A Multi-Axis Cooperative Motion Synchronization Control Method Based on Grating Encoder

By converting the position feedback of the multi-axis motion control system from the time domain to the phase domain for unified alignment, the position error problem caused by transmission path and clock jitter is solved, realizing high-precision multi-axis cooperative motion synchronization control and improving the synchronization accuracy and reliability of the system.

CN122902290BActive Publication Date: 2026-11-17HOPU TECH (NINGBO) CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202611400108.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2026-09-10
Publication Date
2026-11-17
Estimated Expiration
2046-09-10

AI Technical Summary

Technical Problem

In multi-axis motion control systems, the different signal transmission path lengths of each axis encoder and the inconsistency in position feedback data caused by controller sampling clock jitter lead to sub-micron level position synchronization errors in high-speed and high-precision coordinated motion, affecting the machining contour accuracy.

Method used

By converting the position feedback of each axis from the time domain to the phase domain for unified alignment, the instantaneous phase angle is calculated using orthogonal sinusoidal signals, and first-order phase linear interpolation correction is performed using virtual phase grid points to generate phase domain alignment position values. Synchronous control is achieved through a ring data buffer, a dynamic coupling model between multiple input and multiple output axes, and geometric constraint equations.

Benefits of technology

It eliminates the problem of inconsistent sampling times in multi-axis position feedback, realizes high-precision multi-axis cooperative motion synchronization control, and improves the cooperative accuracy and operational reliability of the system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122902290B_ABST
    Figure CN122902290B_ABST
Patent Text Reader

Abstract

This invention discloses a multi-axis cooperative motion synchronization control method based on a grating encoder. The method calculates the instantaneous phase angle of each axis based on orthogonal sinusoidal signals. Using virtual phase grid points as a reference, it performs first-order phase linear interpolation correction on the position count values ​​based on the instantaneous velocity estimate, generating phase-domain aligned position values. These values ​​are stored in a circular buffer, and the multi-axis position sequences are extracted. Velocity and acceleration sequences are obtained through differential operations. Then, based on the inter-axis dynamic coupling model, the recursive least squares method is used to identify the inter-axis dynamic coupling coefficients online and superimpose compensation onto the control commands of the corresponding motion axes. The constraint residuals are calculated using geometric constraint equations. When a threshold is exceeded, an abnormal axis is identified and weighted suppression is performed. The position of the abnormal axis is estimated based on the feedback from the normal axis and the constraint equations as temporary feedback. This invention eliminates the problem of inconsistent sampling times in multi-axis position feedback and achieves high-precision cooperative control.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of multi-axis motion control technology, and specifically to a multi-axis cooperative motion synchronization control method based on an optical encoder. Background Technology

[0002] Multi-axis cooperative motion control systems are the core of high-end equipment such as CNC machine tools and industrial robots for achieving high-precision contour machining and trajectory tracking. They typically rely on optical encoders for position feedback on each motion axis, and the synchronization and accuracy of the feedback data from each axis directly determine the overall cooperative accuracy of the system. In practical applications, multi-axis motion control systems generally use a unified clock in the controller to periodically sample the optical encoders of each axis; that is, the position feedback data of each axis is read using the controller's sampling clock as the time reference. However, due to the different physical path lengths of the encoder signals from the sensor to the controller, the varying interface circuit delays, and the clock jitter in the controller's sampling trigger for each axis, the position feedback data of each axis actually correspond to different physical times. In high-speed, high-precision cooperative motion, even microsecond-level time deviations can translate into sub-micron-level position synchronization errors. This error is decomposed along the tangential and normal components of the trajectory during contour interpolation, directly degrading the machining contour accuracy. Summary of the Invention

[0003] In view of the above problems, the present invention provides a multi-axis cooperative motion synchronization control method based on grating encoder, which realizes the synchronous acquisition of multi-axis position data under a unified phase reference by converting the position feedback of each axis from time domain sampling to phase domain unified alignment.

[0004] To achieve the above objectives, this application provides a multi-axis cooperative motion synchronization control method based on a grating encoder, comprising:

[0005] Obtain the quadrature sine signals and position count values ​​output by the grating encoders corresponding to each motion axis;

[0006] The instantaneous phase angle of each motion axis is calculated based on the orthogonal sinusoidal signal. Using the virtual phase grid point as a unified alignment reference, and based on the phase deviation between the instantaneous phase angle and the virtual phase grid point, combined with the instantaneous velocity estimate of the motion axis, the position count value is corrected by first-order phase linear interpolation to generate the phase domain alignment position value of each motion axis.

[0007] The phase domain aligned position values ​​are stored in a circular data buffer. The multi-axis position sequence within a preset time window is extracted from the circular data buffer, and the difference operation is performed on the multi-axis position sequence to generate the velocity sequence and acceleration sequence of each motion axis.

[0008] Based on the multi-input multi-output axis dynamic coupling model, the acceleration sequence is used as the model output, and the velocity sequence and the control commands of each motion axis are used as the model input. The recursive least squares method is used to identify the dynamic coupling coefficient between axes online.

[0009] The coupling compensation amount is calculated based on the dynamic coupling coefficient and the real-time motion state of each motion axis, and the coupling compensation amount is superimposed on the control command of the corresponding motion axis.

[0010] Based on the geometric topology of the multi-axis mechanical structure, establish the geometric constraint equations between the feedback positions of each motion axis, and substitute the phase domain alignment position values ​​into the geometric constraint equations to calculate the constraint residuals.

[0011] When the constraint residual continuously exceeds the preset anomaly detection threshold, it is determined that there is an abnormal encoder feedback axis, and weighted suppression processing is performed on the feedback data of the abnormal axis.

[0012] The estimated position value of the abnormal axis is estimated based on the feedback data of other normal axes and the geometric constraint equations. The estimated position value is then used as temporary feedback data for the abnormal axis in the coordinated control.

[0013] In some embodiments, the instantaneous phase angle of each motion axis is calculated based on orthogonal sinusoidal signals, and virtual phase grid points are used as a unified alignment reference. Based on the phase deviation between the instantaneous phase angle and the virtual phase grid points, combined with the instantaneous velocity estimate of the motion axis, a first-order phase linear interpolation correction is performed on the position count value to generate the phase domain alignment position value of each motion axis, including:

[0014] A coordinate rotation digital calculation algorithm is executed on the orthogonal sinusoidal signal to extract the phase angle of the orthogonal signal of each motion axis grating encoder at the current moment as the instantaneous phase angle;

[0015] Virtual phase grid points are pre-divided with equal phase intervals of the grating pitch, and the next virtual phase grid point that is closest to the instantaneous phase angle is selected as the unified alignment reference.

[0016] The difference between the instantaneous phase angle and the unified alignment reference is calculated as the phase deviation, and the instantaneous velocity estimate of the motion axis in the current control cycle is obtained.

[0017] Perform zero-crossing detection or frequency measurement algorithms on the quadrature sinusoidal signal to extract the quadrature signal angular frequency of each motion axis grating encoder at the current moment. Based on the ratio between the phase deviation and the quadrature signal angular frequency, calculate the time deviation required for the motion axis to reach the unified alignment reference from the current sampling moment.

[0018] Based on the product relationship between the time deviation and the instantaneous velocity estimate, the change in position of the motion axis within the time deviation is calculated as the position correction amount;

[0019] The position correction value is superimposed with the position count value to obtain the equivalent position value of the motion axis at the time corresponding to the unified alignment reference, which is used as the phase domain alignment position value.

[0020] When the motion axis is stationary or quasi-stationary, the instantaneous velocity estimate approaches zero, the position correction approaches zero, and the phase domain aligned position value approaches the position count value.

[0021] In some embodiments, a zero-crossing detection or frequency measurement algorithm is performed on the quadrature sinusoidal signal to extract the quadrature signal angular frequency of each motion axis grating encoder at the current moment. Based on the ratio between the phase deviation and the quadrature signal angular frequency, the time deviation required for the motion axis to reach the unified alignment reference from the current sampling moment is calculated, including:

[0022] A zero-crossing detection algorithm is applied to the quadrature sinusoidal signal to detect the rising zero-crossing point when the quadrature sinusoidal signal crosses from the negative half-cycle to the positive half-cycle, and the falling zero-crossing point when the signal crosses from the positive half-cycle to the negative half-cycle.

[0023] Record the current detected rising zero-crossing time and the previous rising zero-crossing time, calculate the time difference between two adjacent rising zero-crossing times, and use the time difference as the current signal period of the orthogonal sinusoidal signal;

[0024] Using the constant 2π as the numerator and the current signal period as the denominator, calculate the ratio between the numerator and the denominator, and use the ratio as the current angular frequency of the orthogonal sinusoidal signal;

[0025] The ratio between the dividend and the divisor is calculated by using the phase deviation as the dividend and the current signal angular frequency as the divisor.

[0026] The ratio is used as the phase difference time required for the motion axis to reach the unified alignment reference from the current sampling time. The phase difference time characterizes the time required for the grating encoder signal of the motion axis to change from the current phase position to the phase position corresponding to the virtual phase grid point at the current speed.

[0027] The phase difference corresponds to the time as the time deviation.

[0028] In some embodiments, the method further includes:

[0029] Real-time parameter extraction is performed on the orthogonal sinusoidal signal. The real-time parameters include signal amplitude, DC offset, orthogonal phase difference, and signal harmonic component content. The real-time parameters are compared with the preset nominal parameter range. When any parameter deviates from the nominal parameter range by more than a preset threshold, the subdivision compensation coefficient is updated based on the ellipse fitting model or compensation matrix model, and the updated subdivision compensation coefficient is applied to the calculation of the instantaneous phase angle.

[0030] In some embodiments, real-time parameter extraction is performed on the orthogonal sinusoidal signal, including:

[0031] Multiple sampling points are acquired within one grating signal period. An extreme value detection algorithm is executed on the sampling points to extract the maximum and minimum values ​​of the orthogonal sinusoidal signal. The signal amplitude is calculated based on the difference between the maximum and minimum values, and the DC offset of the signal is calculated based on the average value between the maximum and minimum values.

[0032] A zero-crossing detection algorithm is applied to the two orthogonal sinusoidal signals to extract the zero-crossing time of each signal. Based on the ratio of the time difference between the two zero-crossing times to the period of the grating signal, the orthogonal phase difference between the two orthogonal signals is calculated.

[0033] The Discrete Fourier Transform algorithm is applied to the orthogonal sinusoidal signal to extract the amplitude of the fundamental component and the amplitude of each harmonic component. Based on the ratio between the amplitude of each harmonic component and the amplitude of the fundamental component, the harmonic component content of the signal is calculated.

[0034] The real-time parameters are compared with the preset nominal parameter range, including:

[0035] Establish the first nominal parameter range corresponding to the signal amplitude, the second nominal parameter range corresponding to the signal DC offset, the third nominal parameter range corresponding to the quadrature phase difference, and the fourth nominal parameter range corresponding to the signal harmonic component content;

[0036] The signal amplitude is compared with the first nominal parameter range, the signal DC offset is compared with the second nominal parameter range, the quadrature phase difference is compared with the third nominal parameter range, and the signal harmonic component content is compared with the fourth nominal parameter range.

[0037] When any parameter of the signal amplitude, signal DC offset, quadrature phase difference, or signal harmonic component content deviates from the corresponding nominal parameter range, adaptive subdivision compensation is triggered.

[0038] In some embodiments, based on a multi-input multi-output (MIMO) axis dynamic coupling model, the acceleration sequence is used as the model output, and the velocity sequence and control commands for each motion axis are used as the model input. The recursive least squares method is employed to identify the axis dynamic coupling coefficients online, including:

[0039] Establish the deviation expression between the actual acceleration and the nominal acceleration of each motion axis, attribute the deviation to the superposition effect of the velocity state and control command of other motion axes through the coupling coefficient to be identified, and form the mathematical structure of the dynamic coupling model between multiple input and multiple output axes.

[0040] Define the error function of the recursive least squares method, take the acceleration sequence as the expected output of the model, combine the velocity sequence with the control commands of each motion axis into a regression vector, take the linear combination of the regression vector and the coupling coefficient to be identified as the model prediction output, and take the difference between the expected output of the model and the predicted output of the model as the error function.

[0041] The velocity sequence, acceleration sequence, and control commands for each motion axis are obtained from the circular data buffer. The velocity sequence and control commands for each motion axis are combined into a regression vector, and the acceleration sequence is used as the expected output of the model.

[0042] Based on the recursive update formula of the recursive least squares method, the estimated value of the coupling coefficient at the previous time is corrected using the regression vector and error function at the current time, and the estimated value of the coupling coefficient at the current time is used as the inter-axis dynamic coupling coefficient.

[0043] The recursive update formula includes a forgetting factor, which is used to control the weight of the influence of historical data on the current estimate, enabling the recursive least squares method to track the time-varying characteristics of the inter-axis dynamic coupling coefficient.

[0044] In some embodiments, an expression for the deviation between the actual acceleration and the nominal acceleration of each motion axis is established, which is represented by formula (1), as follows:

[0045] ;

[0046] In formula (1), For the first axis at time The actual acceleration, For the first axis at time The nominal acceleration generated by the control command of this axis For the first Axis to the first The velocity coupling coefficient to be identified on the shaft. For the first Axis to the first The command coupling coefficient of the axis, For the first axis at time speed, For the first axis at time Control commands;

[0047] The error function of the recursive least squares method is defined by formula (2), which is as follows:

[0048] ;

[0049] In formula (2), For a moment The error value, This is the transpose of the regression vector. For a moment The vector of estimated coupling coefficients;

[0050] The regression vector is represented by formula (3), which is as follows:

[0051] ;

[0052] In formula (3), This represents the total number of motion axes.

[0053] Based on the recursive update formula of the recursive least squares method, the estimated value of the coupling coefficient at the previous time is corrected using the regression vector and error function at the current time, and the estimated value of the coupling coefficient at the current time is used as the dynamic coupling coefficient between axes, which is expressed by formula (4), as follows:

[0054] ;

[0055] In formula (4), For a moment The gain matrix, , For a moment The covariance matrix, Forgetting factor, .

[0056] In some embodiments, geometric constraint equations are established between the feedback positions of each motion axis based on the geometric topology of the multi-axis mechanical structure, and the phase domain alignment position values ​​are substituted into the geometric constraint equations to calculate the constraint residuals, including:

[0057] Based on the geometric topology of the multi-axis mechanical structure, determine the type of geometric constraint that each motion axis should satisfy under the ideal no-deformation state. The geometric constraint types include constant distance constraint, constant angle constraint, or trajectory equation constraint.

[0058] Based on the type of geometric constraint, a mathematical expression is established between the theoretical positions of each motion axis as the geometric constraint equation. The geometric constraint equation contains fixed geometric parameters determined by the mechanical structure, which are expressed by formula (5). Formula (5) is as follows:

[0059] ;

[0060] In formula (5), For the first axis at time The theoretical position, For the first axis at time The theoretical position, The first is determined by the mechanical structure Axis and the first Theoretical spacing between axes;

[0061] Extract the phase domain alignment position value of each motion axis at the current moment from the phase domain alignment position value generated in each control cycle;

[0062] Substitute the phase domain alignment position value of each motion axis at the current moment into the geometric constraint equation, and calculate the deviation between the actual position of each motion axis and the theoretical constraint described by the geometric constraint equation as the constraint residual.

[0063] When the multi-axis mechanical structure is a gantry dual-drive platform, the geometric constraint type is a constant distance constraint. The geometric constraint equation is that the difference between the positions of the two axes is equal to the theoretical spacing determined by the crossbeam mechanical structure. The constraint residual is the absolute value of the deviation between the actual difference between the phase domain alignment position values ​​of the two axes and the theoretical spacing.

[0064] In some embodiments, when the constraint residual continuously exceeds a preset anomaly detection threshold, it is determined that there is an encoder feedback anomaly axis, and weighted suppression processing is performed on the feedback data of the anomaly axis, including:

[0065] Set an anomaly detection counter, which monitors the comparison result between the constraint residual and the preset anomaly detection threshold in each control cycle;

[0066] When the constraint residual exceeds the preset anomaly detection threshold, the anomaly detection counter increments; when the constraint residual does not exceed the preset anomaly detection threshold, the anomaly detection counter is reset to zero.

[0067] When the cumulative value of the anomaly detection counter reaches the preset continuous cycle number threshold, it is determined that there is an encoder feedback abnormal axis, and the axis identifier of the abnormal axis is recorded.

[0068] The feedback weight of the abnormal axis in the collaborative control closed loop is reduced from the standard weight to a preset suppression weight. The suppression weight is less than the standard weight, which is used to reduce the impact of the feedback data of the abnormal axis on the generation of collaborative control commands.

[0069] Meanwhile, based on the geometric constraint equation, the phase domain alignment position values ​​of other normal axes are substituted into the geometric constraint equation as known quantities, and the theoretical position values ​​of the abnormal axes are obtained by inverse solution as estimated position values.

[0070] The estimated position value is used as temporary feedback data for the abnormal axis, replacing the original phase domain aligned position value of the abnormal axis in the coordinated control calculation of subsequent control cycles.

[0071] In some embodiments, the estimated position value is used as temporary feedback data for the abnormal axis in the collaborative control, including:

[0072] The estimated position value is written into the feedback data register corresponding to the abnormal axis, replacing the phase domain aligned position value of the abnormal axis, and serving as the position feedback input of the abnormal axis in the cooperative control closed loop;

[0073] In the coordinated control operation of the control cycle, the central controller reads the estimated position value from the feedback data register and performs closed-loop regulation based on the estimated position value to keep the motion trajectory of the abnormal axis consistent with the theoretical constraints described by the geometric constraint equation.

[0074] When the constraint residual recovers to below the preset anomaly judgment threshold and remains stable for a preset number of recovery cycles, the abnormal axis is judged to have returned to normal. The feedback weight of the abnormal axis is restored to the standard weight, and the original phase domain alignment position value of the abnormal axis is switched back as the feedback input.

[0075] Unlike existing technologies, the above-mentioned technical solution provides a multi-axis cooperative motion synchronization control method based on a grating encoder. It calculates the instantaneous phase angle of each axis based on orthogonal sinusoidal signals, uses virtual phase grid points as a reference, and performs first-order phase linear interpolation correction on the position count values ​​in conjunction with instantaneous velocity estimates to generate phase-domain aligned position values. These values ​​are then stored in a circular buffer, and the multi-axis position sequences are extracted. After differential operations, velocity and acceleration sequences are obtained. Based on the inter-axis dynamic coupling model, a recursive least squares method is used to identify the inter-axis dynamic coupling coefficients online and superimpose compensation onto the control commands of the corresponding motion axes. Constraint residuals are calculated using geometric constraint equations. When a threshold is exceeded, an abnormal axis is identified and weighted suppression is performed. The position of the abnormal axis is estimated based on the feedback from the normal axis and the constraint equations as temporary feedback. This invention eliminates the problem of inconsistent sampling times in multi-axis position feedback and achieves high-precision cooperative control.

[0076] The above description of the invention is merely an overview of the technical solution of this application. In order to enable those skilled in the art to better understand the technical solution of this application and to implement it based on the description and drawings, and to make the above-mentioned objectives and other objectives, features and advantages of this application easier to understand, the following description is provided in conjunction with the specific embodiments and drawings of this application. Attached Figure Description

[0077] The accompanying drawings are only used to illustrate the principles, implementation methods, applications, features, and effects of specific embodiments of the present invention and other related contents, and should not be considered as limitations on this application.

[0078] In the accompanying drawings of the instruction manual:

[0079] Figure 1 This is a schematic diagram illustrating steps S101 to S108 of the method described in a specific implementation.

[0080] Figure 2 This is a schematic diagram illustrating steps S201 to S206 of the method described in a specific implementation.

[0081] Figure 3 The diagram illustrates steps S301 to S304 of the method described in the specific implementation. Detailed Implementation

[0082] To illustrate the possible application scenarios, technical principles, implementable specific solutions, and achievable objectives and effects of this application in detail, the following description, in conjunction with the listed specific embodiments and accompanying drawings, provides a detailed explanation. The embodiments described herein are merely illustrative of the technical solutions of this application and are therefore intended to limit the scope of protection of this application.

[0083] In this document, the term "embodiment" means that a specific feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The term "embodiment" appearing in various places throughout the specification does not necessarily refer to the same embodiment, nor does it specifically limit its independence or connection with other embodiments. In principle, in this application, as long as there are no technical contradictions or conflicts, the technical features mentioned in each embodiment can be combined in any way to form corresponding implementable technical solutions.

[0084] Unless otherwise defined, the technical terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the use of related terms herein is merely for the purpose of describing particular embodiments and is not intended to limit this application.

[0085] In the description of this application, the term "and / or" is used to describe the logical relationship between objects, indicating that three relationships can exist. For example, A and / or B means: A exists, B exists, and A and B exist simultaneously. Additionally, the character " / " in this document generally indicates that the preceding and following objects have an "or" logical relationship.

[0086] In this application, terms such as “first” and “second” are used only to distinguish one entity or operation from another, and do not necessarily require or imply any actual quantity, hierarchy or order relationship between these entities or operations.

[0087] Without further limitations, the use of terms such as “comprising,” “including,” “having,” or other similar open-ended expressions in this application is intended to cover non-exclusive inclusion, which does not exclude the presence of additional elements in a process, method, or product that includes the stated elements, such that a process, method, or product that includes a list of elements may include not only those defined elements but also other elements not expressly listed, or elements inherent to such a process, method, or product.

[0088] As understood in the Examination Guidelines, in this application, expressions such as "greater than," "less than," and "exceeding" are understood to exclude the stated number; expressions such as "above," "below," and "within" are understood to include the stated number. Furthermore, in the description of the embodiments in this application, "multiple" means two or more (including two), and similar expressions related to "multiple" are also understood in this way, such as "multiple groups" and "multiple times," unless otherwise explicitly specified.

[0089] In the description of the embodiments of this application, the space-related expressions used, such as "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "vertical," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," and "circumferential," indicate the orientation or positional relationship based on the orientation or positional relationship shown in the specific embodiments or drawings. They are only for the purpose of describing the specific embodiments of this application or for the reader's understanding, and do not indicate or imply that the device or component referred to must have a specific position, a specific orientation, or be constructed or operated in a specific orientation. Therefore, they should not be construed as limitations on the embodiments of this application.

[0090] The processor described in the embodiments of this application can be implemented by hardware, firmware, software, or a combination thereof. It can be a circuit, one or more of an application-specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field-programmable gate array (FPGA), a central processing unit (CPU), a controller, a microcontroller, or a microprocessor. It also includes other physical, biological, or chemical structures that can implement the same or equivalent functions as the processors listed above, such as biological neurons, quantum computing units, DNA computing units, etc., so that the processor can execute some or all of the steps in the computer program or method involved in the various embodiments of this application, or any combination of the steps mentioned therein.

[0091] The computer program involved in the embodiments can be stored in a computer device readable storage medium, which includes, but is not limited to, disks, magnetic tapes, magnetic cards, floppy disks, flash memory, optical disks, optical cards, read-only memory (ROM), random access memory (RAM), erasable programmable ROM (EPROM), and electrically erasable programmable ROM (EEPROM), etc., and also includes other biological, physical, or chemical structures that can achieve the same or equivalent functions as the storage media listed above, such as DNA, RNA, proteins, and other units with information storage capabilities. In specific embodiments, the storage medium involved can be one of the above-mentioned media types, or a combination of the above-mentioned media types. In different embodiments, the computer program involved in the embodiments can be centrally stored in a single medium, or distributed and stored in multiple media. The memory containing the computer device readable storage medium can be non-volatile memory or random access memory. These computer device readable storage media can be built into the device, or can be connected to the device involved in the embodiments as an external device or part of an external device. In some embodiments, the memory having a computer device readable storage medium is deployed locally; in other embodiments, the memory may be deployed remotely from the processor, for example, as a network-attached memory accessed via RF circuitry or an external port and a communication network, wherein the communication network may be the Internet, one or more intranets, a local area network (LAN), a wide area network (WLAN), a storage area network (SAN), or a suitable combination thereof, as long as computer device access to the memory is enabled. Furthermore, the computer program involved in the embodiments may be stored in plaintext / ciphertext form, or it may be designed as training data, integrated and recombined through model training and implicitly stored in the parameter states of a deep neural network or other machine learning model.

[0092] Please see Figure 1 This embodiment provides a multi-axis cooperative motion synchronization control method based on a grating encoder, including:

[0093] S101. Obtain the quadrature sine signal and position count value output by the grating encoder corresponding to each motion axis;

[0094] S102. Calculate the instantaneous phase angle of each motion axis based on the orthogonal sinusoidal signal, and use the virtual phase grid point as a unified alignment reference. Based on the phase deviation between the instantaneous phase angle and the virtual phase grid point and the instantaneous velocity estimate of the motion axis, perform first-order phase linear interpolation correction on the position count value to generate the phase domain alignment position value of each motion axis.

[0095] S103. Store the phase domain alignment position value into the circular data buffer, extract the multi-axis position sequence within the preset time window from the circular data buffer, and perform differential operation on the multi-axis position sequence to generate the velocity sequence and acceleration sequence of each motion axis.

[0096] S104. Based on the multi-input multi-output axis dynamic coupling model, the acceleration sequence is used as the model output, and the velocity sequence and the control commands of each motion axis are used as the model input. The recursive least squares method is used to identify the dynamic coupling coefficient between axes online.

[0097] S105. Calculate the coupling compensation amount based on the dynamic coupling coefficient and the real-time motion state of each motion axis, and add the coupling compensation amount to the control command of the corresponding motion axis.

[0098] S106. Based on the geometric topology of the multi-axis mechanical structure, establish the geometric constraint equations between the feedback positions of each motion axis, and substitute the phase domain alignment position values ​​into the geometric constraint equations to calculate the constraint residuals.

[0099] S107. When the constraint residual continuously exceeds the preset anomaly judgment threshold, it is determined that there is an encoder feedback abnormal axis, and weighted suppression processing is performed on the feedback data of the abnormal axis.

[0100] S108. Estimate the estimated position value of the abnormal axis based on the feedback data of other normal axes and the geometric constraint equations, and use the estimated position value as temporary feedback data of the abnormal axis to participate in the collaborative control.

[0101] In step S101, the grating encoder generates two quadrature sinusoidal signals with equal amplitude and a 90-degree phase difference through the relative motion of the reading head and the grating ruler. These signals are transmitted to the servo driver via differential cables. After shaping and filtering, one signal outputs a position count value through a subdivision circuit, and the other outputs a digitized amplitude value to the central controller through an analog sampling channel. The position count value is an integer obtained by accumulating the signal period by a counter, representing the coarse position of the motion axis; the quadrature sinusoidal signal carries fine position information within one signal period, and the two together constitute complete position feedback.

[0102] In step S102, the instantaneous phase angle is obtained through trigonometric function calculations based on the amplitude relationship of the orthogonal sinusoidal signal, representing the precise position of the motion axis within a grating pitch. Virtual phase grid points are pre-defined, equally spaced phase reference points, providing a unified alignment reference for each axis; preferably, the grid is divided by equal phase intervals of the grating pitch. The phase deviation is the difference between the instantaneous phase angle and the grid point, representing the degree of offset between the current phase and the target phase. The instantaneous velocity estimate is obtained by differentially analyzing the position count value, or it can be obtained from the driver's velocity loop output. First-order phase linear interpolation correction estimates the time required to reach the grid point based on the ratio of the phase deviation to the signal angular frequency, calculates the position change within that time using the instantaneous velocity estimate, and superimposes it onto the position count value to obtain the phase domain aligned position value. This step converts the position data of each axis from time domain sampling to unified phase domain alignment, eliminating the problem of inconsistent feedback timing caused by transmission delay differences and sampling clock jitter.

[0103] In step S103, the circular data buffer is a memory area with a circular queue structure, where position data is continuously stored and retrieved through the cyclic movement of read and write pointers. A preset time window is set based on the system's dynamic response characteristics and the data volume requirements of the identification operation, limiting the range of data extracted from the buffer. The multi-axis position sequence is a data set of phase-domain aligned position values ​​of each axis within the window, arranged chronologically. Differential operations are performed on the multi-axis position sequence, sequentially performing first-order and second-order differences. The first-order difference result is a velocity sequence, and the second-order difference result is an acceleration sequence.

[0104] In step S104, the multi-input multi-output (MIMO) axis dynamic coupling model describes the mutual influence between axes, characterizing that the actual acceleration of any axis is the superposition of the nominal acceleration generated by the axis's command and the additional acceleration generated by the motion state of other axes through the coupling coefficient. The axis dynamic coupling coefficient is a set of parameters that quantifies the coupling strength between axes, corresponding to the inertial coupling, elastic coupling, and other effects of the mechanical structure. The recursive least squares method uses new observation data to gradually correct the parameter estimates, enabling real-time parameter updates during operation, unlike offline identification methods. Preferably, a forgetting factor is introduced to control the influence weight of historical data, allowing the estimation to track the time-varying characteristics of the coupling coefficient.

[0105] In step S105, the real-time motion state includes information such as the velocity and acceleration of each axis at the current moment, obtained by differentially aligning the position values ​​in the phase domain. The coupling compensation amount is calculated based on the coupling coefficient and the real-time motion state. Specifically, it involves multiplying the motion states of other axes by their corresponding coupling coefficients and then summing the results to offset the disturbances caused by mechanical structure coupling on the motion of this axis. The compensation amount is then added to the control command before it is sent to the servo driver; preferably, it is added to the speed loop or current loop command.

[0106] In step S106, the geometric topology relationship refers to the spatial positional relationship between the axes in a multi-axis mechanical structure, determined by design parameters. The geometric constraint equation is a mathematical expression for the theoretical positions established based on the geometric topology relationship, reflecting the inherent geometric constraints of the mechanical structure. The constraint residual characterizes the degree of deviation between the actual position and the theoretical constraint, providing a quantitative basis for the consistency verification of multi-axis feedback data.

[0107] In step S107, the preset anomaly judgment threshold is set based on the statistical distribution characteristics of the constraint residuals under normal operating conditions, and is used to distinguish between normal fluctuations and abnormal deviations. An encoder feedback anomalous axis is a motion axis whose feedback data is unreliable, for example, caused by slippage, pulse loss, or signal quality degradation. Weighted suppression processing is used to reduce the weight of the anomalous axis feedback data in the control closed loop, for example, reducing the feedback weight of the anomalous axis from the standard weight to a preset suppression weight to weaken the impact of anomalous feedback on the generation of coordinated control commands.

[0108] In step S108, the feedback data for other normal axes consists of the phase domain alignment position values ​​of the remaining motion axes that have not experienced any anomalies. The estimated position values ​​are calculated based on the feedback data from the normal axes and the geometric constraint equations. The normal axis position values ​​are substituted into the constraint equations as known quantities to solve for the theoretical position values ​​of the abnormal axes under the constraint conditions. Temporary feedback data is used as the estimated position values ​​as the position feedback input for the abnormal axes in the control closed loop, replacing the original feedback in subsequent coordinated control calculations, thus achieving fault-tolerant operation of the system when a single-axis encoder malfunctions.

[0109] This embodiment eliminates the inconsistency in multi-axis position feedback caused by signal transmission delay differences and sampling clock jitter by converting the position feedback of the grating encoders of each motion axis from time-domain sampling to phase-domain unified alignment. Furthermore, it caches the aligned position data through a circular data buffer and generates velocity and acceleration sequences. Based on a multi-input multi-output axis dynamic coupling model, it uses recursive least squares to identify the dynamic coupling coefficients between axes online, achieving real-time compensation for inter-axis coupling disturbances. Simultaneously, it establishes a cross-axis consistency verification mechanism using the geometric constraints of the multi-axis mechanical structure, enabling online identification and fault-tolerant handling of encoder feedback anomalies. Through weighted suppression and estimated position value substitution, it maintains the safe operation of the system under abnormal conditions. This embodiment improves the synchronization accuracy and operational reliability of multi-axis cooperative motion control from three levels: data alignment, coupling compensation, and anomaly fault tolerance.

[0110] Please see Figure 2In some embodiments, the instantaneous phase angle of each motion axis is calculated based on orthogonal sinusoidal signals, and virtual phase grid points are used as a unified alignment reference. Based on the phase deviation between the instantaneous phase angle and the virtual phase grid points, combined with the instantaneous velocity estimate of the motion axis, a first-order phase linear interpolation correction is performed on the position count value to generate the phase domain alignment position value of each motion axis, including:

[0111] S201. Execute the coordinate rotation digital calculation algorithm on the orthogonal sinusoidal signal to extract the phase angle of the orthogonal signal of each motion axis grating encoder at the current moment as the instantaneous phase angle;

[0112] S202. Pre-divide virtual phase grid points with equal intervals of the grating pitch phase, compare the instantaneous phase angle with the phase values ​​of two adjacent virtual phase grid points respectively, and select the next virtual phase grid point that is closest to the instantaneous phase angle as a unified alignment reference, so that the position data of each motion axis are uniformly aligned to the phase time corresponding to the reference.

[0113] S203. Calculate the difference between the instantaneous phase angle and the unified alignment reference as the phase deviation, and obtain the estimated instantaneous speed of the motion axis in the current control cycle;

[0114] S204. Perform zero-crossing detection or frequency measurement algorithm on the quadrature sinusoidal signal, extract the quadrature signal angular frequency of each motion axis grating encoder at the current moment, and calculate the time deviation required for the motion axis to reach the unified alignment reference from the current sampling moment based on the ratio between the phase deviation and the quadrature signal angular frequency.

[0115] S205. Based on the product relationship between the time deviation and the instantaneous velocity estimate, calculate the position change of the motion axis within the time deviation as the position correction amount;

[0116] S206. The position correction amount and the position count value are superimposed to obtain the equivalent position value of the motion axis at the time corresponding to the unified alignment reference, which is used as the phase domain alignment position value.

[0117] When the motion axis is stationary or quasi-stationary, the instantaneous velocity estimate approaches zero, the position correction approaches zero, and the phase domain alignment position value approaches the position count value, thus avoiding phase alignment divergence in a stationary state.

[0118] In step S201, the coordinate rotation digital calculation algorithm approximates the target angle through iterative rotation. Each iteration determines the rotation direction based on the sign of the current vector, rotating the vector towards the target direction by a preset angle. The accumulated rotation angle after each iteration is the desired phase angle. This algorithm only requires shifting and addition / subtraction operations to complete the trigonometric function calculation. The number of iterations is determined based on the required phase angle resolution; the more iterations, the higher the phase angle resolution. The orthogonal signal phase angle is the angle between the synthesized vector of two orthogonal sinusoidal signals and the reference axis, with a value range covering a complete signal cycle. The input data of the coordinate rotation digital calculation algorithm is the digitized amplitude of the orthogonal sinusoidal signal output from the grating encoder via the analog sampling channel, and the output is the orthogonal signal phase angle at the current moment, i.e., the instantaneous phase angle.

[0119] In step S202, the grating pitch is the distance between adjacent lines on the grating ruler, which is an inherent parameter of the grating encoder. The virtual phase grid points are pre-divided at equal intervals of the grating pitch phase. The number of equal divisions is set according to the required phase alignment accuracy. The more equal divisions, the higher the alignment accuracy. The phase value of each grid point is stored in the form of a lookup table.

[0120] In step S203, the phase deviation is the difference between the instantaneous phase angle and the phase value of the unified alignment reference. This difference is processed by modulo operation to ensure that it is within the phase range corresponding to one signal cycle. The current control cycle is the cycle in which the central controller performs one complete control operation. The instantaneous speed estimate can be obtained by dividing the difference between the position count value of the previous control cycle and the current control cycle by the control cycle duration, or it can be directly obtained from the speed loop output of the servo driver.

[0121] In steps S204 to S205, zero-crossing detection monitors the moment when the quadrature sinusoidal signal crosses the zero level using a comparator. The time interval between two adjacent zero-crossing moments in the same direction is the signal period, and the signal angular frequency is the ratio of the constant 2π to the signal period. Frequency measurement obtains the signal frequency by performing spectral analysis or period counting on the quadrature sinusoidal signal, and then converts it into the signal angular frequency. The time deviation is the ratio of the phase deviation to the angular frequency of the quadrature signal, representing the time required for the grating encoder signal of the moving axis to change from the current phase position to the phase position corresponding to the unified alignment reference at the current speed. The position correction is the displacement generated by the moving axis moving at the current instantaneous speed within the time deviation, and the sign of the position correction is consistent with the direction of movement of the moving axis.

[0122] In step S206, the stationary or quasi-stationary state can be determined based on whether the instantaneous speed estimate is lower than a preset speed threshold. The preset speed threshold is determined based on the resolution of the grating encoder and the control cycle length. Specifically, the speed value corresponding to the smallest displacement that the grating encoder can distinguish within one control cycle is used as a benchmark, and this speed value is multiplied by a preset safety factor to obtain the preset speed threshold. When the actual speed of the motion axis is lower than this threshold, the motion axis is considered to be in a stationary or quasi-stationary state.

[0123] This embodiment extracts the instantaneous phase angle from the orthogonal sinusoidal signal using a coordinate rotation digital calculation algorithm. It pre-divides the virtual phase grid points by dividing the phase equally with the grating pitch and selects a unified alignment reference. The time deviation is calculated by the ratio of the phase deviation to the angular frequency of the orthogonal signal. Then, the position correction is calculated by combining the instantaneous velocity estimate. Finally, the position correction is superimposed with the position count value to obtain the phase domain aligned position value. This eliminates the problem of inconsistent position feedback time of each axis caused by differences in signal transmission delay and sampling clock jitter, and provides time-consistent position data for multi-axis collaborative control.

[0124] In some embodiments, a zero-crossing detection or frequency measurement algorithm is performed on the quadrature sinusoidal signal to extract the quadrature signal angular frequency of each motion axis grating encoder at the current moment. Based on the ratio between the phase deviation and the quadrature signal angular frequency, the time deviation required for the motion axis to reach the unified alignment reference from the current sampling moment is calculated, including:

[0125] A zero-crossing detection algorithm is applied to the quadrature sinusoidal signal to detect the rising zero-crossing point when the quadrature sinusoidal signal crosses from the negative half-cycle to the positive half-cycle, and the falling zero-crossing point when the signal crosses from the positive half-cycle to the negative half-cycle.

[0126] Record the current detected rising zero-crossing time and the previous rising zero-crossing time, calculate the time difference between two adjacent rising zero-crossing times, and use the time difference as the current signal period of the orthogonal sinusoidal signal;

[0127] Using the constant 2π as the numerator and the current signal period as the denominator, calculate the ratio between the numerator and the denominator, and use the ratio as the current angular frequency of the orthogonal sinusoidal signal;

[0128] The ratio between the dividend and the divisor is calculated by using the phase deviation as the dividend and the current signal angular frequency as the divisor.

[0129] The ratio is used as the phase difference time required for the motion axis to reach the unified alignment reference from the current sampling time. The phase difference time characterizes the time required for the grating encoder signal of the motion axis to change from the current phase position to the phase position corresponding to the virtual phase grid point at the current speed.

[0130] The phase difference corresponds to the time as the time deviation.

[0131] In this embodiment, the distinction between rising and falling zero-crossing moments is determined based on the direction of the signal slope. When the quadrature sine signal crosses from the negative half-cycle to the positive half-cycle, the signal slope direction is positive, and the corresponding zero-crossing moment is the rising zero-crossing moment. When the quadrature sine signal crosses from the positive half-cycle to the negative half-cycle, the signal slope direction is negative, and the corresponding zero-crossing moment is the falling zero-crossing moment. The zero-crossing detection algorithm captures the zero-crossing moment by monitoring the comparison between the instantaneous amplitude of the quadrature sine signal and the zero level using a comparator. The moment when the comparator output level changes is the zero-crossing moment. Preferably, a glitch filter circuit is set at the comparator output to filter out comparator output jitter caused by signal noise and avoid false triggering of zero-crossing detection. The glitch filter circuit is implemented by setting a minimum pulse width threshold. When the duration of the comparator output level is less than the threshold, it is considered a glitch pulse and is filtered out. When the duration is greater than or equal to the threshold, it is considered a valid zero-crossing event and is retained.

[0132] Recording the zero-crossing moment can be achieved using a hardware timer. The timer counts at a fixed frequency. When a zero-crossing event is detected, the current count value of the latched timer is used as the timestamp corresponding to that zero-crossing moment and stored in a register. The current signal period is the time difference between two adjacent zero-crossing moments, that is, the difference between the timer count value corresponding to the currently detected zero-crossing moment and the timer count value corresponding to the previous zero-crossing moment, multiplied by the timer's counting period to convert it into a time dimension. The current signal period reflects the time required for the encoder signal to complete one full grid pitch at the current speed. The higher the speed, the shorter the current signal period.

[0133] The current signal angular frequency is stored as a floating-point number for subsequent time deviation calculations. The current signal angular frequency is the phase change experienced by the orthogonal sinusoidal signal per unit time, and its value changes in real time with the velocity of the motion axis.

[0134] The ratio of the phase deviation to the current signal angular frequency is the phase difference time, which characterizes the time required for the grating encoder signal of the moving axis to change from the current phase position to the phase position corresponding to the virtual phase grid point at the current speed.

[0135] This embodiment captures the rising zero-crossing moment of an orthogonal sinusoidal signal using a zero-crossing detection algorithm. It calculates the current signal period using the time difference between adjacent rising zero-crossing moments, and then obtains the current signal angular frequency. The ratio of the phase deviation to the current signal angular frequency is then used as the time deviation, achieving a precise conversion of the phase deviation from phase to time dimensions. This provides a time reference for subsequently calculating the position correction based on the product of the time deviation and the instantaneous velocity estimate.

[0136] In some embodiments, the method further includes:

[0137] Real-time parameter extraction is performed on the orthogonal sinusoidal signal. The real-time parameters include signal amplitude, DC offset, orthogonal phase difference, and signal harmonic component content. The real-time parameters are compared with the preset nominal parameter range. When any parameter deviates from the nominal parameter range by more than a preset threshold, the subdivision compensation coefficient is updated based on the ellipse fitting model or compensation matrix model, and the updated subdivision compensation coefficient is applied to the calculation of the instantaneous phase angle.

[0138] In this embodiment, real-time parameter extraction uses the digitized amplitude of the quadrature sine signal output from the grating encoder via the analog sampling channel as the data source. Parameter extraction is completed once per signal cycle. The extracted signal amplitude, DC offset, quadrature phase difference, and signal harmonic component content correspond to the characterization quantities of degradation factors such as changes in grating scale transmittance, light source aging, changes in reading head gap, and grating scale surface contamination, respectively. The nominal parameter range is the fluctuation range of each parameter value calibrated by the grating encoder under standard operating conditions at the time of manufacture. The preset threshold is set proportionally according to the width of the nominal parameter range, or it can be determined based on the statistical distribution of parameter fluctuations during the encoder's historical operation.

[0139] The elliptic fitting model is suitable for situations where both orthogonal signal amplitude drift and DC offset exist simultaneously, while the compensation matrix model is suitable for situations where the orthogonal phase difference deviates from the ideal value. Both models take real-time parameter extraction results as input and output updated subdivision compensation coefficients. The updated subdivision compensation coefficients replace the original coefficients and are applied to the calculation of the instantaneous phase angle, ensuring that the calculation result of the instantaneous phase angle is not affected by signal quality degradation. This embodiment, through real-time parameter extraction and dynamic updating of subdivision compensation coefficients, ensures that the calculation accuracy of the instantaneous phase angle remains stable even when the encoder signal quality degrades.

[0140] In some embodiments, real-time parameter extraction is performed on the orthogonal sinusoidal signal, including:

[0141] Multiple sampling points are acquired within one grating signal period. An extreme value detection algorithm is executed on the sampling points to extract the maximum and minimum values ​​of the orthogonal sinusoidal signal. The signal amplitude is calculated based on the difference between the maximum and minimum values, and the DC offset of the signal is calculated based on the average value between the maximum and minimum values.

[0142] A zero-crossing detection algorithm is applied to the two orthogonal sinusoidal signals to extract the zero-crossing time of each signal. Based on the ratio of the time difference between the two zero-crossing times to the period of the grating signal, the orthogonal phase difference between the two orthogonal signals is calculated.

[0143] The Discrete Fourier Transform algorithm is applied to the orthogonal sinusoidal signal to extract the amplitude of the fundamental component and the amplitude of each harmonic component. Based on the ratio between the amplitude of each harmonic component and the amplitude of the fundamental component, the harmonic component content of the signal is calculated.

[0144] The real-time parameters are compared with the preset nominal parameter range, including:

[0145] Establish the first nominal parameter range corresponding to the signal amplitude, the second nominal parameter range corresponding to the signal DC offset, the third nominal parameter range corresponding to the quadrature phase difference, and the fourth nominal parameter range corresponding to the signal harmonic component content;

[0146] The signal amplitude is compared with the first nominal parameter range, the signal DC offset is compared with the second nominal parameter range, the quadrature phase difference is compared with the third nominal parameter range, and the signal harmonic component content is compared with the fourth nominal parameter range.

[0147] When any parameter of the signal amplitude, signal DC offset, quadrature phase difference, or signal harmonic component content deviates from the corresponding nominal parameter range, adaptive subdivision compensation is triggered.

[0148] In this embodiment, sampling points are collected at equal intervals within one grating signal period. The number of sampling points is determined based on the required parameter extraction accuracy, and a balance is struck between accuracy and computational complexity during actual setup. The extreme value detection algorithm iterates through the sampling point sequence, records the maximum and minimum values ​​in the sequence, and the difference between the maximum and minimum values ​​is the signal amplitude. Half of the sum of the maximum and minimum values ​​is the DC offset of the signal.

[0149] The zero-crossing times of the two orthogonal sinusoidal signals are extracted separately using the aforementioned zero-crossing detection algorithm. The ratio of the time difference between the two zero-crossing times within the same grating signal period to the grating signal period is taken, and then multiplied by 360 degrees. The resulting angle value is the orthogonal phase difference between the two orthogonal signals.

[0150] The Discrete Fourier Transform (DFT) algorithm takes a sequence of sampling points as input and outputs the amplitude of each frequency component. The number of transformation points is related to the frequency resolution; the more transformation points, the higher the frequency resolution. The fundamental frequency corresponds to the reciprocal of the grating signal period, and each harmonic frequency is an integer multiple of the fundamental frequency. The amplitude component corresponding to the fundamental frequency is taken as the fundamental component amplitude, and the amplitude components corresponding to each harmonic frequency within a preset range are taken as the harmonic component amplitudes. The ratio of the sum of the harmonic component amplitudes to the fundamental component amplitude is the harmonic component content of the signal.

[0151] The first nominal parameter range, the second nominal parameter range, the third nominal parameter range, and the fourth nominal parameter range are set according to the signal amplitude, signal DC offset, quadrature phase difference, and signal harmonic component content calibrated by the grating encoder under standard working conditions at the time of manufacture. Each nominal parameter range is a range of the corresponding calibration value fluctuating up or down by a preset ratio.

[0152] This embodiment extracts the amplitude, phase, and spectral parameters of the orthogonal sinusoidal signal through extreme value detection, zero-crossing detection, and discrete Fourier transform, respectively. Each parameter is then independently compared with the nominal parameter range calibrated at the factory, enabling multi-dimensional real-time monitoring of the signal quality of the grating encoder and providing a basis for triggering adaptive subdivision compensation.

[0153] Please see Figure 3 In some embodiments, based on a multi-input multi-output (MIMO) axis dynamic coupling model, the acceleration sequence is used as the model output, and the velocity sequence and control commands for each motion axis are used as the model input. The recursive least squares method is employed to identify the axis dynamic coupling coefficients online, including:

[0154] S301. Establish the deviation expression between the actual acceleration and the nominal acceleration of each motion axis, attribute the deviation to the superposition effect of the velocity state and control command of other motion axes through the coupling coefficient to be identified, and form the mathematical structure of the dynamic coupling model between multiple input and multiple output axes.

[0155] S302. Define the error function of the recursive least squares method, take the acceleration sequence as the expected output of the model, combine the velocity sequence with the control commands of each motion axis into a regression vector, take the linear combination of the regression vector and the coupling coefficient to be identified as the model prediction output, and take the difference between the expected output of the model and the predicted output of the model as the error function.

[0156] S303. Obtain the velocity sequence, acceleration sequence and control commands for each motion axis from the circular data buffer, combine the velocity sequence and control commands for each motion axis into a regression vector, and use the acceleration sequence as the expected output of the model.

[0157] S304. Based on the recursive update formula of the recursive least squares method, the estimated value of the coupling coefficient at the previous time is corrected using the regression vector and error function at the current time, and the estimated value of the coupling coefficient at the current time is used as the inter-axis dynamic coupling coefficient.

[0158] The recursive update formula includes a forgetting factor, which is used to control the weight of the influence of historical data on the current estimate, enabling the recursive least squares method to track the time-varying characteristics of the inter-axis dynamic coupling coefficient.

[0159] In step S301, the nominal acceleration is the acceleration generated by the motion axis under the control command of this axis only, without being affected by the coupling of other axes. It is calculated from the control command of this axis through the servo system model, or it can be calculated from the current feedback of the servo driver through the motor torque constant and load inertia. The deviation expression attributes the difference between the actual acceleration and the nominal acceleration to the superposition effect of the speed state and control command of other motion axes through the coupling coefficients to be identified. This superposition effect is characterized in the form of a linear weighted sum, and each weighting coefficient is the coupling coefficient to be identified.

[0160] In step S302, the regression vector is composed of the velocity sequence and control commands for each motion axis arranged in axial order. The coupling coefficient to be identified is a column vector with the same dimension as the regression vector. The model prediction output is a linear combination of the regression vector and the coupling coefficient to be identified, i.e., the sum of corresponding element-wise multiplication. The error function is the difference between the model's expected output and the model's predicted output, used to quantify the degree of deviation between the current estimated value and the true value of the coupling coefficient.

[0161] In step S303, the data extracted from the circular data buffer has a consistent timestamp to ensure the time alignment between the regression vector and the model's expected output.

[0162] In step S304, the recursive update formula calculates the gain matrix based on the regression vector and error function at the current moment. The gain matrix is ​​then used to correct the coupling coefficient estimate from the previous moment. The corrected result is used as the coupling coefficient estimate for the current moment, i.e., the inter-axis dynamic coupling coefficient. The forgetting factor is a real number between zero and one. Its value is adjusted within a preset range according to the rate of change of operating conditions. A smaller value is used when operating conditions change drastically to accelerate the tracking speed of changes in the coupling coefficient, while a larger value is used when operating conditions are stable to maintain the stability of the estimate.

[0163] This embodiment establishes an expression for the deviation between actual acceleration and nominal acceleration, incorporates the inter-axis coupling effect into the model structure in the form of a linear weighted sum, and uses the recursive least squares method to gradually correct the estimated coupling coefficient by the error between the regression vector and the expected output of the model, thereby realizing the online identification of the dynamic coupling coefficient between axes and providing a real-time basis for the calculation of coupling compensation.

[0164] In some embodiments, an expression for the deviation between the actual acceleration and the nominal acceleration of each motion axis is established, which is represented by formula (1), as follows:

[0165] ;

[0166] In formula (1), For the first axis at time The actual acceleration, For the first axis at time The nominal acceleration generated by the control command of this axis For the first Axis to the first The velocity coupling coefficient to be identified on the shaft. For the first Axis to the first The command coupling coefficient of the axis, For the first axis at time speed, For the first axis at time Control commands;

[0167] The error function of the recursive least squares method is defined by formula (2), which is as follows:

[0168] ;

[0169] In formula (2), For a moment The error value, This is the transpose of the regression vector. For a moment The vector of estimated coupling coefficients;

[0170] The regression vector is represented by formula (3), which is as follows:

[0171] ;

[0172] In formula (3), This represents the total number of motion axes.

[0173] Based on the recursive update formula of the recursive least squares method, the estimated value of the coupling coefficient at the previous time is corrected using the regression vector and error function at the current time, and the estimated value of the coupling coefficient at the current time is used as the dynamic coupling coefficient between axes, which is expressed by formula (4), as follows:

[0174] ;

[0175] In formula (4), The dynamic coupling coefficient between shafts. For a moment The gain matrix, , For a moment The covariance matrix, Forgetting factor, .

[0176] In this embodiment, nominal acceleration By the axis at time The control commands are obtained by conversion from the servo system model, or by conversion from the current feedback of the servo driver through the motor torque constant and load inertia. The summation term in formula (1) is limited. That is, the first The actual acceleration of the axis is only affected by the coupling of the velocity state and control commands of other moving axes; the velocity state and control commands of the axis itself do not generate self-coupling terms.

[0177] In formula (2), the regression vector Dimensions and The dimensions are consistent. The regression vector is composed of the coupling coefficients of the velocity to be identified and the coupling coefficients of the command to be identified, arranged in the order corresponding to the order of the elements in the regression vector. The elements are arranged in order of axis number to represent the speeds of each motion axis, and then in order of axis number to represent the control commands of each motion axis. This arrangement order corresponds one-to-one with the arrangement order of the elements in the coupling coefficient estimation vector to ensure that the elements are correctly matched during linear combination operations.

[0178] In formula (4), the gain matrix For the regression vector Column vectors of the same dimension are used to weigh the contribution of the current observation data to the correction of the coupling coefficient estimate. The larger the value of the gain matrix, the greater the contribution of the current observation data to the correction. Covariance matrix The initial values ​​are set as a diagonal matrix, with diagonal elements taking preset large values ​​to reflect the uncertainty of the initial coupling coefficient estimates. As the number of iterations increases, the values ​​of the covariance matrix gradually converge. Forgetting factor The value of is adjusted within a preset range according to the rate of change of the operating conditions. When the operating conditions change drastically, a smaller value is taken to speed up the tracking speed of the time-varying characteristics of the coupling coefficient, and a larger value is taken when the operating conditions are stable to maintain the stability of the estimation results.

[0179] This embodiment fully presents the mathematical structure of the inter-axis dynamic coupling model and the parameter update mechanism of the recursive least squares method through formulas (1) to (4), clarifies the physical meaning and data source of each symbol, makes the online identification process of the inter-axis dynamic coupling coefficient feasible, and provides a parameter basis for the real-time calculation of the coupling compensation amount.

[0180] In some embodiments, geometric constraint equations are established between the feedback positions of each motion axis based on the geometric topology of the multi-axis mechanical structure, and the phase domain alignment position values ​​are substituted into the geometric constraint equations to calculate the constraint residuals, including:

[0181] Based on the geometric topology of the multi-axis mechanical structure, determine the type of geometric constraint that each motion axis should satisfy under the ideal no-deformation state. The geometric constraint types include constant distance constraint, constant angle constraint, or trajectory equation constraint.

[0182] Based on the type of geometric constraint, a mathematical expression is established between the theoretical positions of each motion axis as the geometric constraint equation. The geometric constraint equation contains fixed geometric parameters determined by the mechanical structure, which are expressed by formula (5). Formula (5) is as follows:

[0183] ;

[0184] In formula (5), For the first axis at time The theoretical position, For the first axis at time The theoretical position, The first is determined by the mechanical structure Axis and the first Theoretical spacing between axes;

[0185] Extract the phase domain alignment position value of each motion axis at the current moment from the phase domain alignment position value generated in each control cycle;

[0186] Substitute the phase domain alignment position value of each motion axis at the current moment into the geometric constraint equation, and calculate the deviation between the actual position of each motion axis and the theoretical constraint described by the geometric constraint equation as the constraint residual.

[0187] When the multi-axis mechanical structure is a gantry dual-drive platform, the geometric constraint type is a constant distance constraint. The geometric constraint equation is that the difference between the positions of the two axes is equal to the theoretical spacing determined by the crossbeam mechanical structure. The constraint residual is the absolute value of the deviation between the actual difference between the phase domain alignment position values ​​of the two axes and the theoretical spacing.

[0188] In this embodiment, the ideal deformation-free state refers to a state in which elastic deformation of the mechanical structure due to factors such as temperature changes, load, or installation stress is not considered, and the positional relationship between each motion axis is completely determined by design parameters. The constant distance constraint is applicable to mechanical structures where the theoretical distance between two parallel linear motion axes remains constant; the constant angle constraint is applicable to mechanical structures where the included angle between two rotary motion axes remains constant; and the trajectory equation constraint is applicable to mechanical structures where the positions of each motion axis need to satisfy a specific trajectory equation during multi-axis linkage.

[0189] Fixed geometric parameters in formula (5) It can be directly read from mechanical design drawings, or it can be calibrated by measuring equipment such as laser interferometers after the mechanical structure is installed. The geometric constraint equation described by formula (5) is the theoretical positional relationship that each motion axis should satisfy in an ideal, undeformed state. When calculating the constraint residual, the phase domain alignment position value of each motion axis at the current moment is substituted as the actual position value. and Position, calculate the difference between the actual position on the left side of the equal sign and the theoretical distance on the right side of the equal sign. The absolute value of the deviation between the two is the constraint residual.

[0190] This embodiment provides a quantitative basis for the consistency verification of multi-axis feedback data by determining the geometric constraint type, establishing geometric constraint equations, and calculating constraint residuals, thus enabling the determination of encoder feedback anomalies to have an implementable judgment criterion.

[0191] In some embodiments, when the constraint residual continuously exceeds a preset anomaly detection threshold, it is determined that there is an encoder feedback anomaly axis, and weighted suppression processing is performed on the feedback data of the anomaly axis, including:

[0192] Set an anomaly detection counter, which monitors the comparison result between the constraint residual and the preset anomaly detection threshold in each control cycle;

[0193] When the constraint residual exceeds the preset anomaly detection threshold, the anomaly detection counter increments; when the constraint residual does not exceed the preset anomaly detection threshold, the anomaly detection counter is reset to zero.

[0194] When the cumulative value of the anomaly detection counter reaches the preset continuous cycle number threshold, it is determined that there is an encoder feedback abnormal axis, and the axis identifier of the abnormal axis is recorded.

[0195] The feedback weight of the abnormal axis in the collaborative control closed loop is reduced from the standard weight to a preset suppression weight. The suppression weight is less than the standard weight, which is used to reduce the impact of the feedback data of the abnormal axis on the generation of collaborative control commands.

[0196] Meanwhile, based on the geometric constraint equation, the phase domain alignment position values ​​of other normal axes are substituted into the geometric constraint equation as known quantities, and the theoretical position values ​​of the abnormal axes are obtained by inverse solution as estimated position values.

[0197] The estimated position value is used as temporary feedback data for the abnormal axis, replacing the original phase domain aligned position value of the abnormal axis in the coordinated control calculation of subsequent control cycles.

[0198] In this embodiment, the anomaly detection counter is initialized to zero before monitoring begins. At the start of each control cycle, it increments or resets based on the comparison result of the previous cycle. The preset anomaly detection threshold is obtained by statistically analyzing the constraint residuals under normal operating conditions. The sum of the mean of the constraint residuals under normal operating conditions and the standard deviation of a preset multiple is taken as the preset anomaly detection threshold. The preset duration of cycles threshold is determined based on the maximum allowable anomaly duration and the control cycle length, i.e., the ratio of the maximum anomaly duration to the control cycle length.

[0199] The standard weights are the weights of the feedback data from each motion axis in the collaborative control closed loop under normal operating conditions, with a default value of one. The suppression weights are adjusted within a preset range based on the degree to which the constraint residuals exceed the preset anomaly judgment threshold, or they can be set as the standard weights multiplied by a preset proportional coefficient. When the geometric constraint equations involve multiple motion axes, the phase domain alignment position values ​​of other normal axes can be substituted into the geometric constraint equations as known quantities to directly solve for the theoretical position values ​​of the abnormal axes. When there are more than one abnormal axis, multiple geometric constraint equations need to be solved simultaneously by combining the data from other normal axes.

[0200] This embodiment continuously monitors the constraint residuals using an anomaly detection counter. When the constraint residuals continuously exceed the preset anomaly detection threshold, the encoder is identified as having an abnormal axis. By reducing the feedback weight and replacing the original feedback data with estimated position values, the system achieves fault-tolerant operation when the single-axis encoder is abnormal.

[0201] In some embodiments, the estimated position value is used as temporary feedback data for the abnormal axis in the collaborative control, including:

[0202] The estimated position value is written into the feedback data register corresponding to the abnormal axis, replacing the phase domain aligned position value of the abnormal axis, and serving as the position feedback input of the abnormal axis in the cooperative control closed loop;

[0203] In the coordinated control operation of the control cycle, the central controller reads the estimated position value from the feedback data register and performs closed-loop regulation based on the estimated position value to keep the motion trajectory of the abnormal axis consistent with the theoretical constraints described by the geometric constraint equation.

[0204] When the constraint residual recovers to below the preset anomaly judgment threshold and remains stable for a preset number of recovery cycles, the abnormal axis is judged to have returned to normal. The feedback weight of the abnormal axis is restored to the standard weight, and the original phase domain alignment position value of the abnormal axis is switched back as the feedback input.

[0205] In this embodiment, the feedback data register is located in the memory-mapped space of the central controller, with each motion axis corresponding to an independent register address. When the estimated position value is written to the feedback data register, it uses the same data format and dimensions as the position value aligned with the phase domain, ensuring that no additional format conversion is required during switching. In the coordinated control operation of each control cycle, the central controller reads the estimated position value from the feedback data register as the position feedback input for the abnormal axis, compares the estimated position value with the commanded position value to obtain the position deviation, and generates control commands through adjustment operations of the position loop and velocity loop, so that the actual motion trajectory of the abnormal axis satisfies the theoretical constraint relationship described by the geometric constraint equation.

[0206] The preset number of recovery cycles is determined based on the minimum allowable stable duration and control cycle length of the system, i.e., the ratio of the minimum stable duration to the control cycle length. When the constraint residual recovers to below the preset anomaly judgment threshold and remains stable for the preset number of recovery cycles, the abnormal axis is judged to have returned to normal. At the beginning of the control cycle, the feedback weight of the abnormal axis is restored to the standard weight, and the feedback input is switched from the estimated position value back to the original phase domain aligned position value of the abnormal axis to ensure the continuity of control commands before and after the switch.

[0207] This embodiment achieves seamless replacement of the estimated position value with the original feedback data of the abnormal axis through the writing and reading mechanism of the feedback data register. Through closed-loop adjustment, the motion trajectory of the abnormal axis during fault-tolerant operation satisfies the geometric constraint equation. After the abnormal axis recovers to normal, it smoothly exits the fault-tolerant mode through the weight recovery and feedback switching mechanism, thus ensuring the operational stability of the multi-axis cooperative control system during encoder abnormality and recovery process.

[0208] By adopting the above technical solutions, this invention differs from existing technologies and possesses the following beneficial effects: By converting the position feedback of each motion axis from time-domain sampling to phase-domain unified alignment, the multi-axis position data possesses consistent time attributes under a unified phase reference, providing high-precision time-aligned position data for subsequent coordinated control; based on the aligned position data, the dynamic coupling coefficient between axes is identified online and compensated to the control commands in real time, effectively suppressing inter-axis coupling disturbances and improving the synchronization accuracy of multi-axis coordinated motion; geometric constraint equations are used to perform cross-axis consistency verification on the feedback data, and when encoder feedback is abnormal, weighted suppression and estimated position value substitution are used to maintain stable system operation, enhancing the system's fault tolerance capability for encoder anomalies. This invention improves the synchronization accuracy and operational reliability of multi-axis coordinated motion control from three levels: data alignment, coupling compensation, and anomaly fault tolerance.

[0209] Finally, it should be noted that although the above embodiments have been described in the text and drawings of this application, this should not limit the scope of patent protection of this application. Any technical solutions that are based on the essential concept of this application and utilize the content described in the text and drawings of this application, resulting in equivalent structural or procedural substitutions or modifications, as well as the direct or indirect application of the technical solutions of the above embodiments to other related technical fields, are all included within the scope of patent protection of this application.

Claims

1. A method for multi-axis cooperative motion synchronization control based on a grating encoder, characterized in that, include: Obtain the quadrature sine signals and position count values ​​output by the grating encoders corresponding to each motion axis; The instantaneous phase angle of each motion axis is calculated based on the orthogonal sinusoidal signal. Using the virtual phase grid point as a unified alignment reference, and based on the phase deviation between the instantaneous phase angle and the virtual phase grid point, combined with the instantaneous velocity estimate of the motion axis, the position count value is corrected by first-order phase linear interpolation to generate the phase domain alignment position value of each motion axis. The phase domain alignment position value is stored in a circular data buffer. The multi-axis position sequence within a preset time window is extracted from the circular data buffer, and a difference operation is performed on the multi-axis position sequence to generate the velocity sequence and acceleration sequence of each motion axis. Based on the multi-input multi-output axis dynamic coupling model, the acceleration sequence is used as the model output, and the velocity sequence and the control commands of each motion axis are used as the model input. The recursive least squares method is used to identify the dynamic coupling coefficient between axes online. The coupling compensation amount is calculated based on the dynamic coupling coefficient and the real-time motion state of each motion axis, and the coupling compensation amount is superimposed on the control command of the corresponding motion axis. Based on the geometric topology of the multi-axis mechanical structure, establish geometric constraint equations between the feedback positions of each motion axis, and substitute the phase domain alignment position values ​​into the geometric constraint equations to calculate the constraint residuals. When the constraint residual continuously exceeds the preset anomaly determination threshold, it is determined that there is an encoder feedback abnormal axis, and weighted suppression processing is performed on the feedback data of the abnormal axis. The estimated position value of the abnormal axis is estimated based on the feedback data of other normal axes and the geometric constraint equation, and the estimated position value is used as temporary feedback data of the abnormal axis to participate in the collaborative control.

2. The multi-axis cooperative motion synchronization control method based on a grating encoder according to claim 1, characterized in that, The instantaneous phase angle of each motion axis is calculated based on the orthogonal sinusoidal signal. Using virtual phase grid points as a unified alignment reference, and based on the phase deviation between the instantaneous phase angle and the virtual phase grid points, combined with the estimated instantaneous velocity of the motion axis, a first-order phase linear interpolation correction is performed on the position count value to generate the phase domain alignment position value of each motion axis, including: A coordinate rotation digital calculation algorithm is executed on the orthogonal sinusoidal signal to extract the phase angle of the orthogonal signal of each motion axis grating encoder at the current moment as the instantaneous phase angle; Virtual phase grid points are pre-divided with equal phase intervals of the grating pitch, and the next virtual phase grid point that is closest to the instantaneous phase angle is selected as the unified alignment reference. The difference between the instantaneous phase angle and the unified alignment reference is calculated as the phase deviation, and the instantaneous velocity estimate of the motion axis in the current control cycle is obtained; A zero-crossing detection or frequency measurement algorithm is performed on the orthogonal sinusoidal signal to extract the angular frequency of the orthogonal signal of each motion axis grating encoder at the current moment. Based on the ratio between the phase deviation and the angular frequency of the orthogonal signal, the time deviation required for the motion axis to reach the unified alignment reference from the current sampling moment is calculated. Based on the product relationship between the time deviation and the instantaneous velocity estimate, the change in position of the motion axis within the time deviation is calculated as a position correction amount; The position correction amount is superimposed with the position count value to obtain the equivalent position value of the motion axis at the time corresponding to the unified alignment reference, which is used as the phase domain alignment position value. When the motion axis is at rest or near rest, the instantaneous velocity estimate approaches zero, the position correction approaches zero, and the phase domain alignment position value approaches the position count value.

3. The multi-axis cooperative motion synchronization control method based on a grating encoder according to claim 2, characterized in that, A zero-crossing detection or frequency measurement algorithm is performed on the orthogonal sinusoidal signal to extract the angular frequency of the orthogonal signal from the encoder of each motion axis at the current moment. Based on the ratio between the phase deviation and the angular frequency of the orthogonal signal, the time deviation required for the motion axis to reach the unified alignment reference from the current sampling moment is calculated, including: A zero-crossing detection algorithm is performed on the orthogonal sinusoidal signal to detect the rising zero-crossing point when the orthogonal sinusoidal signal crosses from the negative half-cycle to the positive half-cycle and the falling zero-crossing point when it crosses from the positive half-cycle to the negative half-cycle. Record the current detected rising zero-crossing time and the previous rising zero-crossing time, calculate the time difference between two adjacent rising zero-crossing times, and use the time difference as the current signal period of the orthogonal sinusoidal signal; Using the constant 2π as the numerator and the current signal period as the denominator, calculate the ratio between the numerator and the denominator, and use the ratio as the current signal angular frequency of the orthogonal sinusoidal signal; Using the phase deviation as the dividend and the current signal angular frequency as the divisor, calculate the ratio between the dividend and the divisor; The ratio is used as the phase difference time required for the motion axis to reach the unified alignment reference from the current sampling time. The phase difference time represents the time required for the grating encoder signal of the motion axis to change from the current phase position to the phase position corresponding to the virtual phase grid point at the current speed. The time corresponding to the phase difference is taken as the time deviation.

4. The multi-axis cooperative motion synchronization control method based on a grating encoder according to claim 2, characterized in that, The method further includes: Real-time parameter extraction is performed on the orthogonal sinusoidal signal. The real-time parameters include signal amplitude, DC offset, orthogonal phase difference, and signal harmonic component content. The real-time parameters are compared with a preset nominal parameter range. When any parameter deviates from the nominal parameter range by more than a preset threshold, the subdivision compensation coefficient is updated based on an elliptic fitting model or a compensation matrix model. The updated subdivision compensation coefficient is then applied to the calculation of the instantaneous phase angle.

5. The multi-axis cooperative motion synchronization control method based on a grating encoder according to claim 4, characterized in that, Real-time parameter extraction is performed on the orthogonal sinusoidal signal, including: Multiple sampling points are acquired within one grating signal period. An extreme value detection algorithm is executed on the sampling points to extract the maximum and minimum values ​​of the orthogonal sinusoidal signal. The signal amplitude is calculated based on the difference between the maximum and minimum values. The DC offset of the signal is calculated based on the average value between the maximum and minimum values. A zero-crossing detection algorithm is applied to the two orthogonal sinusoidal signals to extract the zero-crossing time of each signal. Based on the ratio of the time difference between the two zero-crossing times to the period of the grating signal, the orthogonal phase difference between the two orthogonal signals is calculated. The discrete Fourier transform algorithm is performed on the orthogonal sinusoidal signal to extract the amplitude of the fundamental component and the amplitude of each harmonic component of the orthogonal sinusoidal signal. Based on the ratio between the amplitude of each harmonic component and the amplitude of the fundamental component, the harmonic component content of the signal is calculated. The real-time parameters are compared with a preset nominal parameter range, including: Establish a first nominal parameter range corresponding to the signal amplitude, a second nominal parameter range corresponding to the signal DC offset, a third nominal parameter range corresponding to the orthogonal phase difference, and a fourth nominal parameter range corresponding to the signal harmonic component content; The signal amplitude is compared with the first nominal parameter range, the signal DC offset is compared with the second nominal parameter range, the quadrature phase difference is compared with the third nominal parameter range, and the signal harmonic component content is compared with the fourth nominal parameter range. When any of the following parameters deviates from the corresponding nominal parameter range: signal amplitude, signal DC offset, quadrature phase difference, or signal harmonic component content, adaptive subdivision compensation is triggered.

6. The multi-axis cooperative motion synchronization control method based on a grating encoder according to claim 1, characterized in that, Based on a multi-input multi-output (MIMO) axis dynamic coupling model, the acceleration sequence is used as the model output, and the velocity sequence and control commands for each motion axis are used as the model input. The recursive least squares method is used to identify the axis dynamic coupling coefficients online, including: Establish the deviation expression between the actual acceleration and the nominal acceleration of each motion axis, and attribute the deviation to the superposition effect of the velocity state and control command of other motion axes through the coupling coefficient to be identified, thus forming the mathematical structure of the dynamic coupling model between multiple input and multiple output axes; Define the error function of the recursive least squares method, take the acceleration sequence as the expected output of the model, combine the velocity sequence with the control commands of each motion axis into a regression vector, take the linear combination of the regression vector and the coupling coefficient to be identified as the model prediction output, and take the difference between the expected output of the model and the predicted output of the model as the error function. The velocity sequence, the acceleration sequence, and the control commands for each motion axis are obtained from the circular data buffer. The velocity sequence and the control commands for each motion axis are combined into the regression vector, and the acceleration sequence is used as the expected output of the model. Based on the recursive update formula of the recursive least squares method, the estimated value of the coupling coefficient at the previous time is corrected using the regression vector and the error function at the current time, and the estimated value of the coupling coefficient at the current time is used as the inter-axis dynamic coupling coefficient. The recursive update formula includes a forgetting factor, which is used to control the influence weight of historical data on the current estimate, so that the recursive least squares method can track the time-varying characteristics of the inter-axis dynamic coupling coefficient.

7. The multi-axis cooperative motion synchronization control method based on a grating encoder according to claim 6, characterized in that, An expression for the deviation between the actual acceleration and the nominal acceleration of each motion axis is established, which is expressed by formula (1), as follows: ; In formula (1), For the first axis at time The actual acceleration, For the first axis at time The nominal acceleration generated by the control command of this axis For the first Axis to the first The velocity coupling coefficient to be identified on the shaft. For the first Axis to the first The command coupling coefficient of the axis, For the first axis at time speed, For the first axis at time Control commands; The error function of the recursive least squares method is defined by formula (2), which is as follows: ; In formula (2), For a moment The error value, This is the transpose of the regression vector. For a moment The vector of estimated coupling coefficients; The regression vector is represented by formula (3), which is as follows: ; In formula (3), This represents the total number of motion axes. Based on the recursive update formula of the recursive least squares method, the estimated value of the coupling coefficient at the previous time is corrected using the regression vector and the error function at the current time, and the estimated value of the coupling coefficient at the current time is used as the dynamic coupling coefficient between axes, which is expressed by formula (4), as follows: ; In formula (4), The dynamic coupling coefficient between shafts. For a moment The gain matrix, , For a moment The covariance matrix, Forgetting factor, .

8. The multi-axis cooperative motion synchronization control method based on a grating encoder according to claim 1, characterized in that, Based on the geometric topology of the multi-axis mechanical structure, geometric constraint equations are established between the feedback positions of each motion axis. The phase domain alignment position values ​​are substituted into the geometric constraint equations to calculate the constraint residuals, including: Based on the geometric topology of the multi-axis mechanical structure, determine the type of geometric constraint that each motion axis should satisfy under the ideal no-deformation state. The geometric constraint types include constant distance constraint, constant angle constraint, or trajectory equation constraint. Based on the geometric constraint type, a mathematical expression is established between the theoretical positions of each motion axis as the geometric constraint equation. The geometric constraint equation contains fixed geometric parameters determined by the mechanical structure, which are expressed by formula (5). Formula (5) is as follows: ; In formula (5), For the first axis at time The theoretical position, For the first axis at time The theoretical position, The first is determined by the mechanical structure Axis and the first Theoretical spacing between axes; Extract the phase domain alignment position value of each motion axis at the current moment from the phase domain alignment position value generated in each control cycle; Substitute the phase domain alignment position value of each motion axis at the current time into the geometric constraint equation, and calculate the deviation between the actual position of each motion axis and the theoretical constraint described by the geometric constraint equation as the constraint residual. When the multi-axis mechanical structure is a gantry dual-drive platform, the geometric constraint type is a constant distance constraint, the geometric constraint equation is that the difference between the positions of the two axes is equal to the theoretical spacing determined by the crossbeam mechanical structure, and the constraint residual is the absolute value of the deviation between the actual difference between the phase domain alignment position values ​​of the two axes and the theoretical spacing.

9. The multi-axis cooperative motion synchronization control method based on a grating encoder according to claim 8, characterized in that, When the constraint residual continuously exceeds a preset anomaly detection threshold, it is determined that there is an encoder feedback anomaly axis, and weighted suppression processing is performed on the feedback data of the anomaly axis, including: An anomaly detection counter is set up, which monitors the comparison result between the constraint residual and the preset anomaly detection threshold in each control cycle; When the constraint residual exceeds the preset anomaly detection threshold, the anomaly detection counter increments; when the constraint residual does not exceed the preset anomaly detection threshold, the anomaly detection counter is reset to zero. When the cumulative value of the anomaly detection counter reaches a preset duration period threshold, it is determined that there is an encoder feedback anomaly axis, and the axis identifier of the anomaly axis is recorded. The feedback weight of the abnormal axis in the collaborative control closed loop is reduced from the standard weight to a preset suppression weight, wherein the suppression weight is less than the standard weight, in order to reduce the impact of the feedback data of the abnormal axis on the generation of collaborative control commands. Meanwhile, based on the geometric constraint equation, the phase domain alignment position values ​​of the other normal axes are substituted into the geometric constraint equation as known quantities, and the theoretical position value of the abnormal axis is obtained by inverse solution as the estimated position value. The estimated position value is used as temporary feedback data for the abnormal axis, replacing the original phase domain alignment position value of the abnormal axis in the collaborative control calculation of subsequent control cycles.

10. The multi-axis cooperative motion synchronization control method based on a grating encoder according to claim 9, characterized in that, Using the estimated position value as temporary feedback data for the abnormal axis in collaborative control includes: The estimated position value is written into the feedback data register corresponding to the abnormal axis, replacing the phase domain alignment position value of the abnormal axis, and used as the position feedback input of the abnormal axis in the cooperative control closed loop; In the coordinated control operation of the control cycle, the central controller reads the estimated position value from the feedback data register and performs closed-loop adjustment based on the estimated position value to make the motion trajectory of the abnormal axis consistent with the theoretical constraints described by the geometric constraint equation. When the constraint residual recovers to below the preset anomaly determination threshold and remains stable for a preset number of recovery cycles, it is determined that the anomaly axis has returned to normal. The feedback weight of the anomaly axis is restored to the standard weight, and the original phase domain alignment position value of the anomaly axis is switched back as the feedback input.

Citation Information

Patent Citations

  • Servo position control method and system applied to engraving and milling machine

    CN120630808A

  • Precise control method for PCB multi-axis linkage motion coordination and error suppression

    CN121348960A