A radar sensing method for a vehicle tailgate

Abstract

The present invention relates to the technical field of motion detection, and particularly relates to a radar sensing method for a vehicle tailgate, including: A1: Collecting echo signals for the kicking motion of a user; A2: Generating a sample feature sequence according to the echo signals; A3: Generating a plurality of detection parameters corresponding to the kicking motion according to the sample feature sequence; B1: Obtaining the echo signals and generating a detection feature sequence according to the echo signals; B2: Judging whether the detection feature sequence conforms to the detection parameters; if so, it indicates that the kicking motion is detected. The beneficial effects of the present invention are as follows: By collecting echo signals for the kicking motion of a user and generating a sample feature sequence, the effective recording of the kicking motion is realized; at the same time, the simplified description of the sample feature sequence is realized through the detection parameters, so that the recognition algorithm in the actual use process can be simplified, without using a relatively complex recognition algorithm, greatly reducing the computing power requirement for the sensing radar and reducing the cost of the product.

Description

Technical Field

[0001] The present invention relates to the technical field of motion detection, and particularly to a radar sensing method for a vehicle tailgate. Background Art

[0002] The vehicle tailgate refers to the structure at the rear of the vehicle for opening the trunk or the rear compartment. Generally, in a family car, to achieve better structural utilization, the tailgate is usually set to open upward, so that while it is convenient for users to pick up and place items, the impact of the hinge structure on the space is reduced. However, in actual use, since users often need to store large goods in the trunk, it is difficult for users to open the tailgate. Therefore, the inductive tailgate has become an important comfort configuration in high-end vehicles.

[0003] In the prior art, there has been an automatically opening tailgate triggered based on the radar principle. For example, in the product of a certain automobile manufacturer, a gesture recognition radar pointing to the ground is provided under the rear bumper. The gesture recognition radar recognizes the kicking action of the user by emitting high-frequency scanning signals, and then controls the tailgate motor to open the tailgate.

[0004] However, in the actual implementation process, the inventor found that the above technical solutions often rely on existing technical means, that is, an induction system designed based on gesture recognition. Such systems usually have advantages such as high precision and many extracted feature values, but also have problems such as high power consumption and large computing amount due to their complex recognition algorithms, which in turn lead to an increase in the cost of the product and cannot well meet the user's requirements for low cost and high reliability in the relatively simple scenario of foot kick recognition. Summary of the Invention

[0005] In view of the above problems existing in the prior art, a radar sensing method for a vehicle tailgate is provided now.

[0006] The specific technical solution is as follows:

[0007] A radar sensing method for a vehicle tailgate includes an acquisition stage and a detection stage, and the acquisition stage is pre-executed before the detection stage;

[0008] The acquisition stage includes:

[0009] Step A1: Collect echo signals for the kicking action of the user;

[0010] Step A2: Generate a sample feature sequence according to the echo signals;

[0011] Step A3: Generate a plurality of detection parameters corresponding to the kicking action according to the sample feature sequence;

[0012] The detection stage includes:

[0013] Step B1: Obtain the echo signal and generate a detection feature sequence based on the echo signal;

[0014] Step B2: Determine whether the detection feature sequence conforms to the detection parameters;

[0015] If so, it indicates that the kicking action is detected, and a detection result is generated to control the vehicle tailgate according to the detection result;

[0016] If not, it indicates that the kicking action is not detected.

[0017] Preferably, in the step A1 and the step B1, a tailgate radar is used to collect the echo signal;

[0018] The tailgate radar includes four receiving antennas to receive the echo signal respectively.

[0019] Preferably, the step A2 includes:

[0020] Step A21: Perform two-dimensional Fourier transform on one signal frame in the echo signal to obtain a sample two-dimensional matrix;

[0021] Step A22: Perform non-coherent accumulation according to a plurality of the sample two-dimensional matrices to generate a sample range-Doppler map corresponding to the signal frame;

[0022] Step A23: Generate the sample feature sequence according to a plurality of the sample range-Doppler maps.

[0023] Preferably, the step A23 includes:

[0024] Step A231: Generate sample feature values for the sample range-Doppler map;

[0025] Step A232: Generate the sample feature sequence according to a plurality of the sample feature values in chronological order.

[0026] Preferably, in the step A231, the feature value is a weighted average value, and the generation method of the weighted average value includes:

[0027]

[0028] Wherein, WeightedDoppler is the weighted average value, i is the index of the range-Doppler map, Zi is the non-coherent accumulation value corresponding to the index, and Di is the Doppler value corresponding to the index.

[0029] Preferably, the step A3 includes:

[0030] Obtain the functional properties of the sample feature sequence as the detection parameter;

[0031] The functional properties include at least one of: maximum value, minimum value, monotonicity, product of the maximum value and the minimum value, difference between the maximum value and the minimum value, amplitude before the minimum value, and amplitude after the maximum value.

[0032] Preferably, step B1 includes:

[0033] Step B11: Obtain the echo signal and generate a detection two-dimensional matrix according to the echo signal;

[0034] Step B12: Perform non-coherent accumulation on multiple detection two-dimensional matrices to generate a detection range-Doppler map;

[0035] Step B13: Generate the detection feature sequence according to multiple detection range-Doppler maps.

[0036] Preferably, step B11 includes:

[0037] Step B111: Obtain the echo signal, perform one-dimensional Fourier transform on the echo signal, and clear the first two distance arrays;

[0038] Step B112: Perform two-dimensional Fourier transform on the processed echo signal and remove the zero-velocity dimension to generate the detection two-dimensional matrix.

[0039] Preferably, step B13 includes:

[0040] Step B131: Generate detection feature values for the detection range-Doppler map;

[0041] Step B132: Generate the detection feature sequence according to multiple detection feature values in chronological order.

[0042] The above technical solution has the following advantages or beneficial effects: By collecting the echo signal of the user's kicking action and generating a sample feature sequence, the effective recording of the kicking action is realized; at the same time, the simplified description of the sample feature sequence is realized through the detection parameter, so that the recognition algorithm in the actual use process can be simplified, and there is no need to use a more complex recognition algorithm, greatly reducing the computing power requirement for the induction radar. As a result, the above induction method can be set on a processing device with lower performance, greatly reducing the cost of the product. Description of the Drawings

[0043] Referring to the accompanying drawings, the embodiments of the present invention are described more fully. However, the accompanying drawings are only for illustration and explanation and do not constitute a limitation to the scope of the present invention.

[0044] Figure 1 Schematic diagram of the acquisition stage of the embodiment of the present invention;

[0045] Figure 2 Schematic diagram of the detection stage of the embodiment of the present invention;

[0046] Figure 3 Schematic diagram of the sub-steps of step A2 of the embodiment of the present invention;

[0047] Figure 4 Schematic diagram of the sub-steps of step A23 of the embodiment of the present invention;

[0048] Figure 5 Schematic diagram of the kicking feature of the embodiment of the present invention;

[0049] Figure 6 Schematic diagram of the sub-steps of step B1 of the embodiment of the present invention;

[0050] Figure 7 Schematic diagram of the sub-steps of step B11 of the embodiment of the present invention;

[0051] Figure 8 Schematic diagram of the sub-steps of step B13 of the embodiment of the present invention. Detailed implementation manners

[0052] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0053] It should be noted that, without conflict, the embodiments in the present invention and the features in the embodiments can be combined with each other.

[0054] Next, the present invention will be further described in conjunction with the accompanying drawings and specific embodiments, but it is not limited to the present invention.

[0055] The present invention includes:

[0056] A radar sensing method for a vehicle tailgate, including an acquisition stage and a detection stage, and the acquisition stage is pre-executed before the detection stage;

[0057] As Figure 1 shown, the acquisition stage includes:

[0058] Step A1: Collect echo signals for the kicking action of the user;

[0059] Step A2: Generate a sample feature sequence according to the echo signals;

[0060] Step A3: Generate multiple detection parameters corresponding to the kicking motion according to the sample feature sequence;

[0061] As Figure 2 shown, the detection phase includes:

[0062] Step B1: Obtain the echo signal and generate a detection feature sequence according to the echo signal;

[0063] Step B2: Determine whether the detection feature sequence conforms to the detection parameters;

[0064] If so, it indicates that the kicking motion is detected, and a detection result is generated to control the vehicle tailgate according to the detection result;

[0065] If not, it indicates that the kicking motion is not detected.

[0066] Specifically, for the problem in the prior art that the recognition of the kicking motion depends on the gesture recognition algorithm and has a high demand for computing power, in this embodiment, a sample feature sequence is generated from the echo signal, so as to effectively collect the user's kicking motion, and the sample feature sequence is described by a mathematical method, and then the detection parameters representing the sample feature sequence are generated, effectively describing the signal characteristics of the kicking motion. Furthermore, in the actual detection process, the detection feature sequence can be judged by relatively simplified detection parameters, thereby reducing the computing power required for the recognition of the kicking motion and helping to reduce the product cost.

[0067] In the implementation process, the echo signal is the echo signal collected after the radar emits a signal to the detection area, and it may include multiple groups of echo signals according to different settings of the receiving antenna. The sample feature sequence is a time signal sequence generated after processing the echo signal and used to represent the kicking motion, which is used to describe the signal change situation received by the radar during the user's kicking process. The detection parameter is a parameter used to describe the sample feature sequence, that is, the functional property of the sample feature sequence, such as the maximum distance, minimum distance, monotonicity, etc. The detection feature sequence is a feature sequence generated according to the collected signal in the actual detection process, and it is generated by the same processing method as in the acquisition stage. In the actual implementation process, the detection feature sequence is an array stored in storage devices such as memory and registers, and it is updated sequentially over time, thereby realizing the effective monitoring of the monitoring area. During the continuous monitoring process, when a certain segment in the detection feature sequence conforms to the detection parameters, that is, this segment of the detection feature sequence shows the same functional shape as the sample feature sequence, it indicates that the user's kicking motion is obtained, and then a corresponding signal is generated to trigger the opening of the tailgate.

[0068] In a preferred embodiment, in both Step A1 and Step B1, a tailgate radar is used to collect the echo signal;

[0069] The tailgate radar includes four receiving antennas to receive echo signals respectively.

[0070] Specifically, to achieve a better acquisition effect for the kicking motion, the tailgate radar is provided with four receiving antennas in the monitoring direction to receive echoes simultaneously, thereby achieving a better monitoring effect.

[0071] In a preferred embodiment, as Figure 3 shown, step A2 includes:

[0072] Step A21: Perform two-dimensional Fourier transform on one signal frame in the echo signals to obtain a sample two-dimensional matrix;

[0073] Step A22: Perform non-coherent accumulation according to multiple sample two-dimensional matrices to generate a sample range-Doppler map corresponding to the signal frame;

[0074] Step A23: Generate a sample feature sequence according to multiple sample range-Doppler maps.

[0075] Specifically, to achieve a more accurate judgment effect for the kicking motion in the range direction, in this embodiment, the frequency-modulated continuous wave (FMCW) scheme is selected to transmit signals to the detection area and collect radar echoes. In the process of implementing this scheme, multiple linear frequency-modulated signals (Chirps) will be generated in one signal frame (Frame). In step A21, for each linear frequency-modulated signal in one signal frame, two-dimensional Fourier transform is performed to obtain the signal variation in the range dimension and Doppler dimension, and then a two-dimensional matrix is generated. For the case of multiple receiving antennas, it is necessary to generate its sample two-dimensional matrix for each receiving antenna in one signal frame respectively, and then in step A22, non-coherent accumulation is performed on the sample two-dimensional matrices to obtain the complete sample range-Doppler map in this signal frame. Since the sample range-Doppler map represents the characteristics collected by the radar in a single signal frame, combining multiple sample range-Doppler maps can form a sample feature sequence of the range change caused by the kicking motion in chronological order.

[0076] In a preferred embodiment, as Figure 4 shown, step A23 includes:

[0077] Step A231: Generate sample feature values for the sample range-Doppler map;

[0078] Step A232: Generate a sample feature sequence according to multiple sample feature values in chronological order.

[0079] Specifically, to achieve a complete characterization of the kicking motion, in this embodiment, sample feature values are generated for the sample range-Doppler map of a single signal frame, and then multiple sample feature values are arranged and combined in chronological order to form asFigure 5 The sample feature sequence shown, thus reflecting the change in the monitored distance caused by the kicking motion.

[0080] In a preferred embodiment, in step A231, the eigenvalue is a weighted average value, and the generation method of the weighted average value includes:

[0081]

[0082] Wherein, WeightedDoppler is the weighted average value, i is the index of the range-Doppler diagram, Zi is the non-coherent accumulation value corresponding to the index, and Di is the Doppler value corresponding to the index.

[0083] In a preferred embodiment, step A3 includes:

[0084] Obtaining the functional properties of the sample feature sequence as detection parameters;

[0085] The functional properties include at least one of: maximum value, minimum value, monotonicity, product of the maximum value and the minimum value, difference between the maximum value and the minimum value, amplitude before the minimum value, and amplitude after the maximum value.

[0086] Specifically, to implement the detection of the kicking motion through a relatively simplified algorithm, in this embodiment, the sample feature sequence obtained during the detection process is described by a mathematical method, thereby realizing a simplified representation of the signal change caused by the kicking motion, so that the kicking motion can be quickly judged according to the detection parameters during the actual detection process.

[0087] During the implementation process, one or more of the above detection parameters can be set according to actual needs, and the multiple detection parameters can be in an "AND" relationship, an "OR" relationship, or meet several conditions. For example, in one embodiment, taking Figure 5 the sample feature sequence shown as an example, in this embodiment, four functional properties are selected as detection parameters, including: the product of the maximum value and the minimum value is less than -14, the difference in the index distance between the maximum value and the minimum value is between 8 and 13, it is not monotonically increasing from the minimum value to the maximum value, and there are amplitudes less than 2 in 3 to 8 samples before the minimum value and amplitudes less than 2 in 3 to 8 samples after the maximum value.

[0088] In a preferred embodiment, as Figure 6 shown, step B1 includes:

[0089] Step B11: Obtain the echo signal and generate a detection two-dimensional matrix according to the echo signal;

[0090] Step B12: Perform non-coherent accumulation on multiple detection two-dimensional matrices to generate a detection range-Doppler diagram;

[0091] Step B13: Generate a detection feature sequence based on multiple detection range-Doppler diagrams.

[0092] Specifically, to achieve a better recognition effect of the kicking motion during the detection process, in this embodiment, the echo signals of each receiving antenna on a single signal frame are used to generate a detection two-dimensional matrix through the same signal processing process as the acquisition process, the detection range-Doppler diagrams of a single signal frame are non-coherently accumulated, and the detection range-Doppler diagrams are combined according to the time sequence to generate a detection feature sequence, thereby achieving better processing of the signals generated by the kicking motion and facilitating the recognition of the kicking motion based on the detection parameters.

[0093] During the implementation process, the above detection feature sequence generation process is implemented through a four-channel receiving antenna. In one embodiment, the sampling parameter is set to 128 samples, and a single signal frame contains 256 chirp signals, thereby achieving a better recognition effect.

[0094] In a preferred embodiment, as Figure 7 shown, step B11 includes:

[0095] Step B111: Obtain the echo signal, perform a one-dimensional Fourier transform on the echo signal, and clear the first two distance arrays;

[0096] Step B112: Perform a two-dimensional Fourier transform on the processed echo signal and remove the zero-velocity dimension to generate a detection two-dimensional matrix.

[0097] Specifically, for the relatively complex signal situation in the actual implementation process and the possible situation of a long undetected motion, in this embodiment, better processing efficiency is achieved by pruning some signal data during the generation of the detection two-dimensional matrix.

[0098] During the implementation process, taking the embodiment where the above sampling parameter is set to 128 samples and a single signal frame contains 256 chirp signals as an example. In this embodiment, a one-dimensional Fourier transform of 128 is performed on the echo signal, and then only the first 64 distance arrays (range bins) are retained, while the two distance arrays near zero distance are set to zero. Subsequently, a 256-point two-dimensional Fourier transform is performed, and the zero-velocity dimension is removed, thereby generating a detection two-dimensional matrix.

[0099] In a preferred embodiment, as Figure 8 shown, step B13 includes:

[0100] Step B131: Generate detection feature values for the detection range-Doppler diagram;

[0101] Step B132: Generate a detection feature sequence according to multiple detected feature values in chronological order.

[0102] Specifically, to effectively represent the actions detected by the current tailgate radar, in this embodiment, detection feature values are generated from the detection range-Doppler map of a single signal frame, and then multiple sample feature values are arranged and combined in chronological order to form a detection feature sequence in the time domain direction.

[0103] In the actual implementation process, the detection feature sequence can be embodied as a signal feature function that continuously changes according to the external environment or objects during the overall detection process. At the same time, during the detection process, only a specific number of detection feature values are stored in the register unit, thereby forming a "sliding window" on the signal feature function throughout the cycle. For example, in one embodiment, only the detection feature values of 30 signal frames are stored in the register, and each detection feature value is shifted in the register according to the signal frame to delete the detection feature value of the earliest signal frame and add the detection feature value of the new signal frame. When the detection feature sequence composed of multiple detection feature values in the register exhibits the same properties as the detection parameters at a certain moment, that is, when the detection feature sequence meets multiple detection parameters at the same time, it indicates that a kicking action has been detected. For example, taking the Figure 5 sample detection sequence as an example, when the detection feature sequence composed of the detection feature values in the register meets the following properties at a certain moment: the product of the maximum value and the minimum value is less than -14, the difference in the index distance between the maximum value and the minimum value is between 8 and 13, it does not conform to monotonically increasing between the minimum value and the maximum value, and there are amplitudes less than 2 among 3 to 8 samples before the minimum value and amplitudes less than 2 among 3 to 8 samples after the maximum value, it indicates that a kicking action has been detected. Through the above method, the effective recognition of the forward kicking action can be achieved, thereby avoiding false touches caused by other actions, such as lateral kicking and passing of the human body, and improving the detection accuracy.

[0104] The beneficial effects of the present invention are as follows: By collecting echo signals of the user's kicking action and generating a sample feature sequence, the effective recording of the kicking action is realized; at the same time, the simplified description of the sample feature sequence is achieved through detection parameters, so that the recognition algorithm in the actual use process can be simplified, and there is no need to use a relatively complex gesture recognition algorithm, greatly reducing the computing power requirements for the induction radar. As a result, the above induction method can be set on a processing device with lower performance, greatly reducing the cost of the product.

[0105] The above are only preferred embodiments of the present invention, and do not limit the implementation manners and protection scope of the present invention. For those skilled in the art, it should be realized that all equivalent replacements and obvious changes made by using the description and illustrations of the present invention should be included in the protection scope of the present invention.

Claims

1. A radar sensing method for a vehicle tailgate, characterized in that, It includes a collection stage and a detection stage, and the collection stage is pre-executed before the detection stage; The collection stage includes: Step A1: Collect echo signals of the user's kicking motion; Step A2: Generate a sample feature sequence according to the echo signals; The step A2 includes: Step A21: Perform two-dimensional Fourier transform on a signal frame in the echo signals to obtain a sample two-dimensional matrix; Step A22: Perform non-coherent accumulation according to multiple sample two-dimensional matrices to generate a sample range-Doppler map corresponding to the signal frame; Step A23: Generate the sample feature sequence according to multiple sample range-Doppler maps; The step A23 includes: Step A231: Generate sample feature values for the sample range-Doppler map; Step A232: Generate the sample feature sequence according to multiple sample feature values in chronological order; In the step A231, the feature value is a weighted average, and the generation method of the weighted average includes: ； wherein, is the weighted average value, is the index of the range-Doppler map, is the non-coherent accumulation value corresponding to the index, is the Doppler value corresponding to the index; Step A3: Generate multiple detection parameters corresponding to the kicking motion according to the sample feature sequence; The detection stage includes: Step B1: Obtain the echo signals and generate a detection feature sequence according to the echo signals; Step B2: Determine whether the detection feature sequence conforms to the detection parameters; If so, it indicates that the kicking motion is detected, and a detection result is generated to control the vehicle tailgate according to the detection result; If not, it indicates that the kicking motion is not detected.

2. The radar sensing method according to claim 1, wherein In the step A1 and the step B1, a tailgate radar is used to collect the echo signals; The tailgate radar includes four receiving antennas to receive the echo signals respectively.

3. The radar sensing method according to claim 1, wherein The step A3 includes: Obtain the functional properties of the sample feature sequence as the detection parameters; The functional properties include at least one of maximum value, minimum value, monotonicity, product of the maximum value and the minimum value, difference between the maximum value and the minimum value, amplitude before the minimum value, and amplitude after the maximum value.

4. The radar sensing method according to claim 1, wherein The step B1 includes: Step B11: Obtain the echo signals and generate a detection two-dimensional matrix according to the echo signals; Step B12: Perform non-coherent accumulation according to multiple detection two-dimensional matrices to generate a detection range-Doppler map; Step B13: Generate the detection feature sequence according to multiple detection range-Doppler maps.

5. The radar sensing method according to claim 4, wherein The step B11 includes: Step B111: Obtain the echo signals, perform one-dimensional Fourier transform on the echo signals, and clear the first two distance arrays; Step B112: Perform two-dimensional Fourier transform on the processed echo signals and remove the zero-velocity dimension to generate the detection two-dimensional matrix.

6. The radar sensing method according to claim 4, wherein The step B13 includes: Step B131: Generate detection feature values for the detection range-Doppler map; Step B132: Generate the detection feature sequence according to multiple detection feature values in chronological order.

Images