Robot aluminum upright running stability control method based on vibration detection

Abstract

The present application belongs to the technical field of robot operation control, and particularly relates to a robot aluminum column operation stability control method based on vibration detection, comprising: obtaining robot chassis vibration data, column upper vibration data and aluminum column lifting state, using position-related vibration propagation relationship and lifting motion-related vibration relationship to respectively form chassis propagation vibration action and lifting motion-related vibration action, and combining prediction error limit to determine vibration action path state; accordingly forming a candidate lifting trajectory, respectively calculating chassis propagation vibration cumulative action, lifting motion-related vibration cumulative action and predicted column upper vibration cumulative action of the candidate lifting trajectory and the reference lifting trajectory, and only when the candidate lifting trajectory current dominant vibration cumulative action and the predicted column upper vibration cumulative action are lower than the corresponding values of the reference lifting trajectory, the candidate lifting trajectory is executed, thereby improving the aluminum column lifting operation stability.

Description

Technical Field

[0001] This invention relates to the field of robot operation control technology. More specifically, this invention relates to a method for controlling the operational stability of a robot aluminum column based on vibration detection. Background Technology

[0002] In robots with a chassis and an aluminum column that can extend and retract vertically relative to the chassis, the extension position of the aluminum column can be changed by a lifting drive mechanism, thereby adjusting the working height of the robot's working components located on the upper part of the aluminum column. During robot operation, the movement of the robot chassis, the lifting and lowering movement of the aluminum column, and the extension and retraction structure of the aluminum column itself all affect the vibration state of the upper part of the column. Therefore, the vibration of the upper part of the column is one of the factors affecting the stability of the robot's lifting and lowering operation.

[0003] To mitigate vibration during the lifting process, one control approach is to detect the magnitude of vibration during robot operation and adjust the lifting speed of the aluminum column accordingly. For example, the lifting speed can be reduced when increased vibration is detected to decrease the vibration generated by the lifting motion, or the lifting speed can be adjusted based on different operating conditions to ensure the robot maintains a relatively stable operating state while meeting operational requirements.

[0004] However, during the lifting and lowering of the aluminum column, the operating status of the robot chassis, the extension position of the aluminum column, and the lifting and lowering motion of the aluminum column itself can all affect the vibration level of the upper part of the column. The same adjustment of the lifting speed may have different effects on the vibration of the upper part of the column depending on the extension position and operating status. Therefore, directly adjusting the lifting speed based solely on the currently detected overall vibration magnitude makes it difficult to determine whether the adjusted lifting process can improve the overall vibration of the upper part of the column. Summary of the Invention

[0005] To address the technical problem that existing methods of adjusting lifting speed based on vibration magnitude are insufficient to stably reduce vibration at the top of the robot's aluminum column, this invention provides a vibration detection-based method for controlling the operational stability of the robot's aluminum column, comprising: Vibration data of the chassis and the upper part of the column are collected to obtain the extension position, lifting speed, lifting acceleration, and lifting jerk. Based on the chassis vibration data and the extension position, the vibration action of the chassis is generated through the position-related vibration propagation relationship. The net vibration power spectral density of the upper part of the column is summed at frequency intervals to form the measured vibration action of the upper part of the column. Based on the extension position, lifting speed, lifting acceleration, and lifting jerk, the vibration action of the lifting motion is generated through the lifting motion-related vibration relationship. The vibration path state is determined based on the vibration action of the chassis, the vibration action of the lifting motion-related vibration, and the measured vibration action of the upper part of the column. A baseline lifting trajectory is formed based on the current extension position, lifting speed, lifting acceleration, and target extension position. The lifting speed is then adjusted according to the vibration path state to generate candidate lifting trajectories, ensuring that the candidate lifting trajectories have the same start and end motion states as the baseline lifting trajectory. Based on the execution time from the initial extension position to the predicted termination position for both the baseline and candidate lifting trajectories, the predicted chassis-propagated vibration and lifting motion-related vibration are accumulated to form the cumulative chassis-propagated vibration and lifting motion-related vibration for each of the baseline and candidate lifting trajectories. The sum of the cumulative chassis-propagated vibration and lifting motion-related vibration for the same lifting trajectory is used as the predicted cumulative vibration of the upper part of the column. The upper and lower bounds of the predicted vibration cumulative effect are formed based on the chassis propagation prediction error limit, the lifting motion-related prediction error limit, and the joint prediction error limit. Based on the vibration path state and the predicted upper and lower bounds of the cumulative vibration, execute the candidate rise / fall trajectory or the baseline rise / fall trajectory.

[0006] Preferably, the step of collecting chassis vibration data and upper column vibration data to obtain extension position, lifting speed, lifting acceleration, and lifting jerk includes: A first three-axis acceleration detection unit is set at the connection between the chassis and the aluminum column on the side of the robot chassis, and a second three-axis acceleration detection unit is set at the connection between the uppermost movable structure of the aluminum column and the robot's working parts. The intersection of the effective frequency ranges of the first and second three-axis acceleration detection units is taken as the common effective detection frequency band. The minimum sampling frequency that can cover the upper limit of the common effective detection frequency band is selected. The minimum number of integer sampling points that can cover a complete cycle of the lowest frequency corresponding to the lower limit of the common effective detection frequency band is taken as the number of sampling points of the analysis sub-window. Three consecutive, equal-length and non-overlapping analysis sub-windows constitute a complete vibration observation window. After removing the DC component from the acceleration sequence of the three detection axes in each analysis sub-window, the single-sided power spectral density is calculated. The single-sided power spectral densities of the three detection axes of the same three-axis acceleration detection unit at the same frequency point are added to form the total vibration power spectral density. The extension position, lifting speed and lifting acceleration are recorded synchronously, and the lifting acceleration is formed according to the adjacent lifting acceleration and time interval before and after the time center of the analysis sub-window.

[0007] Preferably, the method for obtaining the position-related vibration propagation relationship is as follows: The lower and upper limits of the working stroke of the aluminum column and the starting and ending points of the extension of each level of the telescopic column are calibrated to form a calibration extension position table. The total vibration power spectral density is collected at each calibration extension position under both static and chassis calibration motion conditions. For each frequency point, in all static states, the maximum value of the total vibration power spectral density of the first triaxial acceleration detection unit at that frequency point is taken as the chassis static background reference power spectral density at that frequency point, and the maximum value of the total vibration power spectral density of the second triaxial acceleration detection unit at that frequency point is taken as the upper static background reference power spectral density of the column at that frequency point. At each frequency point, the average total vibration power spectral density of the first triaxial acceleration detection unit under chassis calibration motion conditions is subtracted from the chassis static background at that frequency point. The reference power spectral density is subtracted, and when the result is less than zero, it is taken as zero to form the net vibration power spectral density of the chassis at that frequency point. At each frequency point, the average total vibration power spectral density of the second and third axis acceleration detection units under the chassis calibration motion state is subtracted from the reference power spectral density of the stationary background of the upper part of the column at that frequency point, and when the subtraction result is less than zero, it is taken as zero to form the net vibration power spectral density of the upper part of the column at that frequency point. The frequency points where the net vibration power spectral density of the chassis and the net vibration power spectral density of the upper part of the column are both non-zero at all calibration extension positions are taken as effective frequency points. The ratio of the net vibration power spectral density of the upper part of the column to the net vibration power spectral density of the chassis at the same calibration extension position is taken as the vibration power propagation coefficient. The vibration power propagation coefficient is linearly interpolated between adjacent calibration extension positions according to the extension position.

[0008] Preferably, the step of forming the chassis-propagated vibration magnitude based on chassis vibration data and extension position through position-related vibration propagation relationships includes: Obtain the net vibration power spectral density of the chassis at the effective frequency point of the analysis sub-window; obtain the vibration power propagation coefficient of the extension position corresponding to the time center of the analysis sub-window at the effective frequency point; sum the products of the net vibration power spectral density of the chassis, the vibration power propagation coefficient, and the frequency interval at each effective frequency point to obtain the chassis propagation vibration effect of the analysis sub-window; take the arithmetic mean of the chassis propagation vibration effects of the three analysis sub-windows within the complete vibration observation window as the chassis propagation vibration effect of the complete vibration observation window; sum the products of the net vibration power spectral density of the chassis at the effective frequency point of the current complete vibration observation window, the vibration power propagation coefficient of the current extension position at the effective frequency point, and the frequency interval as the predicted chassis propagation vibration effect.

[0009] Preferably, the summation of the net vibration power spectral density of the upper part of the column at effective frequency points according to frequency intervals to form the measured vibration action of the upper part of the column includes: Obtain the net vibration power spectral density of the upper part of the column at the effective frequency point of the analysis sub-window; sum the products of the net vibration power spectral density of the upper part of the column at each effective frequency point and the frequency interval to obtain the measured vibration action of the upper part of the column in the analysis sub-window; take the arithmetic mean of the measured vibration action of the upper part of the column in the three analysis sub-windows within the complete vibration observation window as the measured vibration action of the upper part of the column in the complete vibration observation window.

[0010] Preferably, the method for obtaining the vibration relationship associated with the lifting motion is as follows: With the robot chassis stationary, the aluminum column is controlled to execute a lifting calibration trajectory, and lifting motion calibration samples are collected. A stationary state sample is collected at the calibration extension position as a zero-motion calibration sample. Using the square of the lifting velocity, the square of the lifting acceleration, and the square of the lifting jerk as fitting variables, and the measured vibration of the upper part of the column as the fitting target, the fitting coefficients for the constant term, the square of the lifting velocity, the square of the lifting acceleration, and the square of the lifting jerk are determined. The calibration extension position that uniquely determines the four fitting coefficients is taken as the valid calibration position. The four fitting coefficients between adjacent valid calibration positions are linearly interpolated according to the extension position. The lifting calibration trajectory is re-executed to obtain verification data. The lifting direction appearing simultaneously in both the lifting motion calibration samples and the verification data is taken as the verified lifting direction.

[0011] Preferably, determining the vibration path state based on the vibration action amount propagated from the chassis, the vibration action amount associated with the lifting motion, and the measured vibration action amount on the upper part of the column includes: When the complete vibration observation window data is valid and the absolute difference between the measured vibration action of the upper part of the column and the sum of the vibration action of the chassis propagation and the vibration action of the lifting motion is not greater than the joint prediction error limit, the chassis propagation vibration action range is formed by taking the maximum value between the chassis propagation vibration action range minus the chassis propagation prediction error limit and zero as the lower limit and the chassis propagation vibration action range plus the chassis propagation prediction error limit as the upper limit; the lifting motion-related vibration action range is formed by taking the maximum value between the lifting motion-related vibration action range minus the lifting motion-related prediction error limit and zero as the lower limit and the lifting motion-related vibration action range plus the lifting motion-related prediction error limit as the upper limit; when the lower limit of the lifting motion-related vibration action range is greater than the upper limit of the chassis propagation vibration action range, it is determined to be the lifting motion-related dominant state; when the lower limit of the chassis propagation vibration action range is greater than the upper limit of the lifting motion-related vibration action range, it is determined to be the chassis propagation dominant state; otherwise, it is determined to be an indeterminate state.

[0012] Preferably, the step of forming a reference lifting trajectory based on the current extension position, lifting speed, lifting acceleration, and target extension position, and adjusting the lifting speed according to the vibration path state to generate candidate lifting trajectories, so that the candidate lifting trajectories have the same start and end motion states as the reference lifting trajectory, includes: The subsequent lifting direction is determined along the baseline lifting trajectory. When the subsequent lifting direction is a verified lifting direction, the first of the target extension position and the first calibrated extension position reached along the subsequent lifting direction is taken as the predicted termination position. The distance from the initial extension position to the predicted termination position is divided by the time required for the baseline lifting trajectory to reach the predicted termination position for the first time from the start time of the candidate lifting trajectory to obtain the baseline average lifting speed. Under the lifting motion association-dominant state, the lifting speed that is allowed to be set by the lifting drive controller, is consistent with the subsequent lifting direction, and has an absolute value of lifting speed less than the baseline average lifting speed is selected, and the absolute value of the selected lifting speed is taken as the candidate average lifting speed. Under the chassis propagation-dominant state, the lifting speed that is allowed to be set by the lifting drive controller, is consistent with the subsequent lifting direction, and has an absolute value of lifting speed greater than the baseline average lifting speed is selected, and the absolute value of the selected lifting speed is taken as the candidate average lifting speed. The extension position, lifting speed, and lifting acceleration of the baseline lifting trajectory at the start time of the candidate lifting trajectory and when it first reaches the predicted termination position are used as boundary conditions to form the candidate lifting trajectory.

[0013] Preferably, the step of executing the candidate rising / falling trajectory or the reference rising / falling trajectory based on the predicted upper and lower bounds of the vibration action path state and the cumulative vibration action includes: In the state dominated by lifting motion association, if the predicted upper bound of the cumulative vibration of the lifting motion association of the candidate lifting trajectory is less than the predicted lower bound of the cumulative vibration of the lifting motion association of the reference lifting trajectory, and the predicted upper bound of the cumulative vibration of the upper part of the column of the candidate lifting trajectory is less than the predicted lower bound of the cumulative vibration of the upper part of the column of the reference lifting trajectory, the candidate lifting trajectory is executed; otherwise, the reference lifting trajectory is executed. In the state dominated by chassis propagation, if the predicted upper bound of the cumulative vibration of the chassis propagation of the candidate lifting trajectory is less than the predicted lower bound of the cumulative vibration of the chassis propagation of the reference lifting trajectory, and the predicted upper bound of the cumulative vibration of the upper part of the column of the candidate lifting trajectory is less than the predicted lower bound of the cumulative vibration of the upper part of the column of the reference lifting trajectory, the candidate lifting trajectory is executed; otherwise, the reference lifting trajectory is executed.

[0014] Preferably, the method for obtaining the chassis propagation prediction error limit, the lifting motion correlation prediction error limit, and the joint prediction error limit is as follows: The chassis calibration motion is re-executed at each calibration extension position to obtain chassis propagation verification data. The maximum absolute difference between the chassis propagation vibration magnitude and the measured vibration magnitude at the top of the column in the chassis propagation verification data is taken as the chassis propagation prediction error limit. The lifting calibration trajectory is re-executed with the robot chassis stationary to obtain lifting motion correlation verification data. The maximum absolute difference between the lifting motion correlation vibration magnitude and the measured vibration magnitude at the top of the column in the lifting motion correlation verification data is taken as the lifting motion correlation prediction error limit. Joint verification data is obtained under the condition that robot chassis movement and aluminum column lifting exist simultaneously. The maximum absolute difference between the measured vibration magnitude at the top of the column and the sum of the chassis propagation vibration magnitude and the lifting motion correlation vibration magnitude in the joint verification data is taken as the joint prediction error limit.

[0015] The beneficial effects of this invention are as follows: This invention characterizes the vibration effect formed by the propagation of the robot chassis vibration through the current aluminum column extension structure to the upper part of the column and the vibration effect associated with the lifting motion of the aluminum column by the position-related vibration propagation relationship and the lifting motion-related vibration relationship, respectively. It also determines the current vibration path state by combining the measured vibration effect of the upper part of the column and the prediction error limit, so that the lifting control no longer directly selects acceleration or deceleration based solely on the total vibration magnitude of the upper part of the column.

[0016] Furthermore, the present invention uses the vibration path state only as the basis for determining the adjustment direction of the candidate lifting trajectory, and makes the candidate lifting trajectory have the same start and end states as the reference lifting trajectory. Then, the chassis propagation vibration and the lifting motion-related vibration are accumulated along the two lifting trajectories respectively. Thus, in the same candidate adjustment, the lifting motion-related vibration change caused by the change of lifting speed and the cumulative effect of chassis propagation vibration caused by the change of time are considered simultaneously.

[0017] Based on this, the candidate lifting trajectory is only executed when the cumulative vibration effect of the candidate lifting trajectory on the current dominant vibration action path and the total cumulative vibration effect of the predicted upper part of the column are both determined to decrease within the corresponding prediction error range. This reduces the risk that the overall vibration improvement effect will be offset by the decrease in vibration effect on one vibration action path and the increase in vibration effect on another vibration action path. This is beneficial to improving the running stability of the upper part of the column during the lifting of the robot's aluminum column. Attached Figure Description

[0018] Figure 1 This is a flowchart of the method for controlling the running stability of a robot aluminum column based on vibration detection in this invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0021] This embodiment discloses a method for controlling the operational stability of a robot's aluminum column based on vibration detection. It is applicable to robots with a robot chassis, an aluminum column that can extend and retract vertically relative to the robot chassis, robot working components mounted on the upper part of the aluminum column, and a lifting drive mechanism. The lifting drive controller determines the target extension position and the original lifting trajectory based on the robot's task. (Refer to...) Figure 1 This includes steps S1-S6: S1. Collect chassis vibration data and upper column vibration data, and obtain extension position, lifting speed, lifting acceleration and lifting jerk.

[0022] A first three-axis acceleration detection unit is installed on the chassis side of the connection structure between the robot chassis and the aluminum column. A second three-axis acceleration detection unit is installed at the installation connection position between the uppermost movable structure of the aluminum column and the robot's working parts. The three detection axes of the first and second three-axis acceleration detection units correspond to the robot's forward, lateral, and vertical directions, respectively, and use the same acceleration unit. The first three-axis acceleration detection unit is used to collect the vibration input of the chassis before it enters the aluminum column, and the second three-axis acceleration detection unit is used to collect the vibration response of the upper part of the column after the chassis vibration passes through the current telescopic structure, as well as the vibration generated on the upper part of the column by the lifting motion.

[0023] The intersection of the effective frequency ranges of the first and second triaxial acceleration detection units is taken as the common effective detection frequency band. The lower and upper limits of the common effective detection frequency band are taken as the lower and upper limits of analysis, respectively. The effective frequency range of the triaxial acceleration detection unit is the frequency interval of the acceleration signal that the triaxial acceleration detection unit can measure, obtained by referring to the frequency range parameters in the product manual of the corresponding triaxial acceleration detection unit.

[0024] Among the sampling frequencies supported by both the first and second triaxial acceleration detection units, the smallest sampling frequency that can cover the upper limit of the common effective detection frequency band analysis is selected as the sampling frequency of both units. The minimum integer number of sampling points that can cover a complete cycle of vibration at the lowest frequency corresponding to the lower limit of analysis is used as the number of sampling points in the analysis sub-window, and the sampling duration corresponding to this number of sampling points is used as the duration of the analysis sub-window. Three consecutive, equal-length, and non-overlapping analysis sub-windows constitute a complete vibration observation window.

[0025] To enable subsequent comparisons of chassis vibration input and column upper vibration response at the same frequency, for each analysis sub-window, the DC component is removed from the acceleration sequences of the three detection axes of the first and second triaxial acceleration detection units. Then, the periodogram method based on discrete Fourier transform is used to calculate the one-sided power spectral density. The one-sided power spectral densities of the three detection axes of the same triaxial acceleration detection unit at the same frequency are added together to form the total vibration power spectral density corresponding to that analysis sub-window. The arithmetic mean of the total vibration power spectral densities of the three analysis sub-windows at the same frequency within the complete vibration observation window is taken as the total vibration power spectral density of the complete vibration observation window.

[0026] In one embodiment, the method for removing the DC component from the acceleration sequence of the three detection axes of the first and second triaxial acceleration detection units is as follows: calculate the average value of the acceleration sample value of each detection axis of the first and second triaxial acceleration detection units within the analysis sub-window, and subtract the average value of the acceleration sample value of the detection axis within the analysis sub-window from each acceleration sample value of the detection axis to obtain the acceleration sequence of the detection axis after removing the DC component.

[0027] The lifting drive controller synchronously outputs the extension position, lifting speed, and lifting acceleration of the aluminum column. Vibration sampling data and the data output by the lifting drive controller use the same control clock and are recorded synchronously with the same timestamps. This ensures that the start time, time center, and end time of each analysis sub-window correspond to the extension position, and the time center corresponds to the lifting speed and lifting acceleration. The lifting acceleration corresponding to the time center of the analysis sub-window is calculated based on the lifting acceleration and time interval corresponding to adjacent timestamps before and after the time center.

[0028] S2. Establish position-related vibration propagation relationships to form the vibration action of the chassis propagation and the measured vibration action of the upper part of the column, and obtain the chassis propagation prediction error limit.

[0029] After the aluminum column stops lifting, the vibration driven by the lifting mechanism no longer continuously enters the upper part of the column. Therefore, changing the robot chassis movement in this state allows the vibration change in the upper part of the column to primarily reflect the propagation effect of the chassis vibration through the current telescopic structure. To describe the relationship between the propagation effect and the extension structure, this invention calibrates the lower limit and upper limit of the working stroke of the aluminum column, as well as the overall extension position of the aluminum column corresponding to the start or end of extension of each level of telescopic column. The lower and upper limits of the working stroke are taken from the stroke limit positions set by the lifting drive controller. According to the predetermined extension sequence of each level of telescopic column, the stroke length of each level of telescopic column that has already extended before the start of extension of a certain level is added to the overall extension position corresponding to the lower limit of the working stroke of the aluminum column to obtain the overall extension position of the aluminum column corresponding to the start of extension of that level. Then, the stroke length of that level of telescopic column itself is added to this position to obtain the overall extension position of the aluminum column corresponding to the end of extension of that level. The lower limit of the working stroke, the overall extension position of the aluminum column when each level of telescopic column begins or ends to extend, and the upper limit of the working stroke are arranged in ascending order of the overall extension position of the aluminum column to form a calibration extension position table.

[0030] The same chassis calibration motion sequence is used for different calibration extension positions, so the differences between positions mainly come from the aluminum column extension structure. The chassis calibration motion sequence consists of two motion segments: the first motion segment is that the robot chassis moves linearly at the maximum permissible linear velocity for a complete vibration observation window, while maintaining an angular velocity of 0; the second motion segment is that the robot chassis turns in place at the maximum permissible angular velocity for a complete vibration observation window, while maintaining a linear velocity of 0. The maximum permissible linear velocity of the robot chassis is taken from the upper limit of the linear velocity set in the robot chassis controller, and the maximum permissible angular velocity is taken from the upper limit of the angular velocity set in the robot chassis controller.

[0031] At each calibrated extension position, the robot chassis and aluminum column are first brought to a complete vibration observation window. Then, the first and second motion segments of the chassis calibration motion sequence are executed in sequence, and the vibration data of the first and second triaxial acceleration detection units are collected simultaneously.

[0032] After data acquisition is completed for all calibrated extension positions, a complete vibration observation window acquired when the robot chassis and aluminum column are stationary at each calibrated extension position contains three analysis sub-windows. These three analysis sub-windows are each used as a static analysis sub-window. For each frequency point, in all static analysis sub-windows corresponding to all calibrated extension positions, the maximum value of the total vibration power spectral density of the first triaxial acceleration detection unit is taken as the static background reference power spectral density of the chassis, and the maximum value of the total vibration power spectral density of the second triaxial acceleration detection unit is taken as the static background reference power spectral density of the upper part of the column. For any calibration extension position, at each frequency point, the total vibration power spectral density of the first triaxial acceleration detection unit during the chassis calibration motion sequence is averaged across all complete vibration observation windows at that frequency point, and the chassis static background reference power spectral density at that frequency point is subtracted. If the subtraction result is less than 0, it is taken as 0, thus forming the net vibration power spectral density of the chassis at that calibration extension position at that frequency point. Similarly, the total vibration power spectral density of the second triaxial acceleration detection unit during all complete vibration observation windows at that frequency point is averaged, and the static background reference power spectral density of the upper part of the column at that frequency point is subtracted. If the subtraction result is less than 0, it is taken as 0, thus forming the net vibration power spectral density of the upper part of the column at that calibration extension position at that frequency point. Data from the last segment of a vibration observation window that is less than one complete vibration observation window is not included in the calculation.

[0033] For any frequency point within the common effective detection frequency band, if the net vibration power spectral density of the chassis and the net vibration power spectral density of the upper part of the column at all calibrated extension positions are not 0, then the frequency point is taken as an effective frequency point, and all effective frequency points constitute an effective frequency point set; if there is no frequency point that meets this condition, then there is no effective frequency point.

[0034] For each calibrated extension position and effective frequency point, the ratio between the net vibration power spectral density of the upper part of the column and the net vibration power spectral density of the chassis at the same calibrated extension position and effective frequency point is used as the vibration power propagation coefficient corresponding to that calibrated extension position at that effective frequency point. When the current extension position is equal to any calibrated extension position in the calibrated extension position table, the vibration power propagation coefficient corresponding to that calibrated extension position at each effective frequency point is used as the vibration power propagation coefficient corresponding to the current extension position at each effective frequency point. When the current extension position is between two adjacent calibrated extension positions in the calibrated extension position table, the vibration power propagation coefficients corresponding to the two adjacent calibrated extension positions at the same effective frequency point are linearly interpolated according to the ratio of the positional distance between the current extension position and the two adjacent calibrated extension positions, and the interpolation result is used as the vibration power propagation coefficient corresponding to the current extension position at that effective frequency point. When the current extension position is outside the coverage area of ​​the calibrated extension position table, the position-related vibration propagation relationship is not invoked.

[0035] For any analysis sub-window, at each frequency point, the total vibration power spectral density of the first triaxial acceleration detection unit is subtracted from the corresponding chassis static background reference power spectral density, and the total vibration power spectral density of the second triaxial acceleration detection unit is subtracted from the corresponding upper column static background reference power spectral density. If the subtraction result is less than 0, it is set to 0; otherwise, the subtraction result is retained, forming the chassis net vibration power spectral density and the upper column net vibration power spectral density for that analysis sub-window. For a complete vibration observation window, at each effective frequency point, the arithmetic mean of the chassis net vibration power spectral densities of the three analysis sub-windows at that effective frequency point is taken as the chassis net vibration power spectral density of the complete vibration observation window at that effective frequency point.

[0036] For any analysis sub-window, the chassis propagated vibration action is obtained based on the chassis net vibration power spectral density at each effective frequency point of the analysis sub-window, the vibration power propagation coefficient at each effective frequency point corresponding to the extension position of the time center of the analysis sub-window, and the frequency interval between adjacent frequency points in the power spectrum. ; In the formula, For the first The chassis vibration propagation magnitude corresponding to each analysis sub-window represents the average vibration level formed by the chassis vibration propagating to the upper part of the column through the current telescopic structure within the corresponding analysis sub-window, reflecting the strength of the propagation effect of the chassis vibration on the upper part of the column through the current telescopic structure. This indicates the sequence number of the analysis sub-window within the complete vibration observation window; Represents the first in the set of effective frequency points One effective frequency point; Represents the set of effective frequency points; Indicates that the aluminum column is located at the first The time center of each analysis sub-window corresponds to the extended position. At that time, the first The vibration power propagation coefficient corresponding to the effective frequency point reflects the first effective frequency point. The proportion of vibration power propagation from the chassis to the top of the column at each effective frequency point; Indicates the first The position of the protruding aluminum column corresponding to the time center of each analysis sub-window; Indicates the first The analysis sub-window is in the... The net vibration power spectral density of the chassis at each effective frequency point; This represents the frequency interval between adjacent frequency points in the power spectrum.

[0037] For any analysis sub-window, the measured vibration magnitude of the upper part of the column is obtained based on the net vibration power spectral density of the upper part of the column at each effective frequency point and the frequency interval between adjacent frequency points in the power spectrum. ; in, For the first The measured vibration action of the upper part of the column in each analysis sub-window represents the average vibration action level actually measured in the upper part of the column within the corresponding analysis sub-window, reflecting the strength of the actual vibration response of the upper part of the column within the corresponding analysis sub-window. Represents the first in the set of effective frequency points One effective frequency point; Represents the set of effective frequency points; Indicates the first The analysis sub-window is in the... Net vibration power spectral density of the upper part of the column at each effective frequency point; This represents the frequency interval between adjacent frequency points in the power spectrum.

[0038] The arithmetic mean of the chassis-propagated vibration magnitudes from the three analysis sub-windows of the complete vibration observation window is taken as the chassis-propagated vibration magnitude of the complete vibration observation window; the arithmetic mean of the measured vibration magnitudes of the upper part of the column from the three analysis sub-windows of the complete vibration observation window is taken as the measured vibration magnitude of the upper part of the column of the complete vibration observation window.

[0039] To determine the prediction error of position-dependent vibration propagation, the chassis calibration motion sequence was re-executed at each calibration extension position, and verification data was collected. The analysis sub-windows contained within each complete vibration observation window during the re-execution of the chassis calibration motion sequence were used as verification analysis sub-windows. For each verification analysis sub-window, the chassis propagated vibration and the measured vibration on the upper part of the column were calculated, and their absolute difference was calculated. After verification was completed for all calibration extension positions, the maximum value among the absolute differences corresponding to all verification analysis sub-windows was taken as the chassis propagation prediction error limit. When no effective frequency point exists, or when a complete vibration observation window cannot be formed at any calibration extension position to obtain a verification analysis sub-window, the position-dependent vibration propagation relationship is invalid.

[0040] S3. Establish the vibration relationship of lifting and lowering motion to form the vibration action of lifting and lowering motion, and obtain the prediction error limit and joint prediction error limit of lifting and lowering motion.

[0041] After obtaining the vibration power propagation coefficient and chassis propagation prediction error limit corresponding to each calibrated extension position, the robot chassis is kept stationary during the subsequent establishment of the vibration relationship associated with the lifting motion. Only the aluminum column is controlled to lift and lower, so as to avoid the vibration input generated by the robot chassis movement from being mixed into the vibration of the upper part of the column corresponding to the lifting motion. Thus, the vibration relationship associated with the lifting motion is established based on the change of the measured vibration action of the upper part of the column with the lifting and lowering state of the aluminum column.

[0042] For each robot task currently configured, the lifting drive controller generates a corresponding original lifting trajectory. The original lifting trajectory whose position path passes through at least one calibrated extension position is used as the lifting calibration trajectory, maintaining the lifting speed, lifting acceleration, and lifting jerk characteristics of the original lifting trajectory. Each lifting calibration trajectory is executed independently. Before executing the next lifting calibration trajectory, the aluminum column is controlled to reach the initial extension position of the current lifting calibration trajectory. The preparatory motion data of the aluminum column before reaching the initial extension position and the stopping process data after the lifting calibration trajectory is completed are not used as lifting motion calibration samples. During the formal execution of the lifting calibration trajectory, vibration data from the second and third-axis acceleration detection units and the aluminum column extension position, lifting speed, and lifting acceleration output by the lifting drive controller are simultaneously collected. Only the analysis sub-windows that completely fall within the formal execution period of the lifting calibration trajectory are used to form lifting motion calibration samples, and the extension position, lifting speed, lifting acceleration, and lifting jerk data corresponding to each analysis sub-window are obtained.

[0043] To ensure that the vibration relationship associated with the lifting motion covers the static state of the aluminum column, a complete vibration observation window is established at each calibrated extension position, keeping both the robot chassis and the aluminum column stationary. The measured vibration magnitude of the upper part of the column corresponding to this complete vibration observation window is calculated, and the lifting velocity, lifting acceleration, and lifting jerk corresponding to this calibrated extension position are all recorded as 0. At the same time, the measured vibration magnitude of the upper part of the column corresponding to this complete vibration observation window is recorded, thus forming a zero-motion calibration sample.

[0044] The direction of movement for each analysis sub-window is determined by comparing its extension position at the end time with its extension position at the start time: when the extension position at the end time is greater than the extension position at the start time, the direction of movement for that analysis sub-window is the extension direction; when the extension position at the end time is less than the extension position at the start time, the direction of movement for that analysis sub-window is the retraction direction; when the extension position at the end time is equal to the extension position at the start time, the direction of movement for that analysis sub-window is not determined, and in this case, that analysis sub-window is not used for movement calibration. For each calibration extension position, when the extension position corresponding to the time center of a certain analysis sub-window is equal to the calibration extension position, record the lifting speed, lifting acceleration, lifting jerk, measured vibration amount on the upper part of the column, and lifting direction corresponding to the analysis sub-window. When the calibration extension position is located between the extension positions corresponding to the time centers of two adjacent analysis sub-windows with the same lifting direction, perform linear interpolation on the lifting speed, lifting acceleration, lifting jerk, and measured vibration amount on the upper part of the column for the two analysis sub-windows according to the ratio of the positional distance between the calibration extension position and the extension positions corresponding to the time centers of the two analysis sub-windows, and record the interpolation results and the lifting direction corresponding to the two analysis sub-windows. When there are no two analysis sub-windows located on both sides of the calibration extension position with the same lifting direction, do not record the data of the lifting calibration trajectory at the calibration extension position.

[0045] For each calibrated extension position, the vibration relationship between the lifting speed, lifting acceleration, and lifting jerk and the vibration of the upper part of the column is constructed, and its functional expression is: ; In the formula, This indicates that the aluminum column is in the extended position. Lifting and lowering speed Climbing acceleration and acceleration / deceleration At that time, the fitted vibration action corresponding to the lifting motion; Indicates the position where the aluminum column extends; Indicates the rate of acceleration or deceleration; Indicates acceleration due to lifting or sinking; Indicates acceleration or deceleration; Indicates the position of the extension The corresponding constant term fitting coefficients; Indicates the position of the extension The corresponding fitting coefficients for the squared terms of acceleration and deceleration; Indicates the position of the extension The corresponding fitting coefficients for the squared terms of acceleration and deceleration; Indicates the position of the extension The corresponding fitting coefficients for the squared terms of acceleration and deceleration.

[0046] For each calibrated extension position, each set of lifting velocity, lifting acceleration, lifting jerk, and measured vibration of the upper part of the column recorded for different lifting calibration trajectories at that calibrated extension position is taken as a lifting motion calibration sample, and the zero motion calibration sample for that calibrated extension position is added. Based on all lifting motion calibration samples for that calibrated extension position, the squares of the lifting velocity, the squares of the lifting acceleration, and the squares of the lifting jerk are used as fitting variables for the lifting motion-related vibration relationship, and the measured vibration of the upper part of the column is used as the fitting target. The least squares method is used to obtain the value of ... , , and Four fitting coefficients. If all the calibration samples of the lifting and lowering motion can uniquely obtain the four fitting coefficients, the calibration extension position is taken as the valid calibration position and the corresponding four fitting coefficients are retained; otherwise, the calibration extension position is not taken as the valid calibration position.

[0047] For each valid calibration position, the maximum absolute values ​​of the lifting velocity, lifting acceleration, and lifting jerk are recorded in all lifting motion calibration samples. When two adjacent calibration positions are both valid calibration positions, the four fitting coefficients corresponding to these two adjacent calibration positions are linearly interpolated according to the extended position to obtain the four fitting coefficients corresponding to any extended position between the two valid calibration positions. Simultaneously, the minimum of the maximum absolute values ​​of the lifting velocity, lifting acceleration, and lifting jerk corresponding to the two valid calibration positions is taken as the upper limit of the absolute values ​​of the lifting velocity, lifting acceleration, and lifting jerk for that valid position interval. When either of the two adjacent calibration positions is not a valid calibration position, the four fitting coefficients are not interpolated for the extended position between the two adjacent calibration positions, and the lifting motion-related vibration relationship is not invoked within that extended position range.

[0048] When the extension position of the aluminum column is equal to any valid calibration position, the fitted vibration action is calculated based on the extension position, lifting speed, lifting acceleration, and lifting jerk of the aluminum column using four fitting coefficients corresponding to that valid calibration position. When the extension position of the aluminum column is between two adjacent valid calibration positions, and the absolute values ​​of the lifting speed, lifting acceleration, and lifting jerk do not exceed the upper limits of the absolute values ​​of the lifting speed, lifting acceleration, and lifting jerk corresponding to the two adjacent valid calibration positions, the fitted vibration action is calculated based on the extension position, lifting speed, lifting acceleration, and lifting jerk using four fitting coefficients obtained by linear interpolation according to the extension position. When the calculated fitted vibration action is less than 0, the vibration action associated with the lifting motion is set to 0; when the fitted vibration action is greater than or equal to 0, the fitted vibration action is taken as the vibration action associated with the lifting motion.

[0049] Because the lifting velocity, lifting acceleration, and lifting jerk are all calculated using square terms in the function formula related to the lifting motion, the extension and retraction directions of the aluminum column cannot be distinguished solely based on this function. To determine which lifting direction the lifting motion-related vibration relationship can be applied to and to determine its prediction error, each lifting calibration trajectory was re-executed, and verification data not used in the calculation of the four fitting coefficients was collected. During the verification process, the robot chassis remained stationary. For each analysis sub-window formed during the verification process, if the extension position corresponding to the time center of the analysis sub-window is located between two adjacent valid calibration positions, and the absolute values ​​of the lifting velocity, lifting acceleration, and lifting jerk corresponding to the analysis sub-window do not exceed the upper limits of the absolute values ​​of the lifting velocity, lifting acceleration, and lifting jerk between the two adjacent valid calibration positions, respectively, the analysis sub-window is considered a valid verification analysis sub-window. For the extension or retraction direction, the lifting direction is considered a verified lifting direction only if it appears both in the lifting motion calibration samples used to calculate the four fitting coefficients and in a valid verification analysis sub-window. Simultaneously, at each valid calibration position, the robot chassis and aluminum column are kept stationary for a complete vibration observation window. The lifting speed, lifting acceleration, and lifting jerk are all recorded as 0. The measured vibration action on the upper part of the column corresponding to this complete vibration observation window is calculated, forming an independent zero-motion verification sample. This independent zero-motion verification sample is only used to verify the prediction results when the lifting speed, lifting acceleration, and lifting jerk are all 0; it is not used to determine the verified lifting direction. For each valid verification analysis sub-window and each independent zero-motion verification sample, the absolute difference between the lifting motion-related vibration action and the corresponding measured vibration action on the upper part of the column is calculated. The maximum value among all absolute differences is taken as the lifting motion-related prediction error limit. If no valid verification analysis sub-window exists, the lifting motion-related vibration relationship is not invoked for subsequent control.

[0050] After obtaining the prediction error limits for chassis propagation, the prediction error limits for lifting motion association, and at least one verified lifting direction, the prediction error of the sum of the chassis propagation vibration and the lifting motion association vibration on the measured vibration of the upper part of the column is further verified when the robot chassis motion and the aluminum column lifting occur simultaneously. During the joint verification process, the robot chassis cyclically executes the chassis calibration motion sequence, and the aluminum column sequentially executes each lifting calibration trajectory; the preparatory motion data used between two adjacent lifting calibration trajectories to enable the aluminum column to reach the starting extension position of the next lifting calibration trajectory is not included in the joint verification. When the robot chassis completes one chassis calibration motion sequence first, execution restarts from the first motion segment of the chassis calibration motion sequence; when the aluminum column completes all lifting calibration trajectories first, execution restarts from the first lifting calibration trajectory, until the robot chassis completes at least one chassis calibration motion sequence and the aluminum column completes at least one full lifting calibration trajectory. For analysis sub-windows formed during the joint verification process, if the extension position corresponding to the time center of the analysis sub-window is within the coverage area of ​​the calibration extension position table and between two adjacent valid calibration positions, and the absolute values ​​of the lifting velocity, lifting acceleration, and lifting jerk corresponding to the analysis sub-window do not exceed their respective upper limits, and the lifting direction of the analysis sub-window belongs to the verified lifting direction, then the analysis sub-window is designated as a joint verification analysis sub-window. For each joint verification analysis sub-window, the chassis-propagated vibration action, the lifting motion-related vibration action, and the measured vibration action on the upper part of the column are calculated respectively, and the absolute difference between the measured vibration action on the upper part of the column and the sum of the chassis-propagated vibration action and the lifting motion-related vibration action is calculated. The maximum value among all the absolute differences corresponding to the joint verification analysis sub-windows is taken as the joint prediction error limit. If no joint verification analysis sub-window exists, no subsequent vibration action path state judgment is performed.

[0051] S4. Determine the vibration path state based on the vibration action amount propagated from the chassis, the vibration action amount associated with the lifting motion, and the measured vibration action amount on the upper part of the column, and form a reference lifting trajectory and generate candidate lifting trajectories.

[0052] After obtaining the chassis propagation prediction error limit, the lifting motion correlation prediction error limit, and the joint prediction error limit, and with at least one verified lifting direction, online control is initiated. Upon initial entry into online control, the original lifting trajectory of the current robot task is used as the reference lifting trajectory, and execution follows this trajectory until a complete vibration observation window is formed. If the remaining execution time from the current moment to reaching the target extension position according to the reference lifting trajectory is less than the duration of a complete vibration observation window, the current robot task is completed directly according to the reference lifting trajectory. Subsequently, for each complete vibration observation window, the extension position, lifting speed, and lifting acceleration output by the lifting drive controller at the end of the window are read. Using these parameters as starting conditions, the lifting drive controller re-forms the reference lifting trajectory to the target extension position of the current robot task. The data from the recently concluded complete vibration observation window is used to subsequently determine the vibration path state and evaluate whether to execute candidate lifting trajectories.

[0053] Each time a new complete vibration observation window is formed, the data from that window is used to determine the vibration path status. If the complete vibration observation window has missing samples or interrupted timestamps, if any triaxial acceleration detection unit has an acceleration sample value that exceeds its measurement range, or if the extension position corresponding to the time center of any analysis sub-window is outside the coverage area of ​​the calibration extension position table or is not located between two adjacent valid calibration positions, or if the absolute value of the rise / fall velocity, absolute value of the rise / fall acceleration, or absolute value of the rise / fall jerk corresponding to the analysis sub-window exceeds the upper limit of the absolute value between two adjacent valid calibration positions, or if the rise / fall direction of the analysis sub-window cannot be determined, or if the rise / fall direction is not a verified rise / fall direction, then the data from the complete vibration observation window is not used to determine the vibration path status. The vibration path status is determined to be indeterminate, and execution continues according to the baseline rise / fall trajectory.

[0054] When the complete vibration observation window can be used to determine the vibration path state, the arithmetic mean of the chassis-propagated vibration, the arithmetic mean of the lifting motion-related vibration, and the arithmetic mean of the measured vibration on the upper part of the column are calculated for the three analysis sub-windows of the complete vibration observation window. These results are then used as the chassis-propagated vibration, lifting motion-related vibration, and measured vibration on the upper part of the column, respectively, for the complete vibration observation window. The absolute difference between the measured vibration on the upper part of the column and the sum of the chassis-propagated vibration and lifting motion-related vibration is calculated. If this absolute difference exceeds the joint prediction error limit, the vibration path state is determined to be undetermined, and execution continues according to the baseline lifting trajectory.

[0055] When the absolute difference between the measured vibration magnitude of the upper part of the column in the complete vibration observation window and the sum of the vibration magnitude propagated by the chassis and the vibration magnitude associated with the lifting motion is not greater than the joint prediction error limit, the chassis propagation vibration range and the lifting motion associated vibration range are respectively formed. The lower limit of the chassis propagation vibration range is the larger of the value of the chassis propagation vibration magnitude minus the chassis propagation prediction error limit in the complete vibration observation window and 0, and the upper limit is the sum of the chassis propagation vibration magnitude and the chassis propagation prediction error limit; the lower limit of the lifting motion associated vibration range is the larger of the value of the lifting motion associated vibration magnitude minus the lifting motion associated prediction error limit in the complete vibration observation window and 0, and the upper limit is the sum of the lifting motion associated vibration magnitude and the lifting motion associated prediction error limit. When the lower bound of the vibration action range associated with the lifting motion is greater than the upper bound of the vibration action range propagated by the chassis, the vibration action path state is determined to be the lifting motion-dominant state. When the lower bound of the vibration action range propagated by the chassis is greater than the upper bound of the vibration action range associated with the lifting motion, the vibration action path state is determined to be the chassis propagation-dominant state. When the lower bound of the vibration action range associated with the lifting motion is neither greater than the upper bound of the vibration action range propagated by the chassis, nor greater than the upper bound of the vibration action range associated with the lifting motion, the vibration action path state is determined to be an indeterminate state. When the vibration action path state is indeterminate, no candidate lifting trajectory is formed, and execution continues according to the baseline lifting trajectory.

[0056] Once the vibration path state is determined to be either a lifting motion-dominant state or a chassis propagation-dominant state, the extension position, lifting speed, and lifting acceleration output by the lifting drive controller at the end of the newly formed complete vibration observation window are used as the starting extension position, starting lifting speed, and starting lifting acceleration of the candidate lifting trajectory, respectively. Starting from this moment along the baseline lifting trajectory, the subsequent lifting direction is determined based on the first change in extension position. If the baseline lifting trajectory no longer shows any change in extension position, or if the subsequent lifting direction does not belong to a verified lifting direction, no candidate lifting trajectory is formed, and execution continues according to the baseline lifting trajectory. If the subsequent lifting direction belongs to a verified lifting direction, the first calibrated extension position reached after the starting extension position is searched along that subsequent lifting direction; if this calibrated extension position does not exist, no candidate lifting trajectory is formed, and execution continues according to the baseline lifting trajectory. If the target extension position of the current robot task is between the starting extension position and the calibrated extension position, the target extension position is used as the predicted termination position; otherwise, the calibrated extension position is used as the predicted termination position. When the starting extension position and the predicted termination position cannot be simultaneously located between or equal to either of the two adjacent calibration positions, or when one of the two adjacent calibration positions is not a valid calibration position, no candidate elevation trajectory is formed, and execution continues according to the baseline elevation trajectory.

[0057] The baseline average lifting speed is obtained by dividing the distance between the initial extension position and the predicted termination position by the time required for the baseline lifting trajectory to reach the predicted termination position for the first time from the start time of the candidate lifting trajectory. When the current vibration path is dominated by lifting motion association, the lifting speed with an absolute value greater than 0, less than the baseline average lifting speed, and closest to the baseline average lifting speed is selected from the lifting speeds allowed by the lifting drive controller and consistent with the subsequent lifting direction. The absolute value of this lifting speed is then used as the candidate average lifting speed. When the current vibration path is dominated by chassis propagation, the lifting speed with an absolute value greater than the baseline average lifting speed and closest to the baseline average lifting speed is selected from the lifting speeds allowed by the lifting drive controller and consistent with the subsequent lifting direction. The absolute value of this lifting speed is then used as the candidate average lifting speed. If no lifting speed meets the corresponding conditions, no candidate lifting trajectory is formed, and execution continues according to the baseline lifting trajectory. The duration of the candidate lifting trajectory is obtained by dividing the distance between the initial extension position and the predicted termination position by the candidate average lifting speed.

[0058] The initial extension position, initial elevation velocity, and initial elevation acceleration of the candidate elevation trajectory are the same as those of the reference elevation trajectory at the start of the candidate elevation trajectory. Similarly, the final extension position, final elevation velocity, and final elevation acceleration of the candidate elevation trajectory are the same as those of the reference elevation trajectory when it first reaches the predicted final position. Using the initial extension position, initial elevation velocity, initial elevation acceleration, final extension position, final elevation velocity, and final elevation acceleration as six boundary conditions, a fifth-order time polynomial uniquely determined by these six boundary conditions is used to form the candidate elevation trajectory. The elevation velocity, elevation acceleration, and elevation jerk of the candidate elevation trajectory at each time step are obtained from the first, second, and third derivatives of this fifth-order time polynomial.

[0059] After a candidate lifting trajectory is formed, the set maximum permissible lifting speed, maximum permissible lifting acceleration, and maximum permissible lifting jerk are read from the lifting drive controller. Based on the fifth-order time polynomial corresponding to the candidate lifting trajectory and its first, second, and third derivatives, the maximum absolute values ​​of the lifting speed, lifting acceleration, and lifting jerk of the candidate lifting trajectory are calculated over the entire duration. The extension position of the candidate lifting trajectory should always be between the initial extension position and the predicted termination position, and the lifting direction should not reverse. The maximum absolute values ​​of the lifting speed, lifting acceleration, and lifting jerk of the candidate lifting trajectory must not exceed the maximum permissible lifting speed, maximum permissible lifting acceleration, and maximum permissible lifting jerk set by the lifting drive controller, and must also not exceed the upper limits of the absolute values ​​of the lifting speed, lifting acceleration, and lifting jerk corresponding to the two adjacent valid calibration positions at the initial extension position and the predicted termination position, respectively. From the start time of the candidate trajectory to the first arrival at the predicted termination position, the extension position of the baseline trajectory should always be between the same pair of adjacent valid calibration positions. The absolute values ​​of its elevation velocity, elevation acceleration, and elevation jerk should not exceed their respective upper limits, and the elevation direction should be a verified elevation direction. If these conditions are met, the candidate trajectory will be used for subsequent vibration cumulative effect evaluation; otherwise, the baseline trajectory will continue to be used.

[0060] S5. Generate the cumulative action of vibration propagated from the chassis, the cumulative action of vibration associated with lifting and lowering motion, and predict the cumulative action of vibration on the upper part of the column, and form the upper and lower bounds of the predicted cumulative action of vibration.

[0061] The candidate lifting trajectory alters the lifting motion process compared to the reference lifting trajectory. This changes both the associated vibration effects and the time required for the aluminum column to pass through each extension position, thus altering the cumulative effect of chassis-propagated vibration. Therefore, it is necessary to simultaneously evaluate the cumulative effect of chassis-propagated vibration, the cumulative effect of lifting motion-related vibration, and the predicted cumulative effect of upper column vibration corresponding to both the candidate and reference lifting trajectories. The prediction interval for this evaluation is from the current moment when the candidate lifting trajectory is formed to the prediction termination position. Within this prediction interval, the chassis net vibration power spectral density corresponding to the complete vibration observation window that just ended when the candidate lifting trajectory was formed is used as the prediction input for the chassis vibration spectrum. Both the reference and candidate lifting trajectories use this chassis net vibration power spectral density for cumulative vibration evaluation.

[0062] Since the reference lifting trajectory and the candidate lifting trajectory require different times to reach the same predicted termination position from the same initial extension position, lifting speed, and lifting acceleration, the cumulative vibration effect is calculated along the execution time of each of the two lifting trajectories. When executing the lifting trajectory, the lifting drive controller continuously outputs the extension position, lifting speed, and lifting acceleration, using the time interval between two adjacent outputs as the discrete time interval. Starting from the end of the complete vibration observation window that just ended when the candidate lifting trajectory was formed, discrete time points are sequentially formed along both the reference and candidate lifting trajectories until they reach their respective predicted termination positions. When the next discrete time point formed according to the discrete time interval exceeds the time when the corresponding lifting trajectory reaches its predicted termination position, the time when the corresponding lifting trajectory reaches its predicted termination position is directly taken as the last discrete time point. Based on the discrete time points formed by the two lifting trajectories, the cumulative effect of chassis-propagated vibration and the cumulative effect of lifting motion-related vibration are calculated separately.

[0063] Specifically, the rising and falling trajectory The cumulative effect of chassis-borne vibrations satisfies the expression: ; In the formula, Indicates the rising and falling trajectory The cumulative effect of chassis-propagated vibration during the period from the initial extension position to the predicted termination position. This indicates the current evaluation's baseline rise / fall trajectory or candidate rise / fall trajectory; Indicates the rising and falling trajectory The discrete time point index; Indicates the rising and falling trajectory The total number of discrete time intervals from the initial extension position to the predicted termination position, and the rising and falling trajectory. The discrete time points are numbered sequentially from 0 to ; and Representing the rising and falling trajectories respectively The The and the first Predicted chassis vibration propagation at each discrete time point; and Representing the rising and falling trajectories respectively The The and the first The time corresponding to each discrete time point; Indicates the rising and falling trajectory The actual duration between two adjacent discrete time points.

[0064] Among them, the rising and falling trajectory The The and the first The method for obtaining the predicted chassis propagation vibration magnitude at each discrete time point is as follows: based on the... For each discrete time point corresponding to the protrusion position, obtain the vibration power propagation coefficient at each effective frequency point for that protrusion position; at each effective frequency point, compare this vibration power propagation coefficient with the chassis net vibration power spectral density and frequency interval at the same effective frequency point of the complete vibration observation window that just ended when the candidate lifting trajectory was formed. Multiply the results and sum the results over all effective frequency points, and use the result as the first... The predicted chassis propagation vibration effect at each discrete time point. The predicted chassis vibration propagation magnitudes corresponding to each discrete time point are obtained in the same way.

[0065] Lifting and lowering trajectory The cumulative action of the associated vibrations during the lifting and lowering motion satisfies the expression: ; In the formula, Indicates the rising and falling trajectory The cumulative effect of the lifting and lowering motion associated with the vibration during the movement from the initial extension position to the predicted termination position. This indicates the current evaluation's baseline rise / fall trajectory or candidate rise / fall trajectory; Indicates the rising and falling trajectory The discrete time point index; Indicates the rising and falling trajectory The total number of discrete time intervals from the initial extension position to the predicted termination position, and the rising and falling trajectory. The discrete time points are numbered sequentially from 0 to ; , , and Representing the rising and falling trajectories respectively The The extension position, lifting speed, lifting acceleration, and lifting jerk corresponding to each discrete time point; , , and Representing the rising and falling trajectories respectively The The extension position, lifting speed, lifting acceleration, and lifting jerk corresponding to each discrete time point; Indicates based on the rising and falling trajectory The The extension position, lifting speed, lifting acceleration, and lifting acceleration at discrete time points are associated with the vibrational action of the lifting motion; Indicates based on the rising and falling trajectory The The extension position, lifting speed, lifting acceleration, and lifting acceleration at discrete time points are associated with the vibrational action of the lifting motion; and Representing the rising and falling trajectories respectively The The and the first The time corresponding to each discrete time point; Indicates the rising and falling trajectory The actual duration between two adjacent discrete time points.

[0066] The cumulative effect of chassis vibration propagation along the same lifting trajectory and the cumulative effect of vibration associated with lifting motion are added together to form the predicted cumulative effect of vibration on the upper part of the column.

[0067] The chassis propagation prediction error limit, the lifting motion-related prediction error limit, and the joint prediction error limit are obtained from the aforementioned verification process. These limits characterize the maximum verified prediction error between the chassis propagation vibration magnitude, the lifting motion-related vibration magnitude, and the sum of the two types of vibration magnitudes, and the measured vibration magnitude on the upper part of the column. Since the durations of the reference lifting trajectory and the candidate lifting trajectory differ, for each lifting trajectory, the chassis propagation prediction error limit, the lifting motion-related prediction error limit, and the joint prediction error limit are multiplied by the duration of the lifting trajectory to obtain the chassis propagation cumulative error margin, the lifting motion-related cumulative error margin, and the joint cumulative error margin for that lifting trajectory. Adding the chassis propagation cumulative vibration magnitude to the chassis propagation cumulative error margin yields the upper bound of the predicted chassis propagation cumulative vibration magnitude; subtracting the chassis propagation cumulative error margin from the chassis propagation cumulative vibration magnitude, taking 0 if the subtraction result is less than 0, and retaining the subtraction result, yields the lower bound of the predicted chassis propagation cumulative vibration magnitude. Adding the cumulative impact of the vibration associated with the lifting motion to the cumulative error margin associated with the lifting motion yields the predicted upper bound of the cumulative impact of the vibration associated with the lifting motion. Subtracting the cumulative error margin associated with the lifting motion from the cumulative impact of the vibration associated with the lifting motion, setting the result to 0 if the difference is less than 0, otherwise retaining the result, yields the predicted lower bound of the cumulative impact of the vibration associated with the lifting motion. Adding the combined cumulative error margin to the predicted cumulative impact of the vibration on the upper part of the column yields the predicted upper bound of the predicted cumulative impact of the vibration on the upper part of the column. Subtracting the combined cumulative error margin from the predicted cumulative impact of the vibration on the upper part of the column, setting the result to 0 if the difference is less than 0, otherwise retaining the result, yields the predicted lower bound of the predicted cumulative impact of the vibration on the upper part of the column.

[0068] When the current vibration path is in a state dominated by lifting and lowering motion, the candidate lifting and lowering trajectory is executed only if the upper bound of the predicted cumulative vibration amount associated with the lifting and lowering motion of the candidate trajectory is less than the lower bound of the predicted cumulative vibration amount associated with the lifting and lowering motion of the reference trajectory, and the upper bound of the predicted cumulative vibration amount of the upper part of the column under the candidate lifting and lowering trajectory is less than the lower bound of the predicted cumulative vibration amount of the upper part of the column under the reference lifting and lowering trajectory. The former comparison condition is used to determine whether the candidate lifting and lowering trajectory can reduce the cumulative vibration amount associated with the currently dominant lifting and lowering motion, and the latter comparison condition is used to determine whether the cumulative vibration amount of the upper part of the column under the action of the candidate lifting and lowering trajectory is lower than that of the reference lifting and lowering trajectory. In the state dominated by lifting and lowering motion, if either of the above two comparison conditions is not met, the reference lifting and lowering trajectory continues to be executed.

[0069] When the current vibration path is dominated by chassis propagation, the candidate lifting trajectory is executed only if the upper bound of the predicted cumulative vibration amount propagated by the chassis in the candidate lifting trajectory is less than the lower bound of the predicted cumulative vibration amount propagated by the chassis in the reference lifting trajectory, and the upper bound of the predicted cumulative vibration amount on the upper part of the column in the candidate lifting trajectory is less than the lower bound of the predicted cumulative vibration amount on the upper part of the column in the reference lifting trajectory. The former comparison condition is used to determine whether the candidate lifting trajectory can reduce the cumulative effect of the currently dominant chassis propagated vibration, and the latter comparison condition is used to determine whether the cumulative vibration amount on the upper part of the column under the action of the candidate lifting trajectory is lower than that of the reference lifting trajectory. In the chassis propagation-dominated state, if either of the above two comparison conditions is not met, the reference lifting trajectory will still be executed.

[0070] S6. Execute candidate rising or falling trajectories or reference rising or falling trajectories based on the vibration action path state and the predicted upper and lower bounds of the cumulative vibration action.

[0071] Based on the trajectory selection result formed in the previous complete vibration observation window, the candidate or reference trajectory is executed starting from the next complete vibration observation window, while continuing to collect vibration data and the extension position, lifting speed, and lifting acceleration output by the lifting drive controller. The candidate trajectory is executed only up to the predicted termination position used in this cumulative vibration effect evaluation; the lifting speed and acceleration of the candidate trajectory when it reaches the predicted termination position are the same as those of the reference trajectory when it first reaches that predicted termination position. If the candidate trajectory reaches the predicted termination position before a new complete vibration observation window is formed and the current robot task has not yet been completed, the portion of the reference trajectory after the predicted termination position is executed, starting from the moment it reaches the predicted termination position. The current lifting task ends when the executing trajectory reaches the target extension position of the current robot task.

[0072] If the current robot task has not yet ended and a new complete vibration observation window has been formed, the vibration action path status is re-determined using the data from the complete vibration observation window. If a candidate lifting trajectory can be formed, the cumulative vibration effect evaluation of the candidate lifting trajectory and the baseline lifting trajectory continues, and the candidate lifting trajectory is executed or the baseline lifting trajectory is continued to be executed according to the evaluation results. If a candidate lifting trajectory cannot be formed, the baseline lifting trajectory is continued to be executed.

[0073] When the first or second triaxial acceleration detection unit is replaced, or its installation position or detection axis direction changes, the corresponding static background reference power spectral density is re-acquired, and the position-related vibration propagation relationship or the lifting motion-related vibration relationship involving the triaxial acceleration detection unit is re-established. When the connection structure between the robot chassis and the aluminum column, the load of the robot working parts, or the extension structure of the aluminum column changes, the position-related vibration propagation relationship and the lifting motion-related vibration relationship are re-established; when the lifting drive mechanism or lifting drive control parameters change, the lifting motion-related vibration relationship is re-established. After re-establishing the position-related vibration propagation relationship, the chassis propagation prediction error limit is obtained again; after re-establishing the lifting motion-related vibration relationship, the lifting motion-related prediction error limit is obtained again; after any vibration relationship is re-established, joint verification is performed again, and a new joint prediction error limit is obtained.

[0074] When the need to re-establish position-related vibration propagation relationships or lifting motion-related vibration relationships occurs during the current robot task execution, the vibration path state judgment and candidate lifting trajectory formation are stopped. The extension position, lifting speed, and lifting acceleration output by the lifting drive controller at this time are read, and these are used as the starting conditions to re-form the reference lifting trajectory to the target extension position. The current robot task is completed according to the reference lifting trajectory. After the necessary vibration relationships are re-established and verified, online control is resumed.

[0075] In the description of this specification, "multiple" or "several" means at least two, such as two, three or more, unless otherwise expressly and specifically defined.

[0076] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.

Claims

1. A method for controlling the operational stability of a robot aluminum column based on vibration detection, characterized in that, include: Collect chassis vibration data and upper column vibration data to obtain extension position, lifting speed, lifting acceleration and lifting jerk. Based on chassis vibration data and extension position, the vibration propagation magnitude of the chassis is formed through position-related vibration propagation relationships; the net vibration power spectral density of the upper part of the column is summed at frequency intervals to form the measured vibration magnitude of the upper part of the column; the vibration magnitude associated with the lifting motion is formed through lifting motion-related vibration relationships based on extension position, lifting speed, lifting acceleration, and lifting jerk; the vibration path state is determined based on the chassis-propagated vibration magnitude, the lifting motion-related vibration magnitude, and the measured vibration magnitude of the upper part of the column. A baseline lifting trajectory is formed based on the current extension position, lifting speed, lifting acceleration, and target extension position. The lifting speed is then adjusted according to the vibration path state to generate candidate lifting trajectories, ensuring that the candidate lifting trajectories have the same start and end motion states as the baseline lifting trajectory. Based on the execution time from the initial extension position to the predicted termination position for both the baseline and candidate lifting trajectories, the predicted chassis-propagated vibration and lifting motion-related vibration are accumulated to form the cumulative chassis-propagated vibration and lifting motion-related vibration for each of the baseline and candidate lifting trajectories. The sum of the cumulative chassis-propagated vibration and lifting motion-related vibration for the same lifting trajectory is used as the predicted cumulative vibration of the upper part of the column. The upper and lower bounds of the predicted vibration cumulative effect are formed based on the chassis propagation prediction error limit, the lifting motion-related prediction error limit, and the joint prediction error limit. Based on the vibration path state and the predicted upper and lower bounds of the cumulative vibration, execute the candidate rise / fall trajectory or the baseline rise / fall trajectory.

2. The method for controlling the operational stability of a robot aluminum column based on vibration detection according to claim 1, characterized in that, The process of collecting chassis vibration data and upper column vibration data to obtain extension position, lifting speed, lifting acceleration, and lifting jerk includes: A first three-axis acceleration detection unit is set at the connection between the chassis and the aluminum column on the side of the robot chassis, and a second three-axis acceleration detection unit is set at the connection between the uppermost movable structure of the aluminum column and the robot's working parts. The intersection of the effective frequency ranges of the first and second three-axis acceleration detection units is taken as the common effective detection frequency band. The minimum sampling frequency that can cover the upper limit of the common effective detection frequency band is selected. The minimum number of integer sampling points that can cover a complete cycle of the lowest frequency corresponding to the lower limit of the common effective detection frequency band is taken as the number of sampling points of the analysis sub-window. Three consecutive, equal-length and non-overlapping analysis sub-windows constitute a complete vibration observation window. After removing the DC component from the acceleration sequence of the three detection axes in each analysis sub-window, the single-sided power spectral density is calculated. The single-sided power spectral densities of the three detection axes of the same three-axis acceleration detection unit at the same frequency point are added to form the total vibration power spectral density. The extension position, lifting speed and lifting acceleration are recorded synchronously, and the lifting acceleration is formed according to the adjacent lifting acceleration and time interval before and after the time center of the analysis sub-window.

3. The method for controlling the operational stability of a robot aluminum column based on vibration detection according to claim 2, characterized in that, The method for obtaining the position-related vibration propagation relationship is as follows: The lower and upper limits of the working stroke of the aluminum column and the starting or ending points of the extension of each telescopic column are calibrated to form a calibration extension position table. The total vibration power spectral density is collected at each calibration extension position under both static and chassis calibration motion states. For each frequency point, in all static states, the maximum value of the total vibration power spectral density of the first triaxial acceleration detection unit at that frequency point is taken as the chassis static background reference power spectral density at that frequency point, and the maximum value of the total vibration power spectral density of the second triaxial acceleration detection unit at that frequency point is taken as the upper static background reference power spectral density of the column at that frequency point. At each frequency point, the average total vibration power spectral density of the first triaxial acceleration detection unit under chassis calibration motion state is subtracted from the chassis static background reference power spectral density at that frequency point; if the subtraction result is less than zero, it is taken as zero to form the chassis net vibration power spectral density at that frequency point. At each frequency point, the average total vibration power spectral density of the second and third axis acceleration detection units under the chassis calibration motion state is subtracted from the static background reference power spectral density of the upper part of the column at that frequency point. If the subtraction result is less than zero, it is taken as zero to form the net vibration power spectral density of the upper part of the column at that frequency point. The frequency points where both the net vibration power spectral density of the chassis and the net vibration power spectral density of the upper part of the column are not zero at all calibration extension positions are taken as effective frequency points. The ratio of the net vibration power spectral density of the upper part of the column to the net vibration power spectral density of the chassis at the same calibration extension position is taken as the vibration power propagation coefficient. The vibration power propagation coefficient is linearly interpolated between adjacent calibration extension positions according to the extension position.

4. The method for controlling the operational stability of a robot aluminum column based on vibration detection according to claim 3, characterized in that, The process of generating the chassis-propagated vibration magnitude based on chassis vibration data and extension position through position-related vibration propagation relationships includes: Obtain the net vibration power spectral density of the chassis at the effective frequency point of the analysis sub-window; obtain the vibration power propagation coefficient of the extension position corresponding to the time center of the analysis sub-window at the effective frequency point; sum the products of the net vibration power spectral density of the chassis, the vibration power propagation coefficient, and the frequency interval at each effective frequency point to obtain the chassis propagation vibration effect of the analysis sub-window; take the arithmetic mean of the chassis propagation vibration effects of the three analysis sub-windows within the complete vibration observation window as the chassis propagation vibration effect of the complete vibration observation window; sum the products of the net vibration power spectral density of the chassis at the effective frequency point of the current complete vibration observation window, the vibration power propagation coefficient of the current extension position at the effective frequency point, and the frequency interval as the predicted chassis propagation vibration effect.

5. The method for controlling the operational stability of a robot aluminum column based on vibration detection according to claim 3, characterized in that, The summation of the net vibration power spectral density of the upper part of the column at effective frequency points according to frequency intervals forms the measured vibration action of the upper part of the column, including: Obtain the net vibration power spectral density of the upper part of the column at the effective frequency point of the analysis sub-window; sum the products of the net vibration power spectral density of the upper part of the column at each effective frequency point and the frequency interval to obtain the measured vibration action of the upper part of the column in the analysis sub-window; take the arithmetic mean of the measured vibration action of the upper part of the column in the three analysis sub-windows within the complete vibration observation window as the measured vibration action of the upper part of the column in the complete vibration observation window.

6. The method for controlling the operational stability of a robot aluminum column based on vibration detection according to claim 3, characterized in that, The method for obtaining the vibration relationship associated with the lifting and lowering motion is as follows: With the robot chassis stationary, the aluminum column is controlled to execute a lifting calibration trajectory, and lifting motion calibration samples are collected. A stationary state sample is collected at the calibration extension position as a zero-motion calibration sample. Using the square of the lifting velocity, the square of the lifting acceleration, and the square of the lifting jerk as fitting variables, and the measured vibration of the upper part of the column as the fitting target, the fitting coefficients for the constant term, the square of the lifting velocity, the square of the lifting acceleration, and the square of the lifting jerk are determined. The calibration extension position that uniquely determines the four fitting coefficients is taken as the valid calibration position. The four fitting coefficients between adjacent valid calibration positions are linearly interpolated according to the extension position. The lifting calibration trajectory is re-executed to obtain verification data. The lifting direction appearing simultaneously in both the lifting motion calibration samples and the verification data is taken as the verified lifting direction.

7. The method for controlling the operational stability of a robot aluminum column based on vibration detection according to claim 6, characterized in that, The determination of the vibration path state based on the vibration transmitted from the chassis, the vibration associated with the lifting motion, and the measured vibration on the upper part of the column includes: When the complete vibration observation window data is valid and the absolute difference between the measured vibration action of the upper part of the column and the sum of the vibration action of the chassis propagation and the vibration action of the lifting motion is not greater than the joint prediction error limit, the chassis propagation vibration action range is formed by taking the maximum value between the chassis propagation vibration action range minus the chassis propagation prediction error limit and zero as the lower limit and the chassis propagation vibration action range plus the chassis propagation prediction error limit as the upper limit; the lifting motion-related vibration action range is formed by taking the maximum value between the lifting motion-related vibration action range minus the lifting motion-related prediction error limit and zero as the lower limit and the lifting motion-related vibration action range plus the lifting motion-related prediction error limit as the upper limit; when the lower limit of the lifting motion-related vibration action range is greater than the upper limit of the chassis propagation vibration action range, it is determined to be the lifting motion-related dominant state; when the lower limit of the chassis propagation vibration action range is greater than the upper limit of the lifting motion-related vibration action range, it is determined to be the chassis propagation dominant state; otherwise, it is determined to be an indeterminate state.

8. The method for controlling the operational stability of a robot aluminum column based on vibration detection according to claim 7, characterized in that, The process of forming a baseline lifting trajectory based on the current extension position, lifting speed, lifting acceleration, and target extension position, and adjusting the lifting speed according to the vibration path state to generate candidate lifting trajectories, ensuring that the candidate lifting trajectories have the same start and end motion states as the baseline lifting trajectory, includes: The subsequent lifting direction is determined along the baseline lifting trajectory. When the subsequent lifting direction is a verified lifting direction, the first of the target extension position and the first calibrated extension position reached along the subsequent lifting direction is taken as the predicted termination position. The distance from the initial extension position to the predicted termination position is divided by the time required for the baseline lifting trajectory to reach the predicted termination position for the first time from the start time of the candidate lifting trajectory to obtain the baseline average lifting speed. Under the lifting motion association-dominant state, the lifting speed that is allowed to be set by the lifting drive controller, is consistent with the subsequent lifting direction, and has an absolute value of lifting speed less than the baseline average lifting speed is selected, and the absolute value of the selected lifting speed is taken as the candidate average lifting speed. Under the chassis propagation-dominant state, the lifting speed that is allowed to be set by the lifting drive controller, is consistent with the subsequent lifting direction, and has an absolute value of lifting speed greater than the baseline average lifting speed is selected, and the absolute value of the selected lifting speed is taken as the candidate average lifting speed. The extension position, lifting speed, and lifting acceleration of the baseline lifting trajectory at the start time of the candidate lifting trajectory and when it first reaches the predicted termination position are used as boundary conditions to form the candidate lifting trajectory.

9. The method for controlling the operational stability of a robot aluminum column based on vibration detection according to claim 7, characterized in that, The step of executing candidate or baseline rising / falling trajectories based on the predicted upper and lower bounds of the vibration path state and cumulative vibration amount includes: In the state dominated by lifting motion association, if the predicted upper bound of the cumulative vibration of the lifting motion association of the candidate lifting trajectory is less than the predicted lower bound of the cumulative vibration of the lifting motion association of the reference lifting trajectory, and the predicted upper bound of the cumulative vibration of the upper part of the column of the candidate lifting trajectory is less than the predicted lower bound of the cumulative vibration of the upper part of the column of the reference lifting trajectory, the candidate lifting trajectory is executed; otherwise, the reference lifting trajectory is executed. In the state dominated by chassis propagation, if the predicted upper bound of the cumulative vibration of the chassis propagation of the candidate lifting trajectory is less than the predicted lower bound of the cumulative vibration of the chassis propagation of the reference lifting trajectory, and the predicted upper bound of the cumulative vibration of the upper part of the column of the candidate lifting trajectory is less than the predicted lower bound of the cumulative vibration of the upper part of the column of the reference lifting trajectory, the candidate lifting trajectory is executed; otherwise, the reference lifting trajectory is executed.

10. The method for controlling the operational stability of a robot aluminum column based on vibration detection according to claim 6, characterized in that, The methods for obtaining the chassis propagation prediction error limit, the lifting motion correlation prediction error limit, and the joint prediction error limit are as follows: The chassis calibration motion is re-executed at each calibration extension position to obtain chassis propagation verification data. The maximum absolute difference between the chassis propagation vibration magnitude and the measured vibration magnitude at the top of the column in the chassis propagation verification data is taken as the chassis propagation prediction error limit. The lifting calibration trajectory is re-executed with the robot chassis stationary to obtain lifting motion correlation verification data. The maximum absolute difference between the lifting motion correlation vibration magnitude and the measured vibration magnitude at the top of the column in the lifting motion correlation verification data is taken as the lifting motion correlation prediction error limit. Joint verification data is obtained under the condition that robot chassis movement and aluminum column lifting exist simultaneously. The maximum absolute difference between the measured vibration magnitude at the top of the column and the sum of the chassis propagation vibration magnitude and the lifting motion correlation vibration magnitude in the joint verification data is taken as the joint prediction error limit.

Images