Control method of full-angle micro-electromechanical gyroscope based on online learning control

Abstract

This application belongs to the field of gyro control technology and relates to a control method for a full-angle micro-electromechanical (MEMS) gyroscope based on online learning control, comprising: obtaining a vibration signal of the full-angle MEMS gyroscope to obtain a real-time control force; performing measurement and control system identification based on the real-time control force to obtain a stiffness error; performing instantaneous stiffness control calculation based on the stiffness error to obtain an instantaneous stiffness control variable; using the instantaneous stiffness control variable as input, establishing an online learning control model for stiffness error correction and obtaining the current stiffness tuning; using the instantaneous stiffness control variable and the current stiffness tuning as input, using an online learning mechanism to perform a cyclic calculation to obtain real-time stiffness tuning; applying the real-time control force and the real-time stiffness tuning together to the full-angle MES gyroscope to update the vibration signal and achieve control of the full-angle MES gyroscope. The application can achieve high-precision control and high-performance self-correction.

Description

Technical Field

[0001] The present application relates to the technical field of gyroscope control, and in particular to a control method of a full-angle micro-electromechanical gyroscope based on online learning control. Background Art

[0002] In modern navigation and positioning systems, microelectromechanical (MEMS) gyroscopes (Gyroscopes) are widely used due to their small size, low cost, and high integration. These gyroscopes detect angular rate inputs by sensing the Coriolis force in a resonator. Common types include butterfly wing gyroscopes, ring gyroscopes, and cup gyroscopes.

[0003] Among them, the full-angle micro-electromechanical gyroscope, as a gyroscope working in the rate integration circuit mode, can directly sense the angle of external input, avoid noise accumulation in the integration process, and improve the accuracy of measurement.

[0004] In the prior art, traditional PID controllers are widely used in the control of full-angle micro-electromechanical gyroscopes.

[0005] However, traditional PID control uses fixed control parameters, which makes it difficult to adapt to changes in system parameters and external disturbances, resulting in insufficient control accuracy. In addition, PID is limited by its linear characteristics, and its adaptability and robustness to nonlinear and time-varying systems are poor. In addition, PID's performance is limited when facing complex dynamic systems such as full-angle micro-electromechanical gyroscopes, and it cannot fully utilize the advantages of the full-angle mode, achieve high-precision control of the full-angle gyroscope, and even more so, cannot achieve high-performance self-correction. Summary of the Invention

[0006] Based on this, it is necessary to provide a control method for a full-angle micro-electromechanical gyroscope based on online learning control to address the above technical problems, which can meet the performance requirements of the full-angle micro-electromechanical gyroscope for error identification and stiffness correction, improve the control accuracy and performance self-correction accuracy, and realize high-precision control and high-precision performance self-correction of the full-angle micro-electromechanical gyroscope.

[0007] The control method of the full-angle micro-electromechanical gyroscope based on online learning control includes:

[0008] Obtain the vibration signal of the full-angle micro-electromechanical gyroscope, perform measurement and control state calculation, and obtain the initial control quantity; perform instantaneous control calculation of the control force based on the initial control quantity to obtain the instantaneous control quantity of the control force; obtain the real-time control force based on the instantaneous control quantity of the control force;

[0009] According to the real-time control force, the measurement and control system is identified to obtain the stiffness error; according to the stiffness error, the stiffness instantaneous control calculation is performed to obtain the stiffness instantaneous control value;

[0010] Taking the instantaneous stiffness control variable as input, an online learning mechanism is used to compensate the stiffness error correction control loop, establish an online learning control model for stiffness error correction, and obtain the current stiffness tuning. Taking the instantaneous stiffness control variable and the current stiffness tuning as input, an online learning mechanism is used to perform cyclic calculations to obtain real-time stiffness tuning.

[0011] The real-time control force and the real-time stiffness tuning are applied to the full-angle micro-electro-mechanical gyroscope to update the vibration signal and realize the control of the full-angle micro-electro-mechanical gyroscope.

[0012] In one embodiment, obtaining the real-time control force according to the instantaneous control amount of the control force includes:

[0013] Taking the instantaneous control quantity of the control force as input, the online learning mechanism is adopted to compensate the control force control loop, establish an online learning control model of the control force, and obtain the current control force; taking the instantaneous control quantity of the control force and the current control force as input, the online learning mechanism is adopted to perform cyclic calculations to obtain the real-time control force.

[0014] In one embodiment, the instantaneous control value of the control force is used as input, and an online learning mechanism is used to compensate the control force control loop, establish an online learning control model of the control force, and obtain the current control force. The instantaneous control value of the control force and the current control force are used as input, and an online learning mechanism is used to perform cyclic calculations to obtain the real-time control force, including:

[0015] Taking the instantaneous control amount of the energy control force as input, the online learning mechanism is used to compensate the energy control force control loop, establish an online learning control model of the energy control force, and obtain the current energy control force; taking the instantaneous control amount of the energy control force and the current energy control force as input, the online learning mechanism is used to perform cyclic calculations to obtain the real-time energy control force;

[0016] Taking the instantaneous control quantity of the orthogonal control force as input, an online learning mechanism is adopted to compensate the orthogonal control force control loop, establish an online learning control model of the orthogonal control force, and obtain the current orthogonal control force; taking the instantaneous control quantity of the orthogonal control force and the current orthogonal control force as input, an online learning mechanism is adopted to perform cyclic calculations to obtain the real-time orthogonal control force.

[0017] In one embodiment, the instantaneous control amount of the energy control force is used as input, and an online learning mechanism is used to compensate the energy control force control loop, establish an online learning control model of the energy control force, and obtain the current energy control force, including:

[0018] Taking the instantaneous control value of the energy control force as input, the energy control law, instantaneous energy controller, and energy control force learning deviation of the gyroscope are designed. Based on the energy control law, instantaneous energy controller, and energy control force learning deviation of the gyroscope, an online learning control model of the energy control force is established to obtain the current energy control force.

[0019] Taking the instantaneous control quantity of the orthogonal control force as input, an online learning mechanism is used to compensate the orthogonal control force control loop, establish an online learning control model of the orthogonal control force, and obtain the current orthogonal control force, including:

[0020] Taking the instantaneous control quantity of the orthogonal control force as input, the orthogonal control law, orthogonal instantaneous controller and orthogonal control force learning deviation of the gyroscope are designed; based on the orthogonal control law, orthogonal instantaneous controller and orthogonal control force learning deviation of the gyroscope, an online learning control model of the orthogonal control force is established, and the current orthogonal control force is obtained.

[0021] In one embodiment, an online learning control model of energy control force is established and the current energy control force is obtained, including:

[0022] ;

[0023] Where, is the current energy control force obtained based on the online learning control model of energy control force, and are different control parameters to be adjusted in energy control, and are different control parameters in energy control, is the energy parameter, For time, is the energy error parameter, is the intermediate variable of integration, Learning bias for energy control.

[0024] In one embodiment, an online learning control model of the orthogonal control force is established, and the current orthogonal control force is obtained, including:

[0025] ;

[0026] Where, is the current orthogonal control force obtained based on the online learning control model of the orthogonal control force, and are different control parameters to be adjusted in orthogonal control, and are different control parameters in orthogonal control, are orthogonal parameters, For time, is the orthogonality error parameter, is the intermediate variable of integration, Learn the bias for the orthogonal control force.

[0027] In one embodiment, a stiffness instantaneous control calculation is performed based on the stiffness error to obtain a stiffness instantaneous control value, including:

[0028] Obtain the stiffness correction scenario of the full-angle MEMS gyroscope and establish the dynamic equation of gyroscope stiffness correction;

[0029] According to the dynamic equation of gyro stiffness correction, the stiffness instantaneous controller is designed;

[0030] According to the stiffness error and the stiffness instantaneous controller, the stiffness instantaneous control calculation is performed to obtain the stiffness instantaneous control quantity.

[0031] In one embodiment, a dynamic equation for gyro stiffness correction is established, including:

[0032] ;

[0033] Where, for The derivative of the equivalent frequency under the mode, is the resonant frequency The delay coefficient, is the equivalent frequency, is the equivalent interference, is a constant, To regulate voltage.

[0034] In one embodiment, the instantaneous stiffness control variable is used as input, and an online learning mechanism is used to compensate the stiffness error correction control loop, establish an online learning control model for stiffness error correction, and obtain the current stiffness tuning, including:

[0035] Taking the instantaneous control variable of stiffness as input, the stiffness control law, stiffness instantaneous controller and stiffness error correction learning deviation of the gyroscope are designed; based on the stiffness control law, stiffness instantaneous controller and stiffness error correction learning deviation of the gyroscope, an online learning control model of stiffness error correction is established, and the current stiffness tuning is obtained.

[0036] In one embodiment, establishing an online learning control model for stiffness error correction and obtaining the current stiffness tuning includes:

[0037] ;

[0038] Where, is the current stiffness tuning obtained from the online learning control model for stiffness error correction, and are different control parameters to be adjusted in stiffness error correction, and are different control parameters in stiffness error correction, is the equivalent frequency, For time, is the equivalent frequency error parameter, is the intermediate variable of integration, Correct the learning bias for stiffness errors.

[0039] The above-mentioned control method for a full-angle MEMS gyroscope based on online learning control adopts an online learning control mechanism to perform online learning compensation on the energy control force control loop and the orthogonal control force control loop, continuously learning and optimizing instantaneous control to optimize the control force control loop. This method can more accurately identify errors in the control force loop, improve the accuracy of the error identification results, achieve high-precision closed-loop control of the control force control loop, and provide more accurate data support for the self-correction of the full-angle MEMS gyroscope. After identifying high-precision error information, the online learning control mechanism is used to perform online learning compensation on the stiffness error correction control loop to optimize the stiffness error correction control loop. By improving the controller output accuracy, the self-correction effect of the stiffness error term is directly affected, thereby improving the accuracy of the stiffness error self-correction. This can achieve high-precision control of the stiffness error correction control loop, further improving the self-correction effect of the full-angle MEMS gyroscope, and improving the overall performance of the full-angle MEMS gyroscope.

[0040] In summary, by designing an online learning control mechanism for full-angle gyros, this application not only improves the control accuracy of the control force control loop and the stiffness error correction control loop, that is, improves the control accuracy of multiple control loops, achieving high-precision control of full-angle MEMS gyros, but also optimizes the high-performance self-correction effect of full-angle MEMS gyros. These improvements work together to give this application significant technical advantages. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 1 is a flow chart of a control method of a full-angle micro-electromechanical gyroscope based on online learning control in one embodiment;

[0042] Figure 2 FIG. 4 is a schematic diagram of the architecture of a control method for a full-angle MEMS gyroscope based on online learning control in one embodiment. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the present application is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely for explaining this application and are not intended to limit this application. All other embodiments obtained by persons of ordinary skill in the art based on the embodiments in this application without creative work are within the scope of protection of this application.

[0044] In addition, the terms "first," "second," and so on, used in this application are for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, features specified as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of this application, "multiple groups" means at least two groups, such as two groups, three groups, and so on, unless otherwise specifically defined.

[0045] In this application, unless otherwise specified or limited, the terms "connect," "fix," etc. should be understood in a broad sense. For example, "fix" can mean a fixed connection, a detachable connection, or an integral connection; it can mean a mechanical connection, an electrical connection, a physical connection, or a wireless communication connection; it can mean a direct connection or an indirect connection through an intermediate medium; it can mean internal communication between two elements or an interaction between two elements, unless otherwise specified. Those skilled in the art will understand the specific meanings of the above terms in this application based on the specific circumstances.

[0046] In addition, the technical solutions between the various embodiments of the present application can be combined with each other, but it must be based on the fact that ordinary technicians in this field can implement it. When the combination of technical solutions is mutually contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0047] This application provides a control method for a full-angle micro-electromechanical gyroscope based on online learning control, such as Figure 1 and Figure 2 As shown, in one embodiment, it includes:

[0048] Step 101, obtain the vibration signal of the full-angle micro-electromechanical gyroscope, perform measurement and control state solution, and obtain the initial control quantity; perform instantaneous control calculation of the control force based on the initial control quantity to obtain the instantaneous control quantity of the control force; and obtain the real-time control force based on the instantaneous control quantity of the control force.

[0049] Specifically:

[0050] Obtain the vibration signal of the full-angle MEMS gyroscope, solve the measurement and control state based on the dynamic equation, and obtain the initial control quantity;

[0051] According to the initial control quantity, the instantaneous control calculation of the control force is performed to obtain the instantaneous control quantity of the control force;

[0052] Taking the instantaneous control quantity of the control force as input, the online learning mechanism is adopted to compensate the control force control loop, establish an online learning control model of the control force, and obtain the current control force; taking the instantaneous control quantity of the control force and the current control force as input, the online learning mechanism is adopted to perform cyclic calculations to obtain the real-time control force.

[0053] More specifically:

[0054] Obtain the vibration signal of the full-angle MEMS gyroscope, solve the measurement and control state of the measurement and control system based on the energy dynamics equation, and obtain the initial energy control value;

[0055] According to the initial energy control amount, the energy control force instantaneous control calculation is performed to obtain the energy control force instantaneous control amount;

[0056] Taking the instantaneous control amount of the energy control force as input, the online learning mechanism is used to compensate the energy control force control loop, establish an online learning control model of the energy control force, and obtain the current energy control force; taking the instantaneous control amount of the energy control force and the current energy control force as input, the online learning mechanism is used to perform cyclic calculations to obtain the real-time energy control force;

[0057] Obtain the vibration signal of the full-angle MEMS gyroscope, solve the measurement and control state of the measurement and control system based on the orthogonal dynamic equation, and obtain the orthogonal initial control quantity;

[0058] According to the orthogonal initial control quantity, the orthogonal control force instantaneous control calculation is performed to obtain the orthogonal control force instantaneous control quantity;

[0059] Taking the instantaneous control quantity of the orthogonal control force as input, an online learning mechanism is adopted to compensate the orthogonal control force control loop, establish an online learning control model of the orthogonal control force, and obtain the current orthogonal control force; taking the instantaneous control quantity of the orthogonal control force and the current orthogonal control force as input, an online learning mechanism is adopted to perform cyclic calculations to obtain the real-time orthogonal control force.

[0060] More specifically:

[0061] Obtain the vibration signal of the full-angle MEMS gyroscope. Under normal operating conditions, the elliptical trajectory will approximately degenerate into a straight line, which satisfies:

[0062] (1-1)

[0063] According to the vibration signal of the full-angle MEMS gyroscope, the dynamic equation of energy is initially expressed as:

[0064] (1-2)

[0065] In the actual operation of the gyroscope, the time-varying factors of the parameters are taken into consideration and the dynamic equation of energy is re-described to obtain the final dynamic equation of energy:

[0066] (1-3)

[0067] in,

[0068] (1-4)

[0069] Where, is the major semi-axis of the elliptical trajectory, is the minor semi-axis of the elliptical trajectory; for energy; is an orthogonal quantity; for The first derivative of ; is the decay time constant; is the vibration mode angle; is the damping axis deflection angle; is the resonant frequency; is the energy control law; To calculate the equivalent interference caused by factors such as environmental changes when calculating energy, it is usually bounded, that is, ; The gyro decay time is usually designed to be greater than 1s, so ; In full-width operation mode ,therefore ;

[0070] Based on the energy dynamics model, the measurement and control state of the measurement and control system is solved to obtain the initial energy control quantity;

[0071] According to the initial energy control amount, the energy control force instantaneous control calculation is performed to obtain the energy control force instantaneous control amount;

[0072] Taking the instantaneous control value of the energy control force as input, the energy control law of the gyroscope is designed:

[0073] (1-5)

[0074] Where, Time intervals for learning; is the learning time interval obtained by the online learning control model based on the energy control force Energy control before;

[0075] Design an energy instantaneous controller:

[0076] (1-6)

[0077] in,

[0078] (1-7)

[0079] Where, is the output of the energy instantaneous controller;

[0080] Design energy control force learning deviation:

[0081] (1-8)

[0082] According to the gyroscope's energy control law, energy instantaneous controller, and energy control force learning deviation, the initial expression of the online learning control model of energy control force is obtained as follows:

[0083] (1-9)

[0084] in,

[0085] (1-10)

[0086] Combining the above equations, we can obtain the simplified expression of the online learning control model of energy control force:

[0087] (1-11)

[0088] Here we consider the learning deviation of the energy control force in the gyro measurement and control system Add a clipping procedure so that the energy control force learning deviation is bounded, that is ;

[0089] Combining the above formula, an online learning control model of energy control force is established, and the current energy control force is obtained:

[0090] (1-12)

[0091] Where, is the current energy control force obtained based on the online learning control model of energy control force, and are different control parameters to be adjusted in energy control, and are different control parameters in energy control, is the energy parameter, For time, is the energy error parameter, is the intermediate variable of integration, Learning bias for energy control;

[0092] Taking the instantaneous control amount of energy control force and the current energy control force as input, an online learning mechanism is used to perform cyclic calculations to obtain the real-time energy control force.

[0093] Obtain the vibration signal of the full-angle MEMS gyroscope and assume that the designed energy control loop can converge stably, that is, satisfy:

[0094] (2-1)

[0095] According to the vibration signal of the full-angle MEMS gyroscope, the orthogonal dynamic equation is initially expressed as:

[0096] (2-2)

[0097] In the actual operation of the gyroscope, the orthogonal dynamic equation is re-described by considering the time-varying factors of the parameters to obtain the final orthogonal dynamic equation:

[0098] (2-3)

[0099] in,

[0100] (2-4)

[0101] Where, For energy, is the initial energy; for The first derivative of ; is the stiffness axis deflection angle; is the orthogonal control law; To calculate the equivalent interference caused by factors such as environmental changes when calculating orthogonal quantities, it is usually bounded, that is, ; The gyro decay time is a positive real number, so ; In full-width operation mode a >0, so >0; The factors affecting derive from frequency decomposition, stiffness axis error, and energy control loop. Under the premise of normal operation of the gyro, this quantity is usually bounded, that is, ;

[0102] Based on the orthogonal dynamic equations, the measurement and control state of the measurement and control system is solved to obtain the orthogonal initial control quantity;

[0103] According to the orthogonal initial control quantity, the orthogonal control force instantaneous control calculation is performed to obtain the orthogonal control force instantaneous control quantity;

[0104] Taking the instantaneous control variable of the orthogonal control force as input, the orthogonal control law of the gyroscope is designed:

[0105] (2-5)

[0106] Where, is the learning time interval obtained by the online learning control model based on the orthogonal control force The orthogonal control force before;

[0107] Design a quadrature transient controller:

[0108] (2-6)

[0109] in,

[0110] (2-7)

[0111] Where, is the output of the orthogonal transient controller;

[0112] Design orthogonal control force learning bias:

[0113] (2-8)

[0114] According to the orthogonal control law of the gyroscope, the orthogonal instantaneous controller and the orthogonal control force learning deviation, the initial expression of the online learning control model of the orthogonal control force is obtained as follows:

[0115] (2-9)

[0116] in,

[0117] (2-10)

[0118] Combining the above equations, we can obtain the simplified expression of the online learning control model of the orthogonal control force:

[0119] (2-11)

[0120] Here we consider the learning deviation of the orthogonal control force in the gyro measurement and control system Add a clipping procedure, so the orthogonal control force learns the deviation is bounded, that is ;

[0121] Combining the above formula, an online learning control model of orthogonal control force is established, and the current orthogonal control force is obtained:

[0122] (2-12)

[0123] Where, is the current orthogonal control force obtained based on the online learning control model of the orthogonal control force, and are different control parameters to be adjusted in orthogonal control, and are different control parameters in orthogonal control, are orthogonal parameters, For time, is the orthogonality error parameter, is the intermediate variable of integration, Learning bias for orthogonal control forces;

[0124] Taking the instantaneous control amount of the orthogonal control force and the current orthogonal control force as input, an online learning mechanism is adopted to perform cyclic calculations to obtain the real-time orthogonal control force.

[0125] In this step, since energy is coupled in the orthogonal control force control loop, the online learning control mechanism of the energy control force control loop is first studied. After the energy closed loop converges, it is regarded as a bounded quantity, thereby decoupling the orthogonal control force control loop, and then the online learning control mechanism of the orthogonal control force control loop is studied.

[0126] The online learning control mechanism compensates the energy control force control loop by continuously learning the output of the instantaneous controller of the energy control force control loop, and uses the instantaneous controller output before the learning interval as the learning item. By reasonably adjusting the control parameters, it can provide more accurate input for the measurement and control system identification, thereby obtaining more accurate error identification results and achieving high-precision control.

[0127] The online learning control mechanism compensates the orthogonal control force control loop by continuously learning the output of the instantaneous controller of the orthogonal control force control loop, and uses the instantaneous controller output before the learning interval as the learning item. By reasonably adjusting the control parameters, it can provide more accurate input for the measurement and control system identification, thereby obtaining more accurate error identification results and achieving high-precision control.

[0128] It should be noted that how to solve the measurement and control state of the measurement and control system and perform instantaneous control calculation of the energy control force are all existing technologies and will not be elaborated here.

[0129] It should also be noted that the control loop includes: a control force control loop (including: an energy control force control loop and an orthogonal control force control loop) and a stiffness error correction control loop; among them, the control force control loop refers to a loop that solves and controls the gyroscope's measurement and control system to obtain the control force, and the stiffness error correction control loop refers to a loop that identifies and controls the gyroscope's measurement and control system to obtain stiffness tuning.

[0130] Step 102 : performing measurement and control system identification based on the real-time control force to obtain the stiffness error; performing stiffness instantaneous control calculation based on the stiffness error to obtain the stiffness instantaneous control value.

[0131] Specifically:

[0132] According to the real-time control force, the measurement and control system is identified to obtain the stiffness error;

[0133] Obtain the stiffness correction scenario of the full-angle MEMS gyroscope and establish the dynamic equation of gyroscope stiffness correction;

[0134] According to the dynamic equation of gyro stiffness correction, a stiffness instantaneous controller is designed; according to the stiffness error and the stiffness instantaneous controller, stiffness instantaneous control calculation is performed to obtain the stiffness instantaneous control quantity.

[0135] More specifically:

[0136] According to the real-time control force, the measurement and control system is identified to obtain the stiffness error;

[0137] Obtain and analyze the stiffness correction scenario of the full-angle micro-electromechanical gyroscope to obtain the initial second-order differential equation and the electrostatic excitation physical model; obtain the electrostatic tuning equation based on the electrostatic excitation physical model; obtain the final second-order differential equation based on the initial second-order differential equation and the electrostatic tuning equation; obtain the equivalent frequency of the gyroscope after correction based on the initial second-order differential equation, the electrostatic excitation physical model, the electrostatic tuning equation, and the final second-order differential equation; establish the dynamic equation for gyroscope stiffness correction based on the equivalent frequency of the gyroscope after correction, taking into account parameter time-varying and disturbance factors;

[0138] According to the dynamic equation of gyro stiffness correction, a stiffness instantaneous controller is designed; according to the stiffness error and the stiffness instantaneous controller, stiffness instantaneous control calculation is performed to obtain the stiffness instantaneous control quantity.

[0139] More specifically:

[0140] According to the real-time control force, the measurement and control system is identified to obtain the stiffness error (including: stiffness cracking and stiffness axis deflection );

[0141] Obtain and analyze the stiffness correction scenario of the full-angle MEMS gyroscope, generally using a DC excitation bias voltage Acting on the resonant structure, with AC voltage Acting on the excitation electrode to form an electrostatic excitation force Under the action of electrostatic force, the movement of the plates causes the capacitor spacing to change, and the changing spacing represents the displacement of the driving mode , the initial capacitor spacing is the initial capacitor channel width ;

[0142] Consider a gyro resonator with X mode (subscript x All represent the corresponding parameters under X mode, for example: Represents the damping of the X mode, which will not be described in detail below) as an example, the initial second-order differential equation is described as:

[0143] (3-1)

[0144] Where, is the equivalent mass of the gyroscope, for The second derivative of for The first derivative of is the displacement of the driving mode, is the damping (under X mode), is the mechanical stiffness, is the electrostatic excitation force;

[0145] The electrostatic excitation physical model is obtained as follows:

[0146] (3-2)

[0147] in,

[0148] (3-3)

[0149] Where, is the DC excitation bias voltage, is the amplitude of the AC excitation voltage, is the capacitance value of the excitation electrode channel, is the dielectric constant of the capacitor, S is the area of ​​the capacitor electrode, is the initial capacitance channel width;

[0150] Defining static capacitance , under normal working conditions, the displacement and initial channel width satisfy , at this time the capacitor is considered to be in the linear region, right x The rate of change model can be approximated as:

[0151] (3-4)

[0152] According to the electrostatic excitation physical model, substituting (3-4) into (3-2) yields the initial expression of the electrostatic tuning equation:

[0153] (3-5)

[0154] because Usually than 2 orders of magnitude smaller, x Usually than It is about one order of magnitude smaller, so the related terms are neglected and simplified to obtain the electrostatic tuning equation:

[0155] (3-6)

[0156] in,

[0157] (3-7)

[0158] According to the initial second-order differential equation and the electrostatic tuning equation, substitute (3-6) into (3-1) to obtain the final second-order differential equation:

[0159] (3-8)

[0160] Where, is the electrostatically tuned stiffness;

[0161] It can be seen from the above formula that Changing the equivalent stiffness of the gyro, which is composed of mechanical stiffness and electrostatically tuned stiffness Two parts, therefore, by adjusting the DC signal The purpose of resonant frequency tuning can be achieved by the size of is a positive constant, as As the spring increases, the equivalent stiffness decreases, so this effect is vividly called "spring softening";

[0162] After electrostatic tuning, taking the X-mode frequency as an example, according to the initial second-order differential equation, the electrostatic excitation physical model, the electrostatic tuning equation, the final second-order differential equation, and the simultaneous equations (3-1) to (3-8), the initial expression of the equivalent frequency after gyro correction is obtained:

[0163] (3-9)

[0164] Among them, in the actual gyro measurement and control system, the DC signal is usually composed of a large basic voltage and a smaller regulated voltage The former is a positive constant and the latter is a time variable adjusted by the controller, which satisfies:

[0165] (3-10)

[0166] And satisfied,

[0167] (3-11)

[0168] Where, is the X-mode equivalent stiffness;

[0169] Then the equivalent frequency can be rewritten as:

[0170] (3-12)

[0171] Will right Taking the derivative, we get:

[0172] (3-13)

[0173] Simplifying the above formula, we can get the equivalent frequency after gyro correction:

[0174] (3-14)

[0175] Where, is the equivalent frequency (time variation not considered), To regulate the voltage, is the dielectric constant of the capacitor, S is the area of ​​the capacitor electrode, is the base voltage, is the equivalent mass of the gyroscope, is the initial capacitance channel width, is the mechanical stiffness;

[0176] at this time, is a constant, so adjust the voltage and equivalent frequency into an approximately linear relationship;

[0177] According to the equivalent frequency after gyro correction, considering the time-varying parameters and disturbance factors, the dynamic equation of gyro stiffness correction is established:

[0178] (3-15)

[0179] Where, for X The derivative of the equivalent frequency under the mode; is the resonant frequency The delay coefficient, >1; is the equivalent frequency; is the equivalent interference caused by factors such as environmental changes, which is usually bounded, that is, ; is a constant;

[0180] According to the dynamic equation of gyro stiffness correction, the stiffness instantaneous controller is designed:

[0181] (3-16)

[0182] in,

[0183] (3-17)

[0184] Where, is the output of the stiffness instantaneous controller;

[0185] According to the stiffness error and the stiffness instantaneous controller, the stiffness instantaneous control calculation is performed to obtain the stiffness instantaneous control quantity.

[0186] In this step, the coordinate transformation method is used to simplify the dynamic equation of gyro stiffness correction into a control problem that stabilizes to zero point, thereby simplifying the theoretical model derivation.

[0187] If the adjustment voltage is applied to the Y mode, the same conclusion can be obtained.

[0188] It should be noted that how to identify the measurement and control system and perform stiffness instantaneous control calculations are all existing technologies and will not be elaborated here.

[0189] Step 103, using the instantaneous control value of stiffness as input, adopting the online learning mechanism to compensate the stiffness error correction control loop, establishing the online learning control model of stiffness error correction, and obtaining the current stiffness tuning; using the instantaneous control value of stiffness and the current stiffness tuning as input, adopting the online learning mechanism, performing cyclic calculation, and obtaining real-time stiffness tuning.

[0190] Specifically:

[0191] Taking the instantaneous stiffness control variable as input, the stiffness control law, stiffness instantaneous controller, and stiffness error correction learning deviation of the gyro are designed. Based on the stiffness control law, stiffness instantaneous controller, and stiffness error correction learning deviation of the gyro, an online learning control model for stiffness error correction is established, and the current stiffness tuning is obtained.

[0192] Taking the instantaneous control value of stiffness and the current stiffness tuning as input, an online learning mechanism is adopted to perform cyclic calculations to obtain real-time stiffness tuning.

[0193] More specifically:

[0194] Taking the instantaneous control variable of stiffness as input, the stiffness control law of the gyro is designed:

[0195] (4-1)

[0196] Where, is the learning time interval obtained by the online learning control model based on stiffness error correction Front stiffness tuning;

[0197] Design the stiffness transient controller, see equations (3-16) and (3-17) for details;

[0198] Design stiffness error correction learning bias:

[0199] (4-2)

[0200] According to the gyro stiffness control law, stiffness instantaneous controller and stiffness error correction learning deviation, the initial expression of the online learning control model for stiffness error correction is:

[0201] (4-3)

[0202] in,

[0203] (4-4)

[0204] Combining the above equations, we can obtain the simplified expression of the online learning control model for stiffness error correction:

[0205] (4-5)

[0206] Here we consider the learning deviation for stiffness correction in the gyro measurement and control system Add a clipping procedure so that the stiffness corrects the learning bias is bounded, that is ;

[0207] Combining the above formula, an online learning control model for stiffness error correction is established, and the current stiffness tuning is obtained:

[0208] (4-6)

[0209] Where, is the current stiffness tuning obtained from the online learning control model for stiffness error correction, and are different control parameters to be adjusted in stiffness error correction, and are different control parameters in stiffness error correction, is the equivalent frequency, For time, is the equivalent frequency error parameter, is the intermediate variable of integration, Correcting learning bias for stiffness errors;

[0210] Taking the instantaneous control value of stiffness and the current stiffness tuning as input, an online learning mechanism is adopted to perform cyclic calculations to obtain real-time stiffness tuning.

[0211] In this step, the online learning algorithm continuously learns the output of the instantaneous controller of the stiffness error correction control loop, and uses the controller output before the learning interval as the learning item to compensate for the current stiffness correction control. By reasonably adjusting the control parameters, the effect of high-precision stiffness error self-correction can be achieved.

[0212] If the self-correction tuning voltage is applied to the Y mode, or the self-correction tuning voltage is applied to the stiffness axis alignment tuning electrode, the design process is similar and will not be repeated here.

[0213] Step 104 : Apply the real-time control force and the real-time stiffness tuning to the full-angle MEMS gyroscope to update the vibration signal and realize the control of the full-angle MEMS gyroscope.

[0214] Specifically:

[0215] The real-time control force and the real-time stiffness tuning are applied to the full-angle micro-electro-mechanical gyroscope to update the vibration signal in real time, thereby realizing the real-time control of the full-angle micro-electro-mechanical gyroscope.

[0216] In this step, how the control force and stiffness tuning work together and how the vibration signal is updated are all existing technologies and will not be described in detail here.

[0217] In this embodiment, the cyclic calculation refers to calculation every second, and the cycle does not stop as long as the gyroscope is working.

[0218] The above-mentioned control method for a full-angle MEMS gyroscope based on online learning control adopts an online learning control mechanism to perform online learning compensation on the energy control force control loop and the orthogonal control force control loop, continuously learning and optimizing instantaneous control to optimize the control force control loop. This method can more accurately identify errors in the control force loop, improve the accuracy of the error identification results, achieve high-precision closed-loop control of the control force control loop, and provide more accurate data support for the self-correction of the full-angle MEMS gyroscope. After identifying high-precision error information, the online learning control mechanism is used to perform online learning compensation on the stiffness error correction control loop to optimize the stiffness error correction control loop. By improving the controller output accuracy, the self-correction effect of the stiffness error term is directly affected, thereby improving the accuracy of the stiffness error self-correction. This can achieve high-precision control of the stiffness error correction control loop, further improving the self-correction effect of the full-angle MEMS gyroscope, and improving the overall performance of the full-angle MEMS gyroscope.

[0219] In summary, by designing an online learning control mechanism for full-angle gyros, this application not only improves the control accuracy of the control force control loop and the stiffness error correction control loop, that is, improves the control accuracy of multiple control loops, achieving high-precision control of full-angle MEMS gyros, but also optimizes the high-performance self-correction effect of full-angle MEMS gyros. These improvements work together to give this application significant technical advantages.

[0220] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. In addition, Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0221] The contents not described in detail in this specification belong to the prior art known to professional and technical personnel in this field.

[0222] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0223] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present application, and these modifications and improvements fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be based on the attached application documents.

Claims

1. A control method for a full-angle micro-electromechanical gyroscope based on online learning control, characterized in that: include: Obtain the vibration signal of the full-angle MEMS gyroscope, perform measurement and control state calculation, and obtain the initial control quantity; According to the initial control quantity, the instantaneous control calculation of the control force is performed to obtain the instantaneous control quantity of the control force; According to the instantaneous control amount of the control force, the real-time control force is obtained; According to the real-time control force, the measurement and control system is identified to obtain the stiffness error; according to the stiffness error, the stiffness instantaneous control calculation is performed to obtain the stiffness instantaneous control value; Taking the instantaneous control variable of stiffness as input, an online learning mechanism is adopted to compensate the stiffness error correction control loop, establish an online learning control model for stiffness error correction, and obtain the current stiffness tuning; Taking the instantaneous control value of stiffness and the current stiffness tuning as input, an online learning mechanism is used to perform cyclic calculations to obtain real-time stiffness tuning. Applying real-time control force and real-time stiffness tuning to the full-angle MEMS gyroscope to update the vibration signal and realize the control of the full-angle MEMS gyroscope; According to the instantaneous control amount of the control force, the real-time control force is obtained, including: Taking the instantaneous control amount of the control force as input, the online learning mechanism is used to compensate the control force control loop, establish an online learning control model of the control force, and obtain the current control force; taking the instantaneous control amount of the control force and the current control force as input, the online learning mechanism is used to perform cyclic calculations to obtain the real-time control force; Establish an online learning control model for energy control force and obtain the current energy control force, including: Where, is the current energy control force obtained based on the online learning control model of energy control force, and are different control parameters to be adjusted in energy control, and are different control parameters in energy control, is the energy parameter, For time, is the energy error parameter, is the intermediate variable of integration, Learning bias for energy control; Establish an online learning control model for orthogonal control force and obtain the current orthogonal control force, including: Where, is the current orthogonal control force obtained based on the online learning control model of the orthogonal control force, and are different control parameters to be adjusted in orthogonal control, and are different control parameters in orthogonal control, are orthogonal parameters, For time, is the orthogonality error parameter, is the intermediate variable of integration, Learning bias for orthogonal control forces; An online learning control model for stiffness error correction is established, and the current stiffness tuning is obtained, including: Where, is the current stiffness tuning obtained from the online learning control model for stiffness error correction, and are different control parameters to be adjusted in stiffness error correction, and are different control parameters in stiffness error correction, is the equivalent frequency, For time, is the equivalent frequency error parameter, is the intermediate variable of integration, Correct the learning bias for stiffness errors.

2. The control method of the full-angle MEMS gyroscope based on online learning control according to claim 1, characterized in that: Taking the instantaneous control quantity of the control force as input, the online learning mechanism is adopted to compensate the control force control loop, establish the online learning control model of the control force, and obtain the current control force; Taking the instantaneous control amount of the control force and the current control force as input, an online learning mechanism is used to perform cyclic calculations to obtain the real-time control force, including: Taking the instantaneous control amount of the energy control force as input, the online learning mechanism is used to compensate the energy control force control loop, establish an online learning control model of the energy control force, and obtain the current energy control force; taking the instantaneous control amount of the energy control force and the current energy control force as input, the online learning mechanism is used to perform cyclic calculations to obtain the real-time energy control force; Taking the instantaneous control quantity of the orthogonal control force as input, an online learning mechanism is adopted to compensate the orthogonal control force control loop, establish an online learning control model of the orthogonal control force, and obtain the current orthogonal control force; taking the instantaneous control quantity of the orthogonal control force and the current orthogonal control force as input, an online learning mechanism is adopted to perform cyclic calculations to obtain the real-time orthogonal control force.

3. The control method of the full-angle MEMS gyroscope based on online learning control according to claim 2, characterized in that: Taking the instantaneous control amount of the energy control force as input, the online learning mechanism is used to compensate the energy control force control loop, establish an online learning control model of the energy control force, and obtain the current energy control force, including: Taking the instantaneous control value of the energy control force as input, the energy control law, instantaneous energy controller, and energy control force learning deviation of the gyroscope are designed. Based on the energy control law, instantaneous energy controller, and energy control force learning deviation of the gyroscope, an online learning control model of the energy control force is established to obtain the current energy control force. Taking the instantaneous control quantity of the orthogonal control force as input, an online learning mechanism is used to compensate the orthogonal control force control loop, establish an online learning control model of the orthogonal control force, and obtain the current orthogonal control force, including: Taking the instantaneous control quantity of the orthogonal control force as input, the orthogonal control law, orthogonal instantaneous controller and orthogonal control force learning deviation of the gyroscope are designed; based on the orthogonal control law, orthogonal instantaneous controller and orthogonal control force learning deviation of the gyroscope, an online learning control model of the orthogonal control force is established, and the current orthogonal control force is obtained.

4. The control method of a full-angle micro-electromechanical gyroscope based on online learning control according to any one of claims 1 to 3, characterized in that: According to the stiffness error, the stiffness instantaneous control calculation is performed to obtain the stiffness instantaneous control value, including: Obtain the stiffness correction scenario of the full-angle MEMS gyroscope and establish the dynamic equation of gyroscope stiffness correction; According to the dynamic equation of gyro stiffness correction, the stiffness instantaneous controller is designed; According to the stiffness error and the stiffness instantaneous controller, the stiffness instantaneous control calculation is performed to obtain the stiffness instantaneous control quantity.

5. The control method of the full-angle MEMS gyroscope based on online learning control according to claim 4, characterized in that: The dynamic equations for gyro stiffness correction are established, including: Where, for X The derivative of the equivalent frequency under the mode, is the resonant frequency The delay coefficient, is the equivalent frequency, is the equivalent interference, is a constant, To regulate voltage.

6. The control method of a full-angle MEMS gyroscope based on online learning control according to any one of claims 1 to 3, characterized in that: Taking the instantaneous stiffness control variable as input, an online learning mechanism is used to compensate the stiffness error correction control loop, establish an online learning control model for stiffness error correction, and obtain the current stiffness tuning, including: Taking the instantaneous control variable of stiffness as input, the stiffness control law, stiffness instantaneous controller and stiffness error correction learning deviation of the gyroscope are designed; based on the stiffness control law, stiffness instantaneous controller and stiffness error correction learning deviation of the gyroscope, an online learning control model of stiffness error correction is established, and the current stiffness tuning is obtained.

Images