A machine learning two-wheeler seat occupancy detection method based on millimeter wave radar

Abstract

The application discloses a kind of machine learning two-wheeled vehicle seat detection methods based on millimeter wave radar, belong to vehicle seat detection technical field, millimeter wave radar can be at lower cost, realize personnel target identification based on signal echo signal distribution, and signal timing change seat occupancy state keeps.By the above mode, the method provided by the application makes full use of the characteristics that millimeter wave radar is sensitive to micro-motion, solves the pain point problem that pressure sensor does not distinguish between people, and judges the existence of target based on machine learning, thereby introducing occupancy grid, so that the detection of seat occupancy is more stable.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of vehicle seat detection, in particular to a machine learning two-wheeled vehicle seat detection method based on millimeter wave radar. BACKGROUND

[0002] In recent years, with the acceleration of urbanization and the increasing problem of traffic congestion, two-wheeled electric vehicles have received widespread attention as a green and portable mode of transportation.

[0003] Existing two-wheeled electric vehicles often require users to manually control the power on and off using a key or button, which is not very convenient and can result in power waste due to forgetting to turn off the power. In recent years, seat detection and power start or disconnection methods based on pressure sensors have also emerged.

[0004] However, such solutions also have the following problems:

[0005] 1. High cost;

[0006] 2. False positives caused by placing other items;

[0007] 3. Unstable measured pressure due to bumpy road conditions.

[0008] Therefore, the present application designs a machine learning two-wheeled vehicle seat detection method based on millimeter wave radar to solve the above problems. SUMMARY

[0009] In view of the above shortcomings of the prior art, the present application provides a machine learning two-wheeled vehicle seat detection method based on millimeter wave radar.

[0010] To achieve the above purpose, the present application realizes the following technical solutions:

[0011] A machine learning two-wheeled vehicle seat detection method based on millimeter wave radar, comprising the following steps:

[0012] Step 1: Install the millimeter wave radar inside the seat under the seat cushion, and the millimeter wave radar is used to collect target information in real time;

[0013] Step 2: The millimeter wave radar demodulates the target information while collecting it to obtain the frequency spectrum of the distance dimension;

[0014] Step 3: The millimeter wave radar extracts the phase of the target information according to the DACM algorithm, calculates the difference between the front and rear frames of the phase, and outputs the micro-motion detection frame phase time difference value;

[0015] Step four: the millimeter wave radar determines the existence of the target based on machine learning, inputs the 1D-FFT spectrum information into a random tree training for prediction, uses the weight trained by the random tree to make real-time existence inference, and outputs the inference judgment result;

[0016] Step five: according to the output result of the random tree model and the phase time sequence difference value of the micro-motion detection frame, it is judged whether it meets the person seating standard, and the two-wheeled vehicle is powered on or powered off.

[0017] Further, in step one, the target information is personnel position information.

[0018] Further, obtaining the distance dimension spectrum includes the following steps:

[0019] A1, the millimeter wave radar sends a frequency-modulated continuous pulse signal through a transmitting antenna, receives the reflection echo signal of the target through two receiving antennas, and performs mixing and demodulation processing on the echo signal to obtain two ADC data;

[0020] A2, the two ADC data are respectively subjected to 1D-FFT processing to obtain the distance dimension spectrum.

[0021] Further, the transmitting antenna is a TX antenna.

[0022] Further, the receiving antenna is an RX antenna.

[0023] Further, the frequency-modulated continuous pulse signal contains N linear frequency-modulated signals in one frame period.

[0024] Further, the specific steps of step three are:

[0025] B1, the phase information corresponding to the target information is extracted on each frame complex 1D-FFT spectrum to obtain the real part signal amplitude I(t) and the complex part signal amplitude Q(t);

[0026]

[0027] Wherein, A[t] is the amplitude, is the phase, S[t] is the 1D-FFT function, and j is the imaginary unit;

[0028] B2, the distance dimension spectrum is used to differentiate and then integrate the phase before and after the frame difference, to remove high-frequency noise and restore phase change information, and the calculation formula is:

[0029]

[0030] Wherein, ω(t) is the discretized analog phase angular velocity, is the first derivative of I[t] with respect to t, is the first order derivative of Q[t] with respect to t;

[0031] is restored by integrating the phase angle velocity over N chirps in a frame time, is the phase change information; t is the time series, Delta t is the time series difference, omega is the phase angle velocity, and m is the upper limit of the time series;

[0032] B3, the difference method is used to calculate the difference of the phase of the front and rear frames and output the micro-motion detection frame phase time series difference value, and when the difference value is higher than a threshold value, it indicates that there is micro-motion on the target distance.

[0033] Further, the threshold value depends on the seat environment, the seat cushion thickness and the seat cushion material.

[0034] Further, the specific steps of step four are:

[0035] C1, the 1D-FFT values of N chirps in a frame are added and averaged, the average value is weighted to become an amplitude value, the amplitude value is taken as the 1D-FFT observation amplitude value of the frame, and then the stationary clutter is removed;

[0036] C2, the 1D-FFT mean value and variance in the target distance range of the 1D-FFT observation amplitude value are taken as the characteristic input random tree model;

[0037] C3, the random tree model outputs the inference judgment result.

[0038] Further, the judgment method of the two-wheeled vehicle with a person seated in step five is:

[0039] D1, the random tree model judges that there is a target;

[0040] D2, the micro-motion detection frame phase time series difference value is higher than the threshold value;

[0041] When both conditions are met, it is judged that a person is currently seated, and the two-wheeled vehicle is powered on.

[0042] When any of the two conditions is not met, it is judged that no one is currently seated, and the power supply of the two-wheeled vehicle is turned off.

[0043] Compared with the prior art, the method provided by the present application fully utilizes the characteristics of millimeter wave radar sensitive to micro-motion, and solves the problem of pressure sensor character differentiation.

[0044] 2、The application uses a millimeter wave radar with lower cost to detect point clouds caused by seat deformation due to sitting, and introduces an occupancy grid based on machine learning to determine the existence of a target, so that the detection of sitting and leaving is more stable, and the problem of unstable measured pressure caused by road bumps is avoided. BRIEF DESCRIPTION OF DRAWINGS

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0046] Figure 1 A flow chart of a millimeter wave radar-based machine learning two-wheeled vehicle seat detection method of the present application. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical scheme and advantages of the embodiments of the present application more clear, the technical scheme in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0048] Embodiment one: in some embodiments, referring to the drawings of the specification Figure 1 A millimeter wave radar-based machine learning two-wheeled vehicle seat detection method, comprising the following steps:

[0049] Step one: install the millimeter wave radar inside the seat under the cushion, the millimeter wave radar is used to collect target information in real time, the target information is the position information of the personnel;

[0050] Step two: the millimeter wave radar demodulates the target information while collecting the target information, and obtains the frequency spectrum of the distance dimension;

[0051] The frequency spectrum of the distance dimension includes the following steps:

[0052] A1, the millimeter wave radar sends a frequency-modulated continuous pulse signal through a TX (Transmitter) antenna, receives the reflected echo signal of the target through two RX (Receiver) antennas, and demodulates the echo signal to obtain two-way ADC (Analog to digital Converter) data;

[0053] The frequency-modulated continuous wave signal contains N Chirp signals in a frame period; N>0;

[0054] A2, 1D-FFT (One Dimensional Fast Fourier Transformation) is performed on the two-way ADC data respectively to obtain the frequency spectrum of the distance dimension.

[0055] Step three: the millimeter wave radar extracts the phase of the target information according to the DACM algorithm (Digital Algorithm for Carrier Modulation), calculates the difference between the phases of the front and rear frames, and outputs the micro-motion detection frame phase time sequence difference value;

[0056] The specific steps are as follows:

[0057] B1, the phase information corresponding to the target information is extracted on each frame of complex 1D-FFT spectrum to obtain the real part signal amplitude I(t) and the complex part signal amplitude Q(t);

[0058]

[0059] Wherein, A[t] is the amplitude, is the phase, S[t] is the 1D-FFT function, and j is the imaginary unit;

[0060] B2, the phase difference between the front and rear frames is first differentiated and then integrated using the frequency spectrum of the distance dimension, which removes high-frequency noise and restores the phase change information, and the calculation formula is:

[0061]

[0062] ω(t) is the discretized analog phase angular velocity; is the first derivative of I[t] with respect to t, is the first derivative of Q[t] with respect to t;

[0063]

[0064] Wherein, is restored by integrating the phase angular velocity on N Chirp signals in a frame of time, is the phase change information, t is the time sequence, Δt is the time sequence difference, ω is the phase angular velocity, and m is the upper limit of the time sequence.

[0065] B3, the difference between the front and rear frames is calculated using the difference method, and the micro-motion detection frame phase time sequence difference value is output. Here, when the difference value is higher than a certain threshold, it indicates that there is micro-motion in the target distance.

[0066] The target distance (range bin) is an important concept in radar signal processing, which refers to the discretized distance unit in the radar detection space. In a radar system, the distance of a target is calculated by measuring the time of electromagnetic waves from radar transmission to the target and return. Since the propagation speed of electromagnetic waves in space is fixed, the distance of the target can be determined by time, distance conversion.

[0067] The threshold needs to be adjusted according to different car seat environments, and is related to the thickness of the seat cushion, the material of the seat cushion, etc., and needs to be fine-tuned.

[0068] Step four: the millimeter wave radar judges the existence of the target based on machine learning, inputs the 1D-FFT spectrum information into the random tree training for prediction, uses the weight trained by the random tree to make real-time existence inference, and outputs the inference judgment result;

[0069] The specific steps are:

[0070] C1, the 1D-FFT values of N linear frequency modulation signals in a frame are added and averaged, the average value is weighted to become an amplitude, the amplitude is taken as the 1D-FFT observation amplitude of the frame, and then the stationary clutter is removed;

[0071] C2, the 1D-FFT average value and variance in the target distance range of the 1D-FFT observation amplitude are taken as the feature input random tree model;

[0072] C3, the random tree model outputs the inference judgment result.

[0073] Step five: the millimeter wave radar judges whether it meets the standard of having a person seated according to the output result of the random tree model and the frame phase timing difference value of micro-motion detection.

[0074] The judgment method of two-wheeled vehicle with a person seated is:

[0075] D1, the random tree model judges that there is a target;

[0076] D2, the frame phase timing difference value of micro-motion detection is higher than a certain threshold.

[0077] When both conditions are met, it is judged that a person is currently seated, and the two-wheeled vehicle is powered on.

[0078] When any of the two conditions is not met, it is judged that no one is currently seated, and the power supply of the two-wheeled vehicle is turned off.

[0079] The method provided by the application makes full use of the characteristics of millimeter wave radar sensitivity to micro-motion, and can sensitively identify people or objects, solving the problem of pressure sensor not distinguishing people.

[0080] The application uses a millimeter wave radar with lower cost to detect point clouds caused by deformation of a seat, and judges the existence of a target based on machine learning to introduce an occupancy grid, so that the detection of seat occupancy is more stable, and the problem of unstable measured pressure caused by road bumps is avoided.

[0081] The above examples are only used to illustrate the technical solutions of the present application, but not to limit it; although the present application has been described in detail with reference to the foregoing examples, those skilled in the art should understand that the technical solutions recorded in the foregoing examples can be modified, or some technical features can be replaced by equivalents; and these modifications or replacements will not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A millimeter wave radar based machine learning scooter seat detection method, characterized in that: The method comprises the following steps: Step 1: install a millimeter wave radar in the seat under the seat cushion, the millimeter wave radar is used to collect target information in real time; Step 2: the millimeter wave radar demodulates the target information collected at the same time to obtain the frequency spectrum of the distance dimension; Step 3: the millimeter wave radar extracts the phase of the target information according to the DACM algorithm, calculates the difference between the front and rear frames of the phase, and outputs the micro-motion detection frame phase time difference value; Step 4: the millimeter wave radar judges the existence of the target based on machine learning, inputs the 1D-FFT spectrum information into a random tree training for prediction, uses the weight trained by the random tree to make real-time existence reasoning, and outputs the reasoning judgment result; Step 5: according to the output result of the random tree model and the micro-motion detection frame phase time difference value, it is judged whether it meets the standard of having a person sitting, and the two-wheeled vehicle is powered on or powered off; The specific steps of step 4 are: C1, add the 1D-FFT values of N linear frequency modulation signals in a frame and take the average value, weight the average value to become the amplitude, take the amplitude as the 1D-FFT observation amplitude of the frame, and then remove the stationary clutter; C2, take the 1D-FFT mean value and variance in the target distance range of the 1D-FFT observation amplitude as the feature input of the random tree model; C3, the random tree model outputs the reasoning judgment result; The judgment method of the two-wheeled vehicle with a person sitting in step 5 is: D1, the random tree model judges that there is a target; D2, the micro-motion detection frame phase time difference value is higher than the threshold; When both conditions are met, it is judged that there is a person sitting, and the two-wheeled vehicle is powered on; When any of the two conditions is not met, it is judged that there is no person sitting, and the power supply of the two-wheeled vehicle is turned off.

2. The millimeter wave radar-based machine learning scooter occupancy detection method of claim 1, wherein, In step 1, the target information is personnel position information.

3. The millimeter wave radar based machine learning scooter occupancy detection method of claim 1, wherein, The frequency spectrum of the distance dimension includes the following steps: A1, the millimeter wave radar sends a frequency modulation continuous pulse signal through a transmitting antenna, receives the reflection echo signal of the target through two receiving antennas, mixes the echo signal, and then demodulates and processes to obtain two-way ADC data; A2, the two-way ADC data is processed by 1D-FFT to obtain the frequency spectrum of the distance dimension.

4. The millimeter wave radar-based machine learning scooter seat detection method of claim 3, wherein, The transmitting antenna is a TX antenna.

5. The millimeter wave radar based machine learning scooter seat detection method of claim 3, wherein, The receiving antenna is an RX antenna.

6. The millimeter wave radar-based machine learning scooter seat detection method of claim 3, wherein, The frequency modulation continuous pulse signal contains N linear frequency modulation signals in a frame period.

7. The millimeter wave radar based machine learning scooter seat detection method of claim 1, wherein, The specific steps of step 3 are: B1, extracting phase information corresponding to target information on each frame complex 1D-FFT spectrum, obtaining real part signal amplitude and complex part signal amplitude ; ； where A[t] is the amplitude, is the phase, S[t] is the 1D-FFT function, is the imaginary unit; B2, use the frequency spectrum of the distance dimension to differentiate and integrate the difference between the front and rear frames of the phase, remove high-frequency noise, restore phase change information, and the calculation formula is: ； ； wherein is the phase angle velocity of the discretized simulation; is the first derivative of t, is the first derivative of t, is phase change information, is a time series, is a time series difference, is a phase angular velocity, m is an upper limit of a time series; B3, use the difference method to calculate the difference between the front and rear frames of the phase and output the micro-motion detection frame phase time difference value, when the difference value is higher than the threshold, it means that there is micro-motion in the target distance.

8. The millimeter wave radar-based machine learning scooter seat detection method of claim 7, wherein, The threshold depends on the seat environment, the thickness of the seat cushion and the seat cushion material.

Images