A method and system for determining actuator deceleration point position

Abstract

The application discloses a kind of actuator deceleration point position determination method and system, comprising: S0, obtain initialization parameter, initialization parameter includes initialization contact position estimated value and initialization variance;S1, obtain last round effective contact position measurement;S2, calculate current round adaptive gain according to last round variance and preset measurement noise parameter;S3, calculate and update current round contact position estimated value according to current round adaptive gain;S4, calculate and update current round variance according to last round contact position estimated value and last round effective contact position measurement;S5, calculate and update current round safety margin according to current round variance;S6, calculate and update current round deceleration point position according to current round safety margin and current round contact position estimated value;S7, round number does not reach preset value, enter next round and repeat S1-S6, round number reaches preset value, output current round deceleration point position as target deceleration point position.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of automation control technology, and in particular to a method and system for determining the deceleration point position of an actuator. Background Technology

[0002] In the field of industrial automation, the actuators (such as pressure heads and probes) in contact operation equipment such as sorting machines, spring testing machines and chip pressing equipment are used to apply force or displacement to the workpiece for processing or even precision machining. It is necessary to ensure the accuracy of the contact position of the actuator with the workpiece. Ensuring the accuracy of the contact position mainly relies on the actuator being able to decelerate at the accurate deceleration point and continue to move to the contact position.

[0003] Currently, the main control methods for actuators are to determine the accurate contact position based on manual adjustment or by directly averaging historical contact position measurements.

[0004] However, in actual processing, the properties of the workpiece will change due to the changes in the processing environment, and there will be deviations in each processing. Therefore, the contact position predicted by the actuator in the above way will have a large deviation. As a result, the deceleration will be advanced or delayed according to the predicted deceleration point in actual practice, which will lead to a decrease in production efficiency or cause high-speed impact damage between the mechanism and the workpiece. Summary of the Invention

[0005] This invention provides a method and system for determining the deceleration point position of an actuator. Under the protection of a safety margin, the deceleration point position at which the actuator needs to begin deceleration is calculated through multiple precise learning processes. This provides a learning sample data basis for ensuring that the contact end of the actuator will not damage the workpiece due to high speed when it contacts the workpiece in actual operation, thereby improving the accuracy and safety of machining.

[0006] In a first aspect, embodiments of the present invention provide a method for determining the deceleration point position of an actuator. This method is applied to an actuator, which includes a contact end, and includes: S0. Obtain initialization parameters, including the initial contact position estimate and the initial variance; S1. Obtain the effective contact position measurement value from the previous round; S2. Calculate the adaptive gain for the current round based on the variance of the previous round and the preset measurement noise parameters; S3. Calculate and update the contact position estimate for the current round based on the adaptive gain of the current round; S4. Calculate and update the variance of the current round based on the estimated contact position of the previous round and the measured effective contact position of the previous round; S5. Calculate and update the safety margin for the current round based on the variance of the current round; S6. Calculate and update the deceleration point position of the current round based on the safety margin of the current round and the estimated contact position of the current round; S7. When the number of rounds has not reached the preset value, proceed to the next round and repeat S1 to S6. When the number of rounds reaches the preset value, output the deceleration point position of the current round as the target deceleration point position.

[0007] The method for determining the deceleration point position of the actuator provided in this invention calculates the adaptive gain of the current cycle based on the variance of the previous cycle and preset measurement noise parameters that can affect the learning convergence. This achieves the goal of controlling the learning progress to continuously approach the convergence direction using measurement noise parameters, ensuring the accuracy of the entire learning process, improving the success rate of the learning process, and ensuring that the confidence level of the updated contact position estimate gradually increases. The method quantifies the variance of the current cycle based on the contact position estimate and effective contact position measurement value of the previous cycle, and calculates and updates the safety margin of the current cycle using the variance of each cycle. This solves the problem of existing technologies that only fix the safety margin, which must be set according to the worst-case scenario, or require manual intervention to adjust the safety margin, leading to excessive margin and efficiency loss under normal production conditions. By using the variance, which reflects the uncertainty of the deviation between the contact position estimate and the actual contact position in each cycle, the method can dynamically adjust according to the actual working conditions. The measurement results are superior to traditional methods, and it can significantly shorten the non-productive deceleration stroke and improve the overall production cycle time while ensuring the same level of safety. With the protection of safety margin, the deceleration point where the actuator needs to start slowing down is calculated through multiple precise learning processes. This provides a learning sample data basis for ensuring that the contact end of the actuator will not damage the workpiece due to high speed when it comes into contact with it in actual operation, thereby improving the accuracy and safety of machining.

[0008] Secondly, embodiments of the present invention also provide a system for determining the position of the actuator deceleration point, the system comprising: Host computer and actuator; The actuator includes a contact end and is connected to a host computer. The actuator is used to drive the contact end to move according to the control instructions of the host computer so that the contact end decelerates at the target deceleration point and contacts the workpiece. The host computer includes a data acquisition module, an anomaly detection module, an adaptive filtering module, a safety margin calculation module, and a deceleration point output module; among them, The data acquisition module is used to acquire initialization parameters, including the initial contact position estimate, the initial variance, and the effective contact position measurement values ​​from the previous round. The anomaly detection module calculates the difference between the effective contact position measurement value and the contact position estimate value of the previous round to obtain the position deviation. If the position deviation does not exceed three times the standard deviation of the previous round, the step of calculating the adaptive gain of the current round based on the variance of the previous round and the preset measurement noise parameters continues. If the position deviation exceeds three times the standard deviation of the previous round, the learning parameters of the previous round are determined based on the current outlier count, and the process returns to the step of obtaining the effective contact position measurement value of the previous round. The adaptive filtering module is used to calculate the adaptive gain of the current round based on the variance of the previous round and the preset measurement noise parameters; calculate and update the contact position estimate of the current round based on the adaptive gain of the current round; and calculate and update the variance of the current round based on the contact position estimate of the previous round and the effective contact position measurement value of the previous round. The safety margin calculation module is used to calculate and update the safety margin for the current round based on the variance of the current round. The deceleration point output module is used to calculate and update the deceleration point position of the current round based on the safety margin of the current round and the estimated contact position of the current round. When the number of rounds reaches a preset value, it outputs the deceleration point position of the current round as the target deceleration point position.

[0009] Thirdly, embodiments of the present invention also provide an electronic device, the electronic device comprising: At least one processor; and A memory that is communicatively connected to at least one processor; wherein, The memory stores a computer program that can be executed by at least one processor, such that the at least one processor is able to perform the method for determining the actuator deceleration point position according to any embodiment of the present invention.

[0010] Fourthly, embodiments of the present invention also provide a computer-readable storage medium storing computer instructions, which are used to cause a processor to execute a method for determining the position of the actuator deceleration point according to any embodiment of the present invention.

[0011] Fifthly, embodiments of the present invention also provide a computer program product, including a computer program that, when executed by a processor, implements the method for determining the actuator deceleration point position according to any embodiment of the present invention.

[0012] It should be noted that the aforementioned computer instructions may be stored, in whole or in part, on a computer-readable storage medium. This computer-readable storage medium may be packaged together with the processor (host computer) of the system for determining the actuator deceleration point location, or it may be packaged separately from the processor of the system for determining the actuator deceleration point location; this application does not impose any limitations on this.

[0013] The descriptions of the second, third, fourth, and fifth aspects in this application can be referred to the detailed description of the first aspect; and the beneficial effects of the descriptions of the second, third, fourth, and fifth aspects can be referred to the analysis of the beneficial effects of the first aspect, which will not be repeated here.

[0014] In this application, the name of the system for determining the actuator deceleration point position does not limit the device or functional module itself. In actual implementation, these devices or functional modules may appear under other names. As long as the function of each device or functional module is similar to that of this application, it falls within the scope of the claims of this application and its equivalents.

[0015] These or other aspects of this application will become more readily apparent in the following description. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 A flowchart illustrating a method for determining the deceleration point position of an actuator, provided in an embodiment of the present invention; Figure 2 An example diagram of the abnormal detection process for effective contact position measurement values ​​provided in this embodiment of the invention; Figure 3 Example flowchart of the process for determining the safety margin provided in the embodiments of the present invention; Figure 4 An example diagram of the modules included in an actuator deceleration point position determination system provided in an embodiment of the present invention. Detailed Implementation

[0018] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, the accompanying drawings show only the parts relevant to the present invention, and not all of the structures.

[0019] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0020] The terms “initial” and “target” in the specification and drawings of this application are used to distinguish different objects or to distinguish different treatments of the same object, rather than to describe a specific order of objects.

[0021] Furthermore, the terms "comprising" and "having," and any variations thereof, used in the description of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the steps or units listed, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus.

[0022] Before discussing the exemplary embodiments in more detail, it should be noted that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the operations (or steps) as sequential processes, many of these operations can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but may also have additional steps not included in the figures. The process can correspond to a method, function, procedure, subroutine, subroutine, etc. Moreover, embodiments and features in the embodiments of the present invention can be combined with each other without conflict.

[0023] It should be noted that in the embodiments of this application, the words "exemplary" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the words "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0024] In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0025] Figure 1This is a flowchart illustrating a method for determining the deceleration point position of an actuator according to an embodiment of the present invention. This embodiment is applicable to situations where the host computer automatically learns the actuator deceleration point position multiple times (in practice, at least hundreds of times). Specifically, to ensure that the host computer can accurately predict the estimated contact position of the actuator in actual use, it is necessary to control the actuator and the host computer to learn the prediction according to the method provided in this embodiment. This method can be executed by an actuator deceleration point position determination system, which mainly includes a host computer and an actuator. The host computer is used to automatically learn and determine the actuator deceleration point position multiple times, and the actuator is used to drive the contact end to move according to the control commands of the host computer so that the contact end decelerates at the target deceleration point position and contacts the workpiece. (Continue to refer to...) Figure 1 This embodiment specifically includes: S0. Obtain initialization parameters, including the initial contact position estimate and the initial variance.

[0026] The initial contact position estimate in the initialization parameters mainly comes from three sources in practice: The first is the nominal contact position from the equipment's mechanical design, which refers to the standard position, coordinate point / spatial area where, under ideal conditions, the two components to be contacted (probe / pressure head / clamp / test head and the workpiece being measured) should make contact, as pre-defined in the equipment drawings, mechanical design, and control program; the second is the contact position from the initial manual teaching; and the third is the effective contact position measurement from the first actual contact operation (the effective contact position measurement value from the zeroth round). The initialization variance reflects the high degree of uncertainty regarding the contact position under the initial conditions.

[0027] Specifically, in the initial round, the initialization parameters can be directly used as the parameters to be used in calculations after the zeroth round (initial round); therefore, in this embodiment, the initialization parameters include not only the initial contact position estimate but also... Initialize variance σ 2 0 (set to a large value, such as 0.1 square millimeters), may also include the adaptive gain K1 of the initial round (calculated using the initial variance, the calculation method will be explained in detail later), the preset value (i.e. the upper limit of the number of learning times of this method, usually set between 100 and 500) and the counter (including the round counter = 0 and the outlier counter = 0).

[0028] S1. Obtain the effective contact position measurement value from the previous round.

[0029] The effective contact position measurement value is determined by the position of the actuator recorded by the contact detection algorithm at the instant of contact detection. In specific implementation, the actuator starts from the initial position according to the motion plan, decelerates at the deceleration point, and moves towards the workpiece at a preset low approach speed. When the contact detection algorithm (based on the sudden change of the force sensor signal or the inflection point of the displacement-force curve) detects the instant of contact, it records the absolute position coordinates of the actuator in the equipment coordinate system. In this embodiment, in the initial round, the effective contact position measurement value of the "previous round" is the effective contact position measurement value of the zeroth round. Therefore, it can be seen that the following steps actually correspond to the "first round" and the subsequent round steps.

[0030] For example, after obtaining the effective contact position measurement values ​​from the previous round, the method further includes: (a) Calculate the difference between the effective contact position measurement value of the previous round and the contact position estimate value of the previous round to obtain the position deviation.

[0031] Specifically, after obtaining the effective contact position measurement value from the previous round, in order to determine whether the abnormality of the effective contact position measurement value is due to contact detection mis-triggeredness, workpiece missingness, or sensor failure, the difference between the effective contact position measurement value from the previous round and the contact position estimate from the previous round can be calculated to obtain the position deviation, as shown in the following formula: ; Where e is the positional deviation; "||" represents taking the absolute value; p meas This indicates the measured effective contact position value from the previous round; This represents the estimated contact position value from the previous round; in this embodiment, n represents the current round, and n-1 represents the previous round.

[0032] (ii) If the position deviation does not exceed three times the standard deviation of the previous round, continue to execute the step of calculating the adaptive gain of the current round based on the variance of the previous round and the preset measurement noise parameters.

[0033] Among them, three times the standard deviation of the previous round is the standard deviation σ corresponding to the variance of the previous round. （n-1） The value after multiplying by 3 is 3σ. （n-1） .

[0034] Specifically, if the position deviation does not exceed (is less than or equal to) three times the standard deviation of the previous round, it can be determined that the contact position estimate of the previous round is normal data. Therefore, the step of calculating the adaptive gain of the current round based on the variance of the previous round and the preset measurement noise parameters can continue. At the same time, when it is determined that "the position deviation does not exceed three times the standard deviation of the previous round", the round counter is updated by incrementing by one; for example, if the initial round is round zero, when it is determined that "the position deviation does not exceed three times the standard deviation of the previous round", the first round is entered.

[0035] (iii) If the position deviation exceeds three times the standard deviation of the previous round, the learning parameters of the previous round are determined based on the current outlier count, and the process returns to the step of obtaining the effective contact position measurement value of the previous round.

[0036] The outlier count is updated by the outlier counter when the "position deviation exceeds three times the standard deviation of the previous round" to represent the total number of times outlier data occurs. The learning parameters are the contact position estimate, variance, number of rounds, and outlier count in this embodiment.

[0037] Specifically, Figure 2 This is an example diagram illustrating the abnormal detection process of effective contact position measurements provided in an embodiment of the present invention; as shown below. Figure 2 As shown, after obtaining the effective contact position measurement value, input the effective contact position p. meas ; and calculate the position deviation e according to the formula in (a) above; then, it can be determined whether the position deviation exceeds three times the standard deviation: e > 3σ （n-1） If so, meaning the position deviation exceeds three times the standard deviation of the previous round, the outlier count is updated by incrementing the outlier counter to determine if the current outlier count exceeds a preset outlier value (which can be 3 in practice), as shown in the figure. If it does, the learning parameters are initialized, meaning all learning parameters are initialized and updated to the initial parameters (e.g., the contact position estimate is updated to the initial contact position estimate and the variance is updated to the initial variance). Then, the process returns to the step of obtaining the valid contact position measurement value from the previous round (at this time, the number of rounds and the outlier count are all reset to zero). If it does not exceed the standard deviation, the update is skipped, meaning the learning parameters from the previous round are retained, and the process returns to the step of obtaining the valid contact position measurement value from the previous round. Therefore, in the above, when the position deviation exceeds three times the standard deviation of the previous round, the valid contact position measurement value from the previous round is determined to be an outlier, and it is necessary to obtain the valid contact position measurement value from the previous round again.

[0038] In this embodiment, after obtaining the effective contact position measurement value from the previous round, it is necessary to use the difference between the effective contact position measurement value from the previous round and the estimated contact position value from the previous round to determine whether there is an anomaly in the effective contact position measurement value from the previous round. Different processing methods are selected based on whether the effective contact position measurement value is normal or abnormal. This enables timely monitoring of anomalies in the effective contact position measurement value from each learning round, preventing abnormal data from directly entering the calculation of normal rounds and interfering with the subsequent learning process, thus preventing the error in the contact position obtained in each subsequent round from continuously increasing. Furthermore, when it is determined that the outlier count exceeds a preset anomaly value, the learning parameters are directly reset. This ensures that when multiple consecutive abnormal data occur, the learning parameters are directly reset to prevent the inability to enter the normal learning steps due to the continuous occurrence of abnormal data, thereby realizing learning reset when multiple consecutive abnormal data occur.

[0039] S2. Calculate the adaptive gain for the current round based on the variance of the previous round and the preset measurement noise parameters.

[0040] This step corresponds to the first round and subsequent rounds. Therefore, the variance from the previous round corresponds to the initial variance in the first round. The preset measurement noise parameter is an adjustable design parameter; its larger value results in smoother learning but slower convergence. Therefore, the preset measurement noise parameter can be set according to learning requirements. The adaptive characteristic of the adaptive gain enables the algorithm to converge quickly in the early stages of learning (high variance) and remain stable in the later stages (low variance).

[0041] Specifically, the adaptive gain for the current round can be calculated directly based on the variance of the previous round and the preset measurement noise parameter. For example, the adaptive gain can be calculated directly based on the minimum mean square error criterion, and the setting of the preset measurement noise parameter can be determined by referring to the principle of gain characteristics: (1) When the variance is much greater than the preset measurement noise parameter, the adaptive gain approaches 1. At this time, the characterization learning process trusts the measurement value more and can achieve fast tracking learning; (2) When the variance is much less than the preset measurement noise parameter, the adaptive gain approaches 0. At this time, the characterization learning process trusts the estimate value more and can achieve smooth filtering learning; (3) When the variance is much equal to the preset measurement noise parameter, the adaptive gain is about 0.5. At this time, the characterization learning process is in a balanced state. Therefore, according to the above principles, users can select appropriate preset measurement noise parameters according to their learning needs.

[0042] In this embodiment, since variance can quantify the uncertainty of the deviation between the estimated contact position and the actual contact position, the adaptive gain of the current round is calculated based on the variance of the previous round and the preset measurement noise parameter that can affect the learning convergence. This not only incorporates the uncertainty of the deviation between the estimated contact position and the measured value of the previous round into the adaptive gain, but also enables the measurement noise parameter to control the learning progress to continuously approach the convergence direction, ensuring the accuracy of the entire learning process and improving the success rate of the learning process convergence.

[0043] For example, calculating the adaptive gain for the current round based on the variance of the previous round and preset measurement noise parameters includes: (a) Sum the variance of the previous round with the preset measurement noise parameter to obtain the sum value.

[0044] Specifically, the variance of the previous round can be added to a preset measurement noise parameter to obtain the sum.

[0045] (ii) The adaptive gain of the current round is obtained by dividing the variance of the previous round by the sum.

[0046] Specifically, the adaptive gain for the current round can be obtained by dividing the variance of the previous round by the sum.

[0047] For example, the equivalent formula for the above steps is as follows: ; Among them, K n This is the adaptive gain for the current round n; R is the variance of the previous round; R is the preset measurement noise parameter. In this embodiment, in the first round, since the "variance of the previous round" used is... ,therefore Furthermore, taking the recommended parameters as an example, K1=0.1 / (0.1+0.001)≈0.99, indicating that the algorithm almost completely trusts the measured values ​​during the first update, which is in line with the engineering intuition that "high initial uncertainty should lead to rapid learning".

[0048] In this embodiment, based on the minimum mean square error criterion, by substituting the variance from the previous round and the preset measurement noise parameters, the relationship between the variance and the preset measurement noise parameters at different stages of learning can be quantified more accurately. The adaptive gain determined based on this relationship enables subsequent algorithms to converge quickly when the variance is large in the early stages of learning, and to ensure stability in the later stages when the variance is small. Therefore, in subsequent multi-round cyclic learning to determine the actuator deceleration point position, the adaptive gain can be used to ensure smooth adaptive filtering during the learning process, improving learning convergence efficiency and maintaining learning stability.

[0049] S3. Calculate and update the contact position estimate for the current round based on the adaptive gain of the current round.

[0050] Specifically, after determining the adaptive gain for the current round, a new contact position estimate can be calculated using the adaptive gain for the current round. Therefore, the contact position estimate for the previous round can be updated using the new contact position estimate to obtain the contact position estimate for the current round.

[0051] In this embodiment, the adaptive gain of the current round can be used to calculate and update the contact position estimate of the current round. The updated contact position estimate can be adjusted based on the adaptive gain to ensure that the confidence of the contact position estimate gradually increases, which provides a basis for accurately calculating the deceleration point position based on the contact position estimate with high confidence.

[0052] For example, calculating and updating the contact position estimate for the current round based on the adaptive gain of the current round includes: (a) Calculate the product of the position deviation and the adaptive gain of the current round to obtain the updated estimate.

[0053] Specifically, the positional deviation (not taken as absolute value) can be multiplied by the adaptive gain of the current round to obtain the updated estimate.

[0054] (ii) Sum the updated estimate with the contact position estimate from the previous round to obtain and update the contact position estimate for the current round.

[0055] Specifically, the updated estimate can be added to the contact position estimate from the previous round, and the new contact position estimate is used to update the contact position estimate from the previous round, which is then used as the contact position estimate for the current round.

[0056] For example, the above steps can be used to calculate the estimated contact position for the current round according to the following formula: ; in, The estimated contact position for the current round; This is the estimated contact position from the previous round; k n This is the adaptive gain for the current round; This represents the positional deviation before its absolute value was taken.

[0057] Furthermore, in the first round, the updated contact position estimate for the current round (first round) is, at this time, if If the third option, "using the effective contact position measurement value of the first actual contact operation (effective contact position measurement value of the zeroth round)," is selected, then the above formula automatically degenerates into... .

[0058] In this embodiment, the calculated position deviation can quantitatively reflect the gap between the measured value and the estimated value. Adding the adaptive gain of the current round can quantify the credibility of the gap. Then, adding it to the contact position estimate of the previous round can make the contact position estimate updated in the current round closer to the true value, ensuring that the contact position estimate can gradually approach the actual value and improve the accuracy of the contact position estimate.

[0059] S4. Calculate and update the variance of the current round based on the estimated contact position of the previous round and the effective contact position measurement of the previous round.

[0060] In this embodiment, variance quantification represents the uncertainty of the deviation between the estimated contact position and the actual contact position (not the measured value). This uncertainty comprehensively reflects workpiece dimensional tolerance, fixture positioning error, sensor measurement noise, and contact detection algorithm error. Furthermore, workpiece dimensional tolerance mainly refers to the random fluctuations in the height / thickness of workpieces in the same batch; fixture positioning error mainly refers to the deviation of the workpiece's installation position in the fixture; sensor measurement noise refers to the random measurement error of the force sensor or displacement sensor; and contact detection algorithm error refers to the uncertainty of the moment of contact determination.

[0061] Specifically, since variance can reflect the difference between the estimated contact position and the effective contact position measurement, the variance of the current round can be further determined and updated directly based on the difference between the estimated contact position and the effective contact position measurement of the previous round.

[0062] For example, calculating and updating the variance of the current round based on the contact position estimate and the effective contact position measurement of the previous round includes: (a) Calculate the product of the complementary term of the adaptive gain of the current round and the variance of the previous round to obtain the estimated attenuation term.

[0063] The complementary term of the adaptive gain is 1 - adaptive gain. In this embodiment, the estimated attenuation term reflects the reduction in uncertainty after incorporating new measurements.

[0064] Specifically, the complementary term of the adaptive gain in the current round can be multiplied by the product of the variances of the previous rounds to obtain the estimated attenuation term.

[0065] (ii) Calculate the square of the position deviation to obtain the second power of the measurement innovation, and multiply the adaptive gain of the current round, the complementary term of the adaptive gain of the current round, and the second power of the measurement innovation in sequence to obtain the estimated correction term.

[0066] In this embodiment, the estimation correction term enables the entire algorithm to adapt to the actual measurement noise level. Therefore, the algorithm can adaptively adjust the variance estimation according to the actual observation deviation level, rather than relying solely on the preset noise parameters.

[0067] Specifically, the square of the position deviation can be calculated first to obtain the second power of the measurement information. Then, the adaptive gain of the current round, the complementary term of the adaptive gain of the current round, and the second power of the measurement information can be multiplied in sequence to obtain the estimated correction term.

[0068] (iii) Summing the estimated attenuation term and the estimated correction term, we obtain and update the variance of the current round.

[0069] Specifically, the estimated decay term and the estimated correction term are summed at the end, and the resulting value can be used as the variance of the current round, thus completing the update of the variance of the current round.

[0070] For example, the above steps can be used to calculate the variance of the current round using the following formula: ; in, The variance of the current round; 1-k n This is the complementary term of the adaptive gain for the current round; This represents the variance of the previous round; To estimate the attenuation term; This is the estimated correction term.

[0071] Furthermore, continuing with the example above, in the first round, the variance of the updated current round (first round) is... At this time, if If the third option, "using the effective contact position measurement value from the first actual contact operation (effective contact position measurement value from the zeroth round)," is selected, then the estimation correction term in the above formula is 0. The variance decreased significantly, and normal recursion resumed from the second round of measurements thereafter. This means that the consistency between the measured and estimated values ​​is itself valid information, which enhances the confidence in the current estimate and reduces uncertainty. The algorithm does not need to add extra judgment steps for the case where the innovation is zero, and the formula system has self-consistently and completely handled all cases.

[0072] In this embodiment, the attenuation term is calculated based on the complementary terms of the variance from the previous round and the adaptive gain from the current round, reflecting the degree of uncertainty reduction after incorporating new measurements. The variance correction term based on the innovation is determined based on the adaptive gain and the square of the position deviation, enabling the algorithm to adapt to the actual measurement noise level. Furthermore, since the difference between the effective contact position measurement and the estimated contact position is used as a product term in the estimation correction term, the final estimated variance reflects the update magnitude and confidence level of the final estimated variance due to the difference (innovation) between the effective contact position measurement and the estimated contact position.

[0073] S5. Calculate and update the safety margin for the current round based on the variance of the current round.

[0074] Specifically, since the statistical reliability of the variance in the early stages may be insufficient when using this scheme for multi-round cyclic calculations, the safety margin for the current round can be calculated and updated based on the variance of the current round.

[0075] In this embodiment, the variance of the current round can be quantified based on the estimated contact position and the measured effective contact position of the previous round. The safety margin of the current round is calculated and updated based on the variance of each round. This solves the problem in the prior art where only a fixed safety margin is set, which must be set according to the worst-case scenario, or where manual adjustment of the safety margin requires long-term manual monitoring, resulting in excessive margin and efficiency loss under normal production conditions. Therefore, by using the variance that reflects the uncertainty of the deviation between the estimated contact position and the actual contact position in each round, the system can be dynamically adjusted according to the actual working conditions. The measurement results are superior to those of the traditional method, and the non-productive deceleration stroke can be significantly shortened and the overall production cycle can be improved while ensuring the same level of safety.

[0076] For example, calculating and updating the safety margin for the current round based on the variance of the current round includes: (a) Determine the dynamic margin parameters based on the number of rounds and the variance of the current round.

[0077] In this embodiment, there is no direct functional relationship between the number of rounds and the difference between the estimated contact position and the effective contact position measurement (information). However, its growth rate is mainly affected by factors such as production cycle time, the proportion of abnormal data, and the reset event triggered by continuous abnormal data. The production cycle time means that only one effective contact position measurement is generated after each physical contact operation is completed. Therefore, the growth rate of the number of rounds is related to the production speed of the equipment. The proportion of abnormal data means that when the position deviation exceeds three times the standard deviation of the previous round, it is determined that only the outlier count is updated. At this time, the data is abnormal, so the number of rounds cannot be updated. If abnormalities occur frequently during production (such as a high workpiece missing rate), the number of rounds will increase slowly. The reset event triggered by continuous abnormal data is the reset event triggered when "the current outlier count exceeds the preset abnormal value", which resets the number of rounds to zero.

[0078] Specifically, the number of rounds reflects the number of normal data samples (i.e., the updated contact position estimate for the current round) obtained in the cyclic rounds (i.e., subsequent update steps will only be executed after the above-mentioned anomaly judgment is passed), and the variance of the current round reflects the difference between the current contact position estimate and the true value. For example, the more rounds the current round has, the larger the sample size of normal data obtained, which is statistically significant. At the same time, the smaller the variance of the current round, the higher the reliability of the normal data obtained at this time. Therefore, the number of rounds and the variance of the current round can reflect the reliability of statistical analysis using the currently obtained normal data samples.

[0079] For example, determining the dynamic margin parameter based on the number of rounds in the current round and the variance of the current round includes: (1) Determine whether the number of rounds in the current round is less than the minimum number of rounds; if it is less, execute (2); if it is not less, execute (3).

[0080] In this embodiment, the minimum number of rounds can be 10.

[0081] Specifically, in the early stages of a cycle, the number of normal samples may be small. Even if the estimated variance of the current cycle meets the requirements, the statistical reliability may be insufficient due to the small number of normal samples. Therefore, in practice, it is necessary to first determine whether the number of cycles in the current cycle is less than the minimum number of cycles so that the number of normal samples reaches a level of statistical reliability. Then, the estimated variance corresponding to the normal samples can be used to further determine the dynamic margin parameters corresponding to the normal samples in the current cycle.

[0082] (2) If the number of rounds in the current round is less than the minimum number of rounds, then the dynamic margin parameter is determined as the margin parameter in the initial learning stage.

[0083] In this embodiment, the dynamic margin parameters include a basic safety margin and a confidence coefficient. During the initial learning phase, the basic safety margin is 0.3 mm, and the confidence coefficient is 3.0.

[0084] Specifically, if the number of rounds in the current round is less than the minimum number of rounds, for example, if the minimum number of rounds is 10, then if the number of rounds in the current round is 9 or less, then the dynamic margin parameter at this time can be determined to be the margin parameter in the early learning stage, which is the most conservative margin parameter.

[0085] (3) If the number of rounds in the current round is greater than or equal to the minimum number of rounds, the dynamic margin parameter is determined as the learning mid-term margin parameter when the variance of the current round exceeds the variance confidence threshold, or the dynamic margin parameter is determined as the learning stable period margin parameter when the variance of the current round does not exceed the variance confidence threshold.

[0086] In this embodiment, the basic safety margin in the mid-learning margin parameter is 0.2 mm with a confidence coefficient of 2.5. The basic safety margin in the stable-learning margin parameter is 0.1 mm with a confidence coefficient of 2.0. The variance confidence threshold can be 0.01 square millimeters in this example.

[0087] Specifically, once the number of rounds in the current round is greater than or equal to the minimum number of rounds, it can be determined that the number of normal samples meets the statistical reliability requirements. Therefore, the variance of each round (i.e., the estimated variance corresponding to the normal samples in each round) can be used to further determine whether the contact position estimate of the current round has sufficiently converged, thus determining the dynamic margin parameter at this point. Therefore, it can be further determined whether the variance of the current round exceeds the variance confidence threshold, for example, whether the variance of the current round is greater than 0.01 square millimeters. If it is greater, it can be determined that the estimated variance is still relatively large, the contact position estimate of the current round varies greatly, and the estimated variance has not yet converged sufficiently. Therefore, the dynamic margin parameter is determined as the mid-learning margin parameter. If it is less than or equal to, it can be determined that the estimated variance has converged to a lower level, the contact position estimate of the current round has high confidence, and the dynamic margin parameter is determined as the learning stable period margin parameter.

[0088] In this embodiment, the dynamic margin parameter is not preset, but is adaptively determined according to the learning progress. When the number of rounds is small, the accumulation of effective samples is insufficient. Regardless of the value of the estimated variance, the reliability of the estimated variance is not statistically reliable with respect to the number of normal samples. Therefore, a conservative margin parameter is required in the early stage of learning, which can use a large margin to ensure the safety and stability of the learning process. When the number of rounds reaches the required level, the reliability of the normal data at this time is further determined based on the estimated variance of each round. This realizes the automatic reduction of the margin to improve efficiency based on the estimated variance. At the same time, it can also automatically determine the margin in the middle or stable period of learning, realizing the adaptive two-factor adjustment of the dynamic margin parameter.

[0089] (ii) Calculate and update the safety margin for the current round based on the standard deviation corresponding to the variance of the dynamic margin parameter and the variance of the current round.

[0090] In this embodiment, the standard deviation corresponding to the variance of the current round is the "standard deviation" in the three-times-standard-deviation example mentioned above.

[0091] Specifically, after determining the dynamic margin parameter, a new safety margin can be calculated based on the standard deviation corresponding to the variance of the current round and the dynamic margin parameter. This new safety margin can then be used as the safety margin for the current round.

[0092] For example, the safety margin for the current round is calculated and updated based on the standard deviation corresponding to the variance of the dynamic margin parameter and the current round, including: (1) Calculate the product of the standard deviation and the confidence coefficient corresponding to the variance of the current round to obtain the product value.

[0093] Specifically, the standard deviation corresponding to the variance of the current round can be obtained first, and then the standard deviation can be multiplied by the confidence coefficient in the dynamic margin parameter to obtain the product value.

[0094] (2) Sum the product value with the basic safety margin to obtain and update the safety margin for the current round.

[0095] Specifically, the product value can be added to the basic safety margin in the dynamic margin parameter to obtain the new safety margin, which can then be used as the safety margin for the current round.

[0096] For example, the above steps can be used to calculate the safety margin for the current round as follows: ; Among them, M safety For the safety margin of the current round (the current round number n in the subscript is omitted); M base Basic safety margin; k σ σ is the confidence coefficient; n This represents the standard deviation corresponding to the variance of the current round.

[0097] In this embodiment, the current standard deviation and its corresponding confidence coefficient are used as correction terms for the basic safety margin. By adding the basic safety margin, not only is dynamic correction of the basic safety margin achieved, but the safety margin terms that need to be corrected can also be accurately determined based on the confidence level of the standard deviation.

[0098] For example, Figure 3 Example flowchart of the process for determining the safety margin provided in embodiments of the present invention; as follows: Figure 3 As shown, after determining the variance of the current round, you can "input" the number of rounds and the variance of the current round. Then, first determine if the number of rounds is less than 10. If yes, directly determine the dynamic margin parameter as the initial learning margin parameter. If not, then determine if the variance of the current round is greater than 0.01. If yes, determine the dynamic margin parameter as the mid-learning margin parameter; if not, determine the dynamic margin parameter as the learning stabilization period parameter. After determining the dynamic margin parameter, the safety margin M of the current round can be calculated according to the above formula. safety And output it.

[0099] S6. Calculate and update the deceleration point position of the current round based on the safety margin and the estimated contact position of the current round.

[0100] Specifically, since the contact position estimate of the current round predicts the contact position where the contact end of the actuator touches the component, and the safety margin of the current round can provide a safety margin for the actuator, it is possible to deduce from the contact position estimate and the safety margin which position the actuator can first move parallel to the workpiece before decelerating downward to start contacting the workpiece. That is, it is deduced that the actuator first moves parallel to the workpiece to the deceleration point position, and then the deduced deceleration point position is used as the deceleration point position of the current round.

[0101] For example, calculating and updating the deceleration point position of the current round based on the safety margin of the current round and the estimated contact position of the current round includes: The deceleration point position of the current round is obtained and updated by successively subtracting the estimated contact position, the distance required for deceleration, and the safety margin of the current round.

[0102] The distance required for deceleration is mainly determined by the velocity and acceleration parameters of the motion plan.

[0103] Specifically, after determining the safety margin for the current round, the deceleration point position for the current round can be calculated using the following formula: , where P decel This is the deceleration point position for the current round; D decel The distance required to decelerate.

[0104] S7. When the number of rounds has not reached the preset value, proceed to the next round and repeat S1 to S6. When the number of rounds reaches the preset value, output the deceleration point position of the current round as the target deceleration point position.

[0105] The preset value is a pre-defined threshold for the number of learning iterations.

[0106] Specifically, in the actual learning process, if the number of rounds has not reached the preset value, the next round can be entered and the steps S1-S6 above can be repeated until the number of rounds reaches the preset value. Then, the deceleration point position of the current round can be used as the target deceleration point position.

[0107] In this embodiment, the dynamic safety margin and the estimated contact position of the current round are used to calculate and update the deceleration point position of the current round. When the number of learning rounds reaches a preset value, the deceleration point position of the current round is output as the target deceleration point position. Under the protection of the safety margin, the deceleration point position that the actuator needs to start decelerating can be calculated through multiple accurate learnings. This provides a learning sample data basis for the contact end of the actuator to avoid damaging the workpiece due to high speed when it contacts the workpiece in actual operation, thus improving the accuracy and safety of machining.

[0108] Optionally, after outputting the deceleration point position of the current round as the target deceleration point position, the following is also included: (a) Determine whether the variance of the current round at the time of output is less than or equal to the variance convergence threshold.

[0109] The variance convergence threshold is a pre-set variance convergence limit for successful learning; it can be adjusted according to the user's needs or expectations regarding the accuracy of the contact position estimate in actual practice.

[0110] Specifically, since this embodiment aims to enable the actuator to learn independently and ensure that the estimated contact position during actual operation is closer to the actual contact position, it is necessary to determine whether the learning was successful after each learning process. Therefore, it can be determined whether the variance of the current round at the time of output is less than or equal to the variance convergence threshold.

[0111] (ii) When the variance of the current round at the time of output is less than or equal to the variance convergence threshold, the determination of the deceleration point position is successful.

[0112] Specifically, when the variance of the current round at the time of output is less than or equal to the variance convergence threshold, it can be determined that the deceleration point position has been successfully determined. Therefore, the main parameters obtained from this learning process can be saved so that the contact point position of the actuator can be accurately predicted when the actuator deceleration point position determination system is used in actual practice. The main parameters to be saved may include: the estimated contact position value of the current round at the time of output, the variance of the current round, the safety margin of the current round, and the deceleration point position of the current round.

[0113] Optionally, if the variance of the current round at the time of output is greater than the variance convergence threshold, then the learning can be determined to be converged. Therefore, it is necessary to record the main parameters of the current round and the non-convergence flag, and retrain to ensure that the contact position estimate and deceleration point position can be accurately determined when the actuator is actually used in the future.

[0114] The method for determining the deceleration point position of the actuator provided in this embodiment of the invention quantifies the uncertainty of the deviation between the estimated contact position and the actual contact position by using variance. Therefore, it calculates the adaptive gain of the current round based on the variance of the previous round and the preset measurement noise parameter that can affect the learning convergence. This not only incorporates the uncertainty of the deviation between the estimated contact position and the measured value of the previous round into the adaptive gain, but also enables the measurement noise parameter to control the learning progress to continuously approach the convergence direction, ensuring the accuracy of the entire learning process, improving the successful convergence of the learning process, and ensuring that the confidence of the updated contact position estimate gradually increases, thus providing a basis for accurately calculating the deceleration point position based on the contact position estimate with high confidence. The variance of the current round is quantified by the estimated contact position and the measured effective contact position from the previous round. The safety margin of the current round is calculated and updated by the variance of each round. This solves the problem of existing technologies that only fix the safety margin, which must be set according to the worst-case scenario, or require manual intervention to adjust the safety margin, which requires long-term manual monitoring and leads to excessive margin and efficiency loss under normal production conditions. By reflecting the variance that reflects the uncertainty between the estimated contact position and the actual contact position in each round, the system can dynamically adjust according to the actual working conditions. The measurement results are better than those of traditional methods. Furthermore, it can significantly shorten the non-productive deceleration stroke and improve the overall production cycle time while ensuring the same level of safety. By utilizing dynamic safety margin and contact position estimation, the deceleration point position of the current cycle is calculated and updated. When the number of learning cycles reaches a preset value, the deceleration point position of the current cycle is output as the target deceleration point position. Under the protection of the safety margin, the deceleration point position at which the actuator needs to start decelerating is calculated through multiple accurate learnings. This provides a learning sample data basis for ensuring that the contact end of the actuator will not damage the workpiece due to high speed when it contacts the workpiece in actual operation, thereby improving the accuracy and safety of machining.

[0115] Optionally, the method provided in this embodiment can be applied to multi-station management in practical use. Multi-stations can include sorting machine stations, pressure testing machine stations, and automated equipment stations requiring precise control of the contact process, such as assembly robots. For example, each station (each containing an actuator and its corresponding contact end) can be learned in the above manner, and the learned samples can be used as the station parameters. For instance, after actual testing or experimentation, taking a spring sorting machine as an example, after conducting experiments according to the above steps, the following specific parameter configuration can be obtained: Therefore, in practice, the learning parameters for each workstation in a multi-station system can be determined in the manner described above. For example, the central control unit determines the workstation selection signal, which is then sent by the workstation router to the host computers of different workstations to perform the aforementioned learning and determine the corresponding workstation parameters.

[0116] Furthermore, in the experiment or test, if the horizontal axis represents the number of rounds (i.e., the data sequence number of the sample data) n (taking the 1st to 100th rounds as an example in the experiment), and the vertical axis represents the contact position value (including the effective contact position measurement value of each round, the actual contact position of the actuator's contact end with the workpiece controlled according to the contact position estimate in each round, and the contact position estimate value of each round); for example, a horizontal dashed line represents the line connecting the points of the actual contact position, scattered points represent the effective contact position measurement value, a solid line represents the line connecting the points of the contact position estimate value, and the shaded area represents the confidence interval (which can be set according to needs, for example, set to...). In the initial stage (rounds n=1 to 5), since the initial adaptive gain K1≈0.99 is close to 1, the estimated contact position quickly approaches the actual contact position value of approximately 10.04 mm from the initial set value of 10.8 mm after only one update. At this time, the confidence interval is relatively wide, reflecting the algorithm's conservative handling of initial uncertainties. In the intermediate stage (n=5 to 15), the estimated contact position value gradually approaches the actual contact position value with small adjustments. The adaptive gain gradually decreases from approximately 0.23 to approximately 0.07, and the confidence interval narrows significantly, indicating that the algorithm's estimation accuracy continues to improve. In the stable stage (n>15), the estimated contact position value stabilizes within ±0.001 mm of the actual value of 10.0 mm, the confidence interval tends to be at its minimum width, and the algorithm enters a high-precision steady state. An abnormal contact position measurement value p appeared on the 12th test. meas =12.5 mm, which deviates from the estimated contact position by about 2.5 mm, far exceeding the range of 3 times the standard deviation. Therefore, the abnormal contact position measurement can be eliminated. All curves are not affected by this outlier, which verifies the effectiveness and robustness of the 3 times standard deviation anomaly detection mechanism.

[0117] Furthermore, the adaptive variation process of the estimated variance and safety margin in this embodiment can also be implemented in experiments with the horizontal axis representing the number of rounds (i.e., sample number n) and the left vertical axis representing the estimated variance σ. 2 n (Unit: square millimeters), the right vertical axis represents the safety margin M. safety (Unit: mm). The estimated variance rapidly decreases from the initial variance of 0.1 mm², dropping to the order of 0.0001 mm² after approximately 10 valid updates and continuing to converge, reaching only about 0.000003 mm² by the 100th update, demonstrating the algorithm's characteristic of continuously improving estimation confidence with data accumulation. The safety margin correspondingly increases from the initial approximately 1.25 mm (in M...base =0.3 mm plus k σ =3.0 multiplied by σ) adaptively shrinks to approximately 0.104 mm after stabilization (according to M). base =0.1 mm plus k σ =2.0 multiplied by σ), the reduction rate reached approximately 92%. The stage boundary line at n=10 is clearly visible. The dynamic margin parameter automatically switches from the margin parameter in the early learning stage to the margin parameter in the stable learning stage based on the variance confidence threshold. The distance from the deceleration point to the contact position is shortened from approximately 3.25 mm initially to approximately 2.1 mm, significantly shortening the non-productive deceleration stroke and improving the overall operating efficiency of the equipment.

[0118] Optionally, an embodiment of the present invention provides a system for determining the position of an actuator deceleration point, the system comprising: Host computer and actuator; The actuator includes a contact end and is connected to a host computer. The actuator is used to drive the contact end to move according to the control instructions of the host computer so that the contact end decelerates at the target deceleration point and contacts the workpiece. The host computer includes a data acquisition module, an anomaly detection module, an adaptive filtering module, a safety margin calculation module, and a deceleration point output module; among them, The data acquisition module is used to acquire initialization parameters, including the initial contact position estimate, the initial variance, and the effective contact position measurement values ​​from the previous round. The anomaly detection module calculates the difference between the effective contact position measurement value and the contact position estimate value of the previous round to obtain the position deviation. If the position deviation does not exceed three times the standard deviation of the previous round, the step of calculating the adaptive gain of the current round based on the variance of the previous round and the preset measurement noise parameters continues. If the position deviation exceeds three times the standard deviation of the previous round, the learning parameters of the previous round are determined based on the current outlier count, and the process returns to the step of obtaining the effective contact position measurement value of the previous round. The adaptive filtering module is used to calculate the adaptive gain of the current round based on the variance of the previous round and the preset measurement noise parameters; calculate and update the contact position estimate of the current round based on the adaptive gain of the current round; and calculate and update the variance of the current round based on the contact position estimate of the previous round and the effective contact position measurement value of the previous round. The safety margin calculation module is used to calculate and update the safety margin for the current round based on the variance of the current round. The deceleration point output module is used to calculate and update the deceleration point position of the current round based on the safety margin of the current round and the estimated contact position of the current round. When the number of rounds reaches a preset value, it outputs the deceleration point position of the current round as the target deceleration point position.

[0119] Optionally, the adaptive gain for the current round is calculated based on the variance of the previous round and preset measurement noise parameters. The adaptive filtering module is specifically used for: The variance of the previous round is summed with the preset measurement noise parameter to obtain the sum value; the adaptive gain of the current round is obtained by dividing the variance of the previous round by the sum value.

[0120] Optionally, the adaptive filtering module calculates and updates the contact position estimate for the current round based on the adaptive gain of the current round. Specifically, the adaptive filtering module is used to: Calculate the product of the position deviation and the adaptive gain of the current round to obtain the updated estimate; sum the updated estimate with the contact position estimate of the previous round to obtain and update the contact position estimate of the current round.

[0121] Optionally, the variance of the current round is calculated and updated based on the contact position estimate and the effective contact position measurement of the previous round. Specifically, the adaptive filtering module is used to: Calculate the product of the complementary term of the adaptive gain in the current round and the variance of the previous round to obtain the estimated attenuation term; calculate the square of the position deviation to obtain the second power of the measurement innovation, and multiply the adaptive gain in the current round, the complementary term of the adaptive gain in the current round, and the second power of the measurement innovation in sequence to obtain the estimated correction term; sum the estimated attenuation term and the estimated correction term to obtain and update the variance of the current round.

[0122] Optionally, the safety margin for the current round is calculated and updated based on the variance of the current round. The safety margin calculation module is specifically used for: The dynamic margin parameter is determined based on the number of rounds and the variance of the current round; the safety margin of the current round is calculated and updated based on the standard deviation corresponding to the variance of the current round and the dynamic margin parameter.

[0123] Optionally, the dynamic margin parameter is determined based on the number of rounds in the current round and the variance of the current round. The safety margin calculation module is specifically used for: Determine if the number of rounds in the current round is less than the minimum number of rounds. If the number of rounds in the current round is less than the minimum number of rounds, then the dynamic margin parameter is determined as the learning initial margin parameter. If the number of rounds in the current round is greater than or equal to the minimum number of rounds, then the dynamic margin parameter is determined as the learning mid-term margin parameter when the variance of the current round exceeds the variance confidence threshold, or the dynamic margin parameter is determined as the learning stable period margin parameter when the variance of the current round does not exceed the variance confidence threshold.

[0124] Optionally, the dynamic margin parameters include a basic safety margin and a confidence coefficient; the safety margin for the current round is calculated and updated based on the standard deviation corresponding to the variance of the current round and the dynamic margin parameters. The safety margin calculation module is specifically used for: Calculate the product of the standard deviation and the confidence coefficient corresponding to the variance of the current round to obtain the product value; sum the product value with the basic safety margin to obtain and update the safety margin of the current round.

[0125] Optionally, the deceleration point position of the current round is calculated and updated based on the safety margin of the current round and the estimated contact position of the current round. The safety margin calculation module is specifically used for: The deceleration point position of the current round is obtained and updated by successively subtracting the estimated contact position, the distance required for deceleration, and the safety margin of the current round.

[0126] Optionally, the host computer also includes a parameter storage module. After outputting the deceleration point position of the current round as the target deceleration point position, the parameter storage module is also used for: Determine if the variance of the current round at the time of output is less than or equal to the variance convergence threshold; if the variance of the current round at the time of output is less than or equal to the variance convergence threshold, the determination of the deceleration point position is successful.

[0127] Optionally, Figure 4 An example diagram of the modules included in an actuator deceleration point position determination system provided in an embodiment of the present invention; as shown below. Figure 4 As shown, the host computer includes a data acquisition module, an anomaly detection module, an adaptive filtering module, a safety margin calculation module, and a deceleration point output module. It may also include a multi-station management module and a parameter storage module. The actuator includes a motion control module and a contact detection module. The multi-station management module primarily determines the currently learned station and then informs the adaptive filtering module and parameter storage module of the selected station. The parameter storage module can read the determined effective parameters (such as estimates and variances) from the adaptive filtering module and can also provide historical parameter data to the adaptive filtering module. The motion control module in the actuator, after obtaining the deceleration point position output by the deceleration point output module, controls the contact end to move according to the deceleration point position and contact the workpiece. The contact detection module in the actuator, at the instant contact is detected, records the absolute position coordinates of the actuator in the equipment coordinate system, i.e., the contact position measurement value.

[0128] The actuator deceleration point position determination system provided in this embodiment of the invention can execute the actuator deceleration point position determination method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0129] It is worth noting that in the embodiments of the actuator deceleration point position determination system described above, the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the scope of protection of the present invention.

[0130] The acquisition, storage, use, and processing of data in the technical solution of this invention all comply with the relevant provisions of national laws and regulations.

[0131] Note that the above description is merely a preferred embodiment of the present invention and the technical principles employed. Those skilled in the art will understand that the present invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of the present invention. Therefore, although the present invention has been described in detail through the above embodiments, the present invention is not limited to the above embodiments, and may include many other equivalent embodiments without departing from the concept of the present invention, the scope of which is determined by the scope of the appended claims.

Claims

1. A method for determining the position of the deceleration point of an actuator, characterized in that, Applied to an actuator, the actuator including a contact end, the method includes: S0. Obtain initialization parameters, including the initial contact position estimate and the initial variance; S1. Obtain the effective contact position measurement value from the previous round; S2. Calculate the adaptive gain for the current round based on the variance of the previous round and the preset measurement noise parameters; S3. Calculate and update the contact position estimate for the current round based on the adaptive gain of the current round; S4. Calculate and update the variance of the current round based on the estimated contact position of the previous round and the effective contact position measurement of the previous round; S5. Calculate and update the safety margin for the current round based on the variance of the current round; S6. Calculate and update the deceleration point position of the current round based on the safety margin of the current round and the estimated contact position of the current round; S7. When the number of rounds has not reached the preset value, proceed to the next round and repeat S1 to S6. When the number of rounds reaches the preset value, output the deceleration point position of the current round as the target deceleration point position.

2. The method according to claim 1, characterized in that, After obtaining the effective contact position measurements from the previous round, the following is also included: The position deviation is obtained by calculating the difference between the effective contact position measurement value and the contact position estimate value of the previous round. If the position deviation does not exceed three times the standard deviation of the previous round, then continue to execute the step of calculating the adaptive gain of the current round based on the variance of the previous round and the preset measurement noise parameters; If the position deviation exceeds three times the standard deviation of the previous round, the learning parameters of the previous round are determined based on the current outlier count, and the process returns to the step of obtaining the effective contact position measurement value of the previous round.

3. The method according to claim 1, characterized in that, The step of calculating the adaptive gain for the current round based on the variance of the previous round and preset measurement noise parameters includes: The summation value is obtained by summing the variance of the previous round with the preset measurement noise parameter. The adaptive gain for the current round is obtained by dividing the variance of the previous round by the sum.

4. The method according to claim 2, characterized in that, The step of calculating and updating the contact position estimate for the current round based on the adaptive gain of the current round includes: Calculate the product of the position deviation and the adaptive gain of the current round to obtain the updated estimate; The updated estimate is summed with the contact position estimate from the previous round to obtain and update the contact position estimate for the current round.

5. The method according to claim 2, characterized in that, The step of calculating and updating the variance of the current round based on the contact position estimate and the effective contact position measurement of the previous round includes: Calculate the product of the complementary term of the adaptive gain in the current round and the variance of the previous round to obtain the estimated attenuation term; Calculate the square of the position deviation to obtain the second power of the measurement innovation, and then multiply the adaptive gain of the current round, the complementary term of the adaptive gain of the current round, and the second power of the measurement innovation in sequence to obtain the estimated correction term; The estimated decay term and the estimated correction term are summed to obtain and update the variance of the current round.

6. The method according to claim 1, characterized in that, The step of calculating and updating the safety margin for the current round based on the variance of the current round includes: The dynamic margin parameters are determined based on the number of rounds in the current round and the variance of the current round. The safety margin for the current round is calculated and updated based on the standard deviation corresponding to the variance of the current round and the dynamic margin parameter.

7. The method according to claim 6, characterized in that, The process of determining the dynamic margin parameter based on the number of rounds and the variance of the current round includes: Determine if the number of rounds in the current round is less than the minimum number of rounds. If the number of rounds in the current round is less than the minimum number of rounds, then the dynamic margin parameter is determined to be the margin parameter for the initial learning stage. If the number of rounds in the current round is greater than or equal to the minimum number of rounds, then the dynamic margin parameter is determined to be the mid-learning margin parameter when the variance of the current round exceeds the variance confidence threshold, or the dynamic margin parameter is determined to be the learning stable period margin parameter when the variance of the current round does not exceed the variance confidence threshold.

8. The method according to claim 6, characterized in that, The dynamic margin parameters include the basic safety margin and the confidence coefficient; The step of calculating and updating the safety margin for the current round based on the dynamic margin parameter and the standard deviation corresponding to the variance of the current round includes: Calculate the product of the standard deviation corresponding to the variance of the current round and the confidence coefficient to obtain the product value; Summing the product with the basic safety margin yields and updates the safety margin for the current round.

9. The method according to claim 1, characterized in that, The step of calculating and updating the deceleration point position of the current round based on the safety margin and the estimated contact position of the current round includes: The deceleration point position of the current round is obtained and updated by successively subtracting the estimated contact position, the distance required for deceleration, and the safety margin of the current round.

10. The method according to claim 1, characterized in that, After outputting the current deceleration point position as the target deceleration point position, the following is also included: Determine whether the variance of the current round at the time of output is less than or equal to the variance convergence threshold; If the variance of the current round at the time of output is less than or equal to the variance convergence threshold, the determination of the deceleration point position is successful.

11. A system for determining the position of an actuator deceleration point, characterized in that, include: Host computer and actuator; The actuator includes a contact end and is communicatively connected to the host computer. The actuator is used to drive the contact end to move according to the control command of the host computer so that the contact end decelerates at the target deceleration point and contacts the workpiece. The host computer includes a data acquisition module, an anomaly detection module, an adaptive filtering module, a safety margin calculation module, and a deceleration point output module; wherein, The data acquisition module is used to acquire initialization parameters, which include an initial contact position estimate, an initial variance, and an acquisition of the effective contact position measurement value from the previous round. The anomaly detection module is used to calculate the difference between the effective contact position measurement value and the contact position estimate value of the previous round to obtain the position deviation. If the position deviation does not exceed three times the standard deviation of the previous round, the step of calculating the adaptive gain of the current round based on the variance of the previous round and the preset measurement noise parameters continues. If the position deviation exceeds three times the standard deviation of the previous round, the learning parameters of the previous round are determined based on the current outlier count, and the step of obtaining the effective contact position measurement value of the previous round is returned. The adaptive filtering module is used to calculate the adaptive gain of the current round based on the variance of the previous round and preset measurement noise parameters; calculate and update the contact position estimate of the current round based on the adaptive gain of the current round; and calculate and update the variance of the current round based on the contact position estimate of the previous round and the effective contact position measurement value of the previous round. The safety margin calculation module is used to calculate and update the safety margin of the current round based on the variance of the current round. The deceleration point output module is used to calculate and update the deceleration point position of the current round based on the safety margin of the current round and the estimated contact position of the current round, and outputs the deceleration point position of the current round as the target deceleration point position when the number of rounds reaches a preset value.

Images