A disturbance suppression device and method based on FPGA frequency rapid identification

Abstract

The application discloses a disturbance suppression device and method based on FPGA frequency rapid identification, and belongs to the technical field of tracking control. The device comprises a disturbance tilting mirror, a control tilting mirror, a target, an image sensor and a parallel filtering frequency identification module based on FPGA. When the device is running, the laser emitted by the target is reflected to the control tilting mirror via the disturbance tilting mirror, and then is reflected to the image sensor by the control tilting mirror. The image sensor provides the boresight error. The parallel filtering frequency identification module based on FPGA identifies the time-varying multi-frequency narrow-band disturbance of the boresight error. Then, the control algorithm of the control system containing the error observer calculates the deflection amount of the control tilting mirror. The deflection of the control tilting mirror is driven based on the deflection amount, so that the light beam is kept at the optical axis target position. The disturbance tilting mirror is used to simulate the time-varying multi-frequency narrow-band disturbance from the optical link. The application realizes the rapid and accurate capture of the disturbance characteristics under the condition of time-varying multi-frequency narrow-band disturbance.

Description

Technical Field

[0001] This invention belongs to the field of tracking control technology, specifically relating to a disturbance suppression device and method based on FPGA frequency rapid identification. Background Technology

[0002] FPGA (Field-Programmable Gate Array) allows users to configure hardware circuits according to specific needs to implement customized logic functions, and is therefore widely used in applications requiring highly parallel processing, high-speed data flow, and low latency. Adaptive algorithms in optical axis correction systems often require a large amount of computation, and traditional DSP hardware solutions struggle to guarantee system real-time performance.

[0003] Commonly used adaptive disturbance suppression methods, such as Least Mean Square (LMS), Recursive Least Squares (RMS), and adaptive notch filter methods, achieve adaptive disturbance suppression by iteratively capturing disturbance characteristics. However, these methods capture disturbance characteristics (such as disturbance frequency) through iteration, resulting in long iteration times, and most require additional sensors to measure the disturbance. Furthermore, optical axis correction systems have high real-time requirements, and these algorithms are complex and computationally intensive, making it difficult for digital systems to maintain real-time performance. In addition, due to the delay limitations of optical axis correction systems, most control methods focus on disturbance suppression within the closed-loop bandwidth and cannot effectively suppress disturbances outside the bandwidth. Summary of the Invention

[0004] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0005] A disturbance suppression device based on FPGA-based rapid frequency identification includes: a disturbance tilt mirror, a control tilt mirror, a target, an image sensor, and an FPGA-based parallel filtering frequency identification module.

[0006] When the FPGA-based frequency rapid identification disturbance suppression device is running, the laser emitted by the target is reflected by the disturbance tilt mirror to the control tilt mirror, and then reflected by the control tilt mirror to the image sensor. The image sensor provides the line-of-sight error. The FPGA-based parallel filtering frequency identification module identifies the time-varying multi-frequency narrowband disturbance of the line-of-sight error. The control algorithm of the control system containing the error observer calculates the deflection of the control tilt mirror. Based on the deflection, the control tilt mirror is driven to deflect so that the beam is kept at the target position on the optical axis. At the same time, the disturbance tilt mirror is used to simulate the time-varying multi-frequency narrowband disturbance from the optical link.

[0007] A disturbance suppression method based on FPGA frequency fast identification, used in the aforementioned disturbance suppression device based on FPGA frequency fast identification, includes:

[0008] Step 1: The disturbance suppression device based on FPGA frequency fast identification obtains the line-of-sight error sampling digital signal containing time-varying multi-frequency narrowband disturbances through the image sensor;

[0009] Step 2: The parallel filtering algorithm implemented by the FPGA-based parallel filtering frequency identification module is used to process the line-of-sight error sampled digital signal to obtain a preliminary estimate of the time-varying multi-frequency narrowband disturbance frequency.

[0010] Step 3: Apply the parallel filtering algorithm to process the preliminary estimation results of the time-varying multi-frequency narrowband disturbance frequency obtained in Step 2, construct a frequency storage array, and obtain the final estimated frequency.

[0011] Step 4: Based on PI control, design an error observer to suppress time-varying multi-frequency narrowband disturbances by means of time delay compensation and multi-rate control, according to the final estimated frequency.

[0012] The present invention has the following beneficial effects:

[0013] This invention employs an FPGA-based parallel filtering frequency identification method and an error observer control scheme, which effectively suppresses time-varying multi-frequency disturbances, improves the efficiency of adaptive parameter generation, and achieves good multi-frequency spike identification and significant suppression of time-varying multi-frequency narrowband disturbances.

[0014] (1) This invention identifies signal frequencies through a parallel bandpass filter, thereby enabling rapid and accurate capture of disturbance characteristics under time-varying multi-frequency narrowband disturbance conditions.

[0015] (2) The present invention performs frequency identification of line-of-sight error, thereby suppressing time-varying multi-frequency narrowband disturbances from the base and optical link, thus avoiding the use of additional sensors.

[0016] (3) The present invention uses FPGA to implement parallel filtering, which ensures the real-time performance of the optical axis correction system and increases the applicable scenarios.

[0017] (4) The present invention uses an error observer and achieves full-band spike disturbance suppression through time delay compensation and multi-rate control, effectively improving the ability of the optical axis correction system to suppress external bandwidth disturbances. Attached Figure Description

[0018] Figure 1 This is a structural diagram of the disturbance suppression device based on FPGA frequency fast identification according to the present invention;

[0019] Figure 2 This is a schematic diagram of the parallel filtering algorithm of the present invention;

[0020] Figure 3 This is a schematic diagram of the error observer of the present invention;

[0021] Figure 4The figures show a comparison of the effects of traditional PI feedback control and the disturbance suppression method based on FPGA frequency fast identification of the present invention on the suppression of time-varying multi-frequency narrowband disturbances. Among them, (a) is the time domain diagram of the suppression effect of traditional PI feedback control on time-varying multi-frequency narrowband disturbances, and (b) is the time domain diagram of the suppression effect of the disturbance suppression method based on FPGA frequency fast identification of the present invention on time-varying multi-frequency narrowband disturbances. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0023] like Figure 1 As shown, the present invention proposes a disturbance suppression device (or optical axis correction system, device) based on FPGA frequency fast identification, which consists of a disturbance tilt mirror, a control tilt mirror (fast reflection mirror), a target (simulated by a laser in this invention), an image sensor, and a parallel filtering frequency identification module based on FPGA (Field Programmable Gate Array).

[0024] During operation, the laser emitted by the laser is reflected by a perturbed tilting mirror to a control tilting mirror, and then reflected by the control tilting mirror to an image sensor. The image sensor provides the line-of-sight error. A parallel filtering frequency identification module based on an FPGA identifies the time-varying multi-frequency narrowband perturbation of the line-of-sight error (the difference between the actual position of the beam and the target position, measured by the image sensor). The control algorithm of the control system, which includes an error observer, calculates the deflection of the control tilting mirror as the control variable for the deflection of the control tilting mirror. By driving the deflection of the control tilting mirror based on this deflection variable, the beam is kept at the target position on the optical axis. Simultaneously, a perturbed tilting mirror is used to simulate the time-varying multi-frequency narrowband perturbation from the optical link. The control system includes an FPGA-based parallel filtering frequency identification module and an error observer.

[0025] This invention further proposes a disturbance suppression method based on FPGA frequency fast identification, which is used in the aforementioned disturbance suppression device based on FPGA frequency fast identification. The implementation steps are as follows:

[0026] Step 1: The disturbance suppression device based on FPGA frequency fast identification obtains the line-of-sight error sampling digital signal containing time-varying multi-frequency narrowband disturbances through the image sensor.

[0027] Step 2: The parallel filtering algorithm implemented by the FPGA-based parallel filtering frequency identification module is used to process the line-of-sight error sampled digital signal to obtain a preliminary estimate of the time-varying multi-frequency narrowband disturbance frequency.

[0028] The parallel filtering algorithm includes passing the line-of-sight error sampled digital signal obtained in step 1 through multiple parallel bandpass filters. The multiple parallel bandpass filters are evenly distributed in the frequency band to be identified according to their center frequencies. The output signals of the multiple parallel bandpass filters reflect the frequency components of the line-of-sight error sampled digital signal. The absolute value of the output signals of the multiple parallel bandpass filters is taken. The time-varying multi-frequency narrowband disturbance suppression threshold of the bandpass filters is adjusted according to the disturbances experienced by the optical path and the time-varying multi-frequency narrowband disturbance suppression effect of the FPGA-based frequency fast identification disturbance suppression device. The poles exceeding the time-varying multi-frequency narrowband disturbance peak identification threshold (corresponding to the center frequency of the bandpass filter) are extracted as the preliminary estimation result of the time-varying multi-frequency narrowband disturbance frequency.

[0029] Step 3: Apply the parallel filtering algorithm to process the preliminary estimation results of the time-varying multi-frequency narrowband disturbance frequency obtained in Step 2, construct a frequency storage array, and obtain the final estimated frequency.

[0030] The average number of poles within a window time (generally more than 2 seconds) is used as an estimate of the number of time-varying multi-frequency narrowband perturbation frequencies. The pole determination for each sampling point within a future window time is as follows: when the number of poles equals the estimated number of time-varying multi-frequency narrowband perturbation frequencies, the time-varying multi-frequency narrowband perturbation frequencies are arranged in ascending order into a row vector and stored in a frequency storage array. The most frequent frequency appearing in each column of the array for each window time is used as the final estimated frequency.

[0031] Step 4: Based on PI (proportional-integral) control, design an error observer to suppress time-varying multi-frequency narrowband disturbances by means of time delay compensation and multi-rate control, according to the final estimated frequency.

[0032] like Figure 2 The diagram shown illustrates the principle of the parallel filtering algorithm in step 2, where:

[0033] ;

[0034] Signal to be identified (digital signal sampled for line-of-sight error) Simultaneously, multiple parallel bandpass filters distributed across the entire frequency band are used. The output of the parallel bandpass filter is The absolute value of the output amplitude of the parallel bandpass filter This reflects the frequency domain distribution of the sampled digital signal of the line-of-sight error. Based on this, a preliminary estimate of the frequency of the time-varying multi-frequency narrowband disturbance is obtained by calculating the poles of the output absolute value that are greater than the peak identification threshold of the time-varying multi-frequency narrowband disturbance. Indicates the first A bandpass filter, with values ​​ranging from 1 to... integers, This represents the total number of bandpass filters; For the notch filter transfer function, For discrete domain variables, These are the parameters for the notch filter. These are the center frequencies of the parallel bandpass filters. Figure 2 (The graph shows the absolute value of the amplitude on the vertical axis and the frequency on the horizontal axis, in Hertz.) It can be seen that the absolute value of the bandpass filter output... Arranging the signals according to the center frequencies of the corresponding filters reflects the distribution of the signals in the frequency domain. Construct a frequency storage array and calculate the average number of poles within a window time as an estimate of the number of time-varying multi-frequency narrowband perturbation frequencies. For each sampling point in a future window time, determine the poles: only when the number of poles equals the estimated number of time-varying multi-frequency narrowband perturbation frequencies, arrange the time-varying multi-frequency narrowband perturbation frequencies in ascending order into a row vector and store them in the frequency storage array. For each window time, the most frequent frequency appearing in each column of the array is taken as the final estimated frequency.

[0035] like Figure 3 The diagram shown is a schematic of the error observer in step 4. The target position along the optical axis. For line-of-sight error, It is a basic PI controller. This is the actual delay of the optical axis correction system. The system model for controlling the tilt mirror (fast-reflection mirror) (the system model obtained after system identification of the control tilt mirror). It is a time-varying multi-frequency narrowband perturbation signal. This represents the actual position of the optical axis; the thin solid line indicates a low sampling rate. Thick solid lines use a high update rate , , It is a positive integer greater than 1; The filter's input and output are at high sampling rates. It is an estimate of the delay of the optical axis correction system. It is a model of a control tilting mirror (fast-reflection mirror) system. By estimating the inverse, the sensitivity transfer function of the control system containing the error observer is derived. for:

[0036] ;

[0037] in, This refers to the filter used in the error observer.

[0038] Analysis shows that the error suppression effect of the disturbance suppression method based on FPGA frequency fast identification mainly depends on ,Will Designed as a notch filter to achieve disturbance suppression;

[0039] Filters (multiple bandpass filters) and its corresponding delay components Parallel configuration Filter, where, This is the filter number, with values ​​ranging from 1 to... positive integers, The total number of filters is designed as follows:

[0040] ;

[0041] ;

[0042] This is an intermediate quantity with no physical meaning;

[0043] hour, The filter is designed as a sign-positive bandpass filter. ; hour, The filter is designed as a negative bandpass filter. , ; For bandpass filter Number, with values ​​from 1 to A positive integer used to distinguish bandpass filters; The total number of bandpass filters is consistent with the number of time-varying multi-frequency narrowband perturbation frequencies estimated by the FPGA-based parallel filter frequency identification module. Estimating delay for optical axis correction system frame, yes The number of frames for delay compensation, This indicates that in a control system containing an error observer, the... The delay compensation frame number is ( Values ​​range from 1 to Positive integers (e.g.) Figure 3 As shown, the right side is a parallel filter. The sampling period is For bandpass filter The center frequency, It is the symbol for imaginary numbers; For bandpass filters, It is in S-domain form as a bandpass filter; Affects the bandwidth of the bandpass filter. It is the notch depth. It is an s-domain variable. for Bandpass center frequency, It is a natural number (without physical meaning).

[0044] It can be seen that as long as the design is done at the multi-frequency narrowband disturbance frequency... To achieve the notch characteristic, multi-frequency narrowband disturbance suppression can be realized, and the frequency identification results of the FPGA-based parallel filter frequency identification module can be adjusted in real time. This enables the suppression of time-varying multi-frequency narrowband disturbances.

[0045] To demonstrate the effectiveness of this FPGA-based frequency fast identification disturbance suppression method in improving the suppression capability of time-varying multi-frequency narrowband disturbances, a corresponding controller was designed, and position tracking experiments were conducted. For example... Figure 4 As shown (horizontal axis represents time, vertical axis represents amplitude); Figure 4 The PI and YK lines in the graph are wavy lines. Since there are approximately two hundred sampling points per second, this means there are approximately two hundred broken lines per second. Therefore, individual lines are not visually apparent; they overlap to form color blocks. Furthermore... Figure 4 It mainly presents the maximum amplitude of the signal. Figure 4 The maximum amplitude of the signal in (a) is greater than Figure 4 The maximum amplitude of the signal in (b) indicates that the disturbance signal has been suppressed. Figure 4 This is a comparison chart showing the effect of applying traditional PI feedback control and the disturbance suppression method based on FPGA frequency fast identification of the present invention on the suppression of time-varying multi-frequency narrowband disturbances. Figure 4 (a) is a time-domain plot showing the time-varying multi-frequency narrowband disturbance suppression effect of traditional PI feedback control. Figure 4 (b) is a time-domain diagram showing the effect of the FPGA-based frequency fast identification perturbation suppression method of the present invention on the suppression of time-varying multi-frequency narrowband perturbations. Figure 4 The PI (proportional-integral) line in (a) is a time-domain plot of the time-varying multi-frequency narrowband disturbance suppression effect of traditional PI feedback control. Figure 4 The YK (short for error observer) line in (b) is an example of the effect of applying the FPGA-based frequency fast identification perturbation suppression method of this invention on time-varying multi-frequency narrowband perturbation suppression. Figure 4It can be seen that, since the time-varying multi-frequency narrowband disturbance is distributed across the entire frequency band, the closed-loop error of the optical axis correction system is large and it is difficult to achieve stable image tracking when the method of the present invention is not applied. After applying the method of the present invention, the time-varying multi-frequency narrowband disturbance is suppressed and the closed-loop error is reduced, which proves the effectiveness of the method of the present invention in suppressing time-varying multi-frequency narrowband disturbances across the entire frequency band.

[0046] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0047] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

[0048] The contents not described in detail in this specification are existing technologies known to those skilled in the art.

Claims

1. A disturbance suppression method based on FPGA frequency fast recognition, characterized in that, The application relates to a disturbance suppression device based on FPGA frequency fast identification. The disturbance suppression device based on FPGA frequency fast identification comprises a disturbance tilting mirror, a control tilting mirror, a target, an image sensor and an FPGA-based parallel filtering frequency identification module; during operation, laser emitted by the target is reflected to the control tilting mirror via the disturbance tilting mirror, and then reflected to the image sensor by the control tilting mirror; the image sensor provides a line-of-sight error; the time-varying multi-frequency narrow-band disturbance of the line-of-sight error is identified by the FPGA-based parallel filtering frequency identification module; a control algorithm of a control system containing an error observer is used to calculate the deflection amount of the control tilting mirror; the deflection of the control tilting mirror is driven based on the deflection amount, so that the light beam is kept at the optical axis target position; meanwhile, the disturbance tilting mirror is used to simulate the time-varying multi-frequency narrow-band disturbance from the optical link. In step 2, a parallel filtering algorithm realized by the FPGA-based parallel filtering frequency identification module is used to process the line-of-sight error sampling digital signal, so as to obtain a preliminary estimation result of the time-varying multi-frequency narrow-band disturbance frequency. In step 3, the parallel filtering algorithm is applied to process the preliminary estimation result of the time-varying multi-frequency narrow-band disturbance frequency obtained in step 2, a frequency storage array is constructed, and a final estimation frequency is obtained. In step 4, on the basis of PI control, an error observer is designed; time delay compensation and multi-rate control are used; and the time-varying multi-frequency narrow-band disturbance is suppressed according to the final estimation frequency. In step 2, the parallel filtering algorithm comprises the following steps: the line-of-sight error sampling digital signal obtained in step 1 is input into a plurality of parallel band-pass filters; the plurality of parallel band-pass filters are uniformly distributed at the center frequencies of the to-be-identified frequency band; the absolute values of the output signals of the plurality of parallel band-pass filters are taken; the time-varying multi-frequency narrow-band disturbance peak value identification threshold of the band-pass filter is adjusted according to the disturbance of the optical path and the time-varying multi-frequency narrow-band disturbance suppression effect of the disturbance suppression device based on FPGA frequency fast identification; and the extreme points exceeding the time-varying multi-frequency narrow-band disturbance peak value identification threshold are extracted as the preliminary estimation result of the time-varying multi-frequency narrow-band disturbance frequency; the expression of the filter used by the parallel filtering algorithm is as follows: The control tilting mirror is a fast tilting mirror. ； A to-be-identified visual axis error sampling digital signal Simultaneously through multiple parallel band-pass filters distributed in the full frequency band The extreme point of the absolute value of the output of the parallel band-pass filter being greater than a peak value identification threshold of the time-varying multi-frequency narrow-band disturbance is a preliminary estimation result of the time-varying multi-frequency narrow-band disturbance frequency in, Indicates the first A bandpass filter, with values ​​ranging from 1 to... integers, This represents the total number of bandpass filters; For the notch filter transfer function, For discrete domain variables, These are the parameters for the notch filter. These are the center frequencies of the bandpass filters connected in parallel.

2. The method of claim 1, wherein the method is characterized by, In step 3, the average number of extreme points in a window time is calculated as the estimation of the number of time-varying multi-frequency narrow-band disturbance frequencies, and the extreme points identified for each sampling point in the future window time are determined.

3. The method of claim 1, wherein the method is characterized by, In step 3, the extreme point determination for each sampling point comprises the following steps: when the number of extreme points is equal to the estimated number of time-varying multi-frequency narrow-band disturbance frequencies, the time-varying multi-frequency narrow-band disturbance frequencies are arranged in a row vector in the order from small to large, and stored into a frequency storage array; and the most frequent frequency appearing in each column of the array in each window time is taken as the final estimation frequency.

4. The method of claim 2, wherein the method further comprises: The window time is set to be more than 2 seconds.

5. The method according to claim 3 or 4, characterized in that, ​ 6. The method of claim 5, wherein the method further comprises: In step 4, the sensitivity transfer function of the control system with error observer is: ； wherein, is an estimate of the optical axis correction system delay, is a filter used in the control system with error observer, is a basic PI controller, is a model of the control tilt mirror system, is the actual delay of the optical axis correction system; the input and output of the filter are at a high sampling rate, is an estimate of the inverse of the model of the control tilt mirror system .

Images