Method, device and equipment for identifying control parameters of a device, and medium

Abstract

The application relates to the technical field of automatic control, and discloses a method and device for identifying control parameters of equipment, equipment and a medium. The method for identifying control parameters of equipment provided by the application comprises the following steps: obtaining historical data of equipment under different working conditions; determining function segmentation points of a segmented function model by using error variation rules between predicted data and historical data of the segmented function model; fitting the historical data by using the segmented function model with the function segmentation points determined, and determining function parameters obtained by fitting each function segment as the control parameters of the equipment. The application solves the problem of insufficient model precision by obtaining historical data, automatically determining segmentation points and fitting data, and improves the precision of the control parameters.

Description

Technical Field

[0001] This application relates to the field of automatic control technology, specifically to a method, apparatus, device, and medium for identifying control parameters of a device. Background Technology

[0002] In related technologies, the physical models established for the working devices of loaders (such as buckets and booms) are relatively complex due to their reliance on fluid mechanics and mechanical dynamics. These models contain a large number of parameters that are difficult to measure directly or that drift over time. Furthermore, the control laws of the working devices are not static, which makes the parameter calibration of the physical models built for loaders complex and results in low accuracy. Summary of the Invention

[0003] This application provides a method, apparatus, device, and medium for identifying control parameters of a device. It solves the problems of complex and low-accuracy calibration of physical model parameters for loader construction.

[0004] In a first aspect, this application provides a method for identifying control parameters of a device. The method includes: acquiring historical data of the device under different operating conditions; determining the function segmentation points of the piecewise function model by utilizing the error variation law between the predicted data and historical data of the piecewise function model, wherein the function segmentation points are used to characterize the signal segmentation points of the device's control signal when changing different control laws; fitting historical data to the piecewise function model with the determined function segmentation points, and determining the function parameters obtained by fitting each function segment as the control parameters of the device.

[0005] Based on the above technical means, the inefficiency and non-global optimization problems caused by manually trying to find segment points in traditional loader identification methods are avoided, the model accuracy and robustness are improved, and the sensitivity to noise and outliers in the data is reduced.

[0006] In one optional implementation, the function segmentation points of the piecewise function model are determined by utilizing the error variation pattern between the predicted data and historical data of the piecewise function model. This includes: based on the piecewise function model, calculating the first cost function value at equally spaced segmentation points for each preset time window in the historical data, wherein each preset time window includes a preset number of historical data corresponding to timestamps; reselecting segmentation points and calculating the second cost function value based on the piecewise function model; and updating the segmentation points if the difference between the second cost function value and the first cost function value is greater than a preset threshold.

[0007] Based on the above technical means, an intelligent decision-making mechanism is introduced to ensure that only significant performance improvements are adopted, thereby improving the accuracy of segment point determination and the robustness of the algorithm, avoiding the inefficiency and subjectivity of manual intervention, and enabling the identified control parameters to more accurately reflect the actual dynamic characteristics of the equipment, thereby improving the accuracy and stability of equipment control.

[0008] In one optional implementation, based on a piecewise function model, for each preset time window in historical data, the first cost function value at equally spaced segment points is calculated, including: setting equally spaced segment points for the piecewise function model and solving the piecewise function model based on the set segment points; predicting the historical data corresponding to each timestamp within the preset time window after a preset number of timestamps according to the solved piecewise function model to obtain the corresponding first predicted data; inputting the first predicted data and the corresponding historical data into the cost function and calculating to obtain the first cost function value.

[0009] Based on the aforementioned technical methods, by inputting predicted data and historical data into the cost function and calculating the first cost function value, a stable and noise-robust evaluation benchmark is provided for subsequent segmentation point optimization. This improves the efficiency and accuracy of the initial evaluation, ensures the accuracy and robustness of segmentation point positioning, and enhances the overall performance of equipment control parameter identification.

[0010] In one optional implementation, the piecewise points of the piecewise function model are set at equal intervals, and the piecewise function model is solved based on the set piecewise points, including: setting the piecewise points in the piecewise function model at equal intervals to obtain multiple segmented intervals of the piecewise function model; solving the piecewise function model of the corresponding segmented interval based on the values ​​of the piecewise points of each segmented interval; and merging the corresponding piecewise function models of each segmented interval to obtain the piecewise function model.

[0011] Based on the above technical means, the problems of inaccurate segmentation point setting and low solution efficiency in the identification process are solved, thereby significantly improving the accuracy and stability of model identification.

[0012] In one optional implementation, reselecting segmentation points and calculating a second cost function value based on a piecewise function model includes: reselecting segmentation points of the piecewise function model within a preset numerical range of the segmentation points; re-predicting the historical data corresponding to each timestamp within a preset time window based on the reselected segmentation points and the piecewise function model, obtaining corresponding second predicted data; and inputting the second predicted data and the corresponding historical data into the cost function and calculating to obtain the second cost function value.

[0013] Based on the aforementioned technical methods, the process of determining the segmentation point positions is optimized by iteratively reselecting the segmentation points of the piecewise function model and calculating the second cost function value based on the new segmentation point configuration. This dynamic adjustment of segmentation points enables the piecewise function model to more accurately capture the actual signal segmentation points of the equipment control signal when changing different control laws, effectively solving the problem of inaccurate segmentation point positions that may be caused by setting equally spaced segmentation points.

[0014] In one optional implementation, acquiring historical data of the equipment under different operating conditions includes: acquiring historical data of corresponding component parameters of the equipment under different operations in no-load and loaded states, wherein the operation includes one of raising the boom, lowering the boom, retracting the bucket, and tipping the bucket, and the component parameters include at least one of the following: cylinder linear velocity, cylinder pilot pressure, and handle current; stitching together the historical data of each component parameter under different states and different operations to obtain stitched historical data; smoothing the stitched historical data, and dividing the smoothed stitched historical data according to a preset time window.

[0015] Based on the above-mentioned technical means, this application effectively solves the problems of data mixing, noise and outlier interference in traditional methods by refining the collection and preprocessing of historical data of the equipment.

[0016] Secondly, this application provides a method for identifying control parameters of a loader. The method includes: decomposing the linkage change process between the cylinder linear speed and the handle current signal of the loader to obtain a sub-process of linkage change between the cylinder linear speed and the cylinder pilot pressure, and a sub-process of linkage change between the cylinder pilot pressure and the handle current signal; identifying control parameters for the linkage change sub-process between the cylinder linear speed and the cylinder pilot pressure and the linkage change sub-process between the cylinder pilot pressure and the handle current signal, respectively, to obtain the control parameters of the loader. The identification of the control parameters is based on the control parameter identification method of the above-mentioned equipment.

[0017] Thirdly, this application provides a device for identifying control parameters of a device. The device includes: a data acquisition module for acquiring historical data of the device under different operating conditions; a segmentation point determination module for determining the function segmentation points of the piecewise function model by utilizing the error variation law between the predicted data and historical data of the piecewise function model, wherein the function segmentation points are used to characterize the signal segmentation points of the device's control signal when changing different control laws; and a parameter determination module for fitting historical data through the piecewise function model with the determined function segmentation points, and determining the function parameters obtained by fitting each function segment as the device control parameters.

[0018] Fourthly, this application provides an electronic device, including: a memory and a processor, which are communicatively connected to each other. The memory stores computer instructions, and the processor executes the computer instructions to perform the control parameter identification method of any embodiment corresponding to the first or second aspect described above.

[0019] Fifthly, this application provides a computer-readable storage medium storing computer instructions for causing a computer to execute the control parameter identification method of any embodiment corresponding to the first or second aspect described above.

[0020] In a sixth aspect, this application provides a computer program product, including computer instructions, which are used to cause a computer to execute a method for identifying control parameters corresponding to any of the embodiments of the first or second aspect described above. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the specific embodiments or related technologies of this application, the drawings used in the description of the specific embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0022] Figure 1 This is a schematic flowchart of a first method for identifying control parameters of a device according to an embodiment of this application; Figure 2 This is a schematic diagram of a second method for identifying control parameters of a device according to an embodiment of this application; Figure 3 This is a schematic diagram of the third process for identifying control parameters of a device according to an embodiment of this application; Figure 4 This is a flowchart illustrating a method for identifying control parameters of a loader according to an embodiment of this application. Figure 5 This is a structural block diagram of a device for identifying control parameters of an apparatus according to an embodiment of this application; Figure 6 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] It is understood that before using the technical solutions disclosed in the various embodiments of this application, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in this application in an appropriate manner in accordance with relevant laws and regulations, and user authorization should be obtained.

[0025] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0026] In related technologies, the dynamic characteristic identification method for loader working devices has shortcomings in terms of model accuracy and robustness. It struggles to cover nonlinear phenomena under multiple working conditions, is susceptible to noise and outliers, leading to model divergence. Furthermore, the method relies on manual adjustment of segment points, resulting in low efficiency, difficulty in achieving global optimization, and the need for readjustment after different machine models or system aging.

[0027] In response, this application proposes a method for identifying the control parameters of a device. For example... Figure 1 As shown, the method includes: Step S101: Obtain historical data of the equipment under different operating conditions.

[0028] Specifically, when acquiring historical data of equipment under different operating conditions, one approach is to manually record or directly acquire raw data output from sensors during equipment operation using a simple data acquisition system. This data may be limited to a single or finite operating condition, such as the equipment's operating data under constant load. Another approach is to perform pre-set, repetitive operations on the equipment in a laboratory environment and collect relevant operating data using a basic data logger. This data may be unprocessed and directly reflects the equipment's performance under specific test conditions. Alternatively, historical data from the equipment's actual operation can be collected.

[0029] Step S102: Utilize the error variation pattern between the predicted data and historical data of the piecewise function model to automatically determine the function segmentation points of the piecewise function model.

[0030] The piecewise function model consists of multiple function segments defined in different intervals, each with independent parameters. It is used to describe the behavior of systems with nonlinear or piecewise linear characteristics, and is particularly suitable for systems exhibiting different dynamic responses in different operating regions.

[0031] Function segmentation points refer to the connection points or switching points between different function segments in a piecewise function model. These points define the applicable range of different function expressions in the model and are crucial for capturing nonlinear transitions in system behavior. In this method, function segmentation points are used to characterize the signal segmentation points of the device's control signal when different control laws are applied.

[0032] Error variation pattern refers to the trend of the difference between the predicted data and the actual historical data of the piecewise function model changing over time or with input variables. By analyzing this pattern, we can identify the moments or intervals in which the system behavior changes significantly, thus providing a basis for determining the function's segmentation points.

[0033] Specifically, when determining function segmentation points using the error variation patterns between predicted and historical data from a piecewise function model, one approach is to manually observe the shape of the error curve between the predicted and historical data, and based on experience, identify points of significant error change, thus manually setting the function segmentation points. This method relies heavily on the operator's expertise and experience. Another approach is to use a simple threshold detection method; when the absolute value of the prediction error or its rate of change exceeds a preset fixed threshold, the corresponding time or data point is marked as a potential function segmentation point. Yet another approach is to pre-define a set of fixed, equally spaced function segmentation points and then perform model fitting based on these points.

[0034] Step S103 involves fitting historical data to a piecewise function model with defined function segmentation points, and then determining the function parameters obtained from fitting each function segment as the control parameters of the device.

[0035] Specifically, one approach is to perform linear fitting using a simple least squares method on the historical data within each segment after determining the function's segmentation points, thereby obtaining the function parameters for each segment. Another approach is to use basic curve fitting algorithms, such as polynomial fitting, to fit the data for each segment and extract the corresponding polynomial coefficients as control parameters.

[0036] It is understood that the technical solution in this embodiment automatically determines the function segmentation points by acquiring historical data of the equipment under different operating conditions and utilizing the error variation law between the predicted data and historical data of the piecewise function model. This avoids the inefficiency and non-global optimization problems caused by manually trying to find segmentation points in traditional loader identification methods. By fitting historical data with a piecewise function model that has determined segmentation points, the inherent nonlinearity of the loader's hydraulic system, the hysteresis of the valve-controlled cylinder link, and the coupling characteristics of the mechanical linkage can be captured more accurately, improving the model's accuracy and robustness while reducing sensitivity to noise and outliers in the data.

[0037] In some of the solutions mentioned above in this application, historical data of the equipment under different operating conditions is proposed to identify control parameters. However, in this process, because the data acquisition does not distinguish the specific working state and operation type of the equipment, the historical data is mixed with information under no-load, load and different actions, which introduces noise and outliers, reduces the data quality, and thus affects the accuracy and robustness of subsequent segment point determination and model fitting.

[0038] In response, this application further proposes a method for identifying the control parameters of a device, such as... Figure 2 As shown, the method includes: Step S201: Obtain historical data of the equipment under different operating conditions, specifically including: Step S2011: Obtain historical data of the corresponding component parameters of the equipment under different operations in no-load and load states.

[0039] The operation includes raising the boom, lowering the boom, retracting the bucket, and tipping the bucket. The component parameters include at least one of the following: cylinder linear velocity, cylinder pilot pressure, and handle current.

[0040] Specifically, acquiring historical data on corresponding component parameters of the equipment under different operations in both unloaded and loaded states is for the purpose of refined data collection and classification to address the complex dynamic characteristics exhibited by the equipment under different working conditions and operating modes. For example, load sensors (such as hydraulic system pressure sensors and engine torque sensors) can be deployed on the equipment to monitor its load status in real time. When the load sensor reading is below a certain preset threshold, the system automatically identifies it as an "unloaded state" and records the data; when the reading is above another preset threshold, it is identified as a "loaded state" and the data is recorded. Simultaneously, the specific operation currently being performed by the equipment can be identified by analyzing the position sensor signals of the control lever, CAN bus data, or using image recognition technology, such as "raising the boom," "lowering the boom," "retracting the bucket," or "tilting the bucket," and only data related to that specific operation can be recorded. For obtaining component parameters, the linear velocity of the hydraulic cylinder can be obtained by measuring the rate of change of piston rod displacement using a displacement sensor (such as a wire-type displacement sensor or a magnetostrictive displacement sensor) installed on the hydraulic cylinder; the pilot pressure of the hydraulic cylinder can be directly measured by a high-precision pressure sensor (such as a piezoresistive or piezoelectric pressure sensor) to measure the pressure value in the hydraulic line; and the handle current can be measured by a current sensor (such as a Hall effect current sensor or a shunt) to measure the electrical signal output by the handle.

[0041] Step S2012: The historical data of the parameters of each component under different states and different operations are spliced ​​together to obtain spliced ​​historical data.

[0042] Specifically, historical data on the parameters of various components under different states and operations are spliced ​​together to obtain spliced ​​historical data. This spliced ​​historical data is used to integrate scattered and categorized data into a continuous and complete sequence, providing a foundation for subsequent unified processing and model fitting. One implementation method is to attach precise timestamps, operating condition identifiers (such as "empty" and "loaded"), and operation identifiers (such as "lifting boom") to each data record during data acquisition. Then, in the data preprocessing stage, based on these identifiers and timestamps, all relevant data are logically merged in chronological order to form a unified multidimensional time series dataset. Another implementation method is to import data from different sources and categories into a unified database through a database management system. Using the database's query and merging functions, the data is correlated and spliced ​​based on timestamps and related identifiers to form a complete historical data stream.

[0043] Step S2013: Smooth the spliced ​​historical data and divide the smoothed spliced ​​historical data according to a preset time window.

[0044] Specifically, smoothing the spliced ​​historical data aims to eliminate potential high-frequency noise and outliers, improving data cleanliness and quality, and preventing these interferences from negatively impacting subsequent model identification processes. For example, moving average filtering can be used to smooth the data by calculating the average value of the data points' neighborhoods, such as simple moving averages or weighted moving averages. Alternatively, Gaussian filtering can be employed, using a Gaussian function to apply a weighted average to more effectively reduce the impact of noise.

[0045] The smoothed, spliced ​​historical data is divided into segments according to preset time windows. This divides the continuous data stream into fixed-length segments, allowing subsequent piecewise function models to process and identify parameters within each independent window. One approach is to use a fixed-length sliding window, for example, extracting a data segment of length M every N time points, where N can be less than M, thus achieving window overlap and increasing data utilization. Another approach is to use non-overlapping fixed windows, extracting a series of non-overlapping data segments from the beginning to the end of the data stream at fixed lengths of M. Furthermore, event-triggered windowing methods can be used. For example, when the start or end signal of a specific operation is detected, this can be used as the boundary for window division, making the data within each window more logically complete.

[0046] Step S202: Utilizing the error variation pattern between the predicted data and historical data of the piecewise function model, the function segmentation points of the piecewise function model are automatically determined. (See details...) Figure 1 Step S102 in the embodiment will not be described again here.

[0047] Step S203 involves fitting historical data to a piecewise function model based on the determined function segmentation points, and then using the function parameters obtained from fitting each segment as the control parameters for the equipment. (See details...) Figure 1 Step S103 in the embodiment will not be described again here.

[0048] It is understood that, through the above technical solution, this application effectively solves the problems of data mixing, noise, and outlier interference in traditional methods by refining the collection and preprocessing of historical equipment data. By distinguishing between no-load and loaded states and limiting data collection to specific operations (such as raising the boom, lowering the boom, retracting the bucket, and tipping the bucket), it ensures that the acquired historical data more accurately reflects the true dynamic characteristics of the equipment under different working conditions and typical actions, avoiding irrelevant information and noise interference caused by working condition switching. Simultaneously, by selecting key component parameters such as cylinder linear velocity, cylinder pilot pressure, and handle current, the core dynamic relationship between control signals and system response is directly captured, improving the representativeness and relevance of the data. By stitching together these categorized data, the data fragmentation problem is solved, forming a coherent and complete dataset. Based on this, smoothing the stitched historical data effectively filters out high-frequency noise and time-varying noise or spike interference mentioned in the background technology, improving data cleanliness and reducing the data's sensitivity to outliers. Finally, the smoothed data is divided according to a preset time window, providing a standardized input for efficient and accurate parameter identification of the subsequent piecewise function model within a fixed interval. This lays the foundation for determining the function segmentation points of the piecewise function model and the fitting function parameters, ultimately improving the accuracy and robustness of equipment control parameter identification.

[0049] In some embodiments, this application further proposes a method for identifying control parameters of a device, such as... Figure 3 As shown, it includes: Step S301: Obtain historical data of the equipment under different operating conditions. Specifically, this includes: Step S3011: Obtain historical data of corresponding component parameters for different operations of the equipment under no-load and load conditions. See details... Figure 2 Step S2011 in the embodiment will not be described again here.

[0050] Step S3012 involves stitching together historical data of various component parameters under different states and operations to obtain stitched historical data. See details... Figure 2 Step S2012 in the embodiment will not be described again here.

[0051] Step S3013: Smooth the spliced ​​historical data and divide the smoothed spliced ​​historical data according to a preset time window. See details... Figure 2 Step S2013 in the embodiment will not be described again here.

[0052] Step S302, using the error variation pattern between the predicted data and historical data of the piecewise function model, determines the function segmentation points of the piecewise function model, specifically including: Step S3021: Based on the piecewise function model, calculate the first cost function value for each preset time window in the historical data at equally spaced segment points.

[0053] Each preset time window includes a preset number of historical data points corresponding to timestamps. This is designed to ensure that there is a sufficient and consistent amount of data for model training and evaluation within each analysis window, avoiding unstable identification results due to data sparsity or overload. The preset number of timestamps can be a fixed value, such as 100 or 200 data points; or it can be a value calculated based on the sampling frequency and window length.

[0054] Piecewise function models are mathematical models used to approximate complex nonlinear relationships. They divide the entire range of input variables into several sub-intervals and use a relatively simple function (such as a linear function, polynomial function, or spline function) to describe the data relationship within each sub-interval. These models effectively capture the nonlinear characteristics of equipment control law switching under different operating conditions, such as dead zones, saturation, or hysteresis. Piecewise function models can take various forms; for example, they can be piecewise linear models, where each piece is represented by a straight line; or piecewise polynomial models, where each piece is represented by a polynomial function.

[0055] Historical data refers to the collection of various operating parameters and component parameters collected by the equipment under different working conditions during actual operation. This data forms the empirical basis for identifying equipment control parameters and reflects the dynamic behavior of the equipment in a real environment. Historical data can come from real-time data collected by the equipment's sensors, such as data on cylinder linear velocity, cylinder pilot pressure, and handle current obtained through onboard sensors; it can also come from test data from simulation environments or experimental benches.

[0056] A preset time window refers to a continuous time series data segment extracted from historical data. It is used to localize continuous raw historical data that may contain various operating conditions and noise, allowing subsequent segmentation point determination and model fitting to focus on specific operational phases or behavioral patterns. The length of the preset time window can be set according to the device's dynamic response characteristics or data sampling frequency; for example, it can be fixed at 1 second, 5 seconds, or a time period containing 100 data points. Alternatively, its start and end points can be dynamically determined based on device operation events (such as changes in handle signals).

[0057] Equally spaced segmentation points refer to a set of initial segmentation points obtained by uniformly dividing the independent variables (such as time or control signal amplitude) of a piecewise function model within a preset time window. These segmentation points serve as the starting points for the iterative optimization process, ensuring initial coverage of the data distribution within the entire time window. Equally spaced segmentation points can be set by dividing the range of independent variables within the preset time window into N equal intervals, thus obtaining N-1 segmentation points; or by determining them according to a preset time interval (e.g., setting one segmentation point every 0.1 seconds).

[0058] The first cost function value is a metric used to quantify the deviation between the prediction results of the piecewise function model and historical data when using equally spaced segment points. It provides a benchmark performance evaluation for subsequent segment point optimization. The first cost function value can be calculated using various error metrics, such as the mean squared error (MSE) or root mean square error (RMSE) between the predicted and actual values, or by using more robust loss functions, such as the Huber loss function or the Cauchy loss function, to reduce the impact of outliers on the evaluation results.

[0059] Step S3022: Reselect the segmentation points and calculate the second cost function value based on the piecewise function model.

[0060] Specifically, this step is crucial for achieving automatic optimization of segmentation points, aiming to overcome the limitations of manual experience-based settings. Methods for reselecting segmentation points can include conducting a local search within a pre-defined small range around each current segmentation point, for example, using gradient descent or heuristic algorithms (such as genetic algorithms or particle swarm optimization) to explore better segmentation point locations.

[0061] The second cost function value is a quantitative indicator of the deviation between the piecewise function model's prediction results and historical data after reselecting the piecewise points. It is used to evaluate the improvement in model performance after reselecting the piecewise points. The calculation method for the second cost function value is the same as that for the first cost function value, ensuring consistency in the evaluation criteria.

[0062] Step S3023: If the difference between the second cost function value and the first cost function value is greater than a preset threshold, then update the segmentation point.

[0063] The preset threshold is a pre-defined value used to determine whether the performance improvement of the model resulting from reselecting segmentation points is significant enough. As a convergence criterion for the iterative optimization process, it avoids unnecessary updates for minor, insignificant improvements, thereby improving the algorithm's efficiency and robustness. The preset threshold can be set according to actual application requirements and data characteristics. For example, it can be set to a small positive number (such as 0.01 or 0.001) representing the minimum absolute amount by which the cost function value needs to be reduced; or it can be set to a percentage reduction in the cost function value.

[0064] Updating the segmentation point means that after reselecting a segmentation point, if the improvement of the second cost function value relative to the first cost function value (i.e., the difference between the two) exceeds a preset threshold, then the current segmentation point is replaced with the newly selected segmentation point. This step ensures that the optimization process can gradually converge to a more accurate segmentation point position, thereby improving the fitting accuracy of the piecewise function model.

[0065] Step S303 involves fitting historical data to a piecewise function model based on the determined function segmentation points, and then using the function parameters obtained from fitting each segment as the control parameters for the equipment. (See details...) Figure 1 Step S103 in the embodiment will not be described again here.

[0066] It is understood that, through the above technical solution, this application provides an automatic and robust method for determining function segmentation points. First, based on a piecewise function model, the first cost function value is calculated for each preset time window in historical data at equally spaced segmentation points, providing a structured benchmark for subsequent optimization. Second, by reselecting segmentation points and calculating the second cost function value, the system can actively explore better segmentation point locations, thereby more accurately capturing the true segmentation points of the equipment control signal when switching between different control laws, effectively avoiding the problem of misjudging noise as nonlinear mutations and setting unnecessary segmentation points. Finally, by comparing whether the difference between the first and second cost function values ​​is greater than a preset threshold to determine whether to update the segmentation points, an intelligent decision-making mechanism is introduced to ensure that only significant performance improvements are adopted, thereby improving the accuracy of segmentation point determination and the robustness of the algorithm, avoiding the inefficiency and subjectivity of manual intervention, and enabling the identified control parameters to more accurately reflect the actual dynamic characteristics of the equipment, thus improving the accuracy and stability of equipment control.

[0067] In some embodiments, this application further proposes a first cost function value based on a piecewise function model, for each preset time window in historical data, when calculating equally spaced segment points, specifically including: Step a1: Set equal intervals for the segmentation points of the piecewise function model, and solve the piecewise function model based on the set segmentation points.

[0068] Specifically, when setting equally spaced segmentation points for a piecewise function model, the input variable can be divided into several intervals based on its value range, with each interval's boundary serving as a segmentation point. For example, if the input variable u ranges from [Umin, Umax] and p segmentation points are needed, the segmentation point can be set as Umin + i^6 (Umax - Umin) / (p+1), where i ranges from 1 to p. Alternatively, the distribution of historical data along the dimension of the input variable u can be statistically analyzed, dividing the number of data points into several equal parts, with each part's boundary point serving as a segmentation point. For instance, all historical data points can be sorted according to the input variable u, and then a segmentation point can be set every N data points, ensuring that each segment interval contains approximately the same number of data points. This aims to provide a structured, unbiased initial distribution of segmentation points, laying the foundation for subsequent model solving and evaluation.

[0069] After setting the equal intervals of the segmentation points, for each segmented interval in the piecewise function model, the least squares method can be used to fit a linear function (w = k) for that interval using historical data within that interval. i 6 u + c i The slope k i and intercept c i Alternatively, the piecewise function model can be viewed as a whole, and given the segmentation points, the slope k of all segments can be iteratively adjusted using optimization algorithms such as gradient descent. i and intercept c i This step aims to minimize the error between the model's predicted values ​​and historical data. Based on the initial segmentation point configuration, this step determines the specific parameters of the piecewise function model, enabling it to initially describe the changing patterns of historical data.

[0070] Step a2: Based on the piecewise function model obtained by the solution, predict the historical data corresponding to each timestamp within the preset time window after a preset number of timestamps, and obtain the corresponding first prediction data.

[0071] Specifically, for each timestamp t within a preset time window, the historical data corresponding to that timestamp can be used as input. Through the solved piecewise function model, the output value for the next timestamp t+1 is predicted, where the "preset number of timestamps" is 1. Alternatively, the output value can be predicted N times in the future (e.g., N=10). This prediction process aims to evaluate the short-term predictive capability of the current model under an equally spaced piecewise configuration, providing a basis for subsequent error quantification.

[0072] Step a3: Input the first predicted data and the corresponding historical data into the cost function and perform calculations to obtain the first cost function value.

[0073] Specifically, the cost function can employ the Huber loss function, which uses squared loss for small errors and linear loss for large errors, thus effectively suppressing the impact of outliers while ensuring fitting accuracy. Alternatively, other error metrics such as mean squared error (MSE) or mean absolute error (MAE) can be used. By quantifying the difference between the model's predicted values ​​and actual historical data, an objective numerical indicator is provided for evaluating the quality of the current segmentation point configuration.

[0074] It is understood that the technical solution in this embodiment, by setting the segmentation points of the piecewise function model at equal intervals, provides a systematic and unbiased initial segmentation point configuration method. Solving the piecewise function model based on these equally spaced segmentation points and performing predictions after a preset number of timestamps allows for an objective and quantitative evaluation of the model's predictive ability under the current segmentation point configuration. Finally, by inputting the predicted data and historical data into the cost function and calculating the first cost function value, a stable and noise-robust evaluation benchmark is provided for subsequent segmentation point optimization. This improves the efficiency and accuracy of the initial evaluation, ensures the accuracy and robustness of segmentation point positioning, and enhances the overall performance of equipment control parameter identification.

[0075] In some of the solutions described above in this application, a first cost function value is proposed based on a piecewise function model to calculate equally spaced segment points, which is used to determine the function segment points. However, in its implementation, if the segment points are not set accurately or the solution model is not systematic, it may lead to inaccurate model fitting, especially when processing noisy data, which can easily amplify the influence of noise and reduce the accuracy and stability of the identification results. In addition, the lack of specific segment point setting and solution steps may lead to low efficiency and inability to efficiently process historical data within a preset time window.

[0076] In some optional embodiments, the piecewise points of the piecewise function model are set at equal intervals, and the piecewise function model is solved based on the set piecewise points, including: Step b1 involves setting equal intervals between the segmentation points in the piecewise function model to obtain multiple segmentation intervals of the piecewise function model.

[0077] Specifically, this step ensures that the segmentation points are evenly distributed across the entire data range, thereby avoiding biases introduced by setting segmentation points empirically or randomly, and providing a structured foundation for subsequent model solving. For example, the position of each segmentation point can be determined by calculating equally spaced step sizes based on the total range of historical data and the preset number of segments. If the data range is divided into N equal intervals, then N-1 equally spaced segmentation points are set.

[0078] Step b2: Solve the piecewise function model of the corresponding segmented interval based on the values ​​of the segmentation points of each segmented interval.

[0079] Specifically, the purpose of this step is to independently fit the function model that best represents the behavior of each local interval based on its data characteristics, thereby improving the local accuracy of the overall model. For example, for each segmented interval defined by adjacent segment points, all historical data points falling within that interval can be extracted, and then statistical regression methods such as least squares, weighted least squares, or robust regression can be used to solve for the parameters of the piecewise function model within that interval, such as the slope and intercept. Alternatively, optimization algorithms, such as gradient descent or genetic algorithms, can be used to iteratively adjust the model parameters within each segmented interval to minimize the error between the model's predicted values ​​and the actual historical data values ​​within that interval, until a preset convergence condition is met.

[0080] Step b3: Merge the corresponding piecewise function models of each segmented interval to obtain the piecewise function model.

[0081] Specifically, this step aims to integrate all locally fitted function models into a complete, continuous piecewise function model, enabling it to describe the control behavior of the device across its entire operating range. For example, the function expressions obtained from solving each segmented interval and their corresponding segment point ranges can be logically combined to form a unified conditional expression. That is, when the input value falls within a certain interval, the function model for that interval is applied for calculation. Alternatively, the parameters (such as slope and intercept) and segment point values ​​of all piecewise functions can be stored in an ordered data structure. When it is necessary to predict a certain input value, the segmented interval to which the input value belongs is first determined, and then the parameters of the corresponding interval are extracted from the data structure for calculation.

[0082] It is understood that the technical solution in this embodiment solves the problems of inaccurate segmentation point setting and low solution efficiency during the identification process, thereby significantly improving the accuracy and stability of model identification. Specifically, setting the segmentation points of the piecewise function model at equal intervals ensures the uniformity of the segmentation point distribution, avoids deviations that may be introduced by empirical settings, and reduces the risk of high-frequency noise being misjudged as high-frequency dynamics of the system or time-varying noise / spiking interference being misjudged as nonlinear abrupt changes, thus improving the rationality of the initial model structure. On this basis, solving the piecewise function model based on the set segmentation points makes the solution process consistent and structured. By solving the model parameters independently for each segment interval, local data characteristics can be captured more accurately, reducing the accumulation of overall errors and thus improving the reliability of the overall identification. Merging the corresponding piecewise function models of each segment interval forms a unified and complete model, enhancing the consistency and applicability of the identification results. This allows the calculation of the first cost function value when dividing equally spaced segments to be based on a more accurate and stable piecewise function model, thus providing a more reliable benchmark for subsequent segment optimization (such as updating segment points).

[0083] In some of the embodiments described above in this application, a first cost function value is calculated based on equally spaced segment points to determine the segment points of the piecewise function model. However, the equally spaced setting may not be able to adapt to the changing patterns of actual data, resulting in inaccurate segment point positions, affecting the model fitting accuracy and efficiency, and thus reducing the robustness and global optimality of the identification results.

[0084] In some optional embodiments, the segmentation points are reselected, and a second cost function value is calculated based on the piecewise function model, including: Step c1: Within the preset range of segmentation points, reselect the segmentation points of the piecewise function model.

[0085] Reselecting segmentation points refers to adjusting and optimizing the positions of segmentation points in the piecewise function model to more accurately reflect the actual signal segmentation points when the equipment control signal changes under different control laws. Reselection can employ various optimization strategies. For example, gradient descent or heuristic search methods can be used to iteratively adjust the segmentation points by calculating the gradient of the cost function with respect to the segmentation point position, or small-scale perturbations can be made near the current segmentation point and the cost function value evaluated to select a better segmentation point position. Furthermore, statistical analysis of historical data can be performed, such as applying change point detection algorithms like CUSUM or EWMA, to identify regions in the data that may exhibit significant changes and to reposition the segmentation points near these regions.

[0086] The preset numerical range defines the search space for reselecting segment points. Its setting can be determined based on the actual value range of historical data, physical constraints (such as sensor range and actuator stroke), or engineering experience. For example, if the independent variable is time, the segment point range should be within the time span of the historical data. Alternatively, a percentage or fixed offset can be set as the preset numerical range based on the initial equally spaced segment point positions, allowing each segment point to move within ±X% of its initial position.

[0087] Step c2: Based on the reselected segmentation points and piecewise function model, re-predict the historical data corresponding to each timestamp within the preset time window after a preset number of timestamps, and obtain the corresponding second prediction data.

[0088] Specifically, this step aims to evaluate the model's predictive performance using the new segmentation point configuration. After reselecting the segmentation points, it is typically necessary to refit the piecewise function model with parameters based on the segmentation intervals defined by the new segmentation points, and then use the fitted model to predict historical data within a preset time window. If the structure of the piecewise function model (such as the number of segments and function type) remains unchanged, and only the segmentation point positions change, then new segments can be defined directly using the new segmentation points, and predictions can be made based on the function form of each segment. Predictions are usually based on historical data at the current timestamp, predicting data for a preset number of subsequent timestamps.

[0089] Step c3: Input the second predicted data and the corresponding historical data into the cost function and perform calculations to obtain the second cost function value.

[0090] Specifically, this step quantifies the error or difference between the model's predicted values ​​and the actual historical data under the new segmentation point configuration. The cost function can use the mean squared error (MSE), which is the average of the sum of squares of the differences between the predicted and actual values, as a commonly used metric for measuring the model's prediction accuracy. Considering the possibility of outliers in historical data, the cost function can also employ a robust loss function less sensitive to outliers, such as Huber loss or Tukey's Biweight loss, to improve the robustness of the identification results. Furthermore, different weights can be assigned to the errors of different timestamps or different types of data based on their importance or reliability, forming a weighted cost function.

[0091] It is understood that, through the above technical solution, this application optimizes the process of determining the segmentation point position by iteratively reselecting the segmentation points of the piecewise function model and calculating the second cost function value based on the new segmentation point configuration. This dynamic adjustment of segmentation points enables the piecewise function model to more accurately capture the actual signal segmentation points of the equipment control signal when changing different control laws, effectively solving the problem of inaccurate segmentation point positions that may be caused by setting equally spaced segmentation points. By inputting the second predicted data and historical data into the cost function to calculate the second cost function value, a quantitative basis is provided for comparing the effects of different segmentation point settings, and reliable guidance is provided for segmentation point updates and optimization. This significantly improves the accuracy and efficiency of model fitting, enhances the adaptability of the identification process to data changes, and thus improves the robustness and global optimality of the equipment control parameter identification results.

[0092] In some optional embodiments, the piecewise function model includes: .

[0093] In this model, 'u' serves as the input value (input variable), representing a measurable input signal or state variable within the equipment control system. It forms the basis of the model, driving its response and ensuring that the model can simulate the behavior of the actual equipment under different input conditions. For example, 'u' can be a control signal of the equipment, such as a handle current signal or a valve opening signal, which directly affects the equipment's operation; or 'u' can be a critical state variable within the equipment, such as the pilot pressure of a hydraulic cylinder or the speed of a motor, which reflects the immediate effect of the control signal.

[0094] As the output value, 'w' is the dependent variable calculated by the piecewise function model based on the input value 'u', representing the response or output state of the equipment under a specific input. It provides the model's prediction of the equipment's behavior, which is used to compare with actual historical data to evaluate the model's fitting accuracy and determine control parameters. For example, 'w' can be the actual motion parameters of the equipment, such as the linear velocity of the hydraulic cylinder, the displacement or speed of the actuator, etc., which directly reflect the dynamic response of the equipment; or 'w' can be intermediate physical quantities within the equipment, such as the flow rate or pressure of the hydraulic system, which are direct results of the control signal.

[0095] d1,d2...d pPiece points are threshold values ​​for the input value u in a piecewise function model, dividing the entire input range into p+1 distinct intervals. At these piecewise points, the slope or intercept of the function may change to capture the switching of the device's control law or abrupt changes in its nonlinear characteristics. These piecewise points characterize the signal segmentation points of the device's control signal when changing different control laws, and are crucial for the model to capture the nonlinear dynamic characteristics of the device, accurately dividing different operating modes or response stages. Piece points can be determined by analyzing the relationship between the control signal and the response signal in historical data to identify the input value thresholds where the response characteristics change significantly, such as when the handle current reaches a certain value, the cylinder's response speed changes from linear to saturation; or they can be initially set through expert experience or understanding of the device's working principle based on a physical model, and then finely adjusted through optimization algorithms.

[0096] k1,k2...k p+1 The slopes of each segment are the slope parameters of the linear function within each segment interval of the piecewise function model. They describe the rate of change of the output value w when the input value u changes within that interval. These slope parameters quantify the dynamic response sensitivity of the equipment in different control laws or operating ranges, and are an important component of the equipment control parameters, reflecting the gain characteristics of the control system to the input. The slopes can be solved by fitting historical data within each segment interval using the least squares method or other regression analysis methods; or they can be obtained by iteratively adjusting the model predictions to minimize the error between the model predictions and historical data given the segment points using optimization algorithms such as gradient descent.

[0097] c1, c2...c p+1 The intercepts are the parameters for the linear function within each segmented interval of the piecewise function model. They represent the offset of the output value w when the input value u is zero (or the starting point of the interval). These intercept parameters supplement the model's description of the device's static characteristics or baseline output, refine the model's fitting details to the behavior within each segmented interval, and ensure the model's continuity and accuracy across the entire input range. The intercepts can be determined together with the slope through linear regression analysis on historical data within each segmented interval; alternatively, they can be obtained by calculating the average offset of historical data within each segmented interval after determining the slope, to ensure the model's overall fit within that interval.

[0098] It is understandable that the above technical solution clearly defines the specific mathematical structure of the piecewise function model, thereby solving the problems of insufficient fitting accuracy, susceptibility to noise interference, and low efficiency in searching for piecewise points caused by the ambiguity of the model form. Specifically, the clear definition of the input value u and the output value w ensures that the model can accurately map the control signal and the device response, avoiding fitting bias caused by unclear definitions of the dependent variable. The precisely defined piecewise points d1, d2...d...p This allows the model to effectively capture abrupt changes in the equipment control signal during switching between different control laws, thereby better handling the inherent nonlinearity and hysteresis of hydraulic systems and significantly reducing the risk of misinterpreting noise as high-frequency dynamic or nonlinear abrupt changes in the system. Furthermore, the slopes k1, k2...k of each segment... p+1 and the intercepts of each segment c1, c2...c p+1 The clear definition of parameters provides clear parameters for quantifying the dynamic response of equipment in different operating ranges, enabling the model to fit historical data more precisely and ensuring the integrity of parameter identification. The technical solution in this embodiment optimizes the subsequent piecewise point update and cost function calculation processes, reducing the model's sensitivity to data noise. This results in more efficient, stable, and robust control parameter identification when using a piecewise function model to fit historical data and determine the equipment's control parameters.

[0099] In some optional embodiments of this application, the cost function includes: .

[0100] The cost function is a mathematical expression used to quantify the difference between model predictions and actual observations. Its core function is to provide a quantifiable objective for model optimization. By minimizing the cost function value, model parameters are adjusted so that the model can better fit historical data. In the parameter identification process, the cost function is a key tool for evaluating model performance and guiding iterative parameter updates.

[0101] in, i This represents the prediction result made by the piecewise function model for the i-th historical data point under given input conditions. It is the output of the model's simulation of system behavior based on current parameters, reflecting the model's understanding and expression of the system's dynamic characteristics. i This represents the i-th actual observed historical data point, i.e., the system's true behavior or output. It serves as the benchmark for model predictions, used to compare the predicted values ​​with the actual data. The cost function is compared with i to calculate the error. δ is a preset threshold or adjustment parameter used to control the sensitivity and handling of the cost function to errors. For example, δ can be a positive real number, the magnitude of which determines the range of error within which the cost function exhibits quadratic penalty characteristics, and the range outside which it switches to linear penalty characteristics. The setting of δ can be optimized based on experience, data distribution characteristics, or through methods such as cross-validation. ρ H ( i y iThe value represents the cost function calculated based on the Huber loss function. The Huber loss function is a hybrid loss function that uses squared error when the error is small (e.g., when the absolute value of the error is less than or equal to δ) to maintain sensitivity to normal fluctuations; while it uses linear error when the error is large (e.g., when the absolute value of the error is greater than δ) to reduce the contribution of outliers to the total cost function, thereby effectively suppressing the negative impact of noise and outliers on the identification of model parameters.

[0102] It is understood that, through the above technical solution, this application uses the Huber loss function as the cost function, effectively solving the problem that traditional least squares methods are prone to inaccurate model identification results and poor stability when dealing with data containing noise and outliers. Specifically, when the model predicts values... When the error between i and the historical true value yi is small, the cost function employs a quadratic penalty mechanism, which accurately captures normal system fluctuations and ensures the model's fitting accuracy to conventional data. When the error is large, i.e., when noise or outliers may exist, the cost function switches to a linear penalty mechanism, significantly reducing the weight of these outliers on the total cost function. This avoids the situation in traditional methods where outliers are excessively amplified, leading to model divergence or parameters deviating from the true values. The introduction of the parameter δ allows the cost function to flexibly adjust its sensitivity to errors according to the characteristics of the actual data, further enhancing the robustness of the identification process and its adaptability to different noise levels. Therefore, in the process of fitting historical data using a piecewise function model and determining control parameters, this cost function can provide a more stable and accurate error assessment, thereby obtaining more reliable equipment control parameters and improving the anti-interference capability and parameter identification accuracy of the entire identification method.

[0103] In some optional embodiments, this application proposes a method for identifying loader control parameters. For example... Figure 4 As shown, the method includes: Step S401: Decompose the linkage change process between the cylinder linear speed of the loader and the handle current signal to obtain the linkage change sub-process between the cylinder linear speed and the cylinder pilot pressure, and the linkage change sub-process between the cylinder pilot pressure and the handle current signal.

[0104] Step S402: Identify the control parameters for the linkage change sub-process of the hydraulic cylinder linear velocity and the hydraulic cylinder pilot pressure and the linkage change sub-process of the hydraulic cylinder pilot pressure and the handle current signal, respectively, to obtain the control parameters of the loader. The identification of the control parameters is based on the identification method of the control parameters of the above-mentioned equipment.

[0105] Specifically, during the identification process, for the linked changes in the cylinder linear velocity and the cylinder pilot pressure, the aforementioned identification method is used to automatically determine the segmentation points and fit the parameters. Similarly, the same method is applied to the linked changes in the cylinder pilot pressure and the handle current signal. This approach avoids the inefficiency and subjectivity of manually trial-and-error segmentation points, improves identification efficiency, and facilitates the acquisition of the global optimal solution. Furthermore, the aforementioned identification method effectively handles data noise, ensuring the robustness and accuracy of the identification process and overcoming the limitations of noise sensitivity and outlier influence.

[0106] It is understandable that this method effectively reduces the complexity of identification by decomposing the complete control chain into two relatively independent sub-processes. The sub-process of the linkage between the cylinder linear velocity and the cylinder pilot pressure mainly characterizes the dynamic response characteristics of the hydraulic cylinder, while the sub-process of the linkage between the cylinder pilot pressure and the handle current signal mainly characterizes the control characteristics of the pilot valve. This decomposition makes it easier to handle and model the nonlinear characteristics of each sub-process separately, thus avoiding model inaccuracies caused by the overall nonlinearity of the system under a wide range of operating conditions. Furthermore, the control parameter identification for each of the two sub-processes allows for independent optimization of the identification parameters for the characteristics of each sub-process. Since the dynamic behavior of each sub-process is relatively simple, the identification process is more efficient and accurate.

[0107] In one example, a more specific case will be used to illustrate the above technical solution in greater detail: In the scenario of identifying control parameters for a loader, it is necessary to accurately acquire the dynamic characteristics of the loader's working device to optimize its control performance. Related technologies face challenges in handling the inherent nonlinearity of the loader's hydraulic system, the hysteresis of valve-controlled cylinders, and the coupling characteristics of mechanical linkages. These challenges include insufficient model accuracy, sensitivity to data quality, and low parameter identification efficiency. For example, the complete control chain from the handle current signal to the final cylinder speed contains multiple dynamic elements that are difficult to describe with precise physical equations. Traditional mechanistic modeling is complex and difficult to be accurate, while simplified linear models cannot cover the wide range of working conditions and drastic load changes of the loader. Especially during low-speed micro-motion and direction switching, the system exhibits nonlinear phenomena such as dead zones, saturation, and hysteresis, leading to a significant decrease in control performance. Furthermore, high-frequency noise may be misidentified as high-frequency dynamics, and time-varying noise or spike interference may be misidentified as nonlinear abrupt changes, causing the automatic segmentation point search algorithm to set unnecessary segmentation points at noisy locations. If the data points used by the identification algorithm contain strong noise or outliers, it may cause the model to diverge, reducing the accuracy and stability of the identification results. Meanwhile, the traditional method of manually adjusting segment points using a trial-and-error approach is inefficient and cannot achieve global optimality.

[0108] To address the aforementioned issues, this solution provides a method for identifying loader control parameters. First, the linkage change process between the loader's cylinder linear velocity and the handle current signal is broken down into sub-processes involving the linkage change between the cylinder linear velocity and the cylinder pilot pressure, and sub-processes involving the linkage change between the cylinder pilot pressure and the handle current signal. The method for identifying the control parameters in the sub-process involving the linkage change between the cylinder pilot pressure and the handle current signal is used as an example for illustration.

[0109] The first step is to acquire historical data of the loader under different operating conditions. Specifically, historical data on the handle current signal and hydraulic cylinder pilot pressure are acquired when the loader performs different operations such as raising the boom, lowering the boom, retracting the bucket, and tipping the bucket under both unloaded and loaded conditions. For example, when raising the boom under unloaded conditions, the handle current signal and the corresponding hydraulic cylinder pilot pressure data are recorded; when tipping the bucket under loaded conditions, the corresponding handle current signal and hydraulic cylinder pilot pressure data are also recorded. Then, these historical data of various component parameters acquired under different conditions and operations are stitched together to form a comprehensive stitched historical data set. To improve data quality, the stitched historical data is smoothed to effectively suppress the influence of high-frequency noise and outliers, preventing them from being misjudged as system dynamic or nonlinear mutations, thereby improving the robustness of subsequent model identification. The smoothed stitched historical data is then divided according to a preset time window to prepare for subsequent piecewise function model processing.

[0110] The second step is to determine the function segmentation points of the piecewise function model by utilizing the error variation pattern between the predicted data and historical data. These function segmentation points are used to characterize the signal segmentation points of the loader control signal when different control laws are applied, such as the dead zone, linear zone, or saturation zone that the handle current signal may correspond to in different operating ranges. The specific operation is as follows: For each preset time window in the historical data, the segmentation points in the piecewise function model are first set at equal intervals. For example, the input range of the handle current signal is divided into several intervals to obtain multiple segmented intervals of the piecewise function model. Based on the segmentation point values ​​of each segmented interval, the function model parameters of the corresponding segmented interval are solved, and the corresponding piecewise function models of each segmented interval are merged to obtain a complete initial piecewise function model.

[0111] Based on the initial piecewise function model obtained from the solution, predictions are made for the historical data corresponding to each timestamp within a preset time window after a preset number of timestamps, resulting in the corresponding first predicted data. The first predicted data and the corresponding historical data are then input into a cost function and calculated to obtain the first cost function value. This cost function is robust and can effectively reduce the impact of outliers on the identification results, preventing model divergence.

[0112] Next, within the preset numerical range of the segmentation points, the segmentation points of the piecewise function model are reselected. For example, a small-scale perturbation is performed near the initial equally spaced segmentation points, or an optimization algorithm is used for searching. Based on the reselected segmentation points and the piecewise function model, a preset number of timestamps are used to predict the historical data corresponding to each timestamp within the preset time window, resulting in the corresponding second predicted data. The second predicted data and the corresponding historical data are input into the cost function and calculated to obtain the second cost function value.

[0113] By comparing the difference between the second cost function value and the first cost function value, if the difference is greater than a preset threshold, the segmentation point is updated to continuously optimize the selection of the segmentation point. The iterative process continuously adjusts the segmentation point until the cost function value converges or meets the preset conditions, thereby automatically finding the optimal function segmentation point. This automated segmentation point determination method avoids the inefficiency and non-global optimum problems of traditional "trial and error" parameter tuning, significantly improving identification efficiency and model accuracy.

[0114] The third step involves fitting historical data to a piecewise function model with determined function segmentation points, and then using the function parameters obtained from fitting each segment as the control parameters for the loader. Once the optimal function segmentation points are determined, the piecewise function model can more accurately capture the nonlinear relationship between the handle current signal and the cylinder pilot pressure. By fitting each segment interval, the slope and intercept of each segment can be obtained. These parameters precisely characterize the dynamic response characteristics of the loader in different control signal ranges, such as the dead zone characteristics that may be exhibited during handle micro-movements, and the linear or saturation characteristics during wide-range operation. These identified parameters are the control parameters for the linkage between the loader cylinder pilot pressure and the handle current signal.

[0115] By employing the methods described above, this scheme overcomes the shortcomings of traditional methods in handling complex nonlinear dynamics of loaders, data noise and outliers, and parameter identification efficiency. Through data smoothing and a robust cost function, the impact of data quality on the identification results is effectively reduced. An automated segmented point optimization mechanism improves identification efficiency and model accuracy, enabling the identified control parameters to more accurately reflect the true dynamic characteristics of the loader under a wide range of operating conditions, providing a reliable basis for the design of advanced controllers.

[0116] This embodiment also provides a device for identifying control parameters of a device, which is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the term "module" can refer to a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0117] This embodiment provides a device for identifying the control parameters of a device, such as... Figure 5 As shown, it includes: The data acquisition module 501 is used to acquire historical data of the equipment under different operating conditions.

[0118] The segmentation point determination module 502 is used to determine the function segmentation points of the piecewise function model by utilizing the error variation law between the predicted data and historical data of the piecewise function model. The function segmentation points are used to characterize the signal segmentation points of the equipment's control signal when different control laws are changed.

[0119] The parameter determination module 503 is used to fit historical data by using a piecewise function model with determined function segmentation points, and to determine the function parameters obtained by fitting each function segment as equipment control parameters.

[0120] In some optional implementations, the segment point determination module 502 includes: The first unit is used to calculate the first cost function value for each preset time window in historical data based on the piecewise function model, where each preset time window includes a preset number of historical data corresponding to timestamps.

[0121] The second unit is used to reselect the segmentation points and calculate the second cost function value based on the piecewise function model.

[0122] The third unit is used to update the segmentation point if the difference between the second cost function value and the first cost function value is greater than a preset threshold.

[0123] This embodiment provides a device for identifying control parameters of a loader, including: The first module is used to break down the linkage change process between the cylinder linear velocity and the handle current signal of the loader, and obtain the linkage change sub-process between the cylinder linear velocity and the cylinder pilot pressure, and the linkage change sub-process between the cylinder pilot pressure and the handle current signal.

[0124] The second module is used to identify the control parameters of the linkage change sub-process of the hydraulic cylinder linear velocity and the hydraulic cylinder pilot pressure and the linkage change sub-process of the hydraulic cylinder pilot pressure and the handle current signal, respectively, so as to obtain the control parameters of the loader. The identification of the control parameters is based on the identification method of the equipment's control parameters.

[0125] The device for identifying control parameters of a device provided in this application can execute the device control parameter identification method provided in any embodiment of this application, and has the corresponding functional modules and beneficial effects for executing the method. Further functional descriptions of the above modules and units are the same as those in the corresponding embodiments described above, and will not be repeated here.

[0126] Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0127] The following is a detailed reference. Figure 6 This diagram illustrates a suitable structural schematic for implementing the electronic device described in the embodiments of this application. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 601, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 602 or a program loaded from memory 608 into random access memory (RAM) 603. The RAM 603 also stores various programs and data required for the operation of the electronic device. The processor 601, ROM 602, and RAM 603 are interconnected via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0128] Typically, the following devices can be connected to I / O interface 605: input devices 606 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 607 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 608 including, for example, magnetic tapes, hard disks, etc.; and communication devices 609. Communication device 609 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 6 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0129] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication device 609, or installed from memory 608, or installed from ROM 602. When the computer program is executed by processor 601, it performs the functions defined in the method for identifying control parameters of the device according to embodiments of this application.

[0130] Figure 6 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0131] This application also provides a computer-readable storage medium. The methods described in this application can be implemented in hardware or firmware, or implemented as recordable on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the method for identifying control parameters of the device shown in the above embodiments is implemented.

[0132] A portion of this application can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to this application through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0133] Although embodiments of this application have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of this application, and all such modifications and variations fall within the scope defined by the appended claims.

Claims

1. A method for identifying control parameters of a device, characterized in that, The method includes: Acquire historical data of the equipment under different operating conditions; By utilizing the error variation pattern between the predicted data and the historical data of the piecewise function model, the function segmentation point of the piecewise function model is determined. The function segmentation point is used to characterize the signal segmentation point of the equipment's control signal when different control laws are changed. The historical data is fitted by the piecewise function model with the function segmentation points determined, and the function parameters obtained by fitting each function segment are determined as the control parameters of the equipment.

2. The method according to claim 1, characterized in that, The step of determining the function segmentation points of the piecewise function model by utilizing the error variation pattern between the predicted data and the historical data of the piecewise function model includes: Based on the piecewise function model, for each preset time window in the historical data, the first cost function value at equally spaced segment points is calculated, wherein each preset time window includes a preset number of timestamps corresponding to the historical data. The segmentation point is reselected, and the second cost function value is calculated based on the piecewise function model; If the difference between the second cost function value and the first cost function value is greater than a preset threshold, then the segmentation point is updated.

3. The method according to claim 2, characterized in that, The step of calculating the first cost function value at equally spaced segment points for each preset time window in the historical data, based on the piecewise function model, includes: The piecewise function model is set with equal intervals at its segmentation points, and the piecewise function model is solved based on these set segmentation points. Based on the piecewise function model obtained by the solution, the historical data corresponding to each timestamp within the preset time window is predicted after the preset number of timestamps to obtain the corresponding first prediction data. The first predicted data and the corresponding historical data are input into the cost function and calculated to obtain the first cost function value.

4. The method according to claim 3, characterized in that, The step of setting equal intervals for the segmentation points of the piecewise function model and solving the piecewise function model based on the set segmentation points includes: By setting the segmentation points in the piecewise function model at equal intervals, multiple segmentation intervals of the piecewise function model are obtained; Based on the values ​​of the segmentation points of each segmented interval, solve the piecewise function model for the corresponding segmented interval; The corresponding piecewise function models of each segmented interval are merged to obtain the piecewise function model.

5. The method according to claim 3, characterized in that, The step of reselecting the segmentation point and calculating the second cost function value based on the piecewise function model includes: Within the preset numerical range of the segmentation points, the segmentation points of the piecewise function model are reselected; Based on the reselected segmentation point and the piecewise function model, the historical data corresponding to each timestamp within the preset time window is re-predicted after the preset number of timestamps to obtain the corresponding second prediction data. The second predicted data and the corresponding historical data are input into the cost function and calculated to obtain the value of the second cost function.

6. The method according to claim 1, characterized in that, The acquisition of historical data of the device under different operating conditions includes: Historical data of corresponding component parameters of the equipment under different operations in no-load and loaded states are obtained respectively. The operation includes raising the boom, lowering the boom, retracting the bucket, and tipping the bucket. The component parameters include at least one of the following: hydraulic cylinder linear velocity, hydraulic cylinder pilot pressure, and handle current. By stitching together the historical data of the parameters of each component under different states and different operations, the stitched historical data is obtained. The spliced ​​historical data is smoothed, and the smoothed spliced ​​historical data is divided according to a preset time window.

7. A method for identifying control parameters of a loader, characterized in that, The method includes: The linkage change process between the linear velocity of the hydraulic cylinder and the current signal of the handle of the loader is broken down to obtain the linkage change sub-process between the linear velocity of the hydraulic cylinder and the pilot pressure of the hydraulic cylinder, and the linkage change sub-process between the pilot pressure of the hydraulic cylinder and the current signal of the handle. The control parameters of the loader are obtained by identifying the control parameters of the linkage change sub-process of the hydraulic cylinder linear velocity and the hydraulic cylinder pilot pressure and the linkage change sub-process of the hydraulic cylinder pilot pressure and the handle current signal, respectively. The identification of the control parameters is based on the control parameter identification method of any one of the equipment as described in claims 1-6.

8. A device for identifying control parameters of an equipment, characterized in that, The device includes: The data acquisition module is used to acquire historical data of the equipment under different operating conditions; The segmentation point determination module is used to determine the function segmentation points of the piecewise function model by utilizing the error variation law between the predicted data and the historical data of the piecewise function model. The function segmentation points are used to characterize the signal segmentation points of the device's control signal when different control laws are changed. The parameter determination module is used to fit the historical data by the piecewise function model with the function segmentation points determined, and to determine the function parameters obtained by fitting each function segment as the equipment control parameters.

9. An electronic device, characterized in that, include: The device includes a memory and a processor, which are interconnected. The memory stores computer instructions, and the processor executes the computer instructions to perform the method for identifying control parameters of the device according to any one of claims 1 to 6 or the method for identifying control parameters of the loader according to claim 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the method for identifying control parameters of the device according to any one of claims 1 to 6 or the method for identifying control parameters of the loader according to claim 7.

Images