CN111289966B - Motion Information Measurement Method Based on Coherent Phase Tracking of MIMO FMCW Radar - Google Patents

Abstract

A motion information measurement method based on coherent phase tracking of MIMO FM continuous wave radar, by reconstructing the received signal of the MIMO FM continuous wave radar system into a complex number domain beat signal, and further obtaining the motion trajectory of the detected target through a coherent phase target tracking algorithm . The invention can realize the measurement of motion information of sub-millimeter level, and at the same time, the MIMO radar with M sending and N receiving is equivalent to a SIMO radar with 1 sending and M×N receiving, which greatly improves the spatial resolution of the radar and realizes the detection of different directions. Accurate measurement of target movement information.

Description

Motion Information Measurement Method Based on Coherent Phase Tracking of MIMO FMCW Radar

technical field

The present invention relates to a technology in the field of wireless communication, in particular to a method for accurate relative motion measurement based on millimeter-wave radar, which is not only applicable to MIMO-FMCW radar systems, but also can be used without using the virtual array method. It is suitable for single-transmit-single-receive-FMCW radar system and SIMO-FMCW radar system.

Background technique

Since the 20th century, frequency-modulated continuous wave (FMCW) radar has been widely used in road vehicle monitoring and recording systems, automobile anti-collision radar, traffic flow detectors, automatic driving, gesture interaction, medical and other civil fields. FM continuous wave radar has the advantages of simple structure, easy modulation and low cost. The signal modulation of FMCW radar mainly includes three modes: triangular wave modulation, sawtooth wave modulation, and sine wave modulation. Among them, sine wave modulation is mostly used for single target detection, while multi-target detection requires sawtooth wave or triangular wave modulation. Among them, the triangular wave modulation can measure the distance and speed information of the object at the same time, while the traditional sawtooth wave modulation can only measure the distance of the object. The ranging accuracy of traditional FMCW radar is inversely proportional to its modulation bandwidth B, and the accuracy is poor, usually at the centimeter-decimeter level.

After searching the prior art, Guochao Wang et al. proposed in "Application of Linear-Frequency-Modulated Continuous-Wave (LFMCW) radars for tracking of vital signs" To achieve non-contact range tracking of vital signs (such as respiration) using linear frequency modulated continuous wave (LFMCW) radar, the combination of hardware simplicity and tracking accuracy outperforms other remote sensing methods in the biomedical scenarios addressed. However, this prior art has a low working frequency band and is not sensitive enough to small movements; no phase compensation method is proposed for additional changes in phase measurement, and serious deviations may occur during measurement; no time domain filtering is performed, which is greatly affected by transceiver coupling; Measures the angle of an object relative to the radar.

SUMMARY OF THE INVENTION

Aiming at the above shortcomings of the prior art, the present invention proposes a motion information measurement method based on coherent phase tracking of MIMO frequency-modulated continuous wave radar. It realizes the measurement of motion information at the sub-millimeter level, and it also breaks through the limitation that the traditional sawtooth wave modulation cannot measure the speed of the object. In addition, the invention also processes the MIMO radar signal, and converts the MIMO radar with M transmissions and N receptions to be equivalent to a SIMO radar with 1 transmission and M×N reception (also called virtual array antenna method), which greatly improves the radar performance. The spatial resolution is high, and the accurate measurement of the motion information of the target in different directions is realized.

The present invention is achieved through the following technical solutions:

The invention relates to a motion information measurement method based on coherent phase tracking of a MIMO frequency-modulated continuous wave radar. The received signal of a MIMO frequency-modulated continuous wave radar system is reconstructed to obtain a beat frequency signal in a complex number domain, and a detection target is further obtained through a coherent phase target tracking algorithm. movement trajectory.

The reconstruction refers to obtaining a quadrature signal after the received signal is sampled by a multiplier, amplifying and analog-to-digital conversion, and reconstructing a complex domain beat signal.

The coherent phase target tracking algorithm refers to: performing time-domain filtering, phase analysis, and phase compensation on the complex-domain beat signal, so as to obtain the motion trajectory of the detection target.

The MIMO FM continuous wave radar system includes: a phase-locked loop, a radio frequency module, an intermediate frequency amplifying module, an ADC, and an MCU, wherein: the MCU generates different modulation waveforms by configuring the phase-locked loop to realize the corresponding modulation mode, and the phase-locked loop generates the The modulated waveform is input to the voltage-controlled oscillator and generates a transmission signal. The transmission signal is transmitted to the detection target through the transmission antenna and scatters to generate a reflected signal modulated with the movement information of the detection target. The reception signal obtained by each receiving antenna passes through the transmission signal respectively. The quadrature differential multiplier obtains several quadrature signals, and after amplification and analog-to-digital conversion sampling, the MCU restores the motion information of the detection target.

technical effect

The invention as a whole solves the technical problems of the accurate measurement of the motion information of the target object by the MIMO frequency-modulated continuous wave radar and the measurement of the angle of the target object relative to the radar.

Compared with the prior art, the present invention has a higher working frequency band, is very sensitive to small movements, and can realize the measurement of motion information at the sub-millimeter level; the time domain filtering is used to further increase the measurement accuracy; the time domain filtering can not only filter out the transmission and reception In theory, the coupled signal can also filter out the reflected signal in the environment by subtracting the no-load signal in the same environment (referred to as no-load when there is no detection target), which greatly improves the adaptability of the FMCW radar system to the environment; phase The adoption of the compensation method greatly improves the measurement accuracy; the processing of the MIMO radar signal not only reduces the cost but also improves the performance of the system.

Description of drawings

Figure 1 is a schematic diagram of a MIMO FM continuous wave radar system;

Figure 2 is a schematic diagram of a 2×2 radar antenna array;

3 is a schematic diagram of a virtual receiving antenna;

4 is a schematic diagram of palm detection according to an embodiment;

5 is a schematic diagram of palm detection data of an embodiment;

6 is a schematic diagram of the detection of the appearance of a human body in an embodiment;

FIG. 7 is a schematic diagram of human body appearance detection data according to an embodiment.

Detailed ways

As shown in FIG. 1 , the MIMO FM continuous wave radar system involved in this embodiment includes: a phase-locked loop, a radio frequency module, an intermediate frequency amplifying module, an ADC, and an MCU, wherein the MCU generates different modulation waveforms by configuring the phase-locked loop to achieve Corresponding to the modulation mode, the modulation waveform generated by the phase-locked loop is input to the voltage-controlled oscillator (VCO) and generates a transmission signal. The transmission signal is transmitted to the detection target through the transmission antennas TX1, TX2...TXM, and the scattering phenomenon occurs, resulting in the modulation of the detection target movement. The reflected signal of the information is received by the receiving antennas RX1, RX2...RXN. The receiving signal obtained by each receiving antenna and the transmitting signal respectively obtain N-way orthogonal signals through the orthogonal difference multiplier. After amplification and analog-to-digital conversion sampling, the The MCU restores the motion information of the detected target.

Said configuration refers to: sawtooth wave or triangular wave modulation, center frequency f _c , modulation bandwidth B and pulse repetition period (PRT) t ₀ .

其中：A₀是信号幅度，f_c是中心频率，B是调制带宽，t₀是PRT,φ₀初始相位，t是“fast time”(即一个周期内的时间)，

The transmitted signal is approximately

where: A ₀ is the signal amplitude, f _c is the center frequency, B is the modulation bandwidth, t ₀ is the PRT, φ ₀ is the initial phase, and t is the "fast time" (that is, the time in one cycle),

The received signal is approximately s _r (t)=ρA ₀ s _t (t-Δt), where: ρ is the amplitude relationship between the transmitted and received signals, which is mainly related to the transmission loss and the radar scattering area of the detection target. The transmit signal is mixed with the receive signal to obtain the beat frequency signal

其中：c是光速，R(t′)是t′时刻目标距离雷达的距离，t′是“slow time”(即以一个周期为一个离散的时间点，连起来的时间)，该频域形式中，

表示|S_b(f)|的横坐标，可转化得到探测目标的距离，

可转化得到亚毫米级的探测目标的位置变化信息。The frequency domain form of the beat signal is approximated as:

Among them: c is the speed of light, R(t') is the distance between the target and the radar at t', t' is "slow time" (that is, a period is a discrete time point, connected time), the frequency domain form middle,

Represents the abscissa of |S _b (f)|, which can be transformed to obtain the distance of the detected target,

It can be converted to obtain the position change information of the detection target at the sub-millimeter level.

分别是不同发射天线被不同接收天线接收到的电磁波信号的相位。如图3所示，根据电磁波传输距离的不同，

近似为

其中：i＝1，2，j＝1，2，说明每个

都可以等效为一个虚拟的接收天线RX_ij且RX_ij的位置矢量为Ti+Rj，从而二元接收天线阵被扩展为了四元虚拟天线阵。用同样的原理即可将M发N收的MIMO雷达等效为1发M×N收的SIMO雷达。As shown in Figure 2, it is the principle of the virtual array antenna method: taking a 2×2 radar antenna array as an example, the signals transmitted by TX1 and TX2 are all received by RX1 and RX2. In the figure, θ is the azimuth angle of the detection target; I ₁ and I ₂ are the excitation of TX1 and TX2 respectively; k is the wave vector, and the direction is directed by the antenna to the detection target;

are the phases of the electromagnetic wave signals received by different transmitting antennas by different receiving antennas, respectively. As shown in Figure 3, according to the difference of electromagnetic wave transmission distance,

approximately

Where: i=1, 2, j=1, 2, indicating each

Both can be equivalent to a virtual receiving antenna RX _ij and the position vector of RX _ij is Ti+Rj, so the binary receiving antenna array is extended to a quad virtual antenna array. Using the same principle, the MIMO radar with M sending and N receiving can be equivalent to a SIMO radar with 1 sending and M×N receiving.

This embodiment relates to the coherent phase target tracking algorithm of the above system, and the specific steps include:

1) Acquisition of beat frequency signal in complex domain: take the quadrature signals s _bI (p) and s _bQ (p) obtained by a certain channel of RX receiving circuit that is sampled by the ADC and entered into the MCU, where p is an integer. The complex domain beat signal s _b (p)=s _bI (p)+js _bQ (p) is reconstructed.

2) Time domain filtering: ① Take a segment (such as 20 seconds) of the beat frequency signal s _bn (p) when the temporarily sampled or stored system has no detection target (referred to as no-load); ② take s _bn (p) and points 1 to M (=t ₀ f _s ) of s _b (p) to obtain s _bn (p _m ) and s _b (p _m ). Subtract the two to get s _bx (p _m ). ③ Perform fast Fourier transform (FFT) on s _bx (pm ) to obtain S _bx (f), and take its maximum value S _bx ( _f ) _max in the range of 0～2PRT ^-1 . ④ Move all the values of s _bn (p) forward by one place to obtain s _bn1 (p)=s _bn (p+1). Take s _bn1 (p) as the new system no-load signal, and repeat steps ② and ③ to obtain a new maximum value. By analogy, repeat M times to obtain M maximum values, and form them into an array P (serial number 1-M) in sequence. ⑤ Take the sequence number p _x corresponding to the minimum value in P to obtain the beat signal after time domain filtering: s _bx (p)=s _b (p)-s _bn (p+p _x -1).

其中：f_d(t₀n)＝m_n″f_s，m_n″为F(n,m′)_N×M′中第n行最大值的横坐标。②找到频域矩阵F(n,m′)_N×M′中的每行的最大值，并将这些最大值的横坐标另存入向量ix_N×1，去掉向量ix_N×1中的最大值和最小值后取剩余ix_(N-2)×1的平均值并四舍五入为整数m₀。③取频域矩阵F(n,m′)_N×M′中的第m₀列的各点的复相位向量iy_N×1，则初步得到的目标运动轨迹信息r(t₀n)＝iy(n)c/(4πf_c)。3) Phase analysis: ① Take each M point in the beat signal s _bx (p) as a row to obtain the detection matrix R(n,m) _N×M , and perform the detection on each row of the detection matrix R(n,m) _N×M . Fast Fourier transform (FFT) to obtain frequency domain matrix F(n,m') _N×M' , M' depends on the length of fast Fourier transform, p is an integer, sampling rate is f _s , M=t ₀ f _s . At time t ₀ n, the distance of the detectable target is

Wherein: f _d (t _{0 n} )=mn ″f _s , where mn ″ is the abscissa of the maximum value of the _nth row in F(n, m′) _N×M′ . ②Find the maximum value of each row in the frequency domain matrix F(n,m′) _N×M′ , and save the abscissa of these maximum values into the vector ix _N×1 , and remove the maximum value in the vector ix _N×1 After summing the minimum value, take the average value of the remaining ix _(N-2)×1 and round to the integer m ₀ . ③ Take the complex phase vector iy _N×1 of each point in the m _0th column in the frequency domain matrix F(n,m′) _N×M′ , then the initially obtained target motion trajectory information r(t ₀ n)=iy (n) _c /(4πfc).

4) Phase compensation: Due to the asynchronous problem of the circuit system itself, that is, the sampling clock and the FMCW pulse cycle are not synchronized, the small asynchronous error will accumulate over time, causing a linear deviation to be added to the phase information. If no compensation is performed, the measurement will be lose accuracy. For a specific system, the additional linear change of each frequency point can be measured first or temporarily measured. The specific measurement methods are as follows: ① Take the beat signal at no-load for a period of time (for example, 5s), and measure the trajectory r(t ₀ n)=iy(n)c at a certain m ₀ through steps 1) to 3). /(4πf _c ), note that m ₀ in step 2) is specified at this time. ② Take two points p1 and p2 on the trajectory of r(t ₀ n)=iy(n)c/(4πf _c ), and the slope k can be obtained. ③ By changing m ₀ , the value of k at different m ₀ can be measured, denoted as k(m ₀ ). For a certain trajectory r(t ₀ n)=iy(n)c/(4πf _c ), and its corresponding column is m ₀ , the phase compensation result is: r′(t ₀ n)=iy(n)c/( 4πf _c )-k(m ₀ )n. That is, the final detected target motion trajectory information.

This embodiment proves the accuracy of the above method by detecting the micro-motion of the palm:

Parameter configuration: sawtooth wave modulation, center frequency f _c =80GHz, modulation bandwidth B=4GHz, and pulse repetition period (PRT) t ₀ =6ms.

As shown in Figure 4, place the radar board on the desktop, place the palm at about 10cm directly above the radar board and move it up and down slightly, with an amplitude of about 15mm, stop after 7s, and start to move again after 13s, for a total of 15s, using the phase method The measurement target motion information algorithm processes the collected data, and obtains the experimental results shown in Figure 5.

As shown in Figure 5, for the palm motion trajectory obtained by using the coherent phase target tracking algorithm, it can be seen that the motion amplitude, motion time, and stationary time of the measured motion trajectory are basically the same.

The present embodiment proves the accuracy of the above method by detecting the appearance of a human body:

Parameter configuration: sawtooth wave modulation, center frequency f _c =80GHz, modulation bandwidth B=4GHz, and pulse repetition period (PRT) t ₀ =6ms.

As shown in Figure 6, place the radar board on the table and there is no object above the radar board. The human body leans on the edge of the table and bends down until the body enters about 20cm above the radar board for 2 to 3 seconds; stand up straight to the radar board again. There is no object above, and the distance map between the human body and the radar board is obtained by processing the collected data.

As shown in Figure 7, when the human body appears, it can be detected that the human body is about 20cm away from the radar, and the process of the human body leaving can be detected. When the human body does not appear, because the energy detected by the system is very small and there are many interferences, when the detected energy is too small, it can be judged that the human body does not appear, and the detection distance is set to 0, indicating that the human body does not appear.

The same beat frequency signal is processed by using the coherent phase target tracking algorithm proposed by the present invention, and the sub-millimeter level motion information tracking is realized. At the same time, combined with the MIMO radar signal processing, the measurement of the target relative radar angle and the improvement of the spatial resolution are realized.

After specific practical experiments, the above-mentioned device is operated with the parameters of sawtooth wave modulation, 4GHz scanning bandwidth, and 80GHz center frequency.

Compared with the prior art, the method has a higher working frequency band and is more sensitive to micro-motion; the time domain filtering is used to further increase the measurement accuracy; The no-load signal of the system in the same environment filters out the reflected signals in the environment, which greatly improves the adaptability of the FMCW radar system to the environment; the use of the phase compensation method greatly improves the measurement accuracy; generally, the sub-millimeter level is achieved. The above motion measurement accuracy; MIMO radar signal processing not only reduces the cost but also improves the performance of the system.

The above-mentioned specific implementation can be partially adjusted by those skilled in the art in different ways without departing from the principle and purpose of the present invention. The protection scope of the present invention is subject to the claims and is not limited by the above-mentioned specific implementation. Each implementation within the scope is bound by the present invention.

Claims (2)

1. A motion information measurement method based on MIMO frequency modulation continuous wave radar coherent phase tracking is characterized in that a complex domain beat signal is reconstructed from a received signal of an MIMO frequency modulation continuous wave radar system, and a motion track of a detection target is further obtained through a coherent phase target tracking algorithm;

the reconstruction means that: obtaining orthogonal signals and reconstructing a complex domain beat frequency signal after the received signals are subjected to multiplier, amplification and analog-to-digital conversion sampling;

the coherent phase target tracking algorithm is as follows: performing time domain filtering, phase analysis and phase compensation on the complex domain beat frequency signal to obtain a motion track of a detection target;

the MIMO frequency modulation continuous wave radar system comprises: phase-locked loop, radio frequency module, intermediate frequency amplification module, ADC and MCU, wherein: the MCU generates different modulation waveforms by configuring a phase-locked loop to realize a corresponding modulation mode, the modulation waveforms generated by the phase-locked loop are input to the voltage-controlled oscillator and generate a transmitting signal, the transmitting signal is transmitted to a detection target through a transmitting antenna and generates a scattering phenomenon, a reflection signal with modulated detection target motion information is generated, a receiving signal obtained by each receiving antenna and the transmitting signal respectively pass through an orthogonal differential multiplier to obtain a plurality of orthogonal signals, and the orthogonal signals are amplified, subjected to analog-to-digital conversion and sampling and then restored by the MCU to obtain the motion information of the detection target;

the complex-domain beat signal is s_b(p)＝s_bI(p)+js_bQ(p) wherein: p is an integer, s_bI(p) and s_bQ(p) obtaining an orthogonal signal obtained by taking a certain path of RX receiving circuit which enters the MCU through ADC sampling;

the time-domain filtering specifically includes:

firstly, taking a section of beat frequency signal s when the system is idle_bn(p)；

② take s_bn(p) and s_b1 to M points of (p) to obtain s_bn(p_m) And s_b(p_m) And the two are subtracted to obtain s_bx(p_m) Wherein: m ═ t₀f_s；

(iii) to s_bx(p_m) Fast Fourier transform is carried out to obtain S_bx(f) Taking it at 0-2 PRT^-1Maximum value S within the range_bx(f)_max；

Fourthly, s is_bn(p) all values are shifted one bit forward to get s_bn1(p)＝s_bn(p +1), mixing s_bn1(p) repeating the third step as a new signal when the system is in no-load to obtain a new maximum value; repeating the operation for M times by analogy, and forming an array P with the obtained M maximum values in sequence, wherein the sequence number of the array P is 1-M;

taking the number P corresponding to the minimum value in the array P_xAnd obtaining a beat frequency signal after time domain filtering: s_bx(p)＝s_b(p)-s_bn(p+p_x-1)；

The phase analysis specifically includes:

firstly, beat frequency signal s_bx(p) obtaining a detection matrix R (n, M) with every M dots as a row_N×MFor the detection matrix R (n, m)_N×MPerforms fast Fourier transform on each row to obtain a frequency domain matrix F (n, m')_N×M′Wherein: m' depends on the length of the fast Fourier transform, p is an integer, and the sampling rate is f_s，M＝t₀f_sAt t₀The distance of the target detected at the n moment is

Wherein: f. of_d(t₀n)＝m_n″f_s，m_n"is F (n, m')_N×M′The abscissa of the maximum value of the nth row;

② find the frequency domain matrix F (n, m')_N×M′And stores the abscissa of these maxima in a vector ix_N×1Removing vector ix_N×1Taking the rest ix after the maximum value and the minimum value in_(N-2)×1Is rounded off to an integer m₀；

③ frequency domain matrix F (n, m')_N×M′M of (1)₀Complex phase vector iy of points of a column_N×1Then the preliminarily obtained target motion track information r (t)₀n)＝iy(n)c/(4πf_c)；

The phase compensation specifically comprises:

firstly, a beat frequency signal of a period of system no-load time is taken, and m of the beat frequency signal is calculated through time domain filtering and phase analysis₀Trace r (t) of₀n)＝iy(n)c/(4πf_c)；

② get r (t)₀n)＝iy(n)c/(4πf_c) Obtaining a slope k at two points p1 and p2 on the track;

(iii) Change m₀And measuring the k value at different positions and recording the k value as k (m)₀)；

For a certain track r (t)₀n)＝iy(n)c/(4πf_c) Its corresponding column is m₀Then the phase compensation result is: r' (t)₀n)＝iy(n)c/(4πf_c)-k(m₀) n, namely the finally detected target motion track information.

2. The method of claim 1, wherein the signals of different transmitting antennas are distinguished at each receiving antenna, each signal is taken as a new virtual receiving signal, and then the resolution of the target with respect to the radar angle measurement is improved by digital beam forming of the virtual receiving signal.

Images