlation.

10. A compensation circuit based on the millimeter-wave radar noise compensation method based on an idle sampling window as described in any one of claims 1 to 9, characterized in that, It includes a millimeter-wave radar receiving link, a radar frame timing interface, a target detection interface, an interference source timing interface, an idle sampling window matching unit, a noise acquisition unit, a background noise parameter estimation unit, an injection path calibration and storage unit, a disturbed node selection unit, a control unit, a compensation signal generation unit, an amplitude and phase adjustment unit, and a compensation injection unit. The millimeter-wave radar receiving link is equipped with a monitoring node and multiple candidate interference nodes, including an intermediate frequency bias node, an analog-to-digital conversion reference node, and a ground reference sensitive node. The radar frame timing interface is used to acquire radar frame timing information, the target detection interface is used to acquire target detection information, and the interference source timing interface is used to acquire interference source timing information corresponding to the internal periodic electromagnetic interference source. The radar frame timing interface, the target detection interface, and the interference source timing interface are respectively connected to the idle sampling window matching unit. The idle sampling window matching unit is respectively connected to the sampling control terminal of the noise acquisition unit and the control unit. The idle sampling window matching unit is used to establish corresponding idle sampling window groups with the same intra-frame sampling position identifier in multiple consecutive radar frames according to the radar frame timing information and the target detection information, and output the target status and intra-frame sampling position identifier of the corresponding idle sampling window group to the control unit. The input terminal of the noise acquisition unit is connected to the monitoring node, and the output terminal of the noise acquisition unit is connected to the background noise parameter estimation unit. The background noise parameter estimation unit is connected to the idle sampling window matching unit, the interference source timing interface, and the control unit, respectively. The noise acquisition unit is used to acquire background noise samples within the corresponding idle sampling window group. The background noise parameter estimation unit is used to generate the frequency, amplitude, phase and repetition confidence of candidate background noise components based on the background noise samples, and to determine the temporal correlation degree based on the frequency correspondence and intra-frame time correspondence between the candidate background noise components and the temporal information of the interference source. The injection path calibration storage unit stores the path amplitude transfer amount, path phase offset, linear injection range, additional noise amount of compensation branch, and upper limit of safe compensation amplitude corresponding to each candidate disturbed node. The disturbed node selection unit is connected to the background noise parameter estimation unit, the injection path calibration storage unit, and the control unit, respectively, and is used to determine one candidate disturbed node as the disturbed node from the plurality of candidate disturbed nodes based on the candidate background noise components and the injection path calibration data in the injection path calibration storage unit. The output of the compensation signal generation unit is connected sequentially to the amplitude and phase adjustment unit and the compensation injection unit. The compensation injection unit is connected to the multiple candidate disturbed nodes through multiple selectable compensation injection paths. The amplitude and phase adjustment unit includes a controllable amplitude adjustment unit and a controllable phase adjustment unit. The compensation injection unit includes a buffer isolation unit, a DC isolation unit, a limiting protection unit, and a compensation injection switch corresponding to the multiple selectable compensation injection paths. The control unit is connected to the disturbed node selection unit, the compensation signal generation unit, the controllable amplitude adjustment unit, the controllable phase adjustment unit, and the compensation injection switch, respectively. It is used to determine compensation control parameters based on the candidate background noise components and the injection path calibration data corresponding to the disturbed node when the repetition confidence reaches the repetition confidence threshold and the temporal correlation reaches the temporal correlation threshold; control the corresponding compensation injection path to be turned on based on the disturbed node selection result output by the disturbed node selection unit; and determine the upper limit of compensation amplitude and parameter update step size based on the repetition confidence. The noise acquisition unit is further configured to acquire residual noise samples from the corresponding idle sampling window with the intra-frame sampling position identifier in subsequent radar frames after compensation is performed. The background noise parameter estimation unit is further configured to determine the noise energy before compensation based on the background noise samples and the residual noise energy based on the residual noise samples. The control unit is further configured to maintain or correct the compensation control parameters based on the noise energy before compensation and the residual noise energy when the target state indicates that the corresponding idle sampling window is in a state without effective targets and the temporal correlation degree reaches the temporal correlation degree threshold.

Images