A method and system for compensating for moving mirror scanning errors in an FTIR spectrometer

Abstract

The application provides a FTIR spectrometer moving mirror scanning error compensation method and system, belongs to the field of precision optical machine system control technology, is suitable for FTIR spectrometer adopting a flexible transmission moving mirror scanning mechanism, and the method comprises the following steps: a state space model of the flexible transmission moving mirror scanning mechanism is constructed in advance; an interference light intensity signal is acquired to determine a mechanical tracking control component; an optical prediction compensation component is determined according to the state space model; a composite control quantity is constructed according to the mechanical tracking control component, the optical prediction compensation component and a dynamic feedforward control component; the composite control quantity is converted into a motor side driving signal to compensate the scanning error of the moving mirror of the FTIR spectrometer. Through the establishment of an independent optical feedback channel, the dynamic optical deviation caused by the flexible transmission structure can be effectively compensated, and through the multi-source error layered decoupling compensation architecture, the optical modulation stability of the FTIR under the conditions of high-speed scanning and long-term operation can be effectively improved.

Description

Technical Field

[0001] This invention relates to the field of precision optomechanical system control technology, and in particular to a method and system for compensating for scanning errors of a moving mirror in an FTIR spectrometer. Background Technology

[0002] With the continuous increase in demand for environmental monitoring and pollution control, there is an urgent need for rapid and highly sensitive on-site infrared spectroscopy analysis technology in fields such as atmospheric trace gas monitoring, soil organic pollutant screening, and online detection of volatile organic compounds in groundwater. Fourier transform infrared spectrometers (FTIR), with their advantages of requiring no pretreatment, simultaneous detection of multiple components, and rich spectral fingerprint information, have become an important tool for on-site environmental monitoring.

[0003] FTIR spectrometers acquire interference signals by modulating the optical path difference (OPD) and then demodulate them using Fourier transform to obtain the infrared spectrum. The modulation accuracy and stability of the OPD directly affect the quality of the interferogram and the final spectral resolution, making them core factors determining the overall performance of the instrument. Traditional FTIR systems typically employ a rigid transmission structure driven by a linear motor, pushing a moving mirror to reciprocate linearly along the optical axis to achieve OPD scanning. This rigid approach results in a long travel distance, large structural volume, and stringent assembly precision requirements, making it difficult to meet the lightweight requirements of miniaturized, portable, and rapid detection devices.

[0004] To adapt to the trend of equipment miniaturization, a feasible alternative is to use a rotating mirror scanning mechanism with a rotary drive motor and flexible transmission: the rotary drive motor drives the driving wheel, which in turn drives the driven shaft and the moving mirror assembly to reciprocate through a flexible transmission structure such as a steel belt, achieving optical path difference modulation by means of angular displacement. This structure has the advantages of compact size, low moment of inertia, and high scanning speed, and can effectively meet the structural constraints of portable devices.

[0005] However, as an elastic and flexible transmission medium, the steel belt is subject to nonlinear time-varying factors such as elastic elongation, assembly clearance, rim slippage, and temperature drift. This results in a lack of a fixed, rigid mapping relationship between the mechanical rotation angle measured by the motor-side encoder and the optical mirror angle modulated by the actual optical path difference. The transmission process exhibits typical optical-mechanical coupling dynamic deviations: the motor-side angle cannot directly reflect the true angle of the optical mirror on the load side, and the actual change in the optical path difference cannot be uniquely determined by the motor rotation angle. Because the optical mirror angle cannot be directly obtained from the motor encoder, the control system lacks feedback variables that can accurately reflect the optical path modulation state.

[0006] To address the aforementioned issues, existing control schemes for FTIR spectrometers applied to flexible transmission structures typically use mechanical quantities such as motor rotation angle and speed as closed-loop feedback targets. This makes it difficult to simultaneously characterize the dynamic characteristics of the mechanical transmission and the mapping relationship of the optical output within a unified framework. Such purely mechanical closed-loop control can only guarantee the tracking accuracy of the motor's motion trajectory, failing to effectively characterize and compensate for the dynamic optical deviations introduced by the flexible transmission structure. This results in a continuous deviation between the mechanical motion state and the actual modulation state of the optical path, ultimately manifesting as a series of optical performance degradation problems such as interference phase drift, waveform distortion, and a decrease in spectral signal-to-noise ratio.

[0007] Therefore, there is an urgent need to establish an optomechanical integrated control method that takes the actual optical state as the direct control target and can characterize the optical-mechanical coupling characteristics of the flexible transmission structure, so as to break through the inherent limitations of the existing pure mechanical closed loop. Summary of the Invention

[0008] This invention provides a method and system for compensating for the scanning error of a moving mirror in an FTIR spectrometer. This method addresses the shortcomings of existing FTIR spectrometers where purely mechanical closed-loop control cannot effectively characterize and compensate for dynamic optical deviations introduced by flexible transmission structures. It achieves an architectural shift from mechanical motion control to optical state-driven control, thereby improving the optical path difference modulation accuracy and system adaptability.

[0009] In a first aspect, the present invention provides a method for compensating for scanning errors of a moving mirror in an FTIR spectrometer, applicable to FTIR spectrometers employing a flexible drive moving mirror scanning mechanism, the method comprising: A state-space model is pre-constructed based on the state variables of the motor side and load side of the flexible transmission moving mirror scanning mechanism; Acquire the interference light intensity signal, convert the interference light intensity signal into the measured optical path difference value, determine the optical path difference tracking error, and determine the mechanical tracking control component based on the optical path difference tracking error; Based on the measured optical path difference, the look-ahead bias of the optical prediction is determined by the state-space model, and the optical prediction compensation component is determined based on the look-ahead bias of the optical prediction. Based on the inverse dynamic equations that include the state variables of the motor side and the load side, determine the dynamic feedforward control components; A composite control quantity is constructed based on the mechanical tracking control component, the optical prediction compensation component, and the dynamic feedforward control component. The composite control quantity is converted into a motor-side drive signal to compensate for the scanning error of the moving mirror of the FTIR spectrometer.

[0010] According to the present invention, a method for compensating for scanning errors of a moving mirror in an FTIR spectrometer is provided, which pre-constructs a state-space model based on the state variables of the motor side and the load side of the flexible transmission moving mirror scanning mechanism, including: A multidimensional mechanical state vector is constructed based on the state variables of the motor side and the load side of the flexible transmission moving mirror scanning mechanism. Based on the multidimensional mechanical state vector, the augmented mechanical state recursive equation of the flexible transmission moving mirror scanning mechanism is constructed through the torque balance equation of the motor side and the load side. The optical equivalent angle of the moving mirror is determined based on the state variables of the motor side and the load side, and the optical path difference observation equation is constructed based on the optical equivalent angle and the multidimensional mechanical state vector. A state-space model is established based on the augmented mechanical state recursive equation and the optical path difference observation equation.

[0011] According to the present invention, a method for compensating for scanning errors of a moving mirror in an FTIR spectrometer is provided. A multidimensional mechanical state vector is constructed based on the state variables of the motor side and the load side of the flexible drive moving mirror scanning mechanism. Based on the multidimensional mechanical state vector, an augmented mechanical state recursive equation for the flexible drive moving mirror scanning mechanism is constructed through the torque balance equations of the motor side and the load side, including: A continuous domain dynamic equation is constructed by using the torque balance equations of the motor side and the load side of the flexible transmission moving mirror scanning mechanism. Based on the state variables of the motor side and the load side, a basic mechanical state vector is constructed, and the continuous domain dynamic equation is discretized to obtain the basic mechanical state recursive equation. A comprehensive angular displacement disturbance of the flexible transmission moving mirror scanning mechanism is introduced into the basic mechanical state vector to construct a multidimensional mechanical state vector, and a disturbance random decay model based on the comprehensive angular displacement disturbance is established; wherein, the multidimensional mechanical state vector includes: the rotation angle and angular velocity of the motor-side driving wheel, the equivalent rotation angle and angular velocity of the load side, and the comprehensive angular displacement disturbance of the flexible transmission moving mirror scanning mechanism; By combining the basic mechanical state recursive equation with the perturbation stochastic decay model, an augmented mechanical state recursive equation is constructed.

[0012] According to the present invention, a method for compensating for scanning error of a moving mirror in an FTIR spectrometer is provided, wherein the optical equivalent angle is the actual optical rotation angle of the moving mirror participating in optical path modulation, and the optical equivalent angle is determined jointly based on the equivalent rotation angle on the load side and the combined angular displacement disturbance of the flexible transmission moving mirror scanning mechanism.

[0013] According to the present invention, a method for compensating for the scanning error of a moving mirror in an FTIR spectrometer converts the interference light intensity signal into a measured optical path difference value, determines the optical path difference tracking error, and determines a mechanical tracking control component based on the optical path difference tracking error, including: The interference light intensity signal is analyzed and phase unwrapped to obtain the measured optical path difference. The optical path difference tracking error is determined based on the difference between the preset target optical path difference and the measured value of the optical path difference; wherein, the preset target optical path difference is generated by discretizing the standard uniform / variable speed scanning waveform of the moving mirror based on the preset scanning trajectory. Based on the optical path difference tracking error, an incremental PID control law is used to determine the mechanical tracking control component.

[0014] According to the present invention, a method for compensating for the scanning error of a moving mirror in an FTIR spectrometer includes determining the look-ahead bias of optical prediction using the state-space model based on the measured optical path difference, and determining the optical prediction compensation component based on the look-ahead bias of optical prediction, comprising: Using the measured optical path difference at the current moment as the observation, the posterior state estimate at the current moment is obtained recursively through a Kalman filter; Based on the posterior state estimate at the current moment, the predicted value of the multidimensional mechanical state vector at the next moment is derived through the augmented mechanical state recursive equation. Based on the predicted value of the multidimensional mechanical state vector at the next moment, the predicted value of the optical path difference at the next moment is obtained by mapping through the optical path difference observation equation. The look-ahead bias of the optical prediction is determined based on the predicted optical path difference value at the next moment and the preset target optical path difference. Based on the look-ahead bias of the optical prediction, the optical prediction compensation component is determined.

[0015] According to the present invention, a method for compensating for the scanning error of a moving mirror in an FTIR spectrometer is provided, which constructs a composite control quantity based on the mechanical tracking control component, the optical prediction compensation component, and the dynamic feedforward control component, including: The mechanical tracking control component, the optical prediction compensation component, and the dynamic feedforward control component are superimposed to obtain a composite control quantity; wherein, the inverse dynamic equation used to calculate the dynamic feedforward control component is obtained by inversely solving the basic mechanical state recursive equation, and the mechanical tracking control component, the optical prediction compensation component, and the dynamic feedforward control component are independent of each other at their respective error levels.

[0016] Secondly, the present invention also provides an FTIR spectrometer moving mirror scanning error compensation system, comprising: The flexible drive moving mirror optical path module is used to convert the rotational motion of the motor into the reciprocating oscillation of the moving mirror assembly through the flexible drive moving mirror scanning mechanism, so as to complete the physical modulation of the optical path difference; The interference signal acquisition module is used to acquire interference light intensity signals and perform digital processing. The composite control estimation module is used to pre-construct a state-space model based on the state variables of the motor side and load side of the flexible transmission moving mirror scanning mechanism; convert the digitized interference light intensity signal into the measured optical path difference value and determine the optical path difference tracking error; determine the mechanical tracking control component based on the optical path difference tracking error; determine the look-ahead bias of the optical prediction based on the state-space model; determine the optical prediction compensation component based on the look-ahead bias of the optical prediction; determine the dynamic feedforward control component based on the inverse dynamic equation containing the state variables of the motor side and load side; and construct a composite control quantity based on the mechanical tracking control component, the optical prediction compensation component, and the dynamic feedforward control component. The motor drive circuit module is used to convert the composite control quantity into a motor-side drive signal to compensate for the scanning error of the moving mirror of the FTIR spectrometer. The position feedback loop, connected to the drive motor, is used to collect the encoder mechanical rotation angle signal of the drive motor for motor safety monitoring and initial zero-position calibration.

[0017] Thirdly, the present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the FTIR spectrometer moving mirror scanning error compensation method as described above.

[0018] Fourthly, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the FTIR spectrometer moving mirror scanning error compensation method as described above.

[0019] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the FTIR spectrometer moving mirror scanning error compensation method as described above.

[0020] The beneficial effects of the technical solutions provided by some embodiments of the present invention include at least the following: 1) The present invention provides a method and system for compensating for the scanning error of a moving mirror in an FTIR spectrometer. By constructing a state-space model of a flexible moving mirror scanning mechanism to form an independent optical feedback channel, the look-ahead bias of optical prediction is determined, thereby determining the optical prediction compensation component, effectively compensating for the dynamic optical bias introduced by the flexible moving mirror scanning mechanism. In addition, the present invention determines the optical path difference tracking error by collecting interference light intensity signals and constructs a mechanical tracking control component. The mechanical tracking control component, the optical prediction compensation component, and the dynamic feedforward control component are combined to form a multi-source error hierarchical decoupling compensation architecture, thereby constructing a composite control quantity for moving mirror scanning error compensation, effectively improving the optical modulation stability of FTIR under high-speed scanning and long-term operation conditions.

[0021] 2) This invention constructs a state-space model of optomechanical coupling layered mapping. By establishing an augmented mechanical state recursive equation that includes the motion state of the motor side, the motion state of the load side, and the comprehensive angular displacement disturbance of the flexible transmission, the flexible nonlinear deviation caused by the flexible transmission structure is modeled as a first-order slowly varying and estimable disturbance state. Under the condition of using only optical path difference observation information, the effective separation and estimation of the load mechanical rotation angle and the disturbance deviation can be achieved. Furthermore, by combining the collected interference light intensity signal, the actual optical path difference of the optical path or the true angle of the optical mirror is incorporated into the control system modeling and control law design by constructing the optical path difference observation equation. The optical equivalent angle that cannot be directly measured is estimated in real time, so that the control target is changed from mechanical position tracking to optical path difference state tracking, and the optical equivalent angle can be directly used as the control target, which helps to suppress interference phase drift and spectral distortion under high-speed conditions.

[0022] 3) This invention defines the optical equivalent angle as the algebraic derived quantity of the combined angular displacement disturbance of the load-side equivalent rotation angle and the flexible transmission structure, avoiding the modeling defect of the indistinguishable load rotation angle and disturbance in the traditional scheme. It also constructs the optical path difference observation equation by combining the optical mapping coefficient, realizing the distinguishable mapping between the optical modulation state and the mechanical state.

[0023] 4) This invention constructs a three-component parallel compensation architecture for error decoupling: a dynamic feedforward compensation component compensates for deterministic mechanical dynamic lag, an optical prediction compensation component alleviates the accumulated deviation of disturbances, and a mechanical tracking control component eliminates the random error of residual mechanical tracking through incremental PID. The errors of the three control components are calculated independently and only superimposed at the output end. The prediction branch uses the control quantity determined by the control law at the current moment for calculation, which has causal realizability and helps to reduce the accumulated error during long-term continuous scanning.

[0024] 5) The optical prediction branch based on the state-space model added in this invention uses the control quantity that has been determined by the control law at the current moment in the prediction process, without involving unknown future control quantities, and has causal realizability, and helps to reduce the cumulative error in the long-term continuous scanning process.

[0025] 6) This invention does not require replacing the original flexible transmission mechanical structure, but only upgrades the controller algorithm, thus retaining all the advantages of the miniature FTIR spectrometer in terms of portability, battery life, and size. Attached Figure Description

[0026] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0027] Figure 1 This is a schematic diagram of the FTIR spectrometer moving mirror scanning error compensation system provided by the present invention; Figure 2 This is a flowchart illustrating the FTIR spectrometer moving mirror scanning error compensation method provided by the present invention; Figure 3 This is a schematic diagram of the closed-loop control process provided by the present invention; Figure 4 This is a comparison of the optical path difference tracking error between the proposed solution and the traditional mechanical closed-loop solution under the high-speed 5Hz uniform scanning condition provided by the present invention; where (a) is the time-domain optical path difference error of the proposed solution, (b) is the time-domain optical path difference error of the existing traditional mechanical closed-loop solution, and (c) is the statistical curve of the cumulative distribution of absolute error of the two solutions. Figure 5 This is a comparison chart of the single-cycle three-component control torque decomposition result of the present invention and the single PID control torque of the traditional mechanical closed-loop scheme; wherein, (a) is a single complete scan cycle, showing four independent time-series curves of the total control torque, mechanical tracking control component (PID), dynamic feedforward control component, and one-step optical prediction compensation component of the present invention; (b) is the single PID control torque of the traditional mechanical closed-loop scheme under the same time window, without additional feedforward and prediction branches; (c) is a comparison of the power spectral density of the control quantities of the two schemes; Figure 6 This is a comparison chart of the high-speed scanning recovery spectrum and SNR of the solution provided by this invention and the traditional mechanical closed-loop solution. Figure 7 This is a schematic diagram of the structure of the electronic device provided by the present invention.

[0028] Reference numerals: 1-Drive motor, 2-Driving wheel, 3-Steel belt, 4-Driven shaft, 5-Moving mirror group, 6-Infrared light source, 7-Beam splitter, 8-Fixed mirror, 9-Photodetector, 10-Interference signal acquisition module, 11-Optical-mechanical coupling state estimation module, 12-Independent optical path difference timing prediction module, 13-Layered coupling control law calculation module, 14-Motor drive circuit module, 15-Position feedback loop. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0030] This invention addresses the dynamic optical deviation caused by the introduction of a flexible transmission structure in FTIR spectrometers by proposing a method and system for compensating for the scanning error of the moving mirror in FTIR spectrometers. This method is applicable to miniature FTIR spectrometers that employ a flexible transmission moving mirror scanning mechanism.

[0031] It should be noted that in the embodiments of the present invention, the flexible transmission structure is always based on a steel belt. The flexible transmission moving mirror scanning mechanism includes a motor side, a load side and a flexible transmission structure (steel belt). The motor side includes a drive motor and a drive wheel, and the load side includes a driven wheel and a moving mirror assembly. The moving mirror assembly includes a moving mirror and a swing arm. The flexible transmission moving mirror scanning mechanism converts the rotational motion of the motor into the reciprocating swing of the moving mirror assembly via the steel belt, and combines it with other optical elements such as a beam splitter and a fixed mirror to complete the physical modulation of the optical path difference.

[0032] Please see Figure 1 , Figure 1 The present invention provides a schematic diagram of a moving mirror scanning error compensation system for an FTIR spectrometer, which includes a flexible moving mirror optical path module, an interference signal acquisition module, a composite control estimation module, a motor drive circuit module, and a position feedback loop.

[0033] The flexible drive moving mirror optical path module is used to convert the rotational motion of the motor into the reciprocating oscillation of the moving mirror assembly through the flexible drive moving mirror scanning mechanism, so as to complete the physical modulation of the optical path difference.

[0034] Interference signal acquisition module 10 is used to acquire interference light intensity signals and process them digitally.

[0035] The composite control estimation module is used to pre-construct a state-space model based on the state variables of the motor side and load side of the flexible transmission moving mirror scanning mechanism; convert the digitally processed interference light intensity signal into the measured value of optical path difference and determine the optical path difference tracking error; determine the mechanical tracking control component based on the optical path difference tracking error; determine the look-ahead bias of optical prediction based on the state-space model; determine the optical prediction compensation component based on the look-ahead bias of optical prediction; determine the dynamic feedforward control component based on the inverse dynamic equation containing the state variables of the motor side and load side; and construct a composite control quantity based on the mechanical tracking control component, the optical prediction compensation component, and the dynamic feedforward control component.

[0036] The composite control estimation module includes: The optomechanical coupling state estimation module 11 has a built-in Kalman filter. It is used to obtain the posterior state estimate of the current moment by taking the measured optical path difference as the observation and passing it through the Kalman filter recursive. Based on the posterior state estimate of the current moment, the predicted value of the multidimensional mechanical state vector of the next moment is derived by the pre-constructed augmented mechanical state recursive equation. The independent optical path difference timing prediction module 12 is used to predict the optical path difference based on the multidimensional mechanical state vector of the next moment, and to obtain the optical path difference prediction value of the next moment by mapping it through the pre-constructed optical path difference observation equation; and to determine the look-ahead bias of the optical prediction based on the optical path difference prediction value of the next moment and the preset target optical path difference. The hierarchical coupling control law operation module 13 is used to convert the interference light intensity signal into the measured optical path difference value, determine the optical path difference tracking error, determine the mechanical tracking control component based on the optical path difference tracking error, determine the optical prediction compensation component based on the look-ahead deviation of the optical prediction, determine the dynamic feedforward control component based on the basic mechanical state recursive equation, and construct the composite control quantity based on the mechanical tracking control component, the optical prediction compensation component, and the dynamic feedforward control component.

[0037] The motor drive circuit module 14 is used to convert the composite control quantity into a motor-side drive signal to compensate for the scanning error of the moving mirror of the FTIR spectrometer.

[0038] The position feedback loop 15 is connected to the drive motor 1 and is used to collect the encoder mechanical rotation angle signal of the drive motor 1 for motor safety monitoring and initial zero-position calibration.

[0039] Specifically, such as Figure 1As shown in the figure, the flexible transmission moving mirror optical path module includes a drive motor 1, a driving wheel 2, a steel belt 3, a driven shaft 4, a moving mirror group 5, an infrared light source 6, a beam splitter 7, a fixed mirror 8, and a photodetector 9. The infrared light source 6 emits an infrared light beam, which is split by the beam splitter 7 and then incident on the fixed mirror 8 and the moving mirror group 5 respectively. The drive motor 1 outputs a rotational torque, which sequentially drives the moving mirror group 5 to swing reciprocally via the driving wheel 2, the flexible steel belt 3 and the driven shaft 5, so as to change the optical path difference between the two reflected light beams. The two reflected light beams overlap and then interfere with each other, and the interference light intensity is converted into an analog electrical signal by the photodetector 9 for output. Nonlinear disturbances such as elastic elongation, assembly clearance, rim slip and temperature drift exist during the transmission process of the flexible steel belt, and the encoder matched with the motor can only collect the mechanical rotation angle on the motor side, and cannot characterize the actual optical modulation angle of the moving mirror group.

[0040] The interference signal acquisition module 10 is connected to the signal output end of the photodetector 9. The interference signal acquisition module 10 internally adopts a 16-bit or higher high-speed AD sampling chip to collect the analog signal of the interference light intensity. The hardware wiring, data cache area and position feedback loop 15 of the interference signal acquisition module 10 are independent of each other, so as to construct an independent optical feedback channel, and the optical signal is not connected to the inner loop of the motor mechanical feedback. The interference signal acquisition module 10 outputs the sampled digital interference light intensity signal to the opto-mechanical coupling state estimation module 11.

[0041] The opto-mechanical coupling state estimation module 11 is built in with a variable-parameter Kalman filter, pre-stored with motor-side inertia, load-side inertia, equivalent stiffness and damping parameters of the steel belt, and equipped with a pre-constructed state space model. The multi-dimensional mechanical state vector inside the module is:

[0042] Wherein, k is the sampling instant, is the motor-side rotation angle, is the motor-side rotational speed, is the load-side moving mirror rotation angle, is the load-side moving mirror rotational speed, is the comprehensive angular displacement disturbance of the flexible transmission moving mirror scanning mechanism. In the embodiment of the present invention, the optical equivalent angle is defined as an algebraic function of the load-side moving mirror rotation angle , and the comprehensive angular displacement disturbance, and after each step of state update, the optical equivalent angle is updated in real time.

[0043] The state space model carried by the opto-mechanical coupling state estimation module 11 includes an augmented mechanical state recurrence equation and an optical path difference observation equation, wherein the augmented mechanical state recurrence equation is obtained by combining the basic mechanical state recurrence equation and the disturbance random attenuation model:

[0044] in, , These are the multidimensional mechanical state vectors at time k and time k+1, respectively. To control the input, For total process noise, For system matrix, This is the control matrix.

[0045] The optical path difference observation equation uses the optical equivalent angle As the algebraic mapping target, the optical path difference observation equation is obtained through linear mapping based on the optical path geometry:

[0046] in, Let k be the predicted optical path difference value at time k. For observation noise, observation matrix The optomechanical coupling state estimation module 11 generates a posterior state estimate at time k based on a Kalman filter. The result is output to the hierarchical coupling control law calculation module 13; simultaneously, the optomechanical coupling state estimation module 11 generates a multidimensional mechanical state vector prediction value at time k+1 based on the augmented mechanical state recursive equation. The data is then output to the independent optical path difference timing prediction module 12. For example, the optomechanical coupling state estimation module 11 uses a scanning angular velocity... As the operating condition switching threshold, the process noise covariance matrix of the Kalman filter is adaptively adjusted. Observation noise covariance matrix When the scanning angular velocity At that time, the mechanical model had a higher degree of credibility. Take the first fixed interval; when the scanning angular velocity At that time, the flexible disturbance is enhanced. Take a second fixed interval whose value is greater than that of the first interval; The value of varies monotonically based on the signal-to-noise ratio (SNR) of the interference signal; the higher the SNR, the better. The smaller the value; and The interval values ​​are obtained by online estimation using Allen's analysis of variance or covariance matching method. Both Allen's analysis of variance and covariance matching method are mature online estimation methods for noise covariance that can be directly implemented in engineering.

[0047] It is understandable that the Kalman filter in the optomechanical coupling state estimation module 11 can be extended to extended Kalman filter (EKF), unscented Kalman filter (UKF), rolling time domain estimation (MHE), etc., to adapt to the strong nonlinearity of the optical path.

[0048] The independent optical path difference timing prediction module 12 receives the multidimensional mechanical state vector prediction value at time k+1 output by the optomechanical coupling state estimation module 11. This completes one-step look-ahead prediction of optical path difference without relying on the mechanical position signal of the motor encoder. The independent optical path difference timing prediction module 12 reads... Then, the predicted value of the optical path difference is calculated by substituting it into the optical path difference observation equation. Combined with the preset target optical path difference Solving for the look-ahead bias of optical predictions .

[0049] The independent optical path difference timing prediction module 12 only outputs the look-forward bias to the hierarchically coupled control law calculation module 13 as the input error of the optical prediction compensation branch. This independent optical path difference timing prediction module 12 is functionally complementary to the optomechanical coupled state estimation module 11—the optomechanical coupled state estimation module 11 is responsible for the complete five-dimensional joint estimation and filtering of the mechanical state, while the independent optical path difference timing prediction module 12 is responsible for the forward calculation of the optical path difference prediction value. The decoupled design of the two ensures the timing determinism of the control law calculation.

[0050] The hierarchical coupled control law calculation module 13 determines the mechanical tracking control component based on the interference light intensity signal, the optical prediction compensation component based on the look-ahead deviation of the optical prediction, and the dynamic feedforward control component based on the basic mechanical state recursive equation. It then executes a three-component parallel decoupled control architecture. After independent calculation of the three control components, they are superimposed at the drive end to form a composite control quantity, with no coupling at the error layer. The calculation formula can be:

[0051] in, For mechanical tracking control components, the tracking error is determined by real-time optical path difference. The value is obtained through calculation by an incremental PID controller and is used to eliminate the static optical path deviation of the system. This is a dynamic feedforward control component used to compensate for the dynamic inertia of the motor-load mechanical inertia in advance, but it does not include compensation for flexible disturbances. For optical prediction compensation components, satisfy Optical prediction compensation ratio coefficient The value can be switched according to the scanning speed, or based on the posterior state estimation at the current moment. It combines the optical equivalent angle change rate and the disturbance estimate for adaptive adjustment to offset the cumulative error of steel strip disturbance in advance.

[0052] The hierarchical coupling control law calculation module 13 performs independent calculations on the three components and then superimposes them with unified dimensions to obtain the final composite control quantity. Output to motor drive circuit module 14.

[0053] The motor drive circuit module 14 receives the composite control quantity output by the hierarchical coupling control law calculation module 13, converts the control quantity into a PWM drive signal and outputs it to the drive motor 1 to complete torque execution.

[0054] The position feedback loop 15 only collects the encoder mechanical rotation angle signal of the drive motor 1 for motor safety monitoring and initial zero-position calibration. It does not participate in the optical path difference closed-loop control and is completely decoupled from the independent optical feedback channel. The overall signal flow process of the FTIR spectrometer moving mirror scanning error compensation system described above is as follows: The photodetector 9 outputs the interference light intensity analog signal obtained by the flexible transmission moving mirror optical path module, which is digitally sampled by the interference signal acquisition module 10 to obtain the interference light intensity digital signal; the interference light intensity digital signal is input to the optomechanical coupling state estimation module 11, which performs multi-dimensional mechanical state joint estimation based on the Kalman filter and state space model; the optomechanical coupling state estimation module 11 outputs two signals—the posterior state estimate of the current moment is sent to the hierarchical coupling control law operation module 13, and the multi-dimensional mechanical state vector prediction value of the next moment is sent to the independent optical path difference timing prediction module 12; the independent optical path difference timing prediction module 12, based on the multi-dimensional mechanical state vector of the next moment, performs multi-dimensional mechanical state joint estimation based on the posterior state estimate of the current moment. The predicted value is used to calculate the look-ahead deviation, and the look-ahead deviation is output to the hierarchical coupling control law calculation module 13. The hierarchical coupling control law calculation module 13 performs parallel calculations on the three control components and superimposes them to generate a composite control quantity, which is output to the motor drive circuit module 14. The motor drive circuit module 14 converts the composite control quantity into a PWM drive signal and transmits it to the drive motor 1 through the position feedback loop 15. The position feedback loop 15 collects the encoder mechanical rotation angle signal of the drive motor 1 in real time for motor safety monitoring and initial zero-position calibration. This signal does not participate in the optical path difference closed-loop control. The drive motor 1 receives the drive signal and runs, causing the moving mirror group 5 to swing and update the optical path difference. The photodetector 9 collects the new interference light intensity signal in real time, forming a complete optomechanical coupling closed-loop control link.

[0055] Furthermore, the system can also be equipped with a temperature acquisition module to incorporate temperature observations into the comprehensive angular displacement disturbance d(k) to complete joint estimation, thereby further suppressing the optical path difference shift caused by temperature drift.

[0056] Based on the FTIR spectrometer moving mirror scanning error compensation system described above, this invention also proposes an FTIR spectrometer moving mirror scanning error compensation method. Before executing this method, the following offline calibration and parameter loading are completed after the device is powered on: (1) Optical mapping coefficient Calibration: Select no fewer than 10 equally spaced positions within the scanning stroke range of the moving mirror. Record the motor encoder rotation angle and the phase of the interference signal at each position. Obtain the optical mapping coefficient between the optical equivalent angle and the optical path difference through linear fitting using the least squares method.

[0057] (2) Disturbance attenuation coefficient Calibration: With the drive motor stationary and the system in thermal equilibrium, continuously acquire time series of the measured optical path difference of the interference signal, with a sampling duration of no less than 10 seconds, and calculate the first-order autocorrelation coefficient of the sequence as... The initial estimate, The range of values ​​is The basis for this value is: Reflecting the time-domain attenuation characteristics of disturbances in flexible transmission, A value approaching 1 indicates that the disturbance has long-term, slowly varying characteristics. Range of values The corresponding continuous domain time constant is approximately 0.1 to 2 seconds (depending on the sampling period), which aligns with the physical timescale of elastic deformation and thermal drift in steel strips. Furthermore, to more accurately reflect the disturbance characteristics under dynamic operating conditions, optical path difference data during dynamic scanning can be acquired online, and its autocorrelation function can be analyzed, along with the initial... The value is corrected.

[0058] (3) Noise covariance matrix , Determination: Using the covariance matching method, the statistical characteristics of the innovation sequence are collected under stable system operation conditions, and online estimation is performed. and Alternatively, Allen's analysis of variance can be used to determine the noise covariance. Both Allen's analysis of variance and covariance matching are mature online estimation methods for noise covariance that can be directly implemented in engineering.

[0059] (4) Input the moment of inertia of the motor / load , Steel strip torsional stiffness Damping Write the timing target optical path difference corresponding to the preset scan trajectory to the controller memory. .

[0060] After completing the offline calibration and parameter loading, error compensation control is performed using an FTIR spectrometer moving mirror scanning error compensation method proposed in this invention.

[0061] Please see Figure 2 , Figure 2A flowchart illustrating a method for compensating for scanning errors of a moving mirror in an FTIR spectrometer, provided as an embodiment of the present invention, includes: S101. Construct a state-space model in advance based on the state variables of the motor side and load side of the flexible transmission moving mirror scanning mechanism; S102. Obtain the interference light intensity signal, convert the interference light intensity signal into the measured value of optical path difference, determine the optical path difference tracking error, and determine the mechanical tracking control component based on the optical path difference tracking error. S103. Based on the measured optical path difference, determine the look-ahead bias of the optical prediction using the state-space model, and determine the optical prediction compensation component based on the look-ahead bias of the optical prediction. S104. Determine the dynamic feedforward control components based on the inverse dynamic equations that include state variables from the motor side and the load side. S105. Construct a composite control quantity based on the mechanical tracking control component, the optical prediction compensation component, and the dynamic feedforward control component. S106. Convert the composite control quantity into a motor-side drive signal to compensate for the scanning error of the moving mirror of the FTIR spectrometer.

[0062] This invention targets FTIR spectrometers employing flexible drive moving mirror scanning mechanisms. It constructs a state-space model of the flexible drive moving mirror scanning mechanism to form an independent optical feedback channel, thereby determining the look-ahead bias of optical prediction and thus determining the optical prediction compensation component. This effectively compensates for the dynamic optical deviation introduced by the flexible drive structure. Furthermore, this invention determines the optical path difference tracking error by acquiring interference light intensity signals and constructs a mechanical tracking control component. The mechanical tracking control component, optical prediction compensation component, and dynamic feedforward control component form a multi-source error hierarchical decoupled compensation architecture. Mechanical tracking, feedforward, and optical look-ahead compensation correspond to different physical error sources, and various optical and mechanical errors have independent compensation channels. This allows for the construction of a composite control quantity to compensate for moving mirror scanning errors, effectively improving the optical modulation stability of FTIR under high-speed scanning and long-term operation conditions.

[0063] In the above implementation, step S101 mainly involves constructing a state-space model of optomechanical coupling layering.

[0064] Generally, based on the transmission relationship between the motor side and the load side of the flexible transmission moving mirror scanning mechanism, the flexible transmission system composed of the motor side, the flexible transmission structure, and the load side can be equivalently represented as a typical dual-inertia coupling structure connected by two rigid inertia units (motor side and load side) through an elastic damping unit (steel belt). This equivalent dual-inertia coupling structure can characterize the influence of the elastic deformation and damping characteristics of the steel belt on the transmission accuracy. Therefore, a state-space model can be established starting from the continuous domain dynamic equations of this typical dual-inertia coupling structure.

[0065] In some embodiments, a state-space model of the flexible drive moving mirror scanning mechanism is pre-constructed based on the state variables of the motor side and load side of the flexible drive moving mirror scanning mechanism, including: S101-1. Construct a multidimensional mechanical state vector based on the state variables of the motor side and load side of the flexible transmission moving mirror scanning mechanism. Based on the multidimensional mechanical state vector, construct the augmented mechanical state recursive equation of the flexible transmission moving mirror scanning mechanism through the torque balance equation of the motor side and the load side. S101-2. Determine the optical equivalent angle of the moving mirror based on the state variables of the motor side and the load side, and construct the optical path difference observation equation based on the optical equivalent angle and the multidimensional mechanical state vector. S101-3. Establish a state-space model based on the augmented mechanical state recursive equation and the optical path difference observation equation.

[0066] Specifically, in step S101-1 above, based on the motor-load dual inertia dynamic equation and combined with the Kalman filter, the augmented mechanical state recursive equation of the flexible transmission moving mirror scanning mechanism is constructed.

[0067] In some embodiments, a multidimensional mechanical state vector is constructed based on the state variables of the motor side and the load side. Based on the multidimensional mechanical state vector, an augmented mechanical state recursive equation for the flexible transmission moving mirror scanning mechanism is constructed through the torque balance equations of the motor side and the load side, including: The continuous domain dynamic equations are constructed by using the torque balance equations on the motor side and the load side. Based on the state variables of the motor side and the load side, the basic mechanical state vector is constructed, and the continuous domain dynamic equation is discretized to obtain the basic mechanical state recursive equation. A comprehensive angular displacement disturbance of a flexible transmission moving mirror scanning mechanism is introduced into the basic mechanical state vector to construct a multidimensional mechanical state vector, and a disturbance stochastic decay model based on the comprehensive angular displacement disturbance is established. The multidimensional mechanical state vector includes: the rotation angle and angular velocity of the motor-side driving wheel, the equivalent rotation angle and angular velocity of the load side, and the comprehensive angular displacement disturbance of the flexible transmission moving mirror scanning mechanism. By combining the basic mechanical state recursive equations with the perturbation stochastic decay model, an augmented mechanical state recursive equation is constructed.

[0068] Specifically, the torque balance equations on the motor side and the load side together constitute the following dual-inertia continuous-domain dynamic model:

[0069]

[0070] Where t is the continuous-time index. , These are the equivalent moments of inertia on the motor side and the load side, respectively. , These are the equivalent torsional stiffness and damping coefficient of the steel strip, respectively. Let be the motor output torque at time t; Let be the load torque at time t; Let be the rotation angles on the motor side and the load side at time t, respectively. They are respectively The first derivative of represents the angular velocity of the motor side and the load side at time t. They are respectively The second derivative of represents the angular acceleration on the motor side and the load side at time t.

[0071] Constructing a four-dimensional basic mechanical state vector in a continuous domain angular velocity of the motor side angular velocity on the load side The above-described two-inertia continuous-domain dynamic model can be written in the following standard continuous state-space form:

[0072] in, , , For a continuous domain system matrix, It is a continuous domain control matrix; for The first derivative.

[0073] To facilitate real-time recursive calculations by the embedded controller, the above continuous domain equations need to be discretized, defining a discrete four-dimensional basic mechanical state vector. k is the sampling time. The values ​​at time k are the motor-side rotation angle, motor-side angular velocity, load-side rotation angle, and load-side angular velocity, respectively. The forward Euler method (first-order Taylor approximation) is used, with a sampling period of [value missing]. The four-dimensional fundamental mechanical state recurrence equation is obtained as follows:

[0074] in, Let be the four-dimensional basic mechanical state vector at time k+1. Let K be the mechanical process noise at time k (expected to be zero). Let k be the control quantity at time k; It is a four-dimensional state transition matrix. The four-dimensional control matrix has the following specific expansion forms: ,

[0075] The above four-dimensional basic mechanical state recursive equations fully characterize the recursive relationship between the rotation angle and angular velocity of the motor side and the load side in the discrete time domain, which is the basis for further augmented modeling.

[0076] Subsequently, a comprehensive angular displacement disturbance of the flexible transmission moving mirror scanning mechanism is introduced. Its physical meaning is the equivalent angular displacement deviation superimposed on the load side by nonlinear factors such as elastic elongation and clearance slip of the steel belt. This disturbance belongs to the slowly varying comprehensive angular displacement disturbance caused by the flexible transmission of the steel belt, and has slow time-varying evolution characteristics. Therefore, a first-order disturbance stochastic decay model is used to characterize its dynamic change law:

[0077] in, These represent the combined angular displacement disturbances at time k and time k+1, respectively. This is for mechanical process noise (expected to be zero). This is the disturbance attenuation coefficient.

[0078] To establish a unified recursive equation for augmented mechanical states, a five-dimensional mechanical state vector is constructed. The four-dimensional basic mechanical state recursive equations are combined with the disturbance attenuation model and then assembled in a block matrix form:

[0079] Record the total process noise Then the total process noise The expected value is zero, which conforms to the standard Kalman filter zero-mean white noise assumption.

[0080] After simplifying the above equation, we obtain the five-dimensional augmented mechanical state recurrence equation:

[0081] Among them, the system matrix Control matrix It can be derived from simultaneous equations using the forward Euler method, the zero-order hold method, or other discretization methods. Its fifth dimension is determined by the perturbation attenuation coefficient. And zero vector filling.

[0082] In step S101-2 above, the optical path difference observation equation is mainly established based on the geometric relationship of FTIR common optical path interference.

[0083] In the embodiments of the present invention, the optical equivalent angle of the moving mirror is the optical rotation angle of the moving mirror that actually participates in optical path modulation. The optical equivalent angle is determined by the combined angular displacement disturbance of the load-side equivalent rotation angle and the flexible transmission moving mirror scanning mechanism.

[0084] Specifically, this embodiment defines the optical equivalent angle as the load-side mechanical rotation angle based on the FTIR common-path interference geometry. Combined angular displacement disturbance The algebraic sum, i.e., the optical equivalent angle of the moving mirror. .

[0085] Optical path difference The optical equivalent angle of the moving mirror satisfies a fixed linear mapping relationship determined by calibration, that is... ,in, Let be the optical mapping coefficient. This linear mapping is expanded as a five-dimensional mechanical state vector inner product, and observation noise superimposed during the acquisition of the interferometric light intensity signal is introduced. (With an expected value of zero), the optical path difference observation equation can be obtained:

[0086] in, Let k be the optical path difference observation value, and let the observation matrix be... , To observe noise.

[0087] The form of the above optical path difference observation equation is directly derived from The algebraic expansion shows that only the load-side rotation angle Combined angular displacement disturbance The optical path difference output is jointly determined by the motor side rotation angle and speed, which are not directly present in the observation equation. This is entirely consistent with the physical mechanism that "the motor encoder rotation angle under steel belt flexible transmission cannot uniquely represent the actual optical modulation state." Moreover, this approach avoids the problem of algebraic constraints in the state vector preventing independent evolution, allowing the optical path difference observation to simultaneously correct the load side rotation angle. Combined angular displacement disturbance The estimated value achieves good observability of the optical modulation state.

[0088] In step S101-3 above, the augmented mechanical state recursive equation and the optical path difference observation equation are combined to form a state-space model of optomechanical coupling layering.

[0089] In the above implementation, step S102 mainly involves determining the optical path difference tracking error based on the collected interference light intensity signal and determining the mechanical tracking control component.

[0090] In some embodiments, the interference light intensity signal is converted into a measured optical path difference value, and the optical path difference tracking error is determined. Based on the optical path difference tracking error, the mechanical tracking control component is determined, including: S102-1. The interference light intensity signal is analyzed and phase unwrapped to obtain the measured value of the optical path difference; S102-2. Determine the optical path difference tracking error based on the difference between the preset target optical path difference and the measured optical path difference; S102-3. Based on the optical path difference tracking error, the mechanical tracking control component is determined using an incremental PID control law.

[0091] Specifically, the acquired interference light intensity signal satisfies the basic cosine interference model of FTIR:

[0092] in, The average interference light intensity. σ The incident light wavenumber, The optical path difference is changed by the reciprocating oscillation of the moving mirror. This causes the light intensity to oscillate periodically.

[0093] In step S102-1 above, for the acquired interference light intensity signal, the interference phase is analyzed by a digital phase-locked loop algorithm or the phase is demodulated by a Fast Fourier Transform (FFT). After phase unwrapping and conversion processing, the measured optical path difference at the current discrete time k is obtained. :

[0094] in, For the first The interference phase value after unwrapping at the sampling time. The wavelength of the incident infrared light is the center wavelength.

[0095] The conversion formula for the measured optical path difference adopts a common optical path design based on the moving mirror scanning mechanism, and the change in optical path difference... With the displacement of the moving mirror Between satisfy The relationship, combined with the interference phase The derivation is as follows Furthermore, a phase unwrapping algorithm is employed to address situations where the absolute value of the phase difference between adjacent sampling times exceeds [a certain threshold]. At that time, the corresponding additions or subtractions are made according to the positive or negative direction of the phase transition. Phase expansion correction completed.

[0096] In step S102-2 above, based on the preset target optical path difference Determine the real-time optical path difference tracking error:

[0097] Among them, the preset target optical path difference This is generated by discretizing the standard uniform / variable speed scanning waveform of the moving mirror based on a preset scanning trajectory for an FTIR spectrometer, representing each discrete moment within the scanning period. The corresponding unperturbed ideal optical path difference command value. Compared with the measured real-time optical path difference With consistent dimensions, their difference The standard deviation signal that constitutes the closed-loop feedback.

[0098] In steps S102-3 above, the tracking error is based on the real-time optical path difference. Determine the mechanical tracking control components .

[0099] The mechanical tracking control component An incremental PID control law is adopted, with the real-time optical path difference tracking error as the only input. Its discrete recursive expression is as follows:

[0100] in, They are respectively k -1 moment, k The mechanical tracking control component at any given moment; They are respectively k time, k -1 moment, k optical path difference tracking error at time -2; , , These are the proportional, integral, and derivative coefficients, which are preset using conventional PID parameter tuning methods and fixed in the controller before the scan starts. This component only responds to the measured deviation at the current and historical moments, and is a feedback-type error suppression term.

[0101] In the above implementation, step S103 mainly involves constructing the optical path difference time-series prediction branch based on the state-space model, calculating the look-ahead bias of the optical prediction, and determining the optical prediction compensation component.

[0102] In this embodiment, posterior state estimation is performed based on Kalman filtering, and multidimensional mechanical state vector prediction and optical path difference prediction are performed based on the state space model.

[0103] In some embodiments, the look-ahead bias of the optical prediction is determined using a state-space model based on the measured optical path difference, and the optical prediction compensation component is determined based on the look-ahead bias of the optical prediction, including: S103-1. Using the measured optical path difference at the current moment as the observation, the posterior state estimate at the current moment is obtained recursively through a Kalman filter. S103-2. Based on the posterior state estimation at the current moment, derive the predicted value of the multidimensional mechanical state vector at the next moment through the augmented mechanical state recursive equation. S103-3. Based on the predicted value of the multidimensional mechanical state vector at the next moment, the predicted value of the optical path difference at the next moment is obtained by mapping through the optical path difference observation equation. S103-4. Determine the look-ahead bias of optical prediction based on the predicted optical path difference at the next moment and the preset target optical path difference. S103-5. Determine the optical prediction compensation component based on the look-ahead bias of the optical prediction.

[0104] Specifically, in step S103-1 above, based on k Control input at time -1 , k Measured optical path difference at time The result is obtained recursively through the Kalman filter. Posterior state estimation of multidimensional mechanical state vector at time step Specifically, the recursive formula for the Kalman filter is: State prediction:

[0105] Covariance prediction:

[0106] Kalman gain:

[0107] Status Update:

[0108] Covariance update:

[0109] in, for k Posterior state estimate of the Kalman filter output at time -1 for k Prior state estimation at time step; The process noise covariance matrix; To observe the noise covariance matrix; Let K be the Kalman gain matrix at time k; Let be the error covariance matrix. They are respectively k The posterior covariance matrix at time -1 k The prior covariance matrix at time t, k The posterior covariance matrix at time t; A is the identity matrix; B is the system matrix; H is the control matrix; and A is the observation matrix.

[0110] In step S103-2 above, the state is recursively derived using the five-dimensional augmented mechanical state recursive equation to obtain the multidimensional mechanical state prediction value for the next time step:

[0111] in, This represents the multidimensional mechanical state prediction value at time k+1. for Posterior state estimation of multidimensional mechanical state vectors at time points, due to process noise. The expected value is zero, and its expected value is used when making predictions.

[0112] In step S103-3 above, the predicted optical path difference value for the next moment is obtained by mapping through the optical path difference observation equation:

[0113] in, Let H be the predicted optical path difference at time k+1, and H be the observation matrix.

[0114] In steps S103-4 above, the look-ahead bias of the optical prediction is determined based on the difference between the predicted optical path difference at the next moment and the preset target optical path difference.

[0115] in, Let K+1 be the target optical path difference. The look-ahead error of the optical prediction at time k+1 is actually the driving error of the optical prediction compensation component.

[0116] In step S103-5 above, the optical prediction compensation component is determined based on the look-ahead bias of the optical prediction. This component is characterized by leading bias. As input, a proportional compensation law is used, and the expression is:

[0117] in, This is the optical prediction compensation ratio coefficient, a scalar gain. Its value is selected based on the current scanning speed (smaller value for low speed, larger value for high speed), or based on the optical equivalent angle change rate and disturbance estimate in the state estimation results. Adaptive adjustment is performed. This optical prediction compensation component... It belongs to the forward pre-compensation term and is used to offset the cumulative optical path deviation that may be generated by the flexible disturbance of the steel strip in the next moment.

[0118] In the above embodiment, for step S104, the mechanical tracking control component is determined in step S102. Step S103: Determine the optical prediction compensation component. Subsequently, the dynamic feedforward components were further determined. .

[0119] The dynamic feedforward component Based on the undisturbed state The approximate inverse dynamics construction of the fundamental mechanical state recursive equations. Specifically, for the four-dimensional fundamental mechanical state recursive equations... By performing reverse calculations, the inverse dynamic equation is obtained. Based on the desired discrete state trajectories of the load side and the motor side, the required motor torque control quantity is solved by recursively applying the inverse dynamic equation. This is used as the feedforward control component, and the expression for the feedforward control quantity can be obtained:

[0120] in, , The load-side desired rotation angles at time k and k+1 are respectively obtained by converting the preset scan trajectory using calibration coefficients. Furthermore, based on the unperturbed four-dimensional dual-inertia dynamic relationship, the desired rotation angle on the motor side at time k is calculated. , These are the expected rotation angles on the motor side at times k-1 and k+1, respectively.

[0121] This optical prediction compensation component No disturbance This term is independent of the state estimation based on the augmented mechanical state recursive equation. It is only used to compensate for the dynamic lag of the motor-load two-stage inertial system itself and belongs to the open-loop forward torque pre-matching term.

[0122] In the above implementation, for step S105, the three control components—mechanical tracking control component, optical prediction compensation component, and dynamic feedforward control component—are independently calculated at their respective error levels, and then the final composite control quantity is constructed after dimensional unification.

[0123] In some embodiments, a composite control quantity is constructed based on the mechanical tracking control component, the optical prediction compensation component, and the dynamic feedforward control component, including: The mechanical tracking control component, the optical prediction compensation component, and the dynamic feedforward control component are superimposed to obtain the composite control quantity.

[0124] Specifically, the mechanical tracking control component, the optical prediction compensation component, and the dynamic feedforward control component are all converted into equivalent motor torque control quantities or PWM duty cycle increments, and then superimposed according to the following formula to obtain the composite control quantity. :

[0125] It should be noted that the error signals of the three components, especially... and Within each branch, closed-loop iterations are performed without adding at the error level. Instead, the torque commands are algebraically superimposed only at the output drive end, ensuring that error components from different physical sources have independent compensation channels.

[0126] This invention achieves feedforward-feedback-prediction coordinated control based on mechanical tracking control components, optical prediction compensation components, and dynamic feedforward control components. The division of labor logic among the three control components is as follows: 1. Dynamic feedforward control component Based on the absence of perturbation states The design of the four-dimensional basic mechanical state recursive equations is intended to compensate for the pre-calculated, deterministic dynamic lag of the motor-load inertia, and only compensates for the dynamic inertia of the motor-load mechanical system, without including compensation for flexible disturbances. This component mainly targets the trajectory tracking torque requirements under the ideal model, outputting control action before errors occur.

[0127] 2. Optical prediction compensation component The forward bias of the optical path difference is predicted based on a state-space model. At time k+1, the predicted value of the multidimensional mechanical state vector is obtained by calculating the multidimensional mechanical state recursive equation. Afterwards, look-ahead bias That is, the information already obtained at the current time k, used to generate the current time. Pre-compensation control quantity In a physical sense, this manifests as advance compensation for the future state of the system. The prediction process uses the control variables already determined by the control law at the current moment. The calculations do not involve unknown future control quantities and possess causal realizability.

[0128] 3. Mechanical tracking control component Based on real-time optical path difference measurement feedback, its function is to suppress the residual errors that the two open-loop components cannot fully compensate for, as well as the unmodeled high-frequency random disturbances.

[0129] The three control components above have clearly defined functions, forming a multi-source error hierarchical decoupling compensation architecture: dynamic feedforward control component Optical prediction compensation component for handling known deterministic dynamics Handling modelable, slowly varying disturbance deviations, mechanical tracking control components Handling residual uncertainty errors. Disturbance state. The estimated values ​​are mainly used for state updates and prediction calculations, rather than being directly fed into the feedforward channel. This design avoids the high-frequency chattering of control quantities that may be caused by directly introducing noisy estimates into feedforward control.

[0130] In the above embodiment, for step S106, the composite control quantity is... The signal is converted into a motor-side drive signal and fed back to the flexible transmission moving mirror scanning mechanism to compensate for scanning errors of the FTIR spectrometer's moving mirror. The process then returns to step S102, iteratively forming an optically equivalent angle. Optomechanical coupling closed-loop control for core characterization quantities. For example... Figure 3 The diagram shown is a schematic of the closed-loop control process of the present invention.

[0131] The observability principle of this invention is: when the disturbance attenuation coefficient satisfy At that time, the gradually changing composite angular displacement disturbance The time-domain evolution characteristics and the mechanical rotation angle on the load side The controlled dynamic characteristics are significantly distinguishable on the response time scale. The observation matrix, combined with differentiated state transition characteristics, allows for effective bivariate identification through time-series filtering, and the system exhibits good observability under these parameter conditions. Load-side mechanical rotation angle. Due to the rapid response characteristics of the motor torque, and considering the combined angular displacement disturbance... It has first-order slowly varying decay characteristics. (Approaching 1) The Kalman filter can reasonably decompose the total optical angle deviation into load motion and disturbance components by utilizing multiple observations over a time series. If The value approaches 1, considering the overall angular displacement disturbance. Approaching a non-stationary random walk, the load angle and disturbance dynamics converge, leading to a decline in long-term identification accuracy; if If the value is less than 0.95, the disturbance decays too quickly, and long-term observation information is insufficient, making it impossible to reliably separate the two states. Therefore... The interval is a relatively good compromise that balances the slow variation characteristics of the disturbance with the good observability of the system.

[0132] This invention proposes a method for compensating for the scanning error of a moving mirror in an FTIR spectrometer. It does not simply improve the mechanical motion accuracy of the motor; its core lies in establishing a state-space model that includes the motor-side motion state, the load-side motion state, and the comprehensive angular displacement disturbance of the flexible transmission. By coupling the state-space model with the mechanical dynamics model and interferometric optical observation information, various nonlinear factors of the flexible steel belt transmission are incorporated into the system description as estimable states. This achieves the reconstruction of the unmeasurable optical modulation state. Furthermore, combined with the feedback of the interferometric light intensity signal, a recursive state estimation algorithm is used to estimate the optical equivalent angle, which cannot be directly measured, in real time. This transforms the control objective from mechanical position tracking / motor mechanical angle tracking to optical path difference error tracking based on the optical equivalent angle. This invention addresses the technical shortcomings of existing portable FTIR spectrometers with steel belt drives, which are mainly based on closed-loop mechanical quantities of the motor, do not fully consider the optomechanical coupling characteristics of the flexible transmission, and cannot actively suppress the dynamic deviation of the optical path difference. It achieves an architectural shift from mechanical motion control to optical state-driven control, improving the optical path difference modulation accuracy and system adaptability.

[0133] It is understood that the above method embodiments are based on the above system embodiments, and the brief descriptions of different embodiments can be referred to each other.

[0134] The effectiveness of this invention is verified below using some experimental data.

[0135] Experiment 1: Under high-speed 5Hz uniform scanning conditions, the optical path difference tracking error of the traditional mechanical closed-loop scheme and the scheme of the present invention is compared through simulation experiments. The traditional mechanical closed-loop scheme adopts a single control target based on mechanical quantities such as motor angle and speed. The mechanical, noise and disturbance parameters of the two simulation models are completely unified, only the control algorithm is different.

[0136] By combining the temporal error fluctuation curve and the cumulative error distribution, the optical path difference tracking stability of the two schemes was compared, and the results were as follows: Figure 4 The figure shows a comparison of the optical path difference tracking error between the proposed solution and the traditional mechanical closed-loop solution under the high-speed 5Hz uniform scanning condition. In the figure, (a) is the time-domain optical path difference error of the proposed solution, (b) is the time-domain optical path difference error of the existing traditional mechanical closed-loop solution, and (c) is the statistical curve of the cumulative distribution of the absolute error of the two solutions.

[0137] analyze Figure 4 It can be seen that the optical path difference error of the traditional mechanical closed-loop scheme oscillates periodically and significantly, with a peak value close to ±20μm; the optical path difference (OPD) error of the present invention is stably controlled within ±5μm; the error cumulative distribution curve shows that, under the same probability, the optical path difference error of the present invention is much smaller than that of the traditional mechanical closed-loop scheme, and the present invention can significantly suppress optical path difference fluctuations; in addition, the error cumulative distribution curve statistically corroborates that the tracking accuracy of the present invention is better than that of the traditional mechanical closed-loop scheme at all times.

[0138] Experiment 2: Under high-speed 5Hz scanning conditions, the three-component parallel composite control architecture of this invention is compared with the single PID control of the traditional mechanical closed-loop scheme in terms of control structure composition and high-frequency suppression capability. The two schemes are selected with completely identical single scanning time windows and the simulation benchmark parameters are unified.

[0139] like Figure 5 The figure shows a comparison between the single-cycle three-component control torque decomposition result of the present invention and the single PID control torque of the traditional mechanical closed-loop scheme. In this figure, (a) is a single complete scan cycle, which is used to decompose and display four independent time-series curves of the total control torque, mechanical tracking control component (PID), dynamic feedforward control component, and one-step optical prediction compensation component of the present invention; (b) is the single PID control torque of the traditional mechanical closed-loop scheme under the same time window, without additional feedforward and prediction branches; (c) is a comparison of the power spectral density of the control quantity of the two schemes. The horizontal axis is the frequency, the vertical axis is the power spectral density (PSD), and the gray area is the high-frequency disturbance evaluation interval.

[0140] analyze Figure 5 As can be seen, in the three-component parallel composite control architecture of the present invention, the mechanical tracking control component eliminates static offset, the dynamic feedforward control component cancels motion inertia, and the optical prediction compensation component compensates for periodic steel belt disturbances in advance; the traditional mechanical closed-loop scheme has no pre-compensation branch, and the high-frequency interval control excitation is insufficient, making it difficult to suppress high-frequency jitter of the steel belt. Figure 5 The three-component parallel composite control architecture of the present invention intuitively demonstrates the compensation effects of the PID, feedforward, and prediction torques, forming a clear structural distinction from the single PID control of the traditional mechanical closed-loop scheme; (c) The comparison of the power spectral density of the control quantity further verifies that the present invention has a better ability to suppress high-frequency steel strip disturbances, providing direct experimental evidence for the composite control scheme of the present invention.

[0141] Experiment 3: Using high-speed operating conditions and uniform optical path difference time-series data, the impact of the traditional mechanical closed-loop scheme and the scheme of this invention on the final infrared spectrum quality was compared from the perspective of instrument terminal imaging performance. The two schemes used the same spectral reconstruction algorithm without parameter modification.

[0142] Figure 6 The graph shows a comparison of the high-speed scanning restored spectrum and SNR of the present invention and the traditional mechanical closed-loop scheme. The black dashed line is the standard ideal spectrum without scanning error, the blue curve is the restored spectrum obtained by using the interference light intensity signal with error based on the present invention (SNR=5.02dB), and the red curve is the restored spectrum obtained by the traditional mechanical closed-loop scheme (SNR=2.88dB).

[0143] analyze Figure 6It is known that traditional mechanical closed-loop schemes suffer from significant spectral baseline drift and noise spikes, while the scheme of this invention yields a spectral curve that closely matches the standard ideal spectrum, resulting in a significantly improved signal-to-noise ratio. Figure 6 The beneficial effects of this invention can be verified from the perspective of actual detection applications of FTIR spectrometers. Under high-speed scanning conditions, the restored spectrum of this invention is closer to the standard ideal spectrum, the baseline drift and spectral noise are significantly reduced, and the spectral signal-to-noise ratio is significantly improved, proving that the optical path difference compensation scheme of this invention can effectively improve the detection accuracy of infrared spectrometers.

[0144] Figure 7 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 7 As shown, the electronic device may include a processor 710, a communication interface 720, a memory 730, and a communication bus 740, wherein the processor 710, the communication interface 720, and the memory 730 communicate with each other through the communication bus 740. The processor 710 can call logical instructions in the memory 730 to execute an FTIR spectrometer moving mirror scanning error compensation method provided in the above-described method embodiments.

[0145] Furthermore, the logical instructions in the aforementioned memory 730 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0146] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute an FTIR spectrometer moving mirror scanning error compensation method provided in the above-described method embodiments.

[0147] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform an FTIR spectrometer moving mirror scanning error compensation method provided by the methods described above.

[0148] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0149] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0150] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for compensating for scanning error of a moving mirror in an FTIR spectrometer, characterized in that, The method is applicable to FTIR spectrometers employing flexible driven moving mirror scanning mechanisms, and includes: A state-space model is pre-constructed based on the state variables of the motor side and load side of the flexible transmission moving mirror scanning mechanism; Acquire the interference light intensity signal, convert the interference light intensity signal into the measured optical path difference value, determine the optical path difference tracking error, and determine the mechanical tracking control component based on the optical path difference tracking error; Based on the measured optical path difference, the look-ahead bias of the optical prediction is determined by the state-space model, and the optical prediction compensation component is determined based on the look-ahead bias of the optical prediction. Based on the inverse dynamic equations that include the state variables of the motor side and the load side, determine the dynamic feedforward control components; A composite control quantity is constructed based on the mechanical tracking control component, the optical prediction compensation component, and the dynamic feedforward control component. The composite control quantity is converted into a motor-side drive signal to compensate for the scanning error of the moving mirror of the FTIR spectrometer.

2. The FTIR spectrometer moving mirror scanning error compensation method according to claim 1, characterized in that, A state-space model is pre-constructed based on the state variables of the motor side and load side of the flexible transmission moving mirror scanning mechanism, including: A multidimensional mechanical state vector is constructed based on the state variables of the motor side and the load side of the flexible transmission moving mirror scanning mechanism. Based on the multidimensional mechanical state vector, the augmented mechanical state recursive equation of the flexible transmission moving mirror scanning mechanism is constructed through the torque balance equation of the motor side and the load side. The optical equivalent angle of the moving mirror is determined based on the state variables of the motor side and the load side, and the optical path difference observation equation is constructed based on the optical equivalent angle and the multidimensional mechanical state vector. A state-space model is established based on the augmented mechanical state recursive equation and the optical path difference observation equation.

3. The FTIR spectrometer moving mirror scanning error compensation method according to claim 2, characterized in that, A multidimensional mechanical state vector is constructed based on the state variables of the motor side and load side of the flexible transmission moving mirror scanning mechanism. Based on the multidimensional mechanical state vector, an augmented mechanical state recursive equation for the flexible transmission moving mirror scanning mechanism is constructed through the torque balance equations of the motor side and load side, including: A continuous domain dynamic equation is constructed by using the torque balance equations of the motor side and the load side of the flexible transmission moving mirror scanning mechanism. Based on the state variables of the motor side and the load side, a basic mechanical state vector is constructed, and the continuous domain dynamic equation is discretized to obtain the basic mechanical state recursive equation. A comprehensive angular displacement disturbance of the flexible transmission moving mirror scanning mechanism is introduced into the basic mechanical state vector to construct a multidimensional mechanical state vector, and a disturbance random decay model based on the comprehensive angular displacement disturbance is established; wherein, the multidimensional mechanical state vector includes: the rotation angle and angular velocity of the motor-side driving wheel, the equivalent rotation angle and angular velocity of the load side, and the comprehensive angular displacement disturbance of the flexible transmission moving mirror scanning mechanism; By combining the basic mechanical state recursive equation with the perturbation stochastic decay model, an augmented mechanical state recursive equation is constructed.

4. The FTIR spectrometer moving mirror scanning error compensation method according to claim 2, characterized in that, The optical equivalent angle is the actual optical rotation angle of the moving mirror participating in optical path modulation. The optical equivalent angle is determined by the combined angular displacement disturbance of the load-side equivalent rotation angle and the flexible transmission moving mirror scanning mechanism.

5. The FTIR spectrometer moving mirror scanning error compensation method according to claim 1, characterized in that, The interference light intensity signal is converted into a measured optical path difference value, and the optical path difference tracking error is determined. Based on the optical path difference tracking error, the mechanical tracking control component is determined, including: The interference light intensity signal is analyzed and phase unwrapped to obtain the measured optical path difference. The optical path difference tracking error is determined based on the difference between the preset target optical path difference and the measured value of the optical path difference; wherein, the preset target optical path difference is generated by discretizing the standard uniform / variable speed scanning waveform of the moving mirror based on the preset scanning trajectory. Based on the optical path difference tracking error, an incremental PID control law is used to determine the mechanical tracking control component.

6. The FTIR spectrometer moving mirror scanning error compensation method according to claim 2, characterized in that, Based on the measured optical path difference, the look-ahead bias of the optical prediction is determined using the state-space model. Then, the optical prediction compensation component is determined based on the look-ahead bias, including: Using the measured optical path difference at the current moment as the observation, the posterior state estimate at the current moment is obtained recursively through a Kalman filter; Based on the posterior state estimate at the current moment, the predicted value of the multidimensional mechanical state vector at the next moment is derived through the augmented mechanical state recursive equation. Based on the predicted value of the multidimensional mechanical state vector at the next moment, the predicted value of the optical path difference at the next moment is obtained by mapping through the optical path difference observation equation. The look-ahead bias of the optical prediction is determined based on the predicted optical path difference value at the next moment and the preset target optical path difference. Based on the look-ahead bias of the optical prediction, the optical prediction compensation component is determined.

7. The FTIR spectrometer moving mirror scanning error compensation method according to claim 3, characterized in that, Based on the mechanical tracking control component, the optical prediction compensation component, and the dynamic feedforward control component, a composite control quantity is constructed, including: The mechanical tracking control component, the optical prediction compensation component, and the dynamic feedforward control component are superimposed to obtain a composite control quantity; wherein, the inverse dynamic equation used to calculate the dynamic feedforward control component is obtained by inversely solving the basic mechanical state recursive equation, and the mechanical tracking control component, the optical prediction compensation component, and the dynamic feedforward control component are independent of each other at their respective error levels.

8. A system for compensating for scanning error of a moving mirror in an FTIR spectrometer, characterized in that, The system includes: The flexible drive moving mirror optical path module is used to convert the rotational motion of the motor into the reciprocating oscillation of the moving mirror assembly through the flexible drive moving mirror scanning mechanism, so as to complete the physical modulation of the optical path difference; The interference signal acquisition module is used to acquire interference light intensity signals and perform digital processing. The composite control estimation module is used to pre-construct a state-space model based on the state variables of the motor side and load side of the flexible transmission moving mirror scanning mechanism; convert the digitized interference light intensity signal into the measured optical path difference value and determine the optical path difference tracking error; determine the mechanical tracking control component based on the optical path difference tracking error; determine the look-ahead bias of the optical prediction based on the state-space model; determine the optical prediction compensation component based on the look-ahead bias of the optical prediction; determine the dynamic feedforward control component based on the inverse dynamic equation containing the state variables of the motor side and load side; and construct a composite control quantity based on the mechanical tracking control component, the optical prediction compensation component, and the dynamic feedforward control component. The motor drive circuit module is used to convert the composite control quantity into a motor-side drive signal to compensate for the scanning error of the moving mirror of the FTIR spectrometer. The position feedback loop, connected to the drive motor, is used to collect the encoder mechanical rotation angle signal of the drive motor for motor safety monitoring and initial zero-position calibration.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the FTIR spectrometer moving mirror scanning error compensation method as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the FTIR spectrometer moving mirror scanning error compensation method as described in any one of claims 1 to 7.

Images