Millimeter wave wideband parameter adaptive calibration method

Abstract

The application discloses a millimeter wave wide frequency parameter self-adaptive calibration method. The application designs a segmented modulation curve model, and the segmented points are continuous; meanwhile, a frequency spectrum concentration index is defined for quantifying the concentration degree of signal energy in a frequency domain, an optimization objective function is designed according to the frequency spectrum concentration index, and an optimal segmented number and optimal slopes of the segmented modulation curves are automatically obtained according to the characteristics of different VCOs through an optimization algorithm, so that the linear frequency modulation signal generated by the VCO reaches the best linearity, and the limitations of the fixed modulation waveform and the fixed segmented tuning strategy are solved. The application can dynamically adjust the modulation waveform according to the characteristics of different VCOs, thereby improving the linearity of the linear frequency modulation signal and improving the signal quality; and the application can be implemented on a DAC-VCO architecture only through the cooperation of a signal processor, the system design and implementation are simple, and the cost is low.

Description

Technical Field

[0001] The present invention relates to the field of millimeter wave detection technology, and in particular to a millimeter wave broadband parameter adaptive calibration method. Background Art

[0002] Millimeter-wave detectors are widely used in radar, communications, and imaging. Their core principle is to achieve target detection and distance measurement through linear frequency modulation (LFM). Conventional LFM generation typically employs the following architecture: a digital-to-analog converter (DAC) generates the modulation signal to control the frequency of a voltage-controlled oscillator (VCO); the VCO generates the LFM signal based on the control voltage provided by the DAC; and a digital signal processor (DSP) controls the DAC to output a fixed modulation waveform, ensuring the linearity of the LFM signal. This architecture offers the advantages of simplicity, low cost, and ease of digital control. However, VCO linearity remains a key factor impacting system performance.

[0003] The following methods are commonly used to generate linear frequency modulation signals:

[0004] Solution 1: DAC-VCO architecture based on a fixed modulation waveform. The DAC generates a fixed modulation waveform (such as a ramp or sawtooth waveform) based on DSP instructions. The VCO generates an FM signal based on the control voltage provided by the DAC. Linearity calibration: VCO nonlinearity is compensated through predistortion techniques or feedback control (such as a phase-locked loop (PLL)).

[0005] Solution 2: A VCO architecture based on segmented tuning. Segmented tuning divides the VCO's tuning range into multiple subranges, with the control voltage optimized independently for each range. Dynamic switching dynamically switches the tuning range based on the target frequency, ensuring linearity within each range. Temperature compensation incorporates a temperature sensor and compensation circuit to minimize the impact of temperature changes on VCO linearity.

[0006] The above scheme can realize the generation of linear frequency modulation signals to a certain extent, but it still has the following defects: (1) Inconsistent VCO linearity: Due to factors such as the nonlinear characteristics of the varactor diode, resonant circuit design, temperature changes, power supply noise and manufacturing process deviations, the linearity of different VCOs varies. The fixed modulation waveform in the existing scheme cannot adapt to the characteristics of all VCOs, resulting in a decrease in the target signal quality of some detectors. (2) Limitations of fixed modulation waveform: The fixed modulation waveform generated by the DAC cannot dynamically adapt to the individual differences of the VCO, especially in scenarios with wide tuning range and high precision requirements, the linearity problem is more prominent. (3) High calibration complexity: In the existing scheme, the method of improving linearity through pre-distortion or segmented tuning requires a complex calibration process, which increases the difficulty of system design and implementation.

[0007] While existing segmented tuning technology has shown some effectiveness in improving VCO linearity, it suffers from significant drawbacks: First, it lacks adaptability. Using a fixed segmented tuning strategy, it cannot dynamically adjust to the characteristics of different VCOs, making linearity difficult to guarantee under complex conditions. Second, because each segment is tuned independently, segment transitions can produce discontinuities or jumps, reducing tuning accuracy and introducing nonlinear distortion. Finally, segmented tuning requires the implementation of multiple independent control circuits in hardware, increasing the complexity and difficulty of circuit design. This can introduce additional noise and distortion, especially in high-frequency or wide tuning range scenarios. Consequently, existing methods suffer from significant deficiencies in accuracy, adaptability, and hardware implementation. Summary of the Invention

[0008] In view of this, the present invention provides a millimeter-wave broadband parameter adaptive calibration method that can dynamically adapt to modulation waveforms with different VCO linearity characteristics, ensuring that all detectors can generate high-quality linear frequency modulation signals, so that the quality of the target signal is optimized.

[0009] The millimeter wave broadband parameter adaptive calibration method of the present invention includes:

[0010] Step 1: construct a frequency modulation curve model; the frequency modulation curve model consists of M segments, the frequency modulation curve of each segment is a linear function of time, and the frequency end value of each segment is the frequency start value of the next segment;

[0011] Step 2: Construct the optimization objective function as follows:

[0012]

[0013] Where C is the spectrum concentration, Among them, P maxis the maximum value in the VCO spectrum; B is α times the power bandwidth of the VCO spectrum, α < 1; γ is the set scale factor;

[0014] In step 3, with the goal of maximizing the optimization objective function constructed in step 2, an optimization algorithm is used to optimize the number of segments M and the slope of each segment frequency modulation curve in the frequency modulation curve model constructed in step 1, to obtain the optimal number of segments and the optimal slope of each segment, and complete the adaptive calibration of millimeter wave broadband parameters.

[0015] Preferably, in step 1, the frequency modulation curve model is:

[0016]

[0017] Among them, f i , i=1,2,3…M, is the starting value of the i-th segment, satisfying the continuity condition: f i+1 =f i +k i (t i+1 -t i ); f1 is the initial value set.

[0018] Preferably, f1 is the starting frequency of the VCO.

[0019] Preferably, in step 3, the optimization algorithm adopts particle swarm optimization algorithm, genetic algorithm, artificial bee colony algorithm or whale optimization algorithm.

[0020] Beneficial effects:

[0021] The present invention designs a segmented modulation curve model and ensures the continuity of the segmentation points, avoiding the nonlinear distortion problem caused by the discontinuity or jump phenomenon in the interval transition in the segmented tuning method; at the same time, it defines a spectrum concentration index for quantifying the concentration of signal energy in the frequency domain, and designs an objective function so that it can not only represent the linearity of the linear frequency modulation signal but also take into account the selection of the number of segments in the segmented modulation curve. Therefore, through the optimization algorithm, the optimal slope k and the optimal number of segments N of the segmented modulation curve are automatically obtained according to the characteristics of different VCOs, so that the linear frequency modulation signal generated by the VCO achieves the best linearity, solving the limitations of fixed modulation waveforms and fixed segmented tuning strategies. The present invention can dynamically adjust the modulation waveform according to the characteristics of different VCOs, thereby improving the linearity of the linear frequency modulation signal and improving the signal quality; and it can be implemented on a DAC-VCO architecture only through the cooperation of a signal processor, and the system design and implementation are simple and low-cost. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 This is an illustration of the segmented modulation curve approaching the optimal modulation curve.

[0023] Figure 2 The figure shows the comparison of linear and nonlinear VCO intermediate frequency spectra.

[0024] Figure 3 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0025] The present invention is described in detail below with reference to the accompanying drawings and embodiments.

[0026] The present invention provides a method for adaptive calibration of millimeter-wave broadband parameters. The present invention first establishes a segmented modulation curve model and designs a multi-segmented linear function modulation curve. By adjusting the slope of each segment of the linear function, the segmented modulation curve is made to approach the optimal modulation curve of the VCO, thereby improving the linearity of the VCO transmission signal. At the same time, in order to quantify the quality of the linearity, the spectrum concentration is defined as an evaluation index. The particle swarm algorithm is further used to optimize the number of segments and the slope of each segment of the linear function to solve the optimal segmented modulation curve that maximizes the spectrum concentration. Finally, the VCO is modulated based on the optimal segmented modulation curve, so that the system has the ability to adaptively calibrate the linearity of the VCO. This method can effectively improve the linearity of the VCO, enhance the stability of the system and the quality of the target signal.

[0027] The specific steps of the present invention are as follows:

[0028] First, a mathematical model of the frequency segment modulation curve is established as follows:

[0029] Assumption: The time range is t∈[t1,t M+1 ], where t1 is the start time, t M+1 is the end time. M+1 ] is divided into M segments, and the frequency modulation curve of each segment adopts a linear function, then the slope array of the modulation curve is k=[k1,k2,…,k M ], where M is the number of segments and the segment point is t break =[t2,…,t M ].

[0030] Segmented modulation curve f fit The (t,k) model can be expressed as:

[0031]

[0032] Where: f i , (i=1,2,3…M) is the starting value of the i-th segment, satisfying the continuity condition: f i+1 =f i +k i (t i+1 -t i). f1 is the initial value, which can usually be set to f1=0 (or other fixed value).

[0033] Model Description:

[0034] 1. Segment definition: Each segment is a linear function with a slope of k i , the time range is [t i ,t i+1 The function value of each segment is determined by the slope k i , starting value f i and time offset tt i Decide.

[0035] 2. Continuity condition: In order to ensure the continuity of the segmented modulation curve, the starting value f of each segment i Need to meet: f i+1 =f i +k i (t i+1 -t i ), which means that the end value of each segment is the start value of the next segment.

[0036] 3. Initial value: The starting value f1 of the first segment can be set to 0 or other fixed values, depending on the starting frequency of the VCO, and can be equal to the starting frequency of the VCO.

[0037] 4. Segmentation point: segmentation point t break Divide the time range into M segments evenly, each segment corresponds to a slope k i .

[0038] According to the above formula, the segmented modulation curve f fit (t,k) can be directly represented by the slope k and time t. Each segment is a linear function, and the continuity condition is satisfied between adjacent segments. By changing the value of k, the segmented modulation curve can be controlled to be as close to the optimal modulation curve of the VCO as possible, such as Figure 1 shown.

[0039] In this embodiment, in order to facilitate the observation of the difference between the segmented modulation curve and the optimal modulation curve in the figure, the number of segments of the segmented modulation curve is set to 3, that is, Figure 1 The segmented modulation curve has three segments, each with a different slope k i By different slope k i The segmented modulation curve can be better approximated to the optimal modulation curve. In practice, a larger number of segments is selected to better approximate the optimal modulation curve.

[0040] Poor linearity of the VCO transmission signal will lead to problems such as spectrum broadening and energy leakage of the intermediate frequency echo target signal, such as Figure 2 shown.

[0041] from Figure 2 It can be seen that the IF spectrum of the nonlinear VCO shows a significant broadening phenomenon compared to the ideal linear VCO, and the energy of the signal frequency components has been seriously leaked. Therefore, the spectrum concentration is used as an indicator of the linearity of the constant VCO transmission signal. The spectrum concentration is defined as follows:

[0042]

[0043] Where C is the spectrum concentration. The larger it is, the more concentrated the spectrum is, that is, the better the linearity of the VCO is. max Indicates the maximum value in the VCO spectrum, that is, the amplitude value of the highest point in the VCO spectrum. B represents α times the power bandwidth of the VCO spectrum, α < 1. In this embodiment, α is 0.1, that is, B is the amplitude of 0.1P in the VCO spectrum. max The difference in the horizontal coordinates between two points.

[0044] Different VCOs have different optimal modulation curves. To adaptively determine the most suitable segmented modulation curve slope k and number of segments M for each VCO, the present invention employs a particle swarm optimization (PSO) algorithm. PSO is an optimization algorithm based on swarm intelligence that simulates the social behavior of flocks of birds or fish to search for the optimal solution in a search space. This algorithm iteratively updates the positions and velocities of particles, gradually approaching the global optimal solution. It has the advantages of fast convergence, few parameters, and ease of implementation, making it well-suited for solving nonlinear optimization problems.

[0045] A larger number of segments, M, in the segmented modulation curve results in better fitting of the optimal modulation curve. However, a larger number of segments increases the computational complexity. In practice, systems have a certain tolerance for VCO nonlinearity, so setting an excessive number of segments, M, is unnecessary as long as the system requirements are met.

[0046] In order to meet the requirements of linearity reaching the standard without setting too many segments, the objective function of the optimization algorithm is designed. It is shown in the following formula:

[0047]

[0048] In the present application, T is taken as the objective function of the PSO algorithm. The greater the value of the objective function, the more concentrated the energy of the signal in the target frequency band and the fewer the number of segments, thus the higher the quality of the chirp signal and the lower the computational complexity. In the objective function, γ is a scaling factor, which controls the weight of the spectral concentration C and the number of segments M in the objective function by adjusting γ. According to the system requirements, γ is adjusted (if the system requires extremely excellent linearity, γ is adjusted to be large, and if the system requires faster adjustment speed, γ is adjusted to be small), so that the relationship between C and M can be better balanced. By optimizing the objective function, the segment modulation curve slope k and the number of segments M that maximize T can be found.

[0049] By dynamically adjusting the slope k and the number of segments M through the PSO algorithm, the optimal modulation curve of the VCO can be approximated to the maximum extent with the least computational resources. Specifically, the PSO algorithm solves the optimal slope k opt and the optimal number of segments M opt by the following steps:

[0050] 1. Initialize the particle swarm: a set of initial slope k values are randomly generated as the initial positions of the particles.

[0051] 2. Calculate the objective function: for each slope k and number of segments M, calculate the corresponding objective function value.

[0052] 3. Update the particle position: update the position and velocity of the particle according to its own historical optimal solution and the historical optimal solution of the group.

[0053] 4. Iterative optimization: repeat the above process until the optimal slope k opt and the optimal number of segments M opt that maximize the objective function are found.

[0054] The determination of the optimal modulation curve slope k opt and the optimal number of segments M opt is not limited to the PSO algorithm, but other algorithms suitable for nonlinear optimization can also be used. For example, genetic algorithm, which simulates the biological evolution process to realize global search; artificial bee colony algorithm, which simulates the foraging behavior of bees and is suitable for complex nonlinear problems; whale optimization algorithm, which simulates the foraging behavior of whales and is suitable for continuous and nonlinear optimization problems.

[0055] Through the above method, the present application can adaptively adjust the slope k and the number of segments M of the segment modulation curve according to the characteristics of different VCOs, thereby effectively improving the quality of the chirp signal.

[0056] In summary, the working process of the millimeter wave wideband parameter adaptive calibration method can be summarized as Figure 3 .

[0057] To sum up, the above is only the preferred embodiment of the present application, and is not used to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A millimeter wave broadband parameter adaptive calibration method, characterized in that: include: Step 1, constructing a frequency modulation curve model; The frequency modulation curve model consists of M segments, the frequency modulation curve of each segment is a linear function of time, and the frequency end value of each segment is the frequency start value of the next segment; Step 2: Construct the optimization objective function as follows: Where C is the spectrum concentration, Among them, P max is the maximum value in the VCO spectrum; B is α times the power bandwidth of the VCO spectrum, α < 1; γ is the set scale factor; In step 3, with the goal of maximizing the optimization objective function constructed in step 2, an optimization algorithm is used to optimize the number of segments M and the slope of each segment frequency modulation curve in the frequency modulation curve model constructed in step 1, to obtain the optimal number of segments and the optimal slope of each segment, and complete the adaptive calibration of millimeter wave broadband parameters.

2. The method according to claim 1, wherein In step 1, the frequency modulation curve model is: Among them, f i , i=1,2,3…M, is the starting value of the i-th segment, satisfying the continuity condition: f i+1 =f i +k i (t i+1 -t i ); f1 is the initial value set.

3. The method according to claim 2, wherein f1 is the starting frequency of the VCO.

4. The method according to claim 1, wherein In step 3, the optimization algorithm adopts particle swarm optimization algorithm, genetic algorithm, artificial bee colony algorithm or whale optimization algorithm.

Images