Millimeter-wave orthogonal waveform optimization method and vehicle-mounted radar system for autonomous driving

Abstract

The present invention discloses a millimeter wave orthogonal waveform optimization method for autonomous driving and a corresponding vehicle-mounted radar system. The optimization method comprises the following steps: the first transmitting channel of the vehicle-mounted radar system transmits a positive frequency modulation signal, the second transmitting channel transmits a negative frequency modulation signal, the positive frequency modulation signal and the negative frequency modulation signal are transmitted simultaneously and are orthogonal to each other; multiple receiving channels simultaneously receive the signals transmitted by the first transmitting channel and the second transmitting channel, after the signals are amplified by the power pre-stage, the power is divided into two paths and two reference signals are used for matched filtering respectively; the data of each receiving channel is sequentially subjected to PD accumulation, target detection, and target angle measurement, wherein during PD accumulation, the data within multiple frequency modulation cycles are subjected to Doppler dimension FFT to form range image data, based on which the azimuth angle of the target is estimated, and the vehicle-mounted radar system is used to measure the angle of the target. The present invention improves the waveform isolation between the transmitting signals and improves the angular resolution capability of the vehicle-mounted radar system.

Description

Technical Field

[0001] The present invention relates to a millimeter wave orthogonal waveform optimization method, and in particular to a millimeter wave orthogonal waveform optimization method based on positive and negative frequency modulation and phase coding for the needs of autonomous driving, and also to a vehicle-mounted radar system using the method, belonging to the field of autonomous driving technology. Background Art

[0002] At present, the global automotive industry has gradually reached a consensus that autonomous driving represents the future development direction of the automotive industry. In autonomous driving technology, vehicle-mounted sensing technology represented by vehicle-mounted radar systems is the main means of collecting data such as vehicle environment perception and traffic operation environment, and plays an irreplaceable and important role.

[0003] In the vehicle radar system, the forward collision avoidance radar is a microwave radar sensor dedicated to the motor vehicle driving assistance system (ADAS), mainly used for active collision avoidance or pre-collision system (Precrash system), automatic emergency braking system (AEB), adaptive cruise control system (ACC), blind spot detection (BSD), front collision warning (FCW), lane change assist (LCA), deviation warning system (LDWS), safe distance warning (TTC), rear cross traffic alert (RCTA), and assisting motor vehicles to complete obstacle avoidance functions. Due to the complex road conditions on which cars actually drive, the forward collision avoidance radar of the car needs to have high resolution and high refresh rate (low detection cycle) so that it can perceive the driving environment of the car more accurately and quickly.

[0004] In the prior art, automotive forward collision avoidance radars mostly use MIMO (Multi-Input Multi-Output) antenna design, that is, multiple transmission channels are used to send LFMCW (linear frequency modulation continuous wave) signals in time-sharing. However, the existing time-division MIMO technology is not specifically developed for the needs of autonomous driving technology, and there are many deficiencies in specific performance. To this end, technicians have carried out a lot of research work in reducing the detection cycle and optimizing the detection waveform. For example, in the Chinese invention application with application number 201910711214.2, a linear frequency modulation continuous wave waveform optimization method under MIMO technology is disclosed. First, on the basis of transmitting a linear frequency modulation continuous wave, a random phase is added to each pulse of the waveform emitted by each transmitting end, so that the transmitted waveform can be separated in the Doppler dimension. Secondly, the waveform is optimized using the main lobe gain, main-to-side lobe ratio, main lobe width, etc. as evaluation indicators, and the phase sequence that makes the evaluation index reach the optimal value is calculated through the pattern search method, and substituted into the transmitting waveform to achieve optimization. However, when the target speed is large enough, a phase difference will occur when different emission waveforms reach the target, which will eventually cause the target to split during the angle measurement process and affect the waveform optimization effect. Summary of the invention

[0005] The primary technical problem to be solved by the present invention is to provide a millimeter-wave orthogonal waveform optimization method for autonomous driving.

[0006] Another technical problem to be solved by the present invention is to provide a vehicle-mounted radar system adopting the method.

[0007] In order to achieve the above object, the present invention adopts the following technical solutions:

[0008] According to a first aspect of an embodiment of the present invention, a millimeter wave orthogonal waveform optimization method for autonomous driving is provided, comprising the following steps:

[0009] The first transmitting channel of the vehicle-mounted radar system transmits a positive frequency modulation signal, and the second transmitting channel transmits a negative frequency modulation signal, wherein the positive frequency modulation signal and the negative frequency modulation signal are transmitted simultaneously and are orthogonal to each other;

[0010] The multi-channel receiving channel of the vehicle-mounted radar system receives the positive frequency modulation signal and the negative frequency modulation signal simultaneously, and after the power pre-amplification, the power is divided into two channels and the two reference signals are used for matched filtering respectively;

[0011] The data of each receiving channel is subjected to pulse Doppler accumulation, target detection, and target angle measurement in turn. During pulse Doppler accumulation, the data within multiple frequency modulation cycles are subjected to fast Fourier transform in the Doppler dimension to form range image data, based on which the azimuth angle of the target is estimated, thereby realizing the angle measurement of the target by the vehicle-mounted radar system.

[0012] Preferably, the positive frequency modulation signal is divided into a plurality of sub-pulse signals, and a corresponding phase code is superimposed on the initial phase of each sub-pulse signal.

[0013] Preferably, the expression of the positive frequency modulation signal is as follows:

[0014]

[0015] Where A represents the signal modulation amplitude, i∈{1~52} represents the i-th sub-pulse, rect(g) represents the rectangular function, T represents the time length of the frequency modulation cycle, B represents the bandwidth, and Bc i is the ith value of the Barker code sequence, K represents the frequency modulation slope of the linear frequency modulation continuous wave signal, and the calculation method is

[0016] Preferably, the negative frequency modulation signal is not phase encoded.

[0017] Preferably, the expression of the negative frequency modulation signal is as follows:

[0018]

[0019] Where A represents the signal modulation amplitude, i∈{1~52} represents the i-th sub-pulse, rect(g) represents the rectangular function, T represents the time length of the frequency modulation cycle, B represents the bandwidth, and Bc i is the ith value of the Barker code sequence, K represents the frequency modulation slope of the linear frequency modulation continuous wave signal, and the calculation method is

[0020] Preferably, the expressions of the two reference signals REF0 and REF1 are as follows:

[0021]

[0022] Preferably, the expression of the distance profile data is as follows:

[0023]

[0024] Preferably, the reference signal is a power division coupled signal of the transmission signal.

[0025] Preferably, the multiple receiving channels have corresponding receiving antennas respectively, the spacing length d between the receiving antennas is 0.5λ, and the spacing between the transmitting antennas and between the transmitting antennas and the receiving antennas is 2N times the wavelength λ of the transmitting signal, where N is a positive integer.

[0026] According to a second aspect of an embodiment of the present invention, a vehicle-mounted radar system is provided, which adopts the above-mentioned millimeter-wave orthogonal waveform optimization method.

[0027] Compared with the prior art, the millimeter wave orthogonal waveform optimized by the method of the present invention can obtain a waveform isolation index of 60dB, which is higher than the dynamic measurement range of general radars and can fully meet the requirements of vehicle-mounted radar systems, especially automotive forward collision avoidance radars. Compared with time-division MIMO technology, the present invention only needs half the detection time to achieve the same angular resolution; compared with ordinary simultaneous MIMO technology, the present invention has a higher inter-channel distance and reduces the accumulated loss generated by the target cross-distance unit, and has good commercial application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 A schematic diagram of a vehicle-mounted radar system using the millimeter-wave orthogonal waveform optimization method provided by the present invention;

[0029] Figure 2 Schematic diagram of phase encoding of positive frequency modulation signal S_0(t) and negative frequency modulation signal S_1(t);

[0030] Figure 3is a time-frequency diagram of a positive frequency modulation signal S_0(t) and a negative frequency modulation signal S_1(t);

[0031] Figure 4 Schematic diagram of a sawtooth-shaped continuous wave signal with frequency modulation and phase coding transmitted for each transmission channel;

[0032] Figure 5 It is a schematic diagram of the effect of the number of waveform coding cycles on the isolation index;

[0033] Figure 6 For the orthogonal waveforms optimized according to the present invention, CH0 and CH N-1 Schematic diagram of the channel’s range profile signal simulation results. DETAILED DESCRIPTION

[0034] The technical content of the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0035] As mentioned above, the existing time-division MIMO technology is not developed specifically for the needs of autonomous driving technology, and has many deficiencies in specific performance. For example, traditional automotive forward collision avoidance radars mostly use time-division MIMO technology, that is, M (M is a positive integer, the same below) transmission channels transmit signals in sequence, and the receiving channels receive the target echoes generated by different transmission signals in a time-division manner. The disadvantage is that the detection cycle is long, and M transmission cycles are required to output a complete detection result.

[0036] To this end, the present invention first provides a millimeter-wave orthogonal waveform optimization method for autonomous driving needs. This optimization method is used to implement a 2-transmit-N (i.e., M=2, N is a positive integer) vehicle-mounted radar system by designing 2 orthogonal waveforms (positive frequency modulation signals and negative frequency modulation signals, collectively referred to as transmission signals). The consideration of designing only 2 transmission signals here is that when the number of transmission channels is greater than 2, the signal isolation advantages brought by the positive and negative frequency modulation signals cannot be utilized, and the isolation of the generated orthogonal signals is reduced. In addition, since the angular resolution of the radar using MIMO technology is determined by M*N, expanding the number of receiving channels with lower cost and lower heat consumption is a more economical choice.

[0037] In the vehicle-mounted radar system, the voltage ratio of the mismatch signal to the matching signal is defined as waveform isolation, which evaluates the isolation level of the vehicle-mounted radar system. The present invention improves the waveform isolation between the two transmission signals by optimizing the orthogonal characteristics of the transmission signal, thereby improving the angular resolution capability of the vehicle-mounted radar system while maintaining the target detection cycle. The specific implementation process of the optimization method is described in detail below:

[0038] Figure 1The schematic diagram of the vehicle-mounted radar system using the millimeter-wave orthogonal waveform optimization method provided by the present invention. In one embodiment of the millimeter-wave orthogonal waveform optimization method, the vehicle-mounted radar system used is a simultaneous MIMO radar including 2 transmission channels (including transmitting antennas TX0 and TX1) and N receiving channels (referred to as 2 transmit and N receive). Among them, the 2 transmission channels have frequency modulation and phase modulation capabilities respectively; the N receiving channels have receiving antennas RX0~RX N , these receiving antennas RX0~RX N The spacing length d between them is 0.5λ, and the spacing between the transmitting antennas and between the transmitting antenna and the receiving antenna is 2N times the wavelength λ of the transmitting signal.

[0039] exist Figure 1 In the embodiment of the vehicle-mounted radar system shown, two transmitting channels start working at the same time, and waveform modulation is performed on the positive frequency modulation signal and the negative frequency modulation signal respectively transmitted by the two transmitting channels. Among them, the positive frequency modulation signal and the negative frequency modulation signal are orthogonal to each other. At this time, the two transmitting signals need to be waveform modulated by phase encoding. The specific waveform modulation method can be: the first transmitting channel TX0 is designed as a positive frequency modulation signal S_0(t), the frequency modulation slope K is the ratio of the bandwidth B to the frequency modulation period (abbreviated as Prt), and the signal within a Prt period is divided into several sub-pulse signals. In order to ensure the integrity of the Barker code encoding, the number of sub-pulses needs to be an integer multiple of 13, and every 13 sub-pulses is 1 encoding period. The specific encoding method of several periods depends on the specific radar waveform design, but the encoding method has little effect on the isolation index. In the embodiment of the present invention, after the waveform design and isolation simulation, it is found that the isolation achieved by the encoding method of 4 periods is optimal. The corresponding comparison results are shown in FIG. Figure 5 As shown, the embodiment of the present invention takes 4×13=52 sub-pulses as an example, that is, the corresponding 52-bit Barker code (ie, phase encoding, the same below) is superimposed on the initial phase of each sub-pulse signal.

[0040] Correspondingly, the expression of the positive frequency modulation signal S_0(t) is as shown in the formula:

[0041]

[0042] In the above formula, A represents the signal modulation amplitude, i∈{1~52} represents the i-th sub-pulse, rect(g) represents the rectangular function, T represents the time length of the frequency modulation period (Prt), B represents the bandwidth, and Bc i is the ith value of the Barker code sequence, K represents the frequency modulation slope of the LFMCW (linear frequency modulation continuous wave) signal, and the calculation method is

[0043] The phase encoding diagram of the positive frequency modulation signal S_0(t) is as follows: Figure 2 As shown in the upper part, the phase encoding diagram of the negative frequency modulation signal S_1(t) is as follows Figure 2 The lower part shows that one coding cycle is the time length of one Prt, and the minimum coding time unit is Δt, which is the time length T of Prt. The time-frequency diagram of the positive frequency modulation S_0(t) signal is as follows: Figure 3 As shown in the left half, the signal is a LFMCW signal, the frequency modulation frequency is modulated from f1 to f2, the bandwidth B = |f2-f1|, and the time is one frequency modulation cycle, that is, the time length of one Prt T. Here, the positions of f1 and f2 only indicate which is the starting frequency and which is the ending frequency, and the actual values ​​are the same.

[0044] The second transmitting channel TX1 is designed as a negative frequency modulation signal S_1(t), and the frequency modulation slope is the ratio of the bandwidth B to the cycle time Prt. The negative frequency modulation signal S_1(t) is not phase encoded, and the expression is as shown in the formula:

[0045]

[0046] The negative frequency modulation signal S_1(t) is also a LFMCW signal, and its time-frequency diagram is as follows: Figure 3 The right half is shown. The FM frequency is modulated from f2 to f1, the bandwidth B = |f1-f2|, and the time is one FM cycle, that is, the time length T of one Prt.

[0047] In one embodiment of the present invention, Figure 4 As shown, each transmitting channel transmits a zigzag frequency modulated and phase-encoded LFMCW signal, and performs binary phase encoding on the LFMCW signal within a single Prt period. Within a Prt period, it is divided into 4×13=52 equal parts by time, and a binary phase encoding is designed with a length of τ. The Barker code with a length of 13 is (1,1,1,1,1,-1,-1,1,1,-1,1,-1,1), which is extended to 4 times the length to generate a 52-bit Barker code, which can be expressed as:

[0048]

[0049] In the above embodiment, the LFMCW signal with a time length of Prt is divided into 52 shorter, continuous sub-pulses, and the time width of each sub-pulse is Δτ=T / (13×4). The phase of a sub-pulse is -π or π according to the Barker code, and the sub-pulse with a phase of π is recorded as "1", and the sub-pulse with a phase of -π is recorded as "-1". It should be noted that the number of equal parts into which the frequency modulation period (Prt) is divided can be determined by simulation in combination with the specific radar waveform design, and it only needs to be ensured to be an integer multiple of 13.

[0050] When the vehicle-mounted radar system is working, the two transmission channels TX0 and TX1 transmit signals at the same time. The two transmission signals are reflected by the target and reach the N receiving channels. The N receiving channels RX0~RX N Receive signals at the same time.

[0051] For N receiving channels RX0~RX N After the received signal is amplified by the power pre-stage, it is copied (power-divided) into two paths and then matched-filtered using two reference signals REF0 and REF1. The project is implemented by mixing with a mixer, where the source of the reference signal can be the power-divided coupled signal of the transmitted signal as the local oscillator signal. By using the orthogonal characteristics of the two transmitted signals, eight received signals with spatial resolution can be obtained, thereby realizing radar signal detection of the vehicle-mounted radar system.

[0052] Specifically, according to the above formula and, after the two transmitted signals are reflected by the target at a distance R, the echo delay time is c is the speed of light, then the echo signal R0(t) received by receiving channel 0 contains both TX0 and TX1 transmission signals, which can be specifically expressed as:

[0053]

[0054] In order to distinguish the TX0 and TX1 receiving signals from R0(t), it is necessary to use different reference signals and echo signals for matched filtering, which is a mixing process in hardware implementation. The expressions of the two reference signals REF0 and REF1 are shown in formula (5):

[0055]

[0056] When the vehicle-mounted radar system is working normally, N channels receive signals simultaneously. From the basic knowledge of MIMO radar, we know that since the distance between multiple channels is fixed, for any target within the radar field of view, RX0~RX NThe phase of the channel receiving signal differs by a fixed value. The fixed deviation phase is used to estimate the deviation angle of the target compared to the normal direction of the radar antenna. The more channels there are, the higher the angular resolution. When N receiving channels receive signals at the same time, the signal processing method similar to the RX0 channel can be used to obtain the range image signals CH0(f)~CH 2N-1 (f), the specific calculation method is:

[0057]

[0058] The data of each channel is sequentially subjected to PD (pulse Doppler) accumulation, target detection, and target angle measurement. During PD accumulation, the data of multiple Prts are subjected to FFT (fast Fourier transform) in the Doppler dimension, which can separate targets with motion characteristics from stationary targets to form range image data. 2N-1 (f) The signals at the same position of the target are taken out to obtain the complex signals of the 2N targets. The complex signals contain a fixed phase difference formed by the angle of the target deviating from the normal direction of the antenna. By performing FFT on the complex signals of the 2N targets, the angle of the target deviating from the normal direction of the radar antenna can be calculated, and the azimuth angle of the target can be estimated based on this, so as to realize the angle measurement of the target by the vehicle-mounted radar system.

[0059] The following is a detailed description of this: The target detection method can use the traditional Ca-Cfar moving target detection algorithm to extract local peak points on the range Doppler plane. The distance and Doppler position corresponding to the local peak point can be used to calculate the distance and speed of the detected target. Finally, based on the peak point data of 2N channels, the deviation angle of the target relative to the normal of the radar antenna can be calculated. The above method is a commonly used MIMO radar signal processing method and will not be described in detail here.

[0060] In one embodiment of the present invention, after the two received signals are mixed, CH0 and CH N The time domain signals of the two channels can be converted to the frequency domain by FFT to obtain the one-dimensional range image information. The calculation method is expressed as:

[0061]

[0062] The “*” in the formula is the dot product processing, which represents the mixing in hardware implementation, and obtains CH0(t) and CH N The range image expression of the (t) channel is:

[0063]

[0064] In the present invention, waveform isolation is defined as the ratio of the correct matching signal to the waveform leakage power of the adjacent transmission channel, specifically:

[0065]

[0066] The meaning of the waveform isolation index is the logarithm of the ratio of the power of the matching signal to the power of the adaptation signal for a target signal at a certain echo point. This index can be used to evaluate the quality of orthogonal waveform design, that is, the higher the waveform isolation index, the smaller the interference between MIMO radar channels.

[0067] The present invention takes the following radar waveform parameter design as an example to simulate the waveform isolation index that can be achieved by the method proposed by the present invention using positive and negative frequency modulation plus 52-bit intra-pulse phase encoding.

[0068] In an embodiment of the present invention, an example of optimized parameters of the millimeter wave orthogonal waveform is shown in Table 1:

[0069] Table 1

[0070]

[0071]

[0072] The millimeter wave orthogonal waveform optimization method provided by the present invention is used to simulate the isolation results. The simulation results are as follows: Figure 6 As shown. From the simulation results, it can be seen that the present invention designs a millimeter-wave orthogonal waveform based on 2-transmit N-receive, which can obtain a waveform isolation index of 60dB, which is higher than the dynamic measurement range of general radars and can fully meet the use requirements of vehicle-mounted radar systems, especially automotive forward collision avoidance radars. It can be seen that the waveform optimization method of combining frequency orthogonality with phase orthogonality provided by the present invention can significantly improve the waveform isolation of millimeter-wave signals.

[0073] On the other hand, the high isolation transmission waveform optimized by the method of the present invention can replace the commonly used time-division MIMO technology. The commonly used time-division MIMO technology uses a time-division method to turn on the signals of the transmission channels in sequence to avoid signal crosstalk between the two transmission channels. After adopting the waveform optimization method provided by the present invention, the two transmission signals can be turned on at the same time, while the angular resolution of the forward collision avoidance radar of the car remains unchanged, the detection cycle is shortened to half of the time required by the time-division MIMO technology, and the accumulated loss generated by the target cross-distance unit is reduced, which has good commercial application value.

[0074] The above is a detailed description of the millimeter wave orthogonal waveform optimization method and vehicle-mounted radar system for autonomous driving provided by the present invention. For those skilled in the art, any obvious changes made to it without departing from the essence of the present invention will fall within the scope of protection of the patent right of the present invention.

Claims

1. A millimeter wave orthogonal waveform optimization method for autonomous driving, characterized in that The steps include: The first transmitting channel of the vehicle-mounted radar system transmits a positive frequency modulation signal, and the second transmitting channel transmits a negative frequency modulation signal, wherein the positive frequency modulation signal and the negative frequency modulation signal are transmitted simultaneously and are orthogonal to each other; The multi-channel receiving channel of the vehicle-mounted radar system receives the positive frequency modulation signal and the negative frequency modulation signal simultaneously, and after the power pre-amplification, the power is divided into two channels and the two reference signals are used for matched filtering respectively; The data of each receiving channel is subjected to pulse Doppler accumulation, target detection, and target angle measurement in turn. During pulse Doppler accumulation, the data within multiple frequency modulation cycles are subjected to fast Fourier transform in the Doppler dimension to form range image data, based on which the azimuth angle of the target is estimated, thereby realizing the angle measurement of the target by the vehicle-mounted radar system.

2. The millimeter wave orthogonal waveform optimization method according to claim 1, characterized in that: The positive frequency modulation signal is divided into a plurality of sub-pulse signals, and a corresponding phase code is superimposed on the initial phase of each sub-pulse signal.

3. The millimeter wave orthogonal waveform optimization method as claimed in claim 2, characterized in that The expression of the positive frequency modulation signal is as follows: Where A represents the signal modulation amplitude, i∈{1~52} represents the i-th sub-pulse, rect(g) represents the rectangular function, T represents the time length of the frequency modulation cycle, B represents the bandwidth, and Bc i is the ith value of the Barker code sequence, K represents the frequency modulation slope of the linear frequency modulation continuous wave signal, and the calculation method is 4. The millimeter wave orthogonal waveform optimization method according to claim 1, characterized in that: The negative frequency modulated signal is not phase encoded.

5. The millimeter wave orthogonal waveform optimization method as claimed in claim 4, characterized in that The expression of the negative frequency modulation signal is as follows: Where A represents the signal modulation amplitude, i∈{1~52} represents the i-th sub-pulse, rect(g) represents the rectangular function, T represents the time length of the frequency modulation cycle, B represents the bandwidth, and Bc i is the ith value of the Barker code sequence, K represents the frequency modulation slope of the linear frequency modulation continuous wave signal, and the calculation method is 6. The millimeter wave orthogonal waveform optimization method according to claim 3 or 5, characterized in that The expressions of the two reference signals REF0 and REF1 are as follows:

7. The millimeter wave orthogonal waveform optimization method as claimed in claim 6, characterized in that The expression of the distance image data is as follows:

8. The millimeter wave orthogonal waveform optimization method according to claim 1, characterized in that: The reference signal is a power division coupled signal of the transmission signal.

9. The millimeter wave orthogonal waveform optimization method according to claim 1, characterized in that: The multiple receiving channels have corresponding receiving antennas respectively, and the spacing length d between the receiving antennas is 0.5λ. The spacing between the transmitting antennas and between the transmitting antenna and the receiving antenna is 2N times the wavelength λ of the transmitting signal, where N is a positive integer.

10. A vehicle-mounted radar system, characterized in that The millimeter wave orthogonal waveform optimization method described in any one of claims 1 to 9 is adopted.

Images