In-vehicle detection data processing method, controller, vehicle, medium and program product

Abstract

The embodiment of the invention provides an in-vehicle detection data processing method, a controller, a vehicle, a medium and a program product, and the method comprises the steps: carrying out the abnormal value replacement of time dimension data of obtained radar data, and generating a target radar input parameter; and performing living body detection based on the target radar input parameter. According to the embodiment of the invention, the radar data can be optimized, more effective data can be provided for subsequent in-vivo detection, and the misjudgment probability is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle electronic detection, and in particular to a method for processing in-vehicle detection data, a vehicle controller, a vehicle, a computer-readable storage medium, and a computer program product. Background Art

[0002] Currently, radar devices such as ultra-wideband radar (UWB) are commonly used to provide input signals for in-vehicle presence detection. However, when detecting children in a vehicle, these devices are susceptible to electromagnetic interference from complex in-vehicle environments, such as metal accessories and electronic devices. This can interfere with the detection signal and cause abnormal radar output data. The low accuracy of the output data from UWB radars and other radar devices can lead to inaccurate liveness detection data and increase the probability of false positives. Summary of the Invention

[0003] In view of the above problems, embodiments of the present invention are proposed to provide an in-vehicle detection data processing method, a vehicle controller, a vehicle, a computer-readable storage medium and a computer program product that overcome the above problems or at least partially solve the above problems.

[0004] In order to solve the above problems, in a first aspect of the present invention, an embodiment of the present invention discloses a method for processing in-vehicle detection data, comprising:

[0005] Replace the outliers in the time dimension of the acquired radar data to generate the target radar input parameters;

[0006] Liveness detection is performed based on the target radar input parameters.

[0007] Optionally, replacing outliers in time dimension data of the acquired radar data to generate target radar input parameters includes:

[0008] Align the fast time dimension data in the acquired radar data to determine the slow time dimension data;

[0009] Outlier replacement is performed on the slow time dimension data to generate target radar input parameters.

[0010] Optionally, aligning the fast time dimension data in the acquired radar data to determine the slow time dimension data includes:

[0011] interpolating the fast time dimension data to determine a first data sequence;

[0012] The first data sequences are aligned to determine slow-time dimension data.

[0013] Optionally, interpolating the fast time dimension data to determine a first data sequence includes:

[0014] Data interpolation is performed on the fast time dimension data to determine a first data sequence.

[0015] Optionally, performing data interpolation on the fast time dimension data to determine the first data sequence includes:

[0016] Determine the interpolation factor;

[0017] Based on the interpolation multiple, data interpolation is performed on each frame of fast time dimension data to determine a first data sequence.

[0018] Optionally, aligning the first data sequence to determine slow-time dimension data includes:

[0019] Alignment is performed based on the maximum power value in the first data sequence to determine slow-time dimension data.

[0020] Optionally, the performing alignment based on the maximum power value in the first data sequence to determine the slow-time dimension data includes:

[0021] Determining a maximum power value and a corresponding sequence position in the first data sequence;

[0022] Based on the sequence positions, the first data sequences are aligned to determine slow-time dimension data.

[0023] Optionally, the aligning the fast time dimension data in the acquired radar data to determine the slow time dimension data further includes:

[0024] Downsampling the first data sequence to obtain a second data sequence, where the second data sequence has the same sampling rate as the first data sequence;

[0025] The second data sequence is used as the first data sequence, and the steps of aligning the first data sequence and determining slow-time dimension data are performed.

[0026] Optionally, downsampling the first data sequence to obtain a second data sequence includes:

[0027] performing low-pass filtering on the first data sequence;

[0028] The first data sequence after low-pass filtering is downsampled to obtain a second data sequence.

[0029] Optionally, replacing outliers on the slow-time dimension data to generate target radar input parameters includes:

[0030] determining an outlier in the slow time dimension data when the number of sequences in the slow time dimension data reaches a preset number;

[0031] The outliers are replaced based on the slow time dimension data to generate target radar input parameters.

[0032] Optionally, determining an outlier in the slow time dimension data includes:

[0033] Traversing the slow time dimension data to determine target slow time dimension data;

[0034] A box quartile test is performed on the target slow time dimension data to determine outliers.

[0035] Optionally, the screening boundary of the box quartile test is ±1.1 times the interquartile range.

[0036] Optionally, performing outlier replacement on the outlier based on the slow-time dimension data to generate target radar input parameters includes:

[0037] Determining a data characteristic value based on each data in the slow time dimension data;

[0038] The abnormal value is replaced based on the data feature value to generate a target radar input parameter.

[0039] Optionally, the step of replacing outliers on the slow-time dimension data to generate target radar input parameters further includes:

[0040] Filter the slow time dimension data after outlier replacement;

[0041] The filtered slow time dimension data is smoothed to generate target radar input parameters.

[0042] Optionally, filtering the slow time dimension data after outlier replacement includes:

[0043] Kalman filtering is performed on the slow time dimension data after outlier replacement.

[0044] Optionally, the step of replacing outliers on the slow-time dimension data to generate target radar input parameters further includes:

[0045] If the number of sequences of the slow time dimension data does not reach the preset number, the steps of aligning the fast time dimension data in the acquired radar data and determining the slow time dimension data are repeated until the number of sequences of the slow time dimension data reaches the preset number. The second data sequence is used as the first data sequence.

[0046] In the second aspect of the present invention, an embodiment of the present invention discloses a vehicle controller, including a processor, a memory, and a computer program stored in the memory and capable of running on the processor, and when the computer program is executed by the processor, the steps of the in-vehicle detection data processing method as described above are implemented.

[0047] In a third aspect of the present invention, an embodiment of the present invention discloses a vehicle, comprising the vehicle controller as described above.

[0048] In a fourth aspect of the present invention, an embodiment of the present invention discloses a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the in-vehicle detection data processing method described above are implemented.

[0049] In a fifth aspect of the present invention, an embodiment of the present invention discloses a computer program product, including a computer program, which implements the steps of the above-mentioned in-vehicle detection data processing method when executed by a processor.

[0050] The embodiments of the present invention include the following advantages:

[0051] This embodiment of the present invention replaces outliers in the time dimension of acquired radar data to generate target radar input parameters; liveness detection is then performed based on these target radar input parameters. By replacing outliers in the time dimension rather than removing them, data optimization can be performed while maintaining the sampling rate during signal sampling in the time dimension, improving data accuracy. This, in turn, increases the accuracy of the input signal for liveness detection and reduces the probability of false positives. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 is a flowchart of the steps of an embodiment of a method for processing in-vehicle detection data according to the present invention;

[0053] Figure 2 is a flowchart of another embodiment of a method for processing in-vehicle detection data according to the present invention;

[0054] Figure 3 is a flowchart of another embodiment of a method for processing in-vehicle detection data according to the present invention;

[0055] Figure 4 It is a schematic diagram of the installation position of the radar of the present invention;

[0056] Figure 5 is a schematic diagram of the radar architecture of the present invention;

[0057] Figure 6 1 is a schematic diagram of an implementation architecture of an example of an in-vehicle detection data processing method of the present invention;

[0058] Figure 7 This is a flowchart of an example of a method for processing in-vehicle detection data according to the present invention. DETAILED DESCRIPTION

[0059] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0060] Reference Figure 1 , shows a flowchart of an embodiment of a method for processing in-vehicle detection data according to the present invention. The method for processing in-vehicle detection data may specifically include the following steps:

[0061] Step 101: Replace outliers in the time dimension of the acquired radar data to generate target radar input parameters.

[0062] When radar data is acquired, the time dimension data in the radar data can be detected, and the outliers therein can be directly replaced. The radar data after the outliers are replaced can be used as the input parameters of the target radar.

[0063] Step 102: Perform liveness detection based on the target radar input parameters.

[0064] The target radar input parameters are used as input data for liveness detection, thereby performing liveness detection and identifying whether a child is present in the vehicle. Liveness detection methods include frequency domain analysis, time domain analysis, time-frequency analysis, variational mode decomposition, and correlation signal extraction. Frequency domain analysis analyzes signals from a frequency perspective, focusing on the representation of the target radar input parameters in the frequency domain (i.e., sinusoidal waves of different frequencies). The target radar input parameters are converted to a frequency domain representation. Frequency domain characteristics, such as spectrum and frequency distribution, are analyzed. Differences in frequency domain characteristics are used to determine whether the input signal originates from a real living person. If a real living person is present, the presence of a child can be determined. Time domain analysis focuses on the representation of the target radar input parameters in the time domain, namely, how the target radar input parameters change over time. The target radar input parameters can be converted to a time-frequency distribution graph, which is then analyzed for frequency and time characteristics. Differences in time-frequency characteristics are used to determine whether the target radar input parameters originate from a real living person, thereby confirming the presence of a child. Variational mode decomposition (VMD) is a signal decomposition method that decomposes the target radar input parameters into multiple modal components. Each modal component corresponds to a specific frequency component for detection. Preprocessing of the target radar input parameters can be performed, such as denoising and contrast enhancement. VMD is applied to decompose the signal into multiple modal components. The frequency and energy characteristics of the modal components are analyzed to extract signal components related to heartbeats. The extracted heartbeat signal is then used to determine whether the input target radar parameters originate from a live individual. Correlation signal extraction (CSE) is a method for extracting specific signal components. It leverages correlation between signals to separate and extract target signals. In liveness detection, CSE can be used to extract signal components related to vital signs such as heartbeat and respiration in the facial region. The specific process includes: preprocessing the target radar input parameters to improve signal quality; using CSE to separate and extract target signal components, such as heartbeat and respiration signals; and analyzing the extracted target signal components to determine whether the input signal originates from a live individual.

[0065] This embodiment of the present invention replaces outliers in the time dimension of acquired radar data to generate target radar input parameters; liveness detection is then performed based on these target radar input parameters. By replacing outliers in the time dimension rather than removing them, data optimization can be performed while maintaining the sampling rate during signal sampling in the time dimension, improving data accuracy. This, in turn, increases the accuracy of the input signal for liveness detection and reduces the probability of false positives.

[0066] Reference Figure 2 , shows a flowchart of another embodiment of a method for processing in-vehicle detection data according to the present invention. The method for processing in-vehicle detection data may specifically include the following steps:

[0067] Step 201, aligning fast time dimension data in the acquired radar data to determine slow time dimension data;

[0068] Radar data can be divided into fast and slow time dimensions. The fast time dimension primarily records radar echo sampling data within each pulse repetition period. This dimension is often used to obtain distance distribution information about objects in the environment. Specifically, after a radar transmits a pulse, it receives echo signals from objects at varying distances. By sampling and analyzing these echo signals, the distance between the object and the radar can be determined. Since this process occurs within each pulse repetition period, it is considered the fast time dimension. In the fast time dimension, the signal bandwidth is limited by the bandwidth of the transmitted pulse. Therefore, the Nyquist sampling rate in the fast time dimension is at least the bandwidth of the transmitted pulse, typically on the order of MHz. This means that to accurately obtain object distance information, the echo signal must be sampled at a sufficiently high frequency. The slow time dimension corresponds to radar echo data over multiple pulse repetition periods. This dimension is primarily used to analyze the motion of objects over time. When there is relative motion between the radar and the detected target, the phase of successive echoes changes with each sample. This phase change reflects the target's motion. Therefore, by analyzing the signal in the slow time dimension, parameters such as the target's velocity can be determined. In the slow time dimension, the signal sampling rate corresponds to the repetition frequency (PRF) of the transmitted pulse, which is generally on the order of kHz. This means that in order to accurately obtain the target's motion information, the echo signal needs to be sampled and analyzed over multiple pulse repetition periods.

[0069] In an embodiment of the present invention, data generated by a radar device during detection inside a vehicle may be received to obtain radar data, and fast time dimension data in the radar data may be aligned to determine slow time dimension data.

[0070] Step 102: replacing outliers on the slow time dimension data to generate target radar input parameters;

[0071] In the slow time dimension, outlier detection can be performed on the slow time dimension data, and the outliers can be replaced based on the overall trend of the slow time dimension data. The replaced data are the target radar input parameters.

[0072] Step 103: Perform liveness detection based on the target radar input parameters.

[0073] Liveness detection is performed on the determined target radar input parameters to determine whether there is a child in the car.

[0074] The embodiment of the present invention determines the slow time dimension data by aligning the fast time dimension data in the acquired radar data; replaces outliers in the slow time dimension data to generate target radar input parameters; and performs liveness detection based on the target radar input parameters. By aligning the fast time dimension data, the radar data is aligned in the fast time dimension, which eliminates the time difference caused by hardware delays, ensures that the fast time dimension data can be processed at the same time scale, eliminates data anomalies caused by misaligned fast time data, and ensures data accuracy; in the slow time dimension, the data can be replaced with outliers at the same time scale. By replacing outliers instead of eliminating outliers, data optimization can be performed on the basis of maintaining the sampling rate when the signal is sampled in the slow time dimension, thereby improving the accuracy of the data, thereby improving the accuracy of the input signal for liveness detection, and reducing the probability of misjudgment.

[0075] Reference Figure 3 , shows a flowchart of another embodiment of an in-vehicle detection data processing method of the present invention, wherein the in-vehicle detection data processing method may specifically include the following steps: Step 301, interpolating fast time dimension data in the acquired radar data to determine a first data sequence;

[0076] Data interpolation can be performed on the fast time dimension data in the radar data to improve the sampling accuracy of the fast time dimension data. The first data sequence is obtained by performing data interpolation on the fast time dimension data. Data interpolation can be performed on each frame of fast time dimension data to obtain multiple groups of first data sequences. The first data sequence is a data set sequence obtained after data interpolation of the fast time dimension data. In one example of the present invention, the radar can be a UWB (Ultra-Wideband) radar. The deployment location of the radar can refer to Figure 4 The UWB radar transmitting module and the UWB radar receiving module are distributed at both ends of the vertical axis of the car. Children can sit in any seat in the car, and the UWB radar can detect people in the car. Figure 5The radar is primarily divided into a transmitter module and a receiver module. In the transmitter module, a waveform generator produces the required pulse waveform, and a power amplifier modulates this waveform onto the radio frequency (RF) and amplifies it to the required power level. A duplexer connects the transmitter output to the antenna. The receiver module consists of an antenna, waveform generator, duplexer, low-noise RF amplifier, mixer, intermediate frequency amplifier, local oscillator, signal processor, and data processor. The echo signal received by the antenna passes through the duplexer and enters the radar receiver. The receiver utilizes a superheterodyne design. The first stage is typically a low-noise RF amplifier. Subsequent modulation stages (one or more) convert the received signal to a lower intermediate frequency (IF) and ultimately to baseband. The baseband signal is fed into a signal processor, which performs signal processing functions such as pulse compression, matched filtering, Doppler filtering, integration, and motion compensation. The output of the signal processor is also transmitted to the data processor.

[0077] In an optional embodiment of the present invention, interpolating the fast time dimension data to determine the first data sequence includes: performing data interpolation on the fast time dimension data to determine the first data sequence.

[0078] Data interpolation may be performed on each set of fast time dimension data, and the first data sequence may be determined by means of interpolated values.

[0079] In an optional embodiment of the present invention, performing data interpolation on the fast time dimension data to determine the first data sequence includes: determining an interpolation multiple; and performing data interpolation on each frame of fast time dimension data based on the interpolation multiple to determine the first data sequence.

[0080] During the data interpolation process, an interpolation factor can be first determined. The interpolation factor can be determined based on MCU (Microcontroller Unit) resource limitations, for example, 2-5 times. This is not specifically limited in the embodiments of the present invention. Based on the determined interpolation factor, data interpolation is performed on each frame of fast-time dimension data in the radar data to increase data accuracy and generate a first data sequence. By performing data interpolation on each frame of fast-time dimension data, multiple sets of first data sequences are obtained. Data interpolation methods include, but are not limited to, zero-value interpolation, linear interpolation, sinusoidal interpolation, and mixed extraction and interpolation. Zero-value interpolation can insert a zero value between two adjacent points in the fast-time dimension data, thereby increasing the sampling rate to achieve the interpolation factor. Linear interpolation calculates the value of any point between two points in the fast-time dimension data based on their coordinates, and then interpolates a corresponding number of values. Sinusoidal interpolation can intercept a periodic function of the fast-time dimension data in the frequency domain, perform an inverse discrete Fourier transform, and interpolate the corresponding values. Hybrid extraction and interpolation can extract data in the fast time dimension data and then interpolate it.

[0081] Step 302: downsample the first data sequence to obtain a second data sequence, where the second data sequence has the same sampling rate as the first data sequence;

[0082] The first data sequence can be downsampled to restore the original sampling rate to obtain a second data sequence. That is, the second data sequence has the same sampling rate as the first data sequence. By ensuring that the sampling rate is fixed, the processing complexity is reduced, thereby further ensuring the practicality and accuracy of data processing.

[0083] In an optional embodiment of the present invention, downsampling the first data sequence to obtain a third data sequence includes: low-pass filtering the first data sequence; and downsampling the low-pass filtered first data sequence to obtain a second data sequence.

[0084] For the downsampling process, a low-pass filter can be set. The filter function of the low-pass filter can be determined according to the frequency of the radar data. The first data sequence is low-pass filtered through the low-pass filter to smooth the data in the first data sequence. The first data sequence after low-pass filtering is then downsampled to reduce the original sampling rate to obtain a second data sequence. Among them, downsampling methods include direct downsampling (retaining one data point every m data points), average downsampling (grouping m data points and replacing the new value of the group with the average value of every m data points), and performing low-pass filtering before direct downsampling.

[0085] Step 303: Use the second data sequence as the first data sequence, align the first data sequence, and determine slow-time dimension data;

[0086] The second data sequence is used as the first data sequence, an alignment position is found from the first data sequence, alignment is performed based on the alignment position, and slow-time dimension data is determined.

[0087] In an optional embodiment of the present invention, aligning the first data sequence to determine the slow time dimension data includes: aligning based on a maximum power value in the first data sequence to determine the slow time dimension data.

[0088] The maximum power value in the first data sequence can be found, and the first data sequence can be aligned based on the maximum power value to obtain slow-time dimension data, so that the transmitted and received radar data can be aligned, avoiding data anomalies caused by misaligned radar data.

[0089] In an optional embodiment of the present invention, the alignment based on the maximum power value in the first data sequence to determine the slow time dimension data includes: determining the maximum power value and the corresponding sequence position in the first data sequence; and aligning the first data sequence based on the sequence position to determine the slow time dimension data.

[0090] During the data alignment process, the corresponding data peak can be found in the first data sequence, and the peak value is used as the maximum power value. The position corresponding to the peak value is used as the sequence position of the maximum power value. Based on this sequence position as a reference, the first data sequences are aligned to generate slow-time dimension data.

[0091] Step 304: replacing outliers on the slow time dimension data to generate target radar input parameters;

[0092] The outlier value replacement can be performed on the slow time dimension data with outliers, and the replaced slow time dimension data can be used as the input parameter of the target radar to be used as the input data for living body detection such as child in car detection.

[0093] In an optional embodiment of the present invention, the step of replacing outliers on the slow-time dimension data to generate target radar input parameters includes:

[0094] Sub-step S3041, when the number of sequences in the slow time dimension data reaches a preset number, determining an outlier in the slow time dimension data;

[0095] It may be first determined whether the number of sequences in the slow time dimension data reaches a preset number. If the number reaches the preset number, outlier identification is performed based on the slow time dimension data to determine the outliers.

[0096] The determining of abnormal values in the slow time dimension data includes: traversing the slow time dimension data to determine target slow time dimension data; and performing a box line quartile test on the target slow time dimension data to determine abnormal values.

[0097] To determine the outliers, you can traverse all the data sequences in the slow time dimension data, take any data sequence as the target slow time dimension data, perform box line quartile detection on the target slow time dimension data, and determine the outliers therein. Traverse all the data sequences in the slow time dimension data to obtain all the outliers. Specifically, the screening boundary of the box line quartile detection is ±1.1 times the interquartile range. The upper and lower quartile points ±1.1 times the interquartile range are used as the screening boundaries. If any distance exceeds the boundary, the outlier and the outlier data position are marked.

[0098] Sub-step S3042: performing outlier replacement on the outliers based on the slow time dimension data to generate target radar input parameters.

[0099] Then, based on the data sequence with outliers in the slow time dimension data as a whole, the outliers are replaced and repaired by replacement, which can effectively reduce the outliers.

[0100] In an optional embodiment of the present invention, a data characteristic value is determined based on each data in the slow-time dimension data; and the abnormal value is replaced based on the data characteristic value to generate a target radar input parameter.

[0101] Based on each data in the slow time dimension data, the characteristics of the data change are determined, and the data characteristic value corresponding to the abnormal position is removed. The characteristics of the data change are determined based on the replacement method. The replacement method includes but is not limited to mean replacement, median replacement, mode replacement, linear interpolation, maximum likelihood estimation and other methods. The data characteristic value is directly used to replace the abnormal value to generate the target radar input parameter. In an optional embodiment of the present invention, the abnormal value replacement of the slow time dimension data to generate the target radar input parameter also includes: sub-step S3043, filtering the slow time dimension data after the abnormal value replacement;

[0102] The slow-time dimension data after outlier replacement can also be filtered, such as low-pass filtering, to further eliminate the outlier data. In one example of the present invention, filtering the slow-time dimension data after outlier replacement includes: performing Kalman filtering on the slow-time dimension data after outlier replacement. Kalman filtering can be performed on the slow-time dimension data after outlier replacement. Kalman includes two parts: prediction and update. When using the Kalman target state prediction formula, only the state transition matrix is used, without adding an input control matrix, thereby improving processing efficiency.

[0103] Sub-step S3044, smoothing the filtered slow time dimension data to generate target radar input parameters.

[0104] The filtered slow time dimension data is smoothed, such as reverse smoothing, so that the data can further conform to the overall rules and ensure the accuracy of the data.

[0105] In an optional embodiment of the present invention, replacing outliers on the slow-time dimension data to generate target radar input parameters further includes:

[0106] Sub-step S3045, when the number of sequences of the slow time dimension data does not reach the preset number, repeat the steps of aligning the fast time dimension data in the acquired radar data and determining the slow time dimension data until the number of sequences of the slow time dimension data reaches the preset number.

[0107] When the number of sequences of slow time dimension data does not reach the preset number, it means that the current data is insufficient, and it is necessary to repeat the steps of aligning the fast time dimension data in the acquired radar data and determining the slow time dimension data until the number of sequences of slow time dimension data reaches the preset number, and then perform outlier processing to realize data processing in the slow time dimension.

[0108] Step 305: Perform liveness detection based on the target radar input parameters.

[0109] The obtained target radar input parameters are used for liveness detection to identify whether there are children in the car and realize child detection in the car.

[0110] The embodiment of the present invention repairs outliers in radar data from both fast and slow time dimensions, thereby improving sampling accuracy in the fast time dimension and achieving higher distance resolution, while utilizing the maximum power value for data alignment. In the slow time dimension, outliers that exceed the statistical scope are repaired, effectively reducing outliers and simplifying the data analysis process, making feature extraction in subsequent child in-vehicle detection more convenient and efficient, thereby improving the accuracy of child in-vehicle detection.

[0111] In order to make the implementation process of the present invention clear to those skilled in the art, a complete example is used below to illustrate:

[0112] Reference Figure 6 , shows a schematic diagram of an implementation architecture of an example of a vehicle in-vehicle detection data processing method according to the present invention. This implementation architecture may include a microcontroller software driver module, a radar transceiver sensor driver module, a child presence detection module, an application logic processing module, and a microcontroller communication module. The functions and division of labor of each software module are described as follows:

[0113] (1) The microcontroller software driver module is responsible for driving the microcontroller's peripheral circuits, including SPI driver, CAN driver, interface driver, system clock driver, watchdog driver, etc.

[0114] (2) The UWB radar signal transceiver sensor software driver module is a complex driver based on the microcontroller software driver module. Its main function is to drive the UWB radar sensor, complete the sending / receiving of UWB radar data, and perform hardware self-test on the UWB radar driver sensor.

[0115] (3) The UWB radar liveness detection module performs a series of processing on the received CIR data, including packaging, accumulation, filtering, and then performing a detection threshold judgment to determine whether there is a live body.

[0116] (4) The functions of the application logic processing module include turning on and off the liveness detection function; CPD algorithm scheduling and result output.

[0117] (5) The microcontroller communication module is the CAN communication module of the microcontroller. This module outputs the liveness detection results to the vehicle system. Then, it is sent from the vehicle system to the mobile phone through the network to achieve the purpose of reminder.

[0118] The specific implementation process can be referred to Figure 7 , including the following steps:

[0119] Step 1: The radar operates normally and collects CIR (Channel Impulse Response) data from the receiver. Each frame of fast time dimension data has N sampling points, and the modulus value is calculated and recorded as S1.

[0120] Step 2: Interpolate S1 to obtain S2. The interpolation factor m should meet the resource constraints of the selected MCU, such as 2-5 times. Interpolation methods include zero-value interpolation, linear interpolation, sinusoidal interpolation, and mixed extraction and interpolation.

[0121] Step 3: Downsample S2 to restore the original sampling rate to obtain S3. Downsampling methods include direct downsampling (retaining one data point every m data points), average downsampling (grouping m data points and replacing the new value in each group with the average value of every m data points), and low-pass filtering followed by direct downsampling.

[0122] Step 4: Find the exact position of the highest peak (maximum power value) of signal S3, and align the sequence to the same position based on the position of the highest peak (maximum power value), which is recorded as S4.

[0123] Step 5: The processed signal S4 is assembled into a two-dimensional matrix CIR of M*N consisting of fast / slow time dimensions.

[0124] Step 6: Determine whether M groups of fast time dimension data have been collected. If so, proceed to step 7. If not, proceed to step 1.

[0125] Step 7: Determine whether all N groups of slow time dimension data have been processed. If so, proceed to step 11; if not, proceed to step 8.

[0126] Step 8: Use the box-and-line quartile detection method to find outliers in the slow time dimension data of group i (where i is an integer less than or equal to N). The upper and lower quartiles ± 1.1 times the interquartile range are set as the screening boundaries. If any distance exceeds the boundary, the outlier and the outlier data position are marked.

[0127] Step 9: Use the entire set of slow-time dimension data features to replace the marked outliers. The replacement methods include mean, median, mode, linear interpolation, maximum likelihood estimation, etc. The replaced data is recorded as S5.

[0128] Step 10: Use Kalman forward filtering and reverse smoothing to further optimize signal S5, which is recorded as S6. Kalman includes two parts: prediction and update. When using the Kalman target state prediction formula, only the state transfer matrix is used, and there is no need to add the input control matrix.

[0129] It should be noted that for the sake of simplicity, the method embodiments are described as a series of actions. However, those skilled in the art should be aware that the embodiments of the present invention are not limited by the order of the actions described, because according to the embodiments of the present invention, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in this specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

[0130] An embodiment of the present invention also discloses a vehicle controller, comprising a processor, a memory, and a computer program stored in the memory and capable of running on the processor. When the computer program is executed by the processor, the steps of the in-vehicle detection data processing method described above are implemented.

[0131] The memory may include random access memory (RAM) or non-volatile memory, such as at least one disk storage. Alternatively, the memory may be at least one storage device located away from the processor.

[0132] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, and discrete hardware components.

[0133] An embodiment of the present invention further discloses a vehicle, comprising the vehicle controller as described above.

[0134] An embodiment of the present invention further discloses a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the above-mentioned in-vehicle detection data processing method are implemented.

[0135] A computer program product comprises a computer program, wherein when the computer program is executed by a processor, the computer program implements the steps of the above-mentioned method for processing in-vehicle detection data.

[0136] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0137] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, apparatus, or computer program products. Thus, embodiments of the present invention may take the form of a fully hardware embodiment, a fully software embodiment, or an embodiment combining software and hardware. Furthermore, embodiments of the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0138] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the process in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0139] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0140] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device so that a series of operating steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable terminal device to implement the process. Figure 1 a process or multiple processes and / or boxes Figure 1A step that specifies a function in one or more boxes.

[0141] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they become aware of the basic creative concepts. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present invention.

[0142] Finally, it should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or terminal device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of additional identical elements in the process, method, article, or terminal device that includes the element.

[0143] The above is a detailed introduction to an in-vehicle detection data processing method, a vehicle controller, a vehicle, a computer-readable storage medium and a computer program product provided by the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for general technical personnel in this field, according to the idea of the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A method for processing in-vehicle detection data, characterized in that: include: Replace the outliers in the time dimension of the acquired radar data to generate the target radar input parameters; Liveness detection is performed based on the target radar input parameters.

2. The method according to claim 1, characterized in that The step of replacing outliers in the time dimension of the acquired radar data to generate target radar input parameters includes: Align the fast time dimension data in the acquired radar data to determine the slow time dimension data; Outlier replacement is performed on the slow time dimension data to generate target radar input parameters.

3. The method according to claim 2, characterized in that The aligning of fast time dimension data in the acquired radar data to determine slow time dimension data includes: interpolating the fast time dimension data to determine a first data sequence; The first data sequences are aligned to determine slow-time dimension data.

4. The method according to claim 3, characterized in that The interpolating the fast time dimension data to determine a first data sequence includes: Data interpolation is performed on the fast time dimension data to determine a first data sequence.

5. The method according to claim 4, characterized in that The performing data interpolation on the fast time dimension data to determine a first data sequence includes: Determine the interpolation factor; Based on the interpolation multiple, data interpolation is performed on each frame of fast time dimension data to determine a first data sequence.

6. The method according to claim 3, characterized in that The aligning the first data sequence to determine slow-time dimension data includes: Alignment is performed based on the maximum power value in the first data sequence to determine slow-time dimension data.

7. The method according to claim 6, characterized in that The performing alignment based on the maximum power value in the first data sequence to determine the slow time dimension data includes: Determining a maximum power value and a corresponding sequence position in the first data sequence; Based on the sequence positions, the first data sequences are aligned to determine slow-time dimension data.

8. The method according to claim 3, characterized in that The step of aligning the fast time dimension data in the acquired radar data to determine the slow time dimension data further includes: Downsampling the first data sequence to obtain a second data sequence, where the second data sequence has the same sampling rate as the first data sequence; The second data sequence is used as the first data sequence, and the steps of aligning the first data sequence and determining slow-time dimension data are performed.

9. The method according to claim 8, characterized in that The downsampling the first data sequence to obtain a second data sequence includes: performing low-pass filtering on the first data sequence; The first data sequence after low-pass filtering is downsampled to obtain a second data sequence.

10. The method according to claim 2, characterized in that The step of replacing outliers on the slow time dimension data to generate target radar input parameters includes: determining an outlier in the slow time dimension data when the number of sequences in the slow time dimension data reaches a preset number; The outliers are replaced based on the slow time dimension data to generate target radar input parameters.

11. The method according to claim 10, characterized in that Determining an outlier in the slow time dimension data includes: Traversing the slow time dimension data to determine target slow time dimension data; A box quartile test is performed on the target slow time dimension data to determine outliers.

12. The method according to claim 11, characterized in that The screening boundaries of the box quartile test are ±1.1 times the interquartile range.

13. The method according to claim 10, characterized in that Performing outlier replacement on the outliers based on the slow time dimension data to generate target radar input parameters includes: Determining a data characteristic value based on each data in the slow time dimension data; The abnormal value is replaced based on the data feature value to generate a target radar input parameter.

14. The method according to claim 10, characterized in that The step of replacing outliers on the slow time dimension data to generate target radar input parameters further includes: Filter the slow time dimension data after outlier replacement; The filtered slow time dimension data is smoothed to generate target radar input parameters.

15. The method according to claim 14, characterized in that The filtering of the slow time dimension data after the outlier replacement includes: Kalman filtering is performed on the slow time dimension data after outlier replacement.

16. The method according to claim 10, characterized in that The step of replacing outliers on the slow time dimension data to generate target radar input parameters further includes: If the number of sequences of the slow time dimension data does not reach a preset number, the steps of aligning the fast time dimension data in the acquired radar data and determining the slow time dimension data are repeated until the number of sequences of the slow time dimension data reaches a preset number.

17. A vehicle controller, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and capable of running on the processor, wherein when the computer program is executed by the processor, the steps of the in-vehicle detection data processing method according to any one of claims 1 to 16 are implemented.

18. A vehicle, characterized in that: Comprising a vehicle controller as claimed in claim 17.

19. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the in-vehicle detection data processing method according to any one of claims 1 to 16 are implemented.

20. A computer program product, characterized in that The method comprises a computer program, which, when executed by a processor, implements the steps of the in-vehicle detection data processing method according to any one of claims 1 to 16.

Images