A method, medium, and system for compensating for temperature drift in a gantry robot.

Abstract

This application discloses a method, medium, and system for temperature drift compensation of a gantry robot. The method includes the following steps: constructing a standard pose database and a displacement deviation calibration model based on the gantry robot's working area; real-time acquisition of the gantry robot's actual position signal and current ambient temperature signal; calculating the gantry robot's axis offset at the current ambient temperature based on the displacement deviation calibration model and the current ambient temperature signal, and determining the corresponding compensation amount based on the gantry robot's current position; and converting the compensation amount into a compensation command for position correction of the gantry robot after superimposing it with the standard pose database. The medium and system are used to implement the above method. The beneficial effects of this application are: by pre-establishing a standard pose database and a displacement deviation calibration model through the calibration steps, the temperature signal acquisition, offset calculation, and compensation command generation and issuance are automatically completed during subsequent operation, and the entire compensation process does not require manual intervention during machine shutdown.

Description

Technical Field

[0001] This application relates to the field of gantry robot technology, and in particular to a method, medium and system for compensating for temperature drift in gantry robots. Background Technology

[0002] Gantry robots are core equipment widely used in automated production lines for material handling and machine tool loading and unloading. The market offers a variety of product brands and structural forms, but their positioning accuracy varies, making it difficult to meet the long-term stable operation requirements of high-precision manufacturing scenarios.

[0003] Among these factors, structural temperature drift caused by changes in ambient temperature is the main reason for positioning drift and decreased accuracy in gantry robots. In practical industrial applications, the mechanical structures of gantry robots, such as beams, guide rails, and slides, will expand and contract with changes in ambient temperature. The greater the temperature difference, the more obvious the positional deviation of the end effector, directly leading to problems such as inaccurate loading and unloading, misalignment, and alignment failure. Once a deviation occurs, the operator must recalibrate and adjust the robot's position.

[0004] However, in industrial material handling scenarios, gantry robots perform tens of thousands of operation cycles daily, operating under high loads for extended periods. Traditional manual calibration methods cannot match the continuous production rhythm of automated production lines. Each calibration stop takes 20-30 minutes, resulting in 1-2 stoppages per day, severely impacting overall equipment efficiency and production stability. Furthermore, human intervention introduces errors, leading to poor consistency in material handling and a high scrap rate. This contradicts the original intention of automated equipment—high efficiency, unmanned operation, and high stability—reducing gantry robots to "semi-automated" operation in environments with significant temperature variations. Summary of the Invention

[0005] One objective of this application is to provide a gantry robot temperature drift compensation method that can solve at least one of the defects in the above-mentioned background art.

[0006] Another objective of this application is to provide a system capable of implementing a gantry robot temperature drift compensation method that addresses at least one of the deficiencies in the aforementioned background art.

[0007] Another object of this application is to provide a computer-readable storage medium capable of implementing a gantry robot temperature drift compensation method that addresses at least one of the deficiencies in the aforementioned background art.

[0008] To achieve at least one of the above objectives, one aspect of this application provides a method for temperature drift compensation of a gantry robot, comprising the following steps: constructing a standard pose database by performing global point-to-point teaching on the working area of ​​the gantry robot; selecting multiple key points for the gantry robot's movement within the working area, collecting reference offset data of the gantry robot at each key point under different temperatures, and constructing a displacement deviation calibration model; real-time acquisition of the actual position signal of the gantry robot and the current ambient temperature signal; calculating the offset of each axis of the gantry robot under the current ambient temperature based on the displacement deviation calibration model and the current ambient temperature signal, and determining the corresponding compensation amount according to the current position of the gantry robot; superimposing the compensation amount with the reference coordinates in the standard pose database to generate a corrected target position, and converting the corrected target position into a pulse-form compensation command, which is then sent to the servo driver of the gantry robot.

[0009] Preferably, the construction of the displacement deviation calibration model includes the following process: measuring the deviation between the actual position and the target position of the gantry robot at each key point under different temperatures as reference offset data; fitting the reference offset data with a quadratic polynomial model to obtain the displacement deviation calibration model corresponding to each key point, the specific expression of which is as follows: L(T) = a × T 2 +b×T+c; In the formula, L(T) represents the displacement drift of the gantry robot at temperature T; a, b, and c are compensation coefficients, obtained through fitting calculations.

[0010] Preferably, during the operation of the gantry robot, when the temperature difference exceeds a set first threshold or the running time exceeds a set second threshold, a dynamic update of the standard pose database and displacement deviation calibration model is triggered. The specific process of dynamic update is as follows: the pose and position of the gantry robot are measured at some key points within the working area; a new displacement deviation calibration model is refitted and constructed based on the position measurement results, and the deviation between the new displacement deviation calibration model and the old displacement deviation calibration model is calculated. If the deviation is less than a preset third threshold, the new displacement deviation calibration model is adopted; otherwise, the old displacement deviation calibration model is still used. The expression for updating the standard pose database based on the pose measurement results is as follows: P new =λ×P measured +(1-λ) ×P old ; In the formula, P new Let P represent the updated standard pose, λ represent the learning rate, and P represent the standard pose. measured P represents the current measured pose. old This indicates the standard pose before the standard pose database was updated.

[0011] Preferably, the calculation of the offset of each axis includes the following process: substituting the current ambient temperature into the displacement deviation calibration model to calculate the theoretical drift at the current position; and performing spatial interpolation correction on the theoretical drift based on the spatial coordinates of the current position to obtain the actual X-axis, Y-axis, and Z-axis offsets at the current position. The specific calculation expressions are as follows: I(T0) = α I ×L I ×ΔT + β I ×ΔT 2 ; In the formula, I(T0) represents the displacement drift of the gantry robot's I-axis at the current ambient temperature T0, where I = {X, Y, Z}; α I L represents the coefficient of thermal expansion of a material. I The structural length of the I-axis of the gantry robot is represented by ΔT, where ΔT represents the deviation of the current ambient temperature T0 from the reference temperature, and β is the structural length of the I-axis of the gantry robot. I This represents the nonlinear temperature drift coefficient.

[0012] Preferably, when calculating the offset of each axis, α I ×L I The initial parameters are calibrated using the compensation coefficient b fitted by the displacement deviation calibration model; the working area of ​​the gantry robot is divided into multiple sub-regions, and the nonlinear temperature drift coefficient β of the mesh vertices of each sub-region is individually calibrated experimentally. I During the real-time operation of the gantry robot, based on the sub-region where the current position is located, the nonlinear temperature drift coefficient β matching the current position is obtained through trilinear interpolation. I .

[0013] Preferably, the expression for calculating the compensation amount based on the offset of each axis is as follows: Compensate=k correction × I(T0)×S filter ; In the formula, Compensate represents the compensation amount, and k correction Indicates the calibration correction factor. I(T0) represents the displacement drift of the I-axis of the gantry robot at the current ambient temperature T0, S filter This represents the smoothing filter coefficients.

[0014] Preferably, while issuing compensation commands to the servo drive of the gantry robot for real-time position correction, compensation is also made for positioning errors including structural thermal deformation components and transmission chain drift components. The specific compensation process is as follows: the feedback signal from the motor end of the gantry robot is acquired as a semi-closed-loop feedback quantity, and the feedback signal from the end effector of the gantry robot is acquired as a full-closed-loop feedback quantity; Kalman filtering is used to fuse the semi-closed-loop and full-closed-loop feedback quantities of the gantry robot to estimate the structural thermal deformation components and transmission chain drift components; the structural thermal deformation components estimated by Kalman filtering are compensated by correcting the forward kinematics parameters; the transmission chain drift components estimated by Kalman filtering are superimposed on the PID output loop of the servo drive of the gantry robot in the form of feedforward compensation for compensation.

[0015] Preferably, after the compensation command is issued, the compensation effect is verified, which specifically includes the following process: after each compensation command is executed, the feedback position of the end effector of the gantry robot is read; the deviation between the feedback position and the target position is calculated as the residual error; when the absolute value of the residual error is greater than the preset fourth threshold, the accuracy of the working area of ​​the gantry robot is recalibrated for multiple points and multiple temperature gradients, and the displacement deviation calibration model is updated.

[0016] Another aspect of this application provides a system for implementing the above-described temperature drift compensation method for a gantry robot, comprising an operation control module, a detection module, and a compensation algorithm module; the operation control module is configured to allow operators to perform teaching calibration, parameter configuration, compensation enabling, status monitoring, and point storage, and to send compensation commands to the servo driver of the gantry robot; the detection module is configured to collect ambient temperature signals in real time, collect the actual position signals of the gantry robot's end effector in real time, and perform multi-point, multi-temperature gradient global accuracy calibration of the gantry robot's working area; the compensation algorithm module is configured to receive the output results of the detection module, construct a displacement deviation calibration model, and generate compensation commands.

[0017] Another aspect of this application provides a computer-readable storage medium storing a computer program; when the computer program is executed by a processor, it implements the above-described gantry robot temperature drift compensation method.

[0018] Compared with the prior art, the beneficial effects of this application are as follows: (1) This application establishes a standard pose database and displacement deviation calibration model in advance through calibration steps. During subsequent operation, it automatically completes temperature signal acquisition, offset calculation, and compensation command generation and issuance. The entire compensation process does not require manual intervention. Compared with traditional manual calibration methods, it enables the gantry robot to operate continuously and unmanned in a wide temperature range environment, significantly improving the overall efficiency of the equipment and the stability of the production line capacity.

[0019] (2) This application establishes a quantitative displacement deviation calibration model by calibrating the working area at multiple points, so that the compensation calculation has a unified mathematical basis and a repeatable calculation process. Compared with the traditional manual calibration method, it eliminates the problem of inconsistent calibration results caused by differences in human operation, effectively reduces the quality risks such as poor consistency of loading and unloading and high scrap rate caused by calibration deviation, and makes the compensation accuracy quantifiable and traceable.

[0020] (3) This application continuously monitors the ambient temperature and the position signal of the gantry robot, and the compensation calculation steps dynamically respond to changes in ambient temperature, with the compensation command taking effect immediately at the servo level. This enables the gantry robot to maintain high-precision positioning for a long time in industrial scenarios with large day-night temperature differences and significant seasonal temperature changes, without the need for frequent shutdowns for debugging due to temperature fluctuations. This effectively broadens the applicable environment range of the gantry robot and meets the continuous production needs of high-precision, large-scale loading and unloading operations. Attached Figure Description

[0021] Figure 1 This is a schematic diagram illustrating the working steps of the temperature drift compensation method for the gantry robot in this application.

[0022] Figure 2 This is a schematic diagram of the system implementing the temperature drift compensation method for gantry robots in this application. Detailed Implementation

[0023] The present application will now be further described in conjunction with specific embodiments. It should be noted that, in the description of this specification, the use of terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicates that the specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms should not be construed as necessarily referring to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. In addition, those skilled in the art can combine and integrate the different embodiments or examples described in this specification.

[0024] In the description of this application, it should be noted that the terms "center", "lateral", "longitudinal", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", etc., which indicate the orientation and positional relationship based on the orientation or positional relationship shown in the accompanying drawings, are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and should not be construed as limiting the specific protection scope of this application.

[0025] It should be noted that the terms "first," "second," etc., in the specification and claims of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0026] In this application, unless otherwise expressly specified and limited, the terms "installation," "connection," "joining," and "fixing," etc., should be interpreted broadly. For example, they can refer to a connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0027] In this application, unless otherwise expressly specified and limited, "above" or "below" the second feature can include direct contact between the first and second features, or contact between the first and second features through another feature between them. Furthermore, "above," "over," and "on top" of the second feature includes the first feature being directly above or diagonally above the second feature, or simply indicates that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature includes the first feature being directly below or diagonally below the second feature, or simply indicates that the first feature is at a lower horizontal level than the second feature.

[0028] The terms “comprising” and “having”, and any variations thereof, in the specification and claims of this application are intended to cover non-exclusive inclusion, for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or units that are not explicitly listed or that are inherent to such process, method, product, or device.

[0029] To facilitate understanding of the technical solution of this application, the specific structure of the gantry robot will be briefly described below. Specifically, the gantry robot mainly includes columns, beams, rams, end effectors, servo drivers, and transmission mechanisms.

[0030] The uprights are fixed to the workshop floor with anchor bolts at the base, or to the pre-embedded foundation using high-strength connecting plates. To ensure overall stability, four uprights are typically arranged in a rectangular pattern (two at each end), with the tops of the uprights fixed to both ends of the crossbeam via flanges or welding. Leveling bolts are installed between the uprights and the ground to adjust their verticality during installation. The uprights are the fundamental support components of the gantry robot, bearing the weight and dynamic load of the crossbeam and all moving parts, and transferring the load to the foundation.

[0031] Typically, two crossbeams are provided, with each end of a single crossbeam fixedly connected to the top of an end column. The crossbeam serves as a track beam supporting the movement of the ram along the Y-axis, while also providing precise guidance for the horizontal movement of the end effector. The length of the crossbeams determines the Y-axis travel range of the gantry robot, and the distance between the two crossbeams determines the X-axis travel range of the gantry robot.

[0032] The ram is suspended from the guide rail of the crossbeam via a cross slide. A slider that mates with the guide rail is mounted on the back of the cross slide, achieving low-friction sliding by rolling balls or rollers on the guide rail. The ram is the moving component that supports the lifting and lowering motion of the end effector along the Z-axis. The length of the ram determines the Z-axis travel range of the gantry robot.

[0033] The end effector is fixedly mounted to the bottom of the ram via a connecting flange or quick-change interface. The connecting flange has locating pin holes and bolt holes to ensure accurate and repeatable installation of the end effector. The end effector is the actuator that directly contacts the workpiece of the gantry robot, responsible for performing operations such as gripping, picking up, and releasing the workpiece. Common types of end effectors include pneumatic grippers, vacuum suction cups, and electromagnets.

[0034] The transmission mechanism mainly includes an X-axis servo motor, a Y-axis servo motor, a Z-axis servo motor, and several corresponding transmission components. Each axis servo motor drives the end effector to perform movements in the X, Y, and Z axes through its corresponding transmission components. The specific installation method of the transmission mechanism is common knowledge to those skilled in the art, and therefore will not be described in detail here.

[0035] Servo drives are typically installed in a control cabinet near the gantry robot, with the cabinet positioned next to a column or on a workshop wall. The servo drives are electrically connected to the servo motors of each axis via power cables and encoder feedback cables, and connected to the host control system via a communication bus. The servo drives receive control commands from the host control system to control the servo motors of each axis, thereby executing the positional drive of the gantry robot.

[0036] One aspect of this application provides a method for compensating for temperature drift in a gantry robot, such as... Figure 1 As shown, one preferred embodiment includes the following steps: S100: A standard pose database is constructed by teaching the gantry robot to all points in its working area.

[0037] It should be understood that gantry robots are typically used in automated production lines for material handling, machine tool loading and unloading, and other similar scenarios. The working area of ​​a gantry robot refers to the entire spatial range that its end effector can reach during operation. A standard pose database serves as the position reference, and all subsequent compensation amounts are superimposed and corrected with the reference coordinates stored in this database as the zero point; that is, the standard pose database provides a reference for temperature drift compensation.

[0038] S200: Select multiple key points for the gantry robot to move within the work area, collect reference offset data of the gantry robot at each key point at different temperatures, and construct a displacement deviation calibration model.

[0039] It should be understood that key points refer to representative spatial locations within the work area, mainly including the boundary corners of the work area along the X, Y, and Z axes, the geometric center of the work area, and target points that the gantry robot frequently and repeatedly reaches during actual production operations. The selection of these key points can comprehensively reflect the temperature drift characteristics of the entire work area.

[0040] S300: Real-time acquisition of the actual position signal and current ambient temperature signal of the gantry robot.

[0041] S400: Based on the displacement deviation calibration model and the current ambient temperature signal, calculate the offset of each axis of the gantry robot under the current ambient temperature, and determine the corresponding compensation amount according to the current position of the gantry robot.

[0042] It should be understood that the offset of each axis refers to the displacement drift of the X, Y, and Z axes of the gantry robot caused by changes in ambient temperature. Since the three motion axes of the gantry robot differ significantly in structural length, material constraints, and heat source distribution, the thermal deformation characteristics of each axis are different and need to be calculated independently. This step solves the technical problem that a single global model cannot accurately describe the temperature drift characteristics at different spatial locations by applying a discrete displacement deviation calibration model to a continuous workspace and considering the differentiated characteristics of spatial positions, thus improving the compensation accuracy.

[0043] S500: The compensation amount is superimposed with the reference coordinates in the standard pose database to generate the corrected target position, and the corrected target position is converted into a pulse-form compensation command and sent to the servo drive of the gantry robot.

[0044] It should be understood that the superposition of the compensation amount and the standard pose database involves algebraically adding the compensation amount calculated in step S400 to the reference coordinate value of the corresponding target point stored in the standard pose database to generate the corrected target position command. For example, if the target position of a material picking point on the X-axis recorded in the standard pose database is 500.00 mm, and the X-axis compensation amount calculated at the current temperature is +0.07 mm, then the corrected X-axis target position is 500.00 + 0.07 = 500.07 mm. Since the target position command is in units of length, it needs to be converted into pulse commands that the gantry robot's servo drive can recognize and execute, i.e., the required compensation command. The compensation command is sent to the gantry robot's servo drive position loop channel in the form of standard communication protocol frames, such as through real-time industrial Ethernet bus protocols like EtherCAT, PROFINET, and Powerlink. After receiving the compensation command, the servo drive drives the servo motor to move according to the corrected target position in the next control cycle.

[0045] Understandably, this application pre-establishes a standard pose database and displacement deviation calibration model through calibration steps. During subsequent operation, it automatically completes temperature signal acquisition, offset calculation, and compensation command generation and issuance. The entire compensation process requires no manual intervention. Compared to traditional manual calibration methods, this enables the gantry robot to operate continuously and unmanned in a wide temperature range, significantly improving overall equipment efficiency and production line capacity stability.

[0046] This application establishes a quantitative displacement deviation calibration model by performing multi-point calibration within the working area, providing a unified mathematical basis and a repeatable calculation process for compensation calculation. Compared to traditional manual calibration methods, this eliminates the inconsistency in calibration results caused by human error, effectively reducing quality risks such as poor material handling consistency and high scrap rates due to calibration deviations, and making the compensation accuracy quantifiable and traceable.

[0047] This application continuously monitors the ambient temperature and the position signal of the gantry robot, with the compensation calculation step dynamically responding to changes in ambient temperature, and the compensation command taking effect instantly at the servo level. This enables the gantry robot to maintain high-precision positioning for extended periods in industrial scenarios with large day-night temperature differences and significant seasonal temperature variations, eliminating the need for frequent shutdowns for adjustments due to temperature fluctuations. This effectively broadens the applicable environmental range of the gantry robot and meets the continuous production needs of high-precision, large-scale loading and unloading operations.

[0048] In one specific embodiment, when performing step S100, the construction of the standard pose database includes the following process: The operator manipulates the gantry robot to move sequentially to various target points within the work area using a manual or semi-automatic method via the operation control module; these target points include, but are not limited to, material pick-up points, material drop-off points, transfer points, and standby points. When the gantry robot moves to each target point, the spatial coordinate values ​​of each target point in the Cartesian coordinate system are read and recorded, including the X-axis coordinate, Y-axis coordinate, Z-axis coordinate, and possible attitude angles, forming the required standard pose database. This database is stored in the non-volatile memory of the operation control module and supports subsequent read, write, and update operations; the non-volatile memory includes EEPROM, solid-state drives, and industrial-grade memory cards, etc.

[0049] In a specific embodiment, when performing step S200, the construction of the displacement deviation calibration model includes the following process: The operator controls the end effector of the gantry robot to move sequentially to each key point, and measures the actual position coordinates of the end effector when it reaches each key point under multiple different ambient temperature conditions, such as temperature gradients of 10℃, 15℃, 20℃, 25℃, 30℃, 35℃, and 40℃. The actual position coordinates can be read by position detection devices installed on each axis of the gantry robot, including optical scales, magnetic scales, or encoders. The actual position coordinates are compared with the target position coordinates of the key point recorded in the standard pose database, and the deviation value between the two is calculated. This deviation value is used as the reference offset data of the key point at that temperature. By performing curve fitting on the reference offset data collected under the temperature gradient, the law function of the displacement drift of each key point changing with temperature can be obtained, i.e., the required displacement deviation calibration model.

[0050] It is understandable that there are various ways to perform curve fitting on the benchmark offset data, including polynomial fitting, piecewise fitting, and neural network fitting; in this embodiment, a quadratic polynomial model is preferably used to fit the benchmark offset data to obtain the displacement deviation calibration model corresponding to each key point, and the specific expression is as follows: L(T) = a × T2 +b×T+c.

[0051] In the formula, L(T) represents the displacement drift of the gantry robot at temperature T; a, b, and c are all compensation coefficients, obtained through fitting calculations; the least squares method is preferred for fitting, with the goal of minimizing the sum of squared fitting errors at all key points; by solving the objective, the specific values ​​of the compensation coefficients a, b, and c that minimize the overall fitting error can be obtained.

[0052] It is important to know that the quadratic polynomial model can well describe the linear expansion of the corresponding first-order term b×T and the corresponding quadratic term a×T in the thermal deformation of metal structures. 2 The superposition effect of the nonlinear thermal gradient has higher fitting accuracy compared to the linear model, and avoids the risk of overfitting compared to high-order polynomial models. The constant term c is used to compensate for possible system installation deviations at the reference temperature.

[0053] It should be understood that during the long-term operation of a gantry robot, mechanical wear, increased clearance, structural creep, or component replacement can all cause a permanent shift in the correspondence between the target position command of the gantry robot and the actual position of the end effector. Simultaneously, the thermal expansion characteristics of materials, frictional thermal boundary conditions, and lubrication status change slowly over time, rendering the temperature-drift law fitted based on historical data no longer applicable. This leads to a discrepancy between the compensation amount and the actual drift amount, resulting in overcompensation or undercompensation. Therefore, the technical solution of this application establishes a dynamic update mechanism based on temperature changes and operating time for the standard pose database and displacement deviation calibration model. For ease of understanding, this will be described in detail below.

[0054] In one specific embodiment, when the gantry robot is running continuously, if the temperature difference exceeds a set first threshold or the running time exceeds a set second threshold, a dynamic update of the standard pose database and displacement deviation calibration model is triggered. The first threshold refers to the threshold for the change in ambient temperature relative to the reference temperature. The specific value range can be set according to the actual needs of those skilled in the art, for example, it can be 3~6℃; in this embodiment, the first threshold is preferably 5℃. That is, when the temperature difference is within 5℃, the thermal deformation of the gantry robot's structure is usually still within the allowable range of positioning accuracy, and no immediate update is required; when the temperature difference exceeds 5℃, the thermal deformation may exceed the allowable range, requiring an update to maintain model accuracy. The second threshold refers to the threshold for the continuous running time of the gantry robot. The specific value range can be set according to the actual needs of those skilled in the art, for example, it can be 5~10 hours; in this embodiment, the second threshold is preferably 8 hours. That is, after the gantry robot has been running continuously for 8 hours, wear on the mechanical transmission components may cause a slow, irreversible shift in the reference position, requiring an update to eliminate accumulated errors.

[0055] In this embodiment, the specific process of dynamically updating the displacement deviation calibration model is as follows: A portion of key points within the working area are automatically and quickly extracted for position measurement of the gantry robot; typically, the number of key points used for measurement accounts for about 10% of all key points. Based on the position measurement results, a new displacement deviation calibration model is re-fitted and constructed, and the deviation between the new and old displacement deviation calibration models is calculated. If the deviation is less than a preset third threshold, the new displacement deviation calibration model is adopted; otherwise, the old displacement deviation calibration model is still used.

[0056] It is understandable that there are multiple ways to calculate the deviation between the new and old displacement deviation calibration models. In this embodiment, the predicted values ​​of the new and old models are preferred for calculation. Specifically, at multiple temperature points, such as 10℃, 20℃, 30℃, and 40℃, the average or root mean square value of the absolute difference between the predicted values ​​of the new and old models is calculated as the deviation. The specific range of the third threshold can be set according to the actual needs of those skilled in the art. For example, the third threshold can be 0.001~0.01mm. Taking a third threshold of 0.005mm as an example, if the deviation is less than 0.005mm, it indicates that the improvement of the new displacement deviation calibration model compared to the old displacement deviation calibration model is limited, and the new displacement deviation calibration model can be used to replace the old displacement deviation calibration model; if the deviation is greater than or equal to 0.005mm, it indicates that the new displacement deviation calibration model differs significantly from the old displacement deviation calibration model, which may be due to occasional measurement noise or abnormal operating conditions. In this case, the old displacement deviation calibration model is still used to ensure the stability of the system.

[0057] In this embodiment, the specific process of dynamically updating the standard pose database is as follows: Key points within the working area are automatically and quickly extracted for position measurement of the gantry robot; the expression for updating the standard pose database based on the pose measurement results is: P new =λ×P measured +(1-λ) ×P old .

[0058] In the formula, P new P represents the updated standard pose, λ represents the learning rate, and its value can range from [0.1 to 0.5]; measured P represents the current measured pose. old This indicates the standard pose before the standard pose database was updated.

[0059] It is understandable that the learning rate λ controls the magnitude of the correction between the measured value and the standard pose database in a single update. Specifically, a larger learning rate λ, such as 0.4, results in a greater impact of the measured value on the standard pose database, leading to faster updates and making it suitable for conditions with high wear rates. Conversely, a smaller learning rate λ, such as 0.1, results in smoother updates and stronger resistance to noise interference, making it suitable for conditions with slower wear rates or higher sensor noise. By adjusting the value of the learning rate λ, the dynamic update mechanism can achieve optimal performance under different operating conditions.

[0060] In a specific embodiment, during step S300, the actual position signal of the gantry robot is acquired by real-time reading of the actual position of the end effector in each coordinate axis direction using position detection devices installed on each motion axis of the gantry robot. The sampling frequency of the position detection devices is preferably above 1000Hz to ensure timely capture of position changes of the end effector. For the acquisition of the current ambient temperature signal, the ambient temperature value is acquired in real-time using temperature sensors installed at the gantry robot's working site. The temperature sensors are preferably PT100 platinum resistance thermometers, thermocouples, or infrared temperature sensors. The temperature sensors are preferably installed on the gantry robot's crossbeams, columns, or on the slide near the end effector to ensure that the measured temperature accurately reflects the ambient temperature affecting the thermal deformation of the gantry robot structure.

[0061] Understandably, the acquired actual location signal and current ambient temperature signal are preprocessed by the signal conditioning circuit and then transmitted in real time to the compensation algorithm module. The preprocessing methods of the conditioning circuit include amplification, filtering, and analog-to-digital conversion; signal transmission can be achieved via wired or wireless methods; wired methods include RS-485 bus and EtherCAT bus, while wireless methods include industrial Wi-Fi and ZigBee.

[0062] In a specific embodiment, when performing step S400, the calculation of the offset of each axis under the current ambient temperature includes the following process: by substituting the current ambient temperature T0 into the displacement deviation calibration model, the theoretical drift at the current position is calculated. L(T0); then, based on the spatial coordinates of the current position, calculate the theoretical drift amount. L(T0) is used for spatial interpolation correction, for example, by using the trilinear interpolation method to calculate the specific drift of the current point using data from adjacent calibration points, thus obtaining the actual X-axis, Y-axis, and Z-axis offsets at that location. The specific calculation expressions are as follows: I(T0) = α I ×L I ×ΔT + β I ×ΔT 2 ; In the formula, I(T0) represents the displacement drift of the gantry robot's I-axis at the current ambient temperature T0, where I = {X, Y, Z}; α I α represents the coefficient of thermal expansion of a material. For steel truss structures, α is... I Approximately 11 × 10 -6 / ℃;L I ΔT represents the structural length of the I-axis of the gantry robot, ΔT represents the deviation of the current ambient temperature T0 from the reference temperature, and β represents the structural length of the I-axis of the gantry robot. I This represents the nonlinear temperature drift coefficient, which is obtained through experimental calibration.

[0063] It is understandable that in the above expression, α I ×L I ×ΔT is a linear term describing the linear thermal expansion of the structure due to a uniform temperature rise; this linear term is suitable for describing the displacement produced when the structure experiences a uniform temperature rise overall. β I ×ΔT 2 The nonlinear term describes the nonlinear drift component caused by complex factors such as uneven distribution of structural temperature gradients, differences in thermal response rates of different components, and differences in constraint conditions. The introduction of the nonlinear term significantly improves the model's fitting accuracy across the entire temperature range, especially under conditions of large temperature differences, where the nonlinear effect is more pronounced. The reference temperature refers to the ambient temperature value at the time the standard pose database was established; during initial calibration, the ambient temperature at that time is automatically recorded as the initial reference temperature. When the standard pose database is subsequently updated due to mechanical wear, the reference temperature is also updated synchronously to the ambient temperature value at the time of the database update, ensuring that the starting point for ΔT calculation always corresponds to the current baseline state.

[0064] It should be known that the coefficient of thermal expansion of a material is α IThis information can be directly obtained from the materials handbook, but in actual engineering practice, the coefficient of thermal expansion α varies depending on the batch of materials. I The actual values ​​will deviate somewhat from the manual values; furthermore, the effective thermal deformation length of beams and columns is affected by factors such as support method, constraint conditions, and installation stress, and is not simply equal to the geometric length. Therefore, when calculating the offset of each axis, α I ×L I The initial parameters can be calibrated using the compensation coefficient b fitted by the displacement deviation calibration model, thereby reducing the material's thermal expansion coefficient α. I The calculation error caused by the deviation. That is, the compensation coefficient b=α. I ×L I Therefore, when calculating the offset of each axis, the linear term can be directly calculated using b×ΔT.

[0065] It should also be understood that the support stiffness, heat conduction path, and thermal boundary conditions differ at different locations within the gantry robot's working area, leading to variations in temperature drift characteristics at different locations. Therefore, in this embodiment, independent calibration through grid-based partitioning allows each spatial point to obtain a compensation coefficient that precisely matches its local thermal characteristics, thereby improving the overall compensation accuracy across the entire working area; for ease of understanding, a detailed description will follow.

[0066] In this embodiment, the working area of ​​the gantry robot is divided into N, M, and K intervals along the X, Y, and Z axes, respectively, forming N×M×K cubic sub-regions. The selection of the grid size for the sub-regions should balance calibration workload and compensation accuracy: if the grid size is too large, the spatial differences in temperature drift characteristics within the same grid will be ignored, resulting in decreased compensation accuracy; if the grid size is too small, the calibration workload will increase significantly, and insufficient calibration data within each grid may lead to unstable fitting. In a specific example, the grid size is 100mm×100mm×100mm; therefore, for a gantry robot with a working area of ​​2000mm×1500mm×800mm, the number of sub-region grids after division is 20×15×8=2400.

[0067] At each grid vertex in each sub-region, the nonlinear temperature drift coefficient β corresponding to that vertex was experimentally calibrated. I The specific calibration method is as follows: At the vertex position, measure the actual position deviation of each axis at different temperature points, such as 10℃, 20℃, 30℃, and 40℃. Then, perform the inverse calculation using the aforementioned formulas for calculating the offset of each axis, given the known... I(T0), α I L I Given ΔT, we can solve for β inversely. I The value of .

[0068] Specifically, during real-time operation, when the end effector of the gantry robot moves to a certain position within the working area, the upper-level control system first determines the sub-region to which that position belongs; then it reads the β value of the corresponding mesh vertex of that sub-region. I Value; finally, β for the 8 vertices of the grid. I The value is then subjected to trilinear interpolation to obtain the nonlinear temperature drift coefficient β for each axis at that location. X β Y β Z By substituting the nonlinear temperature drift coefficient obtained through interpolation into the offset calculation formula for each axis, the specific offset of that position at the current temperature can be obtained.

[0069] In a specific embodiment, when executing step S400, the expression for calculating the compensation amount based on the offset of each axis is as follows: Compensate=k correction × I(T0)×S filter .

[0070] In the formula, Compensate represents the compensation amount, and k correction Indicates the calibration correction factor. I(T0) represents the displacement drift of the I-axis of the gantry robot at the current ambient temperature T0, S filter This represents the smoothing filter coefficients.

[0071] It is understandable that the calibration correction factor k correction This is a global fine-tuning coefficient used to compensate for potential systematic errors between the theoretical model and the actual physical assembly. For example, during preliminary compensation verification of a gantry robot, if it is found that the measured residual errors of each axis exhibit the same proportional deviation, such as the compensation effect of all axes being approximately 2% off, then k can be adjusted accordingly. correction Adjusting the value to 0.98 achieves a one-time global correction without recalibrating the nonlinear temperature drift coefficients of all mesh vertices. The calibration correction coefficient k is then used. correction The preferred value range is 0.95 to 1.05, and the default value is 1.00 at the factory.

[0072] Smoothing filter coefficients S filter It is a dynamic smoothing coefficient with a value range of [0, 1], used to prevent sudden changes in compensation amount from causing jitter or impact on the gantry robot. In actual working conditions, the ambient temperature may fluctuate instantaneously due to accidental factors such as air conditioning starting and stopping, and workshop doors opening. If the fluctuating temperature value is immediately substituted into the compensation calculation and a large jump compensation command is issued, it will cause the servo motor to frequently accelerate and decelerate, affecting the high-speed stability of loading and unloading. Smoothing filter coefficient S filterIts function is to perform low-pass filtering on temperature signals or compensation quantities, limiting the variation range of a single compensation command.

[0073] It is important to note that the sign of the compensation value varies depending on the operating conditions. Specifically, when the ambient temperature is higher than the reference temperature, the structure expands due to heat, and the end effector drifts in the positive direction; in this case, the compensation value is negative, meaning the end effector is instructed to move in the negative direction to counteract the expansion. When the ambient temperature is lower than the reference temperature, the structure contracts due to cooling, and the end effector drifts in the negative direction; in this case, the compensation value is positive, meaning the end effector is instructed to move in the positive direction to counteract the contraction.

[0074] In a specific embodiment, when executing step S500, the specific process of sending the compensation command to the servo driver is as follows: the corrected target position is converted into a pulse number command that the servo driver can recognize, i.e., the required compensation command, using the following formula; the specific conversion formula is: Pulse = ΔX / (Lead / Pulse) per_revolution ).

[0075] In the formula, Pulse represents the pulse count command, ΔX represents the corrected target position, and Lead represents the lead of the transmission mechanism, the value of which depends on the transmission type. For ball screw drives, Lead is the lead of the screw, that is, the distance the nut or slide moves axially when the screw rotates one revolution; for rack and pinion drives, Lead is the linear distance the rack or the slide connected to it moves when the gear rotates one revolution. per_revolution This indicates the number of pulses output per revolution of the servo motor encoder. Its value depends on the encoder type and its resolution; for example, for incremental encoders, it is usually 2500, 5000, or 10000 pulses / rev.

[0076] After calculating the pulse count command, the pulse count command (Pulse) is encapsulated into a compensation command frame. The preferred format of the command frame includes the following information fields: a frame header for synchronization and identification, an axis number identifying whether the command frame belongs to the X, Y, or Z axis, pulse values ​​indicating the compensation direction and magnitude, and a checksum for detecting communication errors. For multi-axis systems, the compensation command frames for the X, Y, and Z axes are generated and issued independently, without affecting each other.

[0077] Similarly, the corresponding reference coordinates in the standard pose database can be converted into pulse count instructions that the servo driver can recognize. Finally, the pulse count instructions corresponding to the compensation amount and the pulse count instructions corresponding to the reference coordinates are superimposed inside the servo driver to generate the required compensation instructions.

[0078] It is understood that, in another embodiment, for obtaining the compensation command, the compensation amount and the corresponding reference coordinates in the standard pose database can be converted into pulse counts using the above conversion formula to obtain the first pulse count command corresponding to the compensation amount and the second pulse count command corresponding to the reference coordinates; then the first pulse count command and the second pulse count command are superimposed inside the servo driver to generate the required compensation command.

[0079] It should be understood that the compensation command obtained in step S500 is based on feedforward compensation of the displacement deviation calibration model and has predictive properties. Although this feedforward compensation can offset most of the temperature drift error, due to nonlinear factors of the physical structure, such as uneven thermal deformation caused by local frictional heat of the guide rail, nonlinear hysteresis of bearing clearance, and fluctuation of thermal expansion coefficient caused by batch differences in materials, the end effector will still have residual positioning error after feedforward compensation. This positioning error mainly includes structural thermal deformation components and transmission chain drift components. Therefore, in this embodiment, while performing real-time position correction of the gantry robot through compensation commands, it is also necessary to compensate for the positioning error, including structural thermal deformation components and transmission chain drift components. There are various algorithms for compensating for positioning errors, including BP neural networks, fuzzy PID, Kalman filtering, LSTM algorithms, etc. For ease of understanding, the Kalman filtering algorithm will be used as an example to describe the specific compensation process of positioning error in detail below.

[0080] In a specific embodiment, the process of compensating for positioning errors is as follows: (1) Obtain the motor-end feedback signal of the gantry robot as a semi-closed-loop feedback quantity, and simultaneously obtain the end effector feedback signal of the gantry robot as a full-closed-loop feedback quantity. Among them, the motor-end feedback signal refers to the pulse signal output by the rotary encoder installed on the rear shaft of the servo motor. This signal directly reflects the rotation angle and speed of the motor rotor, and the theoretical position of the slide can be indirectly calculated through the transmission ratio. The end effector feedback signal refers to the signal output by the position detection device set near the end effector of the gantry robot or directly detecting its spatial coordinates. This signal directly reflects the actual spatial position of the end effector.

[0081] (2) Kalman filtering is used to fuse the semi-closed-loop feedback and full-closed-loop feedback of the gantry robot to estimate the structural thermal deformation component and the transmission chain drift component.

[0082] It's important to know that Kalman filtering is an optimal state estimation algorithm that can fuse multi-sensor information to provide the optimal estimate of the system state, even in the presence of measurement and process noise. In this embodiment, the specific implementation of Kalman filtering is as follows: Construct the state vector X = [position, velocity, drift] TWhere position represents the position of the end effector, velocity represents the rate of change of the end effector's position, and drift represents the drift of the end effector. Construct the observation vector Z = [P motor P end ] T Among them, P motor P indicates the encoder feedback position at the motor end. end This indicates the position of the end effector's full closed-loop feedback.

[0083] State equation: X(k) = A × X(k-1) + B × u(k) + w(k); where X(k) represents the state vector of the current k-th control cycle, X(k-1) represents the state vector of the (k-1)-th control cycle, A represents the state transition matrix, B represents the control input matrix, u(k) represents the control input of the current k-th control cycle, such as the speed command; w(k) represents the process noise of the current k-th control cycle, which follows a normal distribution with zero mean and covariance Q.

[0084] The observation equation is: Z(k) = H × X(k) + v(k); where Z(k) represents the observation vector of the first k-th control period, H represents the observation matrix, and v(k) represents the observation noise, which follows a normal distribution with zero mean and covariance R.

[0085] The iterative calculation of Kalman filtering includes the following four steps: State prediction: X hat_minus (k)=A×X hat (k-1)+B×u(k); where X hat_minus (k) represents the predicted state vector value for the current k-th control cycle, X hat (k-1) represents the state vector update value in the (k-1)th control cycle.

[0086] Covariance prediction: P(k) = A × P(k-1) × A T +Q; where P(k) and P(k-1) represent the predicted covariance values ​​for the current k-th control cycle and the (k-1)-th control cycle, respectively.

[0087] Kalman gain calculation: K(k) = P(k) × H T ×[H×P(k)×H T +R] -1 Where K(k) represents the Kalman gain of the current k-th control cycle.

[0088] Status Update: X hat (k)= X hat_minus (k)+K(k)×[Z(k)-H×X hat_minus (k)];where Xhat (k) represents the state vector update value of the current k-th control cycle.

[0089] Through the above iterative calculations, the optimal estimates of the actual position and drift of the end effector are obtained, thereby obtaining the positioning error Etotal for the current control cycle.

[0090] (3) Decompose the positioning error Etotal into the structural thermal deformation component E structure and transmission chain drift component E transmission The basis for the decomposition is that different error components have different physical properties and frequency characteristics: the structural thermal deformation component belongs to low-frequency components, while some components in the transmission chain drift component may have higher frequencies.

[0091] It is understandable that the structural thermal deformation component E structure The calculation formula is: E structure =E beam +E column +E slide .

[0092] E beam =alpha beam ×L beam × T×cosθ.

[0093] E column =alpha column ×L column × T×sin .

[0094] E slide =alpha slide ×L slide × T.

[0095] In the formula, E beam E represents the amount of thermal deformation of the beam. column E represents the amount of thermal deformation of the column. slide alpha represents the thermal deformation of the ram. beam alpha column alpha slide L represents the coefficients of thermal expansion of the beam material, column material, and ram material, respectively. beam L represents the effective structural length of the beam. column L represents the effective height of the column. slide Indicates the effective extension length of the bolster. T represents the deviation of the current ambient temperature from the reference temperature, and θ represents the angle between the direction of thermal deformation of the beam and the sensitive direction of the end effector. This indicates the angle between the direction of thermal deformation of the column and the sensitive direction of the end effector.

[0096] Transmission chain drift component E transmission The calculation formula is: E transmission =E rack +E gear +E backlash .

[0097] E rack =alpha rack ×L rack × T.

[0098] E gear =alpha gear ×D× T.

[0099] E backlash = k backlash × T.

[0100] In the formula, E rack E represents the amount of thermal deformation of the rack. gear E represents the change in gear meshing clearance. backlash alpha represents the bearing return error. rack L represents the coefficient of thermal expansion of the rack material. rack alpha represents the effective length of the rack. gear The coefficient of thermal expansion of the gear material is represented by D, where D represents the tooth thickness and k represents the gear material. backlash This represents the temperature coefficient of return error.

[0101] (4) The structural thermal deformation component estimated by Kalman filtering is compensated by correcting the forward kinematics parameters. Specifically, the structural thermal deformation component manifests as the change in the size and shape of each structural component of the gantry robot with temperature. This is equivalent to the slow change in geometric parameters in the robot's kinematic model, such as the length of each axis, offset, and zero joint angle. Therefore, after estimating the structural thermal deformation component by Kalman filtering, it is converted into forward kinematics parameters, such as the correction amount of link length and joint offset in the dh parameters, thus eliminating the influence of structural thermal deformation at the kinematic calculation level. This compensation method does not introduce additional high-frequency noise and is suitable for handling low-frequency structural thermal deformation errors.

[0102] (5) The transmission chain drift component estimated by Kalman filtering is superimposed on the PID output loop of the servo drive of the gantry robot in the form of a feedforward compensation. Specifically, the transmission chain drift component mainly manifests as errors in the transmission mechanism such as rack thermal expansion and gear backlash changes. These errors are characterized by high frequency and small amplitude. By directly superimposing the feedforward correction at the output of the PID controller, the error changes in the transmission chain can be quickly responded to at the control level, achieving high-frequency, real-time compensation. The compensation update frequency of the transmission chain drift component is 1000Hz, that is, the correction is updated every 1ms to match the fast response capability of the servo drive. The correction is smoothed using S-shaped speed planning, with a transition time of 50ms to 100ms, to achieve smooth compensation without impact or jitter.

[0103] It is understandable that the feedforward compensation command obtained through step S500 and the compensation from the Kalman filter feedback are calculated in parallel and implemented independently. That is, the input of the feedforward compensation command is the temperature signal, and its output is a pulse command; the input of the compensation from the Kalman filter feedback is the encoder signal at the motor end and the feedback signal from the end effector, and its output is the correction amount after decomposition of the positioning error. The two are independent of each other and are ultimately algebraically superimposed in the position loop adder of the servo drive.

[0104] In a specific embodiment, after the compensation command is issued, the compensation effect needs to be verified. This includes the following process: after each compensation command is executed, the feedback position of the gantry robot's end effector is read; the deviation between the feedback position and the target position is calculated as the residual error. When the absolute value of the residual error is greater than a preset fourth threshold, it indicates that there is still a residual deviation exceeding the allowable range after the combined action of feedforward compensation and Kalman filter feedback compensation. This residual deviation may originate from: long-term drift of model parameters, sudden external disturbances, sensor performance degradation, etc. At this point, a secondary calibration will be triggered, requiring a multi-point, multi-temperature gradient accuracy calibration of the gantry robot's working area to update the displacement deviation calibration model.

[0105] Understandably, the specific value of the fourth threshold depends on the accuracy requirements of the specific industrial application scenario. For example, for general loading and unloading scenarios, when the positioning accuracy requirement is within ±0.05mm, the fourth threshold can be set to 0.05mm; for high-precision assembly scenarios, when the positioning accuracy requirement is within ±0.01mm, the fourth threshold can be set to 0.01mm. Preferably, the value of the fourth threshold is in the range of 0.01mm to 0.05mm, and the specific value can be determined through experimental calibration.

[0106] The secondary calibration is preferably performed under one of the following circumstances: during production line shift changes, in the next idle cycle after the system detects that the residual error exceeds the threshold, or manually initiated by the operator through the human-machine interface. After the secondary calibration is completed, the system replaces the old displacement deviation calibration model with the new one to restore the compensation accuracy.

[0107] It is important to note that the aforementioned dynamic update mechanism for standard pose data and displacement deviation calibration models is an incremental update, used for local fine-tuning of the model; while the secondary calibration in this embodiment is a global recalibration, used for comprehensive remeasurement and model reconstruction of all key points across the entire working area. The two complement each other: the dynamic update mechanism is suitable for slow drift compensation during daily operation, while the secondary calibration mechanism is suitable for system recovery after residual errors exceed a threshold.

[0108] Another aspect of this application provides a system for implementing the above-described gantry robot temperature drift compensation method, such as... Figure 2 As shown, one preferred embodiment includes an operation control module, a detection module, and a compensation algorithm module.

[0109] The operation control module is configured to allow operators to perform teaching calibration, parameter configuration, compensation enable, status monitoring, and point storage, and to send compensation commands to the servo drive of the gantry robot.

[0110] Specifically, the operation control module can be an industrial computer, a programmable logic controller, an embedded motion controller, or a dedicated CNC system. The operation control module is equipped with a human-machine interface, which can be a touchscreen, host computer software, or a handheld teach pendant. Through this interface, operators can perform the following operations: manually or semi-automatically record the teaching calibration of target point coordinates; set parameter configurations such as mesh size, learning rate λ, and various threshold parameters; enable or disable temperature drift compensation; monitor the status of operating data such as current temperature, axis offset, compensation amount, and residual error in real time; and view and edit the point storage in the standard pose database. The operation control module can receive the compensation amount sent by the compensation algorithm module, superimpose it with the reference coordinates in the standard pose database, and convert it into pulse form to be sent to the servo drive of the gantry robot.

[0111] The detection module is configured to collect ambient temperature signals in real time, collect the actual position signals of the gantry robot's end effector in real time, and perform full-domain accuracy calibration of the gantry robot's working area at multiple points and with multiple temperature gradients.

[0112] Specifically, the detection module includes a sensor unit and a signal processing unit. The sensor unit includes a temperature detection device and a position detection device. The temperature detection device uses PT100 platinum resistance thermometers, K-type thermocouples, and infrared temperature sensors, and is installed on the crossbeams or columns of the gantry robot to measure ambient temperature in real time. The position sensors use incremental optical encoders, absolute magnetic encoders, laser displacement sensors, and incremental or absolute rotary encoders, and are installed at each motion axis and end effector of the gantry robot to measure position signals in real time. The signal processing unit includes a signal conditioning circuit, used to convert the analog or digital signals output by the sensor units into a format that the compensation algorithm module can process.

[0113] The compensation algorithm module is configured to receive the output of the detection module, construct the displacement deviation calibration model, and generate compensation instructions.

[0114] Specifically, the compensation algorithm module can be deployed within the operation control module, such as as an embedded firmware algorithm for a motion controller, running on an ARM or DSP processor; or it can be deployed in a separate computing unit, such as an industrial edge computing device, an industrial PC, or a high-speed computing board based on an FPGA. The compensation algorithm module interacts with the operation control module and the detection module via a communication bus. Internally, the compensation algorithm module includes: a calibration model storage unit, an interpolation calculation unit, a compensation instruction generation unit, a Kalman filtering unit, and a dynamic update unit; wherein, the calibration model storage unit stores the compensation coefficients a, b, and c of each mesh vertex, as well as the nonlinear temperature drift coefficient β. I The interpolation calculation unit performs trilinear interpolation calculations, the compensation command generation unit performs offset calculations for each axis and compensation calculations, the Kalman filtering unit performs semi-closed-loop and full-closed-loop feedback fusion estimation, and the dynamic update unit performs dynamic updates to the model and database. The compensation algorithm module can send the obtained compensation values ​​to the operation control module.

[0115] Understandably, the operation control module, detection module, and compensation algorithm module work together to form a closed-loop temperature drift compensation architecture, enabling the gantry robot to automatically compensate and correct its own positional deviation under different temperature environments, achieving wide temperature range environmental adaptability; the compensation algorithm module calculates the compensation range of the working area based on sensor signals and calibration results, corrects the positioning error of the gantry robot in real time, and improves motion accuracy and environmental adaptability.

[0116] Another aspect of this application provides a computer-readable storage medium, in a preferred embodiment of which a computer program is stored on the storage medium; when the computer program is executed by a processor, it implements the above-described gantry robot temperature drift compensation method.

[0117] The basic principles, main features, and advantages of this application have been described above. Those skilled in the art should understand that this application is not limited to the above embodiments. The embodiments and descriptions in the specification are merely the principles of this application. Various changes and modifications can be made to this application without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claims. The scope of protection claimed by this application is defined by the appended claims and their equivalents.

Claims

1. A method for compensating for temperature drift in a gantry robot, characterized in that, Includes the following steps: A standard pose database is constructed by teaching the gantry robot to all points within its working area. Select multiple key points for the gantry robot to move within the work area, collect reference offset data of the gantry robot at each key point at different temperatures, and construct a displacement deviation calibration model; Real-time acquisition of the actual position signal and current ambient temperature signal of the gantry robot; Based on the displacement deviation calibration model and the current ambient temperature signal, the offset of each axis of the gantry robot under the current ambient temperature is calculated, and the corresponding compensation amount is determined according to the current position of the gantry robot. The compensation amount is superimposed on the reference coordinates in the standard pose database to generate the corrected target position. The corrected target position is then converted into a pulse-form compensation command and sent to the servo drive of the gantry robot.

2. The gantry robot temperature drift compensation method as described in claim 1, characterized in that, The construction of the displacement deviation calibration model includes the following process: The deviation between the actual position and the target position of the gantry robot at each key point is measured at different temperatures and used as the reference offset data; A quadratic polynomial model was used to fit the reference offset data to obtain the displacement deviation calibration model corresponding to each key point. The specific expression is as follows: L(T)=a×T 2 +b×T+c; In the formula, L(T) represents the displacement drift of the gantry robot at temperature T. a, b, and c are compensation coefficients, which are obtained through fitting calculation.

3. The gantry robot temperature drift compensation method as described in claim 2, characterized in that, During the operation of the gantry robot, when the temperature difference exceeds a set first threshold or the running time exceeds a set second threshold, a dynamic update of the standard pose database and displacement deviation calibration model is triggered; the specific process of the dynamic update is as follows: The pose and position of the gantry robot were measured at some key points within the working area. Based on the measured position results, a new displacement deviation calibration model is refitted and constructed. The deviation between the new displacement deviation calibration model and the old displacement deviation calibration model is calculated. If the deviation is less than the preset third threshold, the new displacement deviation calibration model is adopted; otherwise, the old displacement deviation calibration model is still adopted. The expression for updating the standard pose database based on the actual pose measurement results is as follows: P new =λ×P measured +(1-λ) ×P old ; In the formula, P new Let P represent the updated standard pose, λ represent the learning rate, and P represent the standard pose. measured P represents the current measured pose. old This indicates the standard pose before the standard pose database was updated.

4. The gantry robot temperature drift compensation method as described in claim 2, characterized in that, The calculation of the offset for each axis includes the following process: Substitute the current ambient temperature into the displacement deviation calibration model to calculate the theoretical drift at the current position. Based on the spatial coordinates of the current position, spatial interpolation correction is performed on the theoretical drift amount to obtain the actual X-axis, Y-axis, and Z-axis offsets at the current position. The specific calculation expressions are as follows: I(T0) = α I ×L I ×ΔT + β I ×ΔT 2 ; In the formula, I(T0) represents the displacement drift of the gantry robot's I-axis at the current ambient temperature T0, where I = {X, Y, Z}; α I L represents the coefficient of thermal expansion of a material. I The structural length of the I-axis of the gantry robot is represented by ΔT, where ΔT represents the deviation of the current ambient temperature T0 from the reference temperature, and β is the structural length of the I-axis of the gantry robot. I This represents the nonlinear temperature drift coefficient.

5. The gantry robot temperature drift compensation method as described in claim 4, characterized in that, When calculating the offset of each axis, α I ×L I The initial parameters are calibrated by the compensation coefficient b fitted by the displacement deviation calibration model; The workspace of the gantry robot was divided into multiple sub-regions, and the nonlinear temperature drift coefficient β of the mesh vertices in each sub-region was individually calibrated experimentally. I During the real-time operation of the gantry robot, based on the sub-region where the current position is located, the nonlinear temperature drift coefficient β matching the current position is obtained through trilinear interpolation. I .

6. The gantry robot temperature drift compensation method as described in claim 4, characterized in that, The expression for calculating the compensation amount based on the offset of each axis is as follows: Compensate=k correction × I(T0)×S filter ; In the formula, Compensate represents the compensation amount, and k correction Indicates the calibration correction factor. I(T0) represents the displacement drift of the I-axis of the gantry robot at the current ambient temperature T0, S filter This represents the smoothing filter coefficients.

7. The gantry robot temperature drift compensation method according to any one of claims 1-6, characterized in that, While sending compensation commands to the servo drive of the gantry robot for real-time position correction, compensation is also made for positioning errors, including structural thermal deformation components and transmission chain drift components. The specific compensation process is as follows: The feedback signal from the motor end of the gantry robot is obtained as a semi-closed-loop feedback quantity, and the feedback signal from the end effector of the gantry robot is obtained as a full-closed-loop feedback quantity. Kalman filtering is used to fuse the semi-closed-loop and full-closed-loop feedback of the gantry robot to estimate the structural thermal deformation component and the transmission chain drift component. The structural thermal deformation component estimated by Kalman filtering is compensated by correcting the forward kinematics parameters; The drive train drift component estimated by Kalman filtering is superimposed on the PID output loop of the servo driver of the gantry robot in the form of a feedforward compensation.

8. The gantry robot temperature drift compensation method as described in claim 1, characterized in that, After the compensation instruction is issued, the compensation effect is verified, which includes the following process: After each compensation command is executed, read the feedback position of the end effector of the gantry robot; The deviation between the feedback position and the target position is calculated as the residual error; When the absolute value of the residual error is greater than the preset fourth threshold, the working area of ​​the gantry robot is recalibrated with multiple points and multiple temperature gradients to update the displacement deviation calibration model.

9. A system for implementing the temperature drift compensation method for a gantry robot according to any one of claims 1-8, characterized in that, include: Operation control module; The operation control module is configured to allow operators to perform teaching calibration, parameter configuration, compensation enable, status monitoring and point storage, and to send compensation commands to the servo drive of the gantry robot. Detection module; The detection module is configured to collect ambient temperature signals in real time, collect the actual position signals of the gantry robot's end effector in real time, and perform full-domain accuracy calibration of the gantry robot's working area at multiple points and with multiple temperature gradients. The compensation algorithm module is configured to receive the output of the detection module, construct a displacement deviation calibration model, and generate compensation instructions.

10. A computer-readable storage medium, characterized in that, The storage medium stores a computer program; when the computer program is executed by a processor, it implements the gantry robot temperature drift compensation method as described in any one of claims 1-8.

Images