Multi-layer algorithm fusion method and system for controlling actuator and storage medium

Abstract

The invention relates to the technical field of machine control, and discloses a multi-layer algorithm fusion method and system for controlling an actuator and a storage medium, and the method comprises the steps: obtaining a plurality of types of sensor data, distributing the plurality of types of sensor data to a built-in control algorithm, calculating single-line control data, and obtaining a single-line control data matrix according to the plurality of types of single-line control data; forming a sensing data matrix by using various sensing data, and performing calculation by using a fusion algorithm to obtain a fusion control data matrix; calculating a control matrix distance according to the single-line control data matrix and the fusion control data matrix; if the control matrix distance is smaller than the reference matrix distance, a final control data matrix is calculated according to the single-line control data matrix and the fusion control data matrix, and a plurality of actuators are controlled according to the final control data matrix; otherwise, using the single-line control data matrix to control a plurality of actuators; through the synergistic effect of the upgrading algorithm and the curing algorithm, the control precision and stability after vehicle motion control are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of machine control, in particular to a multi-layer algorithm fusion method and system for controlling an actuator and a storage medium. BACKGROUND

[0002] In the control process related to vehicles or robots, corresponding sensors first acquire various types of relevant data in the running process according to the preset rule system set acquisition logic, and then input the directly acquired relevant data into a calculation module. After the calculation module performs operation processing on the original data, control data for controlling the actuator is obtained, and finally the control data is transmitted to the corresponding actuator to drive the actuator to perform the corresponding control action, thereby achieving preliminary regulation and control of the vehicle running state.

[0003] The calculation algorithm for processing the collected data and generating the control data is pre-solidified in the vehicle chip. Through the predetermined operation logic, control data corresponding to different control requirements is obtained. These control data are individually and accurately allocated to various actuators of the vehicle, such as power actuators responsible for power output adjustment of the vehicle, braking actuators for ensuring driving safety, steering actuators for controlling the driving direction of the vehicle, and suspension actuators for maintaining the stability of the vehicle body, and so on. Each actuator completes the corresponding control action according to the received control data.

[0004] However, the algorithm solidified in the vehicle chip has obvious limitations, and it cannot be directly upgraded. With the continuous improvement of vehicle control requirements and the increase of complex driving scenarios, the control data generated by the original solidified algorithm has been difficult to meet the higher precision motion control requirements, resulting in poor control effect. SUMMARY

[0005] In order to be able to perform fusion calculation on the solidified algorithm to upgrade the algorithm, through the synergistic effect of the upgraded algorithm and the solidified algorithm, the control precision and stability after vehicle motion control are improved, the present application provides a multi-layer algorithm fusion method and system for controlling an actuator and a storage medium.

[0006] In a first aspect, the present application provides a multi-layer algorithm fusion method for controlling an actuator, which adopts the following technical scheme: A multi-layer algorithm fusion method for controlling an actuator, comprising the following steps: According to the preset sensor, a plurality of sensing data are acquired, and the plurality of sensing data are allocated to one-to-one built-in control algorithms. The built-in control algorithms calculate single-line control data, and a single-line control data matrix is obtained according to the plurality of single-line control data; A plurality of sensing data are formed into a sensing data matrix, and a fusion control data matrix is calculated according to the preset fusion algorithm based on the sensing data matrix; The control matrix distance is calculated according to the single-line control data matrix and the fusion control data matrix through a matrix distance algorithm; If the control matrix distance is less than a preset reference matrix distance, a final control data matrix is calculated according to the single-line control data matrix and the fusion control data matrix, and the plurality of actuators are controlled according to the final control data matrix; otherwise, the plurality of actuators are controlled using the single-line control data matrix.

[0007] By adopting the technical solution, the data is distributed to the one-to-one built-in control algorithm to generate the single-line control data and form the single-line control data matrix, and the stable output of the original algorithm is retained. The plurality of sensing data is formed into a sensing data matrix, and the fusion control data matrix is calculated through a fusion algorithm. The control matrix distance is calculated through a matrix distance algorithm, and compared with the reference matrix distance. When the control matrix distance is less than the reference matrix distance, the final control data matrix is obtained by combining the single-line control data matrix and the fusion control data matrix to control the plurality of actuators. Otherwise, the single-line control data matrix is used, the organic fusion of the new and old algorithms is realized, and the control effect and reliability are taken into account.

[0008] Optionally, the fusion algorithm comprises: acquiring category data of the actuators; matching the associated control parameters corresponding to the sensing data of different sensors according to the category data; calculating the execution control data using the associated control parameters through a weighting algorithm according to the sensing data of different sensors, wherein the execution control data one-to-one corresponds to the actuators; the plurality of execution control data form the fusion control data matrix.

[0009] By adopting the technical solution, the fusion algorithm is based on the actuator characteristics to accurately associate the sensing data and the control logic, the associated control parameters are more adaptable, the weighting algorithm can be targeted to integrate the multi-source sensing data, and the control effect of the execution control data is improved.

[0010] Optionally, the matrix distance algorithm comprises: calculating the difference value between the corresponding elements of the single-line control data matrix and the fusion control data matrix to obtain a difference value matrix; calculating the average absolute error of the difference value matrix as the control matrix distance; wherein the average absolute error is .

[0011] By adopting the technical solution, the average absolute error is calculated as the control matrix distance through the difference value matrix, and the difference between the two control data matrices can be accurately quantified.

[0012] Optionally, in the step of calculating the final control data matrix according to the single-line control data matrix and the fusion control data matrix, the following sub-steps are further included: The single-line control data matrix is associated with elements at corresponding positions in the fusion control data matrix one by one; The final element is calculated according to the associated elements based on a preset weighted average algorithm; The final element forms a final control data matrix.

[0013] By adopting the above technical solutions, the single-line control data matrix is accurately associated with the elements of the fusion control data matrix, the final element obtained through the weighted average algorithm is more in line with the actual control requirements, and the final control data matrix can take into account the advantages of the two matrices to provide more optimal data for the controller.

[0014] Optionally, within a preset time period, the proportion of the number of times that the control matrix distance is less than the preset reference matrix distance is recorded, and the weighting coefficient of the element at the corresponding position in the fusion control data matrix is adjusted according to the positive correlation of the proportion of the number of times, that is, the greater the proportion of the number of times, the greater the weighting coefficient, and the smaller the proportion of the number of times, the smaller the weighting coefficient.

[0015] By adopting the above technical solutions, the weighting coefficient is dynamically adjusted according to the proportion of the number of times within the time period, the weight is increased when the proportion of the number of times is high to strengthen the fusion effect, and the weight is reduced when the proportion of the number of times is low to reduce the influence of deviation.

[0016] Optionally, within a preset time period, the proportion of the number of times that the control matrix distance is less than the preset reference matrix distance is recorded, and the distance average value is calculated according to the corresponding number of control matrix distances to update the reference matrix distance according to the proportion of the number of times.

[0017] By adopting the above technical solutions, the reference matrix distance is dynamically updated, the average value of the corresponding control matrix distance is selected by the proportion of the number of times, the adaptability of the reference matrix distance is improved, and the subsequent matrix distance comparison is more accurate.

[0018] Optionally, within a preset time period, the proportion of the number of times that the control matrix distance is less than the preset reference matrix distance is recorded, and the reference matrix distance is adjusted according to the inverse correlation of the proportion of the number of times, that is, the greater the proportion of the number of times, the smaller the reference matrix distance, and the smaller the proportion of the number of times, the greater the reference matrix distance.

[0019] By adopting the above technical solutions, the control matrix distance is reduced when the proportion of the number of times is high, which can improve the control threshold of the fusion control data matrix; and the control matrix distance is expanded when the proportion of the number of times is low, which can ensure the continuity of system control and make the control matrix distance judgment more in line with the actual situation.

[0020] Optionally, in the step of controlling a plurality of actuators according to the final control data matrix, the following sub-steps are further included: The associated control parameter is adjusted according to the inverse correlation of the control matrix distance, that is, the greater the control matrix distance, the smaller the associated control parameter, and the smaller the control matrix distance, the greater the associated control parameter.

[0021] By adopting the technical solution, the correlation control parameter is dynamically adjusted according to the control matrix distance; the correlation control parameter is reduced when the control matrix distance is large, so that the influence of deviation is reduced, and the correlation control parameter is increased when the control matrix distance is small, so that the fusion effect is strengthened.

[0022] In a second aspect, the application provides a multi-layer algorithm fusion system for controlling an actuator, which adopts the following technical solution: A multi-layer algorithm fusion system for controlling an actuator, comprising a processor, wherein the processor executes the steps of the multi-layer algorithm fusion method for controlling an actuator according to any one of the above.

[0023] In a third aspect, the application provides a storage medium, which adopts the following technical solution: A storage medium, wherein the storage medium stores a program, and the program is executed by a processor to implement the steps of the multi-layer algorithm fusion method for controlling an actuator according to any one of the above.

[0024] In summary, the application has at least one of the following beneficial technical effects: data is allocated to one-to-one built-in control algorithms to generate single-line control data and form a single-line control data matrix, thereby retaining the stable output of the original algorithm. A plurality of sensing data is formed into a sensing data matrix, and a fusion control data matrix is obtained through a fusion algorithm. The control matrix distance is calculated through a matrix distance algorithm, and compared with a reference matrix distance. When the control matrix distance is less than the reference matrix distance, a final control data matrix is obtained by combining the single-line control data matrix and the fusion control data matrix to control a plurality of actuators. Otherwise, the single-line control data matrix is used, thereby realizing the organic fusion of new and old algorithms and taking into account the control effect and reliability. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 A multi-layer algorithm fusion method for controlling an actuator.

[0026] Figure 2 A sub-step diagram of the fusion algorithm.

[0027] Figure 3 A sub-step diagram of the matrix distance algorithm. DETAILED DESCRIPTION

[0028] Embodiments of the application are described in detail below, and examples of the embodiments are shown in the accompanying drawings.

[0029] In the description of the present specification, the description referring to the terms "certain embodiments", "one embodiment", "some embodiments", "illustrative embodiments", "example", "specific example" or "some examples" means that the particular feature, structure, material or characteristic being described in connection with the embodiment or example is included in at least one embodiment or example of the application. The illustrative appearances of the above-mentioned terms in the description are not necessarily referring to the same embodiment or example. Moreover, the particular features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0030] The embodiment of the present application discloses a multi-layer algorithm fusion method for controlling actuators, referring to Figure 1 , comprising the following steps: Start the preset sensor acquisition process, control the multiple function sensors carried by the vehicle or robot to synchronously acquire the corresponding dimension of the sensing data at the preset sampling frequency (such as 10 Hz); wherein the multiple function sensors are, for example, the vehicle speed sensor for power control, the brake pressure sensor for brake control, the steering angle sensor for steering control, and the suspension displacement sensor for suspension control. During the acquisition process, the data noise is removed by a filtering algorithm, such as Kalman filtering, to ensure that the accuracy error of the sensing data does not exceed the preset threshold, such as ±0.1%. The acquired multiple sensing data is classified and mapped: the vehicle speed data is allocated to the power control algorithm solidified in the chip, the brake pressure data is allocated to the brake control algorithm, the steering angle data is allocated to the steering control algorithm, and the suspension displacement data is allocated to the suspension control algorithm, to realize one-to-one correspondence between the sensing data and the built-in control algorithm. Each built-in control algorithm is based on its own preset operation logic, such as the power control algorithm based on the PID operation formula of the vehicle speed and the target power demand to calculate the input single dimension sensing data, and output the single line control data corresponding to only a single actuator, such as the torque adjustment value of the power motor output by the power control algorithm and the pressure control value of the brake caliper output by the brake control algorithm. Finally, according to the sequence of the power, brake, steering, and suspension actuators, all single line control data are sequentially arranged to construct a single line control data matrix with a dimension of "1x actuator number", each element in the matrix uniquely maps a basic control instruction of an actuator, and the stable output characteristics of the original solidified algorithm are ensured.

[0031] Preprocessing the collected various sensing data: unifying data format, filling in missing data; unifying data format such as converting all data to decimal floating point type, filling in missing data using linear interpolation method to fill in occasional missing sampling points, and then constructing a sensing data matrix according to the rules of sensor type as row and sampling time sequence as column, such as row dimension of "vehicle speed-brake pressure-steering angle-suspension displacement", column dimension of sampling number, to ensure the integrity and dimension consistency of the matrix data. The sensing data matrix is input into the preset fusion algorithm, and the fusion control data matrix is generated through multi-dimensional data collaborative calculation. Referring to Figure 2 , the specific implementation process is as follows: By reading the ECU (Electronic Control Unit) hardware configuration file of the vehicle or robot, the types, models and numbers of the current actuators to be controlled are obtained, such as 1 power actuator, 1 brake actuator, 1 steering actuator and 4 suspension actuators, and the control response characteristics of each actuator are extracted, such as the maximum pressure response speed of the brake actuator and the angle adjustment range of the steering actuator.

[0032] The preset parameter matching database is called, which stores the mapping relationship of "actuator type-sensor type-associated control parameter", such as the associated weight of the steering actuator corresponding to the steering angle sensor is 0.6 and the associated weight of the steering actuator corresponding to the vehicle speed sensor is 0.4; based on the obtained actuator type data, the associated control parameters of each actuator corresponding to different sensor sensing data are matched to ensure that the parameters adapt to the response characteristics of the actuators, such as the displacement sensing data associated parameter of the suspension actuator needs to match the stiffness adjustment demand of the suspension.

[0033] For each row of sensor data in the sensing data matrix, the associated control parameters matched are combined and a preset weighted sum algorithm is used for operation; the weighted sum algorithm such as weight proportion is based on dynamic allocation of actuator priority, and the weight of safety class actuator is higher than that of comfort class actuator. For example, for the steering actuator, the execution control data, i.e. steering angle adjustment amount, is calculated by the formula: steering angle data x 0.6 + vehicle speed data x 0.4; after the calculation of all actuators is completed, the execution control data is arranged into a "1x actuator number" fusion control data matrix according to the same actuator order as the single-line control data matrix, realizing the collaborative integration of multi-source sensing data and improving the scene adaptability of control data.

[0034] The control matrix distance is calculated according to the single-line control data matrix and the fusion control data matrix by matrix distance algorithm; the preset matrix distance algorithm is called, and the matrix distance algorithm based on mean absolute error (MAE) is used, referring to Figure 3 , to realize accurate quantification of the difference between the single-line control data matrix and the fusion control data matrix, and the specific steps are as follows: The single-line control data matrix is denoted as A, with a dimension of m x n, where m is the number of rows and n is the number of columns, and the elements are denoted as aij. The fused control data matrix is denoted as B, with the same dimension as A, and the elements are denoted as bij. The elements of A and B are associated, and the difference is calculated element by element: For the element in the ith row and jth column of the matrix, the difference dij = aij - bij is calculated, or dij = bij - aij, and the absolute value is taken. For all elements in the matrix, i ranges from 1 to m and j ranges from 1 to n. All differences dij are arranged according to the original matrix dimension to construct the difference matrix D (dimension m x n, elements are dij).

[0035] Based on the difference matrix D, the control matrix distance is calculated by the average absolute error formula, which is defined as: ; where |dij| is the absolute value of the difference matrix element, used to eliminate the offsetting effect of positive and negative deviations and highlight the actual magnitude of the deviation; m x n is the total number of matrix elements, which is processed by averaging to adapt the distance comparison of matrices of different dimensions; when calculating, first traverse all elements of the difference matrix to sum , and then divide by the total number of elements m x n to get the average absolute error MAE, which is taken as the final control matrix distance.

[0036] The preset reference matrix distance is retrieved from the parameter storage module, which is obtained through previous vehicle / robot testing and is the reasonable deviation threshold of single-line control data and fused control data. If the preset value is 0.15, compare the calculated control matrix distance with the reference matrix distance, and execute the control according to the following logic: If the control matrix distance is less than the reference matrix distance, it means that the fused control data is reliable, and the generated final control data matrix is called to send control instructions to each actuator, such as sending the final torque adjustment value to the power actuator and the final pressure control value to the brake actuator, to drive the actuators to work together. At the same time, the anti-correlation adjustment mechanism of the associated control parameters is started: calculate the difference between the control matrix distance and the reference matrix distance, such as 0.1 for the control matrix distance and 0.15 for the reference matrix distance, with a difference of 0.05. The larger the difference, the closer the control matrix distance to the reference matrix distance, and the associated weight of the steering actuator is reduced by a preset adjustment step, such as from 0.6 to 0.55, to reduce the influence of the fused data deviation. The smaller the difference, the smaller the control matrix distance compared to the reference matrix distance, and the associated weight of the steering angle is increased by a preset adjustment step, such as from 0.6 to 0.65, to strengthen the effect of the fusion algorithm and achieve dynamic optimization of the associated control parameters.

[0037] The fusion control condition is not met, that is, the control matrix distance ≥ the reference matrix distance: At this time, it is determined that there is a large deviation in the fusion control data. To ensure control safety and continuity, the final control data matrix is ​​abandoned, and the generated single-line control data matrix is ​​directly called to send single-line control instructions to each actuator to ensure that the actuator action complies with the original stable control logic and avoid control errors caused by fusion data deviation.

[0038] Through the above steps, this method achieves the organic synergy between the built-in control algorithm (old algorithm) and the fusion algorithm (new algorithm): it retains the stable output advantage of the old algorithm, and improves the control accuracy through the multi-source data fusion of the new algorithm. At the same time, through matrix distance judgment and dynamic parameter adjustment, it takes into account both control effect and system reliability. It is suitable for the high-precision actuator control needs of vehicles or robots in complex scenarios.

[0039] Within a preset time period, that is, a pre-set statistical cycle, this time period must be determined in combination with the actuator's control response frequency and the actual application scenario. For example, it can be set to 5 minutes or 100 control cycles to ensure that the statistical sample size is sufficient to reflect the data pattern and avoid single deviations affecting the results. During this time period, the control matrix distance calculated by the matrix distance algorithm is recorded in real time each time, and compared one by one with the preset reference matrix distance. Valid events whose control matrix distance is less than the reference matrix distance are screened out. At the same time, the number of occurrences of this valid event (denoted as N) and the total number of calculations of the control matrix distance during this time period (denoted as M) are accumulated. The number of times within this time period is calculated using the formula: Number of times = Number of valid events N / Total number of calculations M. The preset precision is retained during the calculation process, such as retaining two decimal places, to ensure the accuracy of the percentage data.

[0040] After obtaining the proportion of times within the time period, the weight coefficients of the corresponding position elements in the fusion control data matrix are adjusted based on the positive correlation adjustment rule. The specific implementation is as follows: Clarify the correspondence between the elements in the fusion control data matrix; each element in the matrix corresponds to the fusion control data of a specific actuator, and its weighted coefficient is the weight value of the element when weighted calculation is performed with the corresponding element of the single-line control data matrix in the matrix distance algorithm. For example, the weighted coefficient of the element corresponding to the power actuator in the fusion control data matrix is ​​initially set to 0.6.

[0041] If the calculated frequency ratio is larger, such as 90%, it indicates that the fusion control data and single-line control data have small deviation and high reliability, and the weighting coefficient of the corresponding element is increased by a preset adjustment amplitude, such as 10% per time, 0.1 per time, and the upper limit is 0.9, so as to strengthen the influence of the fusion control data in the final control data matrix and further improve the control precision. If the frequency ratio is smaller, such as 30%, it indicates that the fusion control data has high deviation risk, and the weighting coefficient of the corresponding element is decreased by a preset adjustment amplitude, such as 10% per time, 0.1 per time, and the lower limit is 0.3, so as to reduce the influence of the fusion data deviation on the final control result and ensure the control stability.

[0042] The adjusted weighting coefficient is updated to the parameter library of the matrix distance algorithm, and in the next preset time period, the element calculation of the single-line control data matrix and the fusion control data matrix is performed based on the updated weighting coefficient, so as to realize dynamic iterative optimization of the weighting coefficient.

[0043] A preset statistical time period is set, which needs to match the actuator control period and the scene change frequency, such as 3 minutes or 50 control cycles, to ensure that the sample has timeliness and representativeness. In this time period, after each control matrix distance calculation is completed, the distance value is recorded and compared with the preset initial reference matrix distance, and the effective distance data whose control matrix distance is smaller than the initial reference matrix distance is marked, and the number of effective distance data (denoted as A) and the total number of control matrix distance calculations (denoted as B) in this time period are accumulated. The frequency ratio is calculated by the formula: frequency ratio = effective frequency A / total frequency B, and the calculation result is kept to a preset precision, such as one decimal place, to provide a basis for subsequent distance screening.

[0044] According to the frequency ratio obtained above, the number of control matrix distances to be selected is determined; if the frequency ratio is C, such as C=80%, the number of effective distance data whose control matrix distance is smaller than the initial reference matrix distance is selected from all the control matrix distances recorded in this time period, that is, A control matrix distances are selected, to ensure that the selected data has reliability. The A effective distance data is summed and divided by the number of data A to obtain the distance average value, and the precision consistent with the control matrix distance is kept during the calculation process, such as three decimal places, to avoid the influence of calculation error on subsequent update.

[0045] The calculated average distance is used as a new reference matrix distance, replacing the original initial reference matrix distance, and is stored in the algorithm parameter storage module. The updated reference matrix distance will be directly applied to the control matrix distance comparison link in the next preset time period, realizing dynamic iteration of the reference matrix distance. If the frequency ratio changes in the subsequent time period, such as increasing to 90% or decreasing to 60%, repeat the above steps, select the corresponding number of effective distance data based on the new frequency ratio to calculate the average value, continuously update the reference matrix distance, and ensure that it always adapts to the deviation law of the current control scene, improving the accuracy of matrix distance comparison.

[0046] A preset statistical time period is set, which needs to be combined with the control response period of the actuator and the actual scene dynamic change characteristics, such as 4 minutes or 60 control cycles, to ensure that the statistical sample can reflect the control deviation law under the current scene. In this time period, after calculating the control matrix distance by the matrix distance algorithm each time, the distance value is immediately recorded and compared with the currently effective preset reference matrix distance, and the effective events with control matrix distance less than the reference matrix distance are marked, and the number of effective events (denoted as X) and the total number of control matrix distance calculations in this time period (denoted as Y) are accumulated. According to the formula: frequency ratio = effective frequency X / total frequency Y, the frequency ratio of this time period is calculated, and the calculation result is kept to a preset precision, such as two decimal places, to provide a quantitative basis for subsequent reference matrix distance adjustment.

[0047] Based on the above calculated frequency ratio, the reference matrix distance is dynamically adjusted according to the inverse correlation adjustment rule, and the specific implementation is as follows: Adjustment amplitude preset and association: The basic adjustment amplitude of the reference matrix distance is preset, such as 5%~10% of the initial reference matrix distance each time, and the upper and lower limits of the adjusted distance are set to avoid exceeding the safe range of the actuator control, which needs to be combined with the control accuracy requirements of the actuator and the historical deviation data to determine.

[0048] If the frequency ratio is larger, such as 90%, it indicates that the current fusion control data and single-line control data have small deviation and high reliability, and the reference matrix distance is reduced by a preset adjustment range, such as the initial reference matrix distance is 0.2 and the adjustment range is 10%, then the new reference matrix distance = 0.2 x (1-10%) = 0.18; by reducing the control threshold, the accuracy requirement of the fusion control data matrix is further improved, and the control effect is strengthened. If the frequency ratio is smaller, such as 40%, it indicates that the fusion control data has high deviation risk, and the reference matrix distance is increased by a preset adjustment range, such as the initial reference matrix distance is 0.2 and the adjustment range is 10%, then the new reference matrix distance = 0.2 x (1+10%) = 0.22; by expanding the control threshold, the system control interruption caused by frequent triggering of single-line control logic is avoided, and the control continuity is ensured.

[0049] The adjusted new reference matrix distance replaces the original reference matrix distance and is stored in the parameter storage unit of the algorithm, and is directly called in the control matrix distance comparison link of the next preset time period, so as to realize dynamic iterative optimization of the reference matrix distance and ensure that each judgment is consistent with the actual deviation of the current control scene.

[0050] The embodiment of the application further discloses a multi-layer algorithm fusion system for controlling an actuator, comprising a processor, wherein the processor executes the steps of the multi-layer algorithm fusion method for controlling the actuator as described in any one of the above.

[0051] The embodiment of the application further discloses a storage medium, wherein the storage medium stores a program, and the program is executed by a processor to realize the steps of the multi-layer algorithm fusion method for controlling the actuator as described in any one of the above.

[0052] Although the embodiments of the application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the application, and those skilled in the art can make changes, modifications, replacements and variations to the above embodiments within the scope of the application.

Claims

1. A multi-layer algorithm fusion method for controlling an actuator, characterized in that: The steps include: Acquire multiple sensor data based on preset sensors, assign the multiple sensor data to a one-to-one corresponding built-in control algorithm, calculate the single-line control data based on the built-in control algorithm, and obtain a single-line control data matrix based on the multiple single-line control data; The sensor data matrix is ​​formed by a variety of sensor data, and the fusion control data matrix is ​​calculated based on the preset fusion algorithm input into the sensor data matrix; Calculate the control matrix distance using a matrix distance algorithm based on the single-line control data matrix and the fusion control data matrix; If the control matrix distance is less than the preset reference matrix distance, a final control data matrix is ​​calculated based on the single-line control data matrix and the fusion control data matrix, and multiple actuators are controlled based on the final control data matrix; Otherwise, use a single-line control data matrix to control multiple actuators.

2. The multi-layer algorithm fusion method for controlling actuators according to claim 1, characterized in that: Fusion algorithms include: Get the type data of the actuator; Matching associated control parameters corresponding to sensing data of different sensors according to the type data; Calculating execution control data using associated control parameters according to sensing data of different sensors through a weighted algorithm, wherein the execution control data corresponds one-to-one to the actuator; The plurality of execution control data form a fusion control data matrix.

3. The multi-layer algorithm fusion method for controlling actuators according to claim 1, characterized in that: Matrix distance algorithms include: Calculate the difference between the corresponding elements of the single-line control data matrix and the fusion control data matrix to obtain a difference matrix; Calculate the mean absolute error of the difference matrix as the control matrix distance; the mean absolute error is ; Among them, |dij| is the absolute value of the difference matrix element, which is used to eliminate the offsetting effect of positive and negative deviations and highlight the actual magnitude of the deviation; m×n is the total number of matrix elements. Through averaging processing, the result is adapted to the distance comparison of matrices of different dimensions.

4. The multi-layer algorithm fusion method for controlling actuators according to claim 1, characterized in that: The step of calculating the final control data matrix based on the single-line control data matrix and the fused control data matrix also includes the following sub-steps: Associating the elements at corresponding positions in the single-line control data matrix and the fusion control data matrix one by one; The final element is calculated based on the associated elements based on the preset weighted average algorithm; The final elements form the final control data matrix.

5. The multi-layer algorithm fusion method for controlling actuators according to claim 1, characterized in that: Within a preset time period, the proportion of times the control matrix distance is less than the preset reference matrix distance is recorded, and the weighting coefficient of the element at the corresponding position in the fusion control data matrix is ​​adjusted according to the positive correlation between the proportion of times. The greater the proportion of times, the greater the weighting coefficient, and the smaller the proportion of times, the smaller the weighting coefficient.

6. The multi-layer algorithm fusion method for controlling actuators according to claim 1, characterized in that: In a preset time period, the proportion of times the control matrix distance is less than the preset reference matrix distance is recorded, and the corresponding number of control matrix distances is taken according to the proportion of times to calculate the distance average to update the reference matrix distance.

7. The multi-layer algorithm fusion method for controlling actuators according to claim 1, characterized in that: Within a preset time period, the proportion of times the control matrix distance is less than the preset reference matrix distance is recorded, and the reference matrix distance is adjusted inversely according to the proportion of times. The greater the proportion of times, the smaller the reference matrix distance, and the smaller the proportion of times, the larger the reference matrix distance.

8. The multi-layer algorithm fusion method for controlling an actuator according to claim 2, characterized in that: The step of controlling the plurality of actuators according to the final control data matrix further includes the following sub-steps: The correlation control parameter is adjusted according to the anti-correlation of the control matrix distance. The larger the control matrix distance is, the smaller the correlation control parameter is, and the smaller the control matrix distance is, the larger the correlation control parameter is.

9. A multi-layer algorithm fusion system for controlling actuators, characterized in that: The method comprises a processor, wherein the processor executes the steps of the multi-layer algorithm fusion method for controlling an actuator according to any one of claims 1 to 8.

10. A storage medium, characterized in that: The storage medium stores a program, and when the program is executed by the processor, the steps of the multi-layer algorithm fusion method for controlling the actuator according to any one of claims 1 to 8 are implemented.

Images