Method, apparatus, controller and work vehicle for determining work mode

Abstract

Embodiments of the present application provide a method, device, controller and engineering vehicle for determining a work mode. The method comprises: obtaining a work signal of the engineering vehicle during a preset time period, the preset time period comprising a plurality of preset time windows; determining, according to the work signal, a peak value number of envelope peaks corresponding to each preset time window and a characteristic value corresponding to each preset time window; inputting the peak value number and the characteristic value of all the preset time windows into a work recognition model to output, by the work recognition model, a work mode corresponding to the work of the engineering vehicle during the preset time period. Through the above technical solution, the work mode of the engineering vehicle is predicted from the local fluctuation frequency and the overall dispersion degree of the work signal, greatly improving the accuracy and reliability of determining the work mode, the required labor cost and time cost are low, and the work recognition efficiency of the engineering vehicle is effectively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of engineering machinery, in particular to a method and device for determining a work mode, a controller and an engineering vehicle. BACKGROUND

[0002] In the prior art, the work mode of the engineering vehicle is determined by comparing the pressure difference of the main pump of the engineering vehicle during actual work with a preset threshold. However, the threshold comparison for predicting the work mode of the excavator needs to ensure that the pressure data of the engineering vehicle in each work mode has high difference, so as to ensure the accuracy of the prediction.

[0003] Taking the excavator as an example, the work mode of the excavator can include digging work, grading work and breaking work mode. Generally, the randomness of the excavator during grading work is large, so the distribution interval of the corresponding pressure data is relatively wide. At the same time, the corresponding pressure data distribution of the excavator during digging work and breaking work is highly overlapped. Therefore, if the threshold comparison is used to predict the work mode of the excavator, the work mode of the excavator cannot be accurately determined, thereby greatly reducing the work quality and work efficiency of the excavator. SUMMARY

[0004] The purpose of the embodiments of the present application is to provide a method and device for determining a work mode, a controller and an engineering vehicle.

[0005] In order to achieve the above-mentioned purpose, the first aspect of the present application provides a method for determining a work mode, applied to an engineering vehicle, the method comprising:

[0006] obtaining a work signal of the engineering vehicle during work in a preset time period, the preset time period comprising a plurality of preset time windows;

[0007] determining a peak value number of an envelope peak value corresponding to each preset time window and a feature value corresponding to each preset time window according to the work signal;

[0008] inputting the peak value number and the feature value of all the preset time windows into a work recognition model to output a work mode corresponding to the work of the engineering vehicle in the preset time period by the work recognition model.

[0009] In the embodiments of the present application, the engineering vehicle includes a first main pump and a second main pump, the work signal includes a first work signal corresponding to the first main pump and a second work signal corresponding to the second main pump; determining the peak number of the envelope peak corresponding to each preset time window according to the work signal includes: filtering the first work signal and the second work signal respectively; performing differential processing on the filtered first work signal and the filtered second work signal respectively to determine the first envelope peak corresponding to the filtered first work signal and the second envelope peak corresponding to the filtered second work signal respectively; determining the peak number of the first envelope peak and the peak number of the second envelope peak in each preset time window.

[0010] In the embodiments of the present application, the preset time period includes a plurality of time points, the work signal includes a work parameter value corresponding to each time point, and the differential processing of the filtered first work signal and the filtered second work signal to determine the first envelope peak corresponding to the filtered first work signal and the second envelope peak corresponding to the filtered second work signal includes: obtaining the work parameter value corresponding to any one of the time points in the filtered first work signal or the filtered second work signal; determining a first parameter difference between the work parameter value of the current time point and the work parameter value of the previous time point, and a second parameter difference between the work parameter value of the current time point and the work parameter value of the next time point; in the case that the first parameter difference is greater than a preset threshold and the second parameter difference is less than the preset threshold, the work parameter value corresponding to the current time point is determined as the envelope peak corresponding to the work signal.

[0011] In the embodiments of the present application, the engineering vehicle includes a first main pump and a second main pump, the work signal includes a first work signal corresponding to the first main pump and a second work signal corresponding to the second main pump, the preset time period includes a plurality of preset time windows, each preset time window includes a plurality of time points, the work signal includes a work parameter value corresponding to each time point, and the characteristic value includes a coefficient of variation and a variance; determining the characteristic value corresponding to each preset time window according to the work signal includes: for the first work signal or the second work signal, determining a parameter mean of the work parameter value corresponding to each time point in the work signal; determining the variance and the standard deviation of the work parameter value of each preset time window according to the parameter mean and the work parameter value corresponding to each time point; determining the coefficient of variation corresponding to each preset time window according to the parameter mean and the standard deviation corresponding to each preset time window.

[0012] In the embodiments of the present application, the method further comprises a training step of the work recognition model, the training step comprising: obtaining sample signals of the engineering vehicle in a plurality of historical work modes in the process of performing historical work, and taking the sample signals in each historical work mode as a training sample, wherein the engineering vehicle comprises at least a first main pump and a second main pump, the sample signals comprise a first sample signal corresponding to the first main pump and a second sample signal corresponding to the second main pump, and a label of the training sample is the corresponding historical work mode; determining a number of first historical peak values corresponding to a historical time window and a first historical feature value in each historical work mode according to the first sample signal; determining a number of second historical peak values corresponding to the historical time window and a second historical feature value in each historical work mode according to the second sample signal; and inputting the number of first historical peak values, the number of second historical peak values, the first historical feature value and the second historical feature value of all historical time windows into the work recognition model to train the work recognition model.

[0013] In the embodiments of the present application, the method further comprises: for each training of the work recognition model, outputting a plurality of predicted work modes corresponding to the sample signals by the work recognition model; comparing each predicted work mode with an actual work mode to determine a plurality of prediction deviation values of the work recognition model; in a case where an error between adjacent number of prediction deviation values is less than or equal to a preset value, determining that the training of the work recognition model is completed; in a case where the error between adjacent number of prediction deviation values is greater than the preset value, inputting again the number of first historical peak values, the number of second historical peak values, the first historical feature value and the second historical feature value of all historical time windows into the work recognition model to train the work recognition model, until the number of training times of the work recognition model reaches a preset training number or the error between adjacent number of prediction deviation values is less than or equal to the preset value.

[0014] In the embodiments of the present application, the engineering vehicle is an excavator, and the work mode of the engineering vehicle comprises any one of a digging work, a grading work and a breaking work.

[0015] The second aspect of the present application provides a controller configured to execute the above-mentioned method for determining a work mode.

[0016] The third aspect of the present application provides an apparatus for determining a work mode, comprising: a data acquisition device for acquiring work signals generated by an engineering vehicle when working; and the above-mentioned controller.

[0017] The fourth aspect of the present application provides an engineering vehicle, comprising: at least one main pump; and the above-mentioned apparatus for determining a work mode.

[0018] By the technical solution, the peak number and characteristic value of the whole preset time window are input into the work recognition model, the work mode of the engineering vehicle is predicted from the local fluctuation frequency and overall dispersion degree of the work signal, the accuracy and reliability of determining the work mode are greatly improved, the required artificial cost and time cost are low, and the work recognition efficiency of the engineering vehicle is effectively improved.

[0019] Other features and advantages of the embodiments of the present application will be described in detail in the following specific implementation part. BRIEF DESCRIPTION OF DRAWINGS

[0020] The accompanying drawings are included to provide a further understanding of the embodiments of the present application, and constitute a part of the specification, and are used to explain the embodiments of the present application together with the following specific implementation, but do not constitute the limitation to the embodiments of the present application. In the drawings:

[0021] Figure 1 The flowchart of the method for determining the work mode according to the embodiments of the present application is schematically shown;

[0022] Figure 2 The flowchart of the method for determining the work mode according to another embodiment of the present application is schematically shown;

[0023] Figure 3 The flowchart of the model training according to the embodiments of the present application is schematically shown;

[0024] Figure 4 The structural block diagram of the device for determining the work mode according to the embodiments of the present application is schematically shown;

[0025] Figure 5 The internal structure diagram of the computer device according to the embodiments of the present application is schematically shown. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical scheme and advantages of the embodiments of the present application more clear, the technical scheme in the embodiments of the present application will be described clearly and completely below in combination with the drawings in the embodiments of the present application. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiments of the present application, and is not used to limit the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0027] Figure 1 The flowchart of the method for determining the work mode according to the embodiments of the present application is schematically shown. As shown in Figure 1 In an embodiment of the present application, a method for determining a work mode is provided, applied to an engineering vehicle, and the method comprises the following steps:

[0028] At step 101, an operation signal of the engineering vehicle during a preset time period is obtained, and the preset time period includes a plurality of preset time windows.

[0029] At step 102, a peak number of envelope peaks corresponding to each preset time window and a feature value corresponding to each preset time window are determined according to the operation signal.

[0030] At step 103, the peak numbers and the feature values of all the preset time windows are input into an operation identification model, so as to output an operation mode of the engineering vehicle during the preset time period by the operation identification model.

[0031] The engineering vehicle can refer to a vehicle capable of mechanical operation. The operation mode of the engineering vehicle can include excavation operation, grading operation, and crushing operation. The engineering vehicle can include at least one main pump. In order to accurately determine the operation mode of the engineering vehicle, the controller can obtain an operation signal of the engineering vehicle during a preset time period. The preset time period can include a plurality of preset time windows. The operation signal can refer to a signal corresponding to the main pump of the engineering vehicle. The operation signal can be a pressure signal or a current signal.

[0032] In the case of obtaining the operation signal, the controller can determine a peak number of envelope peaks corresponding to each preset time window according to the operation signal. Further, the controller can first determine envelope peaks corresponding to the operation signal according to the operation signal. Then, the controller can further determine the peak number of envelope peaks corresponding to each preset time window. By determining the peak number corresponding to each preset time window, the fluctuation degree of the operation signal can be quantified from a local perspective. In the case of obtaining the operation signal, the controller can determine a feature value corresponding to each preset time window according to the operation signal. The feature value can refer to a statistical feature value. Specifically, the feature value can include variance and coefficient of variation. By determining the feature value, the overall dispersion degree of the operation signal can be effectively reflected, which can provide a reference basis for accurately determining the operation mode of the engineering vehicle.

[0033] In the case of determining the peak number of envelope peaks corresponding to each preset time window and the feature value, the controller can input the peak numbers and the feature values of all the preset time windows into an operation identification model, so as to output an operation mode of the engineering vehicle during the preset time period by the operation identification model. The operation identification model can be a classification model. Specifically, the classification model can be a random forest.

[0034] By the technical solution, the peak number and characteristic value of the whole preset time window are input into the work recognition model to predict the work mode of the engineering vehicle from the local change trend and overall change trend of the work signal, thereby greatly improving the accuracy and reliability of determining the work mode, reducing the required labor cost and time cost, and effectively improving the work quality and work efficiency of the engineering vehicle.

[0035] In one embodiment, the engineering vehicle is an excavator, and the work mode of the engineering vehicle includes any one of a digging work, a grading work, and a breaking work.

[0036] The engineering vehicle can refer to a vehicle capable of mechanical construction. The engineering vehicle can refer to an excavator. The work mode of the engineering vehicle can include any one of a digging work, a grading work, and a breaking work.

[0037] In one embodiment, the engineering vehicle includes a first main pump and a second main pump, and the work signal includes a first work signal corresponding to the first main pump and a second work signal corresponding to the second main pump; determining the peak number of the envelope peak corresponding to each preset time window according to the work signal includes filtering the first work signal and the second work signal respectively; performing differential processing on the filtered first work signal and the filtered second work signal respectively to determine a first envelope peak corresponding to the filtered first work signal and a second envelope peak corresponding to the filtered second work signal respectively; and determining the peak number of the first envelope peak and the peak number of the second envelope peak in each preset time window.

[0038] The engineering vehicle includes a first main pump and a second main pump. The work signal of the engineering vehicle during work in a preset time period can include a first work signal corresponding to the first main pump and a second work signal corresponding to the second main pump. The work signal can be a pressure signal or a current signal. In the case of a pressure signal, the first work signal corresponding to the first main pump can refer to a first pressure signal, and the second work signal corresponding to the second main pump can refer to a second pressure signal.

[0039] After obtaining the first work signal and the second work signal, the controller can filter the first work signal and the second work signal respectively. The filtering method can be S-G envelope filtering. By S-G envelope filtering, the work signal can be approximated and the waveform envelope can be obtained while ensuring that the shape and width of the work signal remain unchanged. After S-G envelope filtering, the signal noise of the work signal can be reduced, and the change trend of the work signal can be extracted, so that the envelope peak can be quickly and accurately determined subsequently.

[0040] The controller can respectively perform differential processing on the filtered first work signal and the filtered second work signal to respectively determine a first envelope peak value corresponding to the filtered first work signal and a second envelope peak value corresponding to the filtered second work signal. After determining all envelope peak values corresponding to the work signal, the controller can further determine a peak number of the first envelope peak value and a peak number of the second envelope peak value in each preset time window. The preset time window can be determined according to actual conditions, or can be determined by comprehensively considering the identification robustness and sensitivity of the work identification model. For example, the preset time window can be a 30s time window. That is, the preset time period can be divided into multiple 30s time windows. The controller can determine the peak number of the first envelope peak value and the peak number of the second envelope peak value in each 30s time window. When determining the work mode subsequently, the peak number of the first envelope peak value and the peak number of the second envelope peak value can be input to the work identification model.

[0041] In one embodiment, the preset time period includes multiple time points, the work signal includes work parameter values corresponding to each time point, and the differential processing of the filtered first work signal and the filtered second work signal to respectively determine the first envelope peak value corresponding to the filtered first work signal and the second envelope peak value corresponding to the filtered second work signal includes: obtaining a work parameter value corresponding to any one of the time points in the filtered first work signal or the second work signal; determining a first parameter difference between the work parameter value of the current time point and the work parameter value of the previous time point, and a second parameter difference between the work parameter value of the current time point and the work parameter value of the next time point; in a case where the first parameter difference is greater than a preset threshold and the second parameter difference is less than the preset threshold, determining the work parameter value of the current time point as the envelope peak value corresponding to the work signal.

[0042] The preset time period can include multiple time points. The work signal can include work parameter values corresponding to each time point. The work signal can be a pressure signal or a current signal. In a case where the work signal is a pressure signal, the work parameter value corresponding to each time point can refer to a pressure value. In a case where the work signal is a current signal, the work parameter value corresponding to each time point can refer to a current value.

[0043] After filtering the first work signal and the second work signal respectively, the controller can obtain a work parameter value corresponding to any one time point in the filtered first work signal or the filtered second work signal. For example, the controller can obtain a work parameter value at a current time point, a work parameter value at a previous time point, and a work parameter value at a next time point in the filtered first work signal or the second work signal. Then, the controller can determine a first parameter difference between the work parameter value at the current time point and the work parameter value at the previous time point, and a second parameter difference between the work parameter value at the current time point and the work parameter value at the next time point.

[0044] In a case where the first parameter difference and the second parameter difference are determined, the controller can compare the first parameter difference with a preset threshold value, and can compare the second parameter difference with the preset threshold value. The preset threshold value can be 0. If the first parameter difference is greater than the preset threshold value and the second parameter difference is less than the preset threshold value, it can be determined that the trend of the pressure signal from the previous time point to the next time point is rising-falling, and at this time the controller can further determine the work parameter value corresponding to the current time point as the envelope peak value corresponding to the work signal. In a case where all envelope peak values corresponding to the work signal in a preset time period are determined, the controller can determine the number of peak values of the envelope peak values in the preset time window, so as to determine the work mode of the engineering vehicle through the work recognition model.

[0045] For example, the controller can obtain a pressure value y0 at a t-1 time point, a pressure value y1 at a t time point, and a pressure value y2 at a t+1 time point in the filtered first pressure signal, and the first pressure difference is y1-y0 and the second pressure difference is y2-y1. In a case where y1-y0>0 and y2-y1<0, the controller can determine the pressure value y1 at the t time point as the envelope peak value corresponding to the first pressure signal. In a case where y1-y0>0 and y2-y1>0 (the trend of the first pressure signal from t-1 to t+1 is rising), y1-y0<0 and y2-y1<0 (the trend of the first pressure signal from t-1 to t+1 is falling), and y1-y0<0 and y2-y1>0 (the trend of the first pressure signal from t-1 to t+1 is falling-rising), the controller can further obtain a pressure value y3 at a t+2 time point to determine the envelope peak value corresponding to the first pressure signal.

[0046] In one embodiment, the engineering vehicle includes a first main pump and a second main pump. The operation signal includes a first operation signal corresponding to the first main pump and a second operation signal corresponding to the second main pump. The preset time period includes multiple preset time windows, each preset time window includes multiple time points, the operation signal includes operation parameter values ​​corresponding to each time point, and the feature values ​​include the coefficient of variation and variance. Determining the feature values ​​corresponding to each preset time window based on the operation signal includes: for the first operation signal or the second operation signal, determining the mean value of the operation parameter values ​​corresponding to each time point in the operation signal; determining the variance and standard deviation of the operation parameter values ​​for each preset time window based on the mean value and the operation parameter values ​​corresponding to each time point; and determining the coefficient of variation corresponding to each preset time window based on the mean value and standard deviation of the parameters corresponding to each preset time window.

[0047] The engineering vehicle includes a first main pump and a second main pump. The operating signals of the engineering vehicle during a preset time period can include a first operating signal corresponding to the first main pump and a second operating signal corresponding to the second main pump. The operating signals can be pressure signals or current signals. When the operating signal is a pressure signal, the first operating signal corresponding to the first main pump can refer to a first pressure signal, and the second operating signal corresponding to the second main pump can refer to a second pressure signal. When the operating signal is a current signal, the first operating signal corresponding to the first main pump can refer to a first current signal, and the second operating signal corresponding to the second main pump can refer to a second current signal. The preset time period can include multiple preset time windows. Each preset time window can include multiple time points. The operating signal can include operating parameter values ​​corresponding to each time point. If the operating signal is a pressure signal, its operating parameter value at each time point refers to the pressure value. If the operating signal is a current signal, its operating parameter value at each time point refers to the current value.

[0048] For either the first or second work signal, the controller can determine the work parameter value corresponding to each time point in the work signal, and average the work parameter values ​​corresponding to all time points to determine the parameter mean. Having determined the parameter mean, the controller can further determine the variance and standard deviation of the work parameter values ​​for each preset time window based on the parameter mean and the work parameter value corresponding to each time point. Then, the controller can determine the coefficient of variation corresponding to each preset time window based on the parameter mean and standard deviation. The coefficient of variation and variance can effectively reflect the overall degree of variation of the work signal.

[0049] In one embodiment, such as Figure 2 As shown, a flowchart illustrating another method for determining the job mode is provided.

[0050] The controller can obtain the pressure signal P1 of the first main pump and the pressure signal P2 of the second main pump when the engineering vehicle is working in a time period. Then, the pressure signal P1 and the pressure signal P2 can be envelope filtered respectively. The controller can perform differential processing on the filtered pressure signal P1 and the filtered pressure signal P2 respectively to determine the local peak value corresponding to the filtered pressure signal P1 and the local peak value corresponding to the filtered pressure signal P2. In the case of determining the local peak value, the controller can divide the time period into a plurality of time windows and determine the envelope peak value quantity of the local peak value in each time window. In the case of obtaining the pressure signal P1 of the first main pump and the pressure signal P2 of the second main pump when the engineering vehicle is working in a time period, the controller can divide the time period into a plurality of time windows and determine the coefficient of variation and the variance corresponding to each time window.

[0051] In the case of determining the coefficient of variation, the variance and the envelope peak value quantity corresponding to the pressure signal P1, and the coefficient of variation, the variance and the envelope peak value quantity corresponding to the pressure signal P2, the controller can train the work recognition model by the above-mentioned parameters. In the case of completing the training of the work recognition model, the controller can obtain the real-time work signal when the engineering vehicle is working in a preset time period, and determine the corresponding coefficient of variation, variance and envelope peak value quantity according to the real-time work signal, so as to output the work mode of the engineering vehicle when working in the preset time period through the trained work recognition model. The work mode can include excavation work, grading work and crushing work.

[0052] In one embodiment, the method further comprises a training step of the work recognition model, and the training step comprises: obtaining sample signals corresponding to a plurality of historical work modes in the process of historical work of the engineering vehicle, and taking the sample signals in each historical work mode as training samples, wherein the engineering vehicle comprises at least a first main pump and a second main pump, the sample signals comprise a first sample signal corresponding to the first main pump and a second sample signal corresponding to the second main pump, and the label of the training sample is the corresponding historical work mode; determining the number of first historical peak values and the first historical characteristic values corresponding to the historical time window in each historical work mode according to the first sample signal respectively; determining the number of second historical peak values and the second historical characteristic values corresponding to the historical time window in each historical work mode according to the second sample signal respectively; inputting the number of first historical peak values, the number of second historical peak values, the first historical characteristic values and the second historical characteristic values of all historical time windows into the work recognition model to train the work recognition model.

[0053] The controller can obtain sample signals of the engineering vehicle in a plurality of historical modes corresponding to historical operation processes, and take the sample signals in each historical operation mode as training samples. The engineering vehicle at least includes a first main pump and a second main pump. The sample signals can include first sample signals corresponding to the first main pump and second sample signals corresponding to the second main pump. The training samples can refer to historical pressure signals or historical current signals in each historical operation mode. The training samples carry labels. The labels of the training samples can be the corresponding historical operation modes. For example, the corresponding historical operation mode of the training sample 1 is grading operation, and the corresponding historical operation mode of the training sample 2 is excavation operation. In the case that the training samples are historical pressure signals in each historical operation mode, the historical pressure signals can include first historical pressure signals corresponding to the first main pump and second historical pressure signals corresponding to the second main pump. In the case that the training samples are historical current signals in each historical operation mode, the historical pressure signals can include first historical current signals corresponding to the first main pump and second historical current signals corresponding to the second main pump. The historical operation modes can include excavation operation, grading operation and breaking operation.

[0054] The controller can determine the number of first historical peaks corresponding to the historical time window in each historical operation mode according to the first sample signals. Specifically, the controller can filter the first sample signals. Then, the controller can perform differential processing on the filtered first sample signals to determine the first historical peaks corresponding to the filtered first sample signals. In the case of determining the first historical peaks, the controller can determine the number of first historical peaks corresponding to the historical time window.

[0055] The controller can determine first historical characteristic values corresponding to the historical time window in each historical operation mode according to the first sample signals. The first sample signals can include first sample values corresponding to each historical time point. The historical time window includes a plurality of historical time points. The first historical characteristic values can refer to the coefficient of variation and the variance corresponding to the historical time window determined according to the first sample signals. Specifically, the controller can determine the mean of the first sample values corresponding to each historical time point in the first sample signals. In the case of determining the mean of the first sample values, the controller can determine the variance and the standard deviation of the first sample values of the historical time window according to the mean of the first sample values and the first sample values corresponding to each historical time point. Then, the controller can further determine the coefficient of variation corresponding to the historical time window according to the mean and the standard deviation of the first sample values corresponding to the historical time window.

[0056] The controller can determine, according to the second sample signal, a peak number of the second historical peak corresponding to the historical time window in each historical operation mode. Specifically, the controller can filter the second sample signal. Then, the controller can perform differential processing on the filtered second sample signal to determine the second historical peak corresponding to the filtered second sample signal. After determining the second historical peak, the controller can determine the number of the second historical peak corresponding to the historical time window.

[0057] The controller can determine, according to the second sample signal, a second historical characteristic value corresponding to the historical time window in each historical operation mode. The second sample signal can include a second sample value corresponding to each historical time point. The second historical characteristic value can refer to a coefficient of variation and a variance corresponding to the historical time window determined according to the second sample signal. Specifically, the controller can determine the mean of the second sample value corresponding to each historical time point in the second sample signal. After determining the mean of the second sample value, the controller can determine the variance and the standard deviation of the second sample value of the historical time window according to the mean of the second sample value and the second sample value corresponding to each historical time point. Then, the controller can further determine the coefficient of variation corresponding to the historical time window according to the mean and the standard deviation of the second sample value corresponding to the historical time window.

[0058] The controller can input the number of the first historical peak, the number of the second historical peak, the first historical characteristic value, and the second historical characteristic value of all the historical time windows to the operation recognition model to train the operation recognition model. The operation recognition model can refer to a classification model. The classification model can be a random forest.

[0059] In one embodiment, the method further comprises: outputting, by the operation recognition model, a plurality of predicted operation modes corresponding to the sample signal for each training of the operation recognition model; comparing each predicted operation mode with the actual operation mode to determine a plurality of prediction deviation values of the operation recognition model; determining that the training of the operation recognition model is completed when an error between adjacent number of prediction deviation values is less than or equal to a preset value; and inputting again the number of the first historical peak, the number of the second historical peak, the first historical characteristic value, and the second historical characteristic value of all the historical time windows to the operation recognition model to train the operation recognition model until the number of training of the operation recognition model reaches a preset training number or the error between adjacent number of prediction deviation values is less than or equal to the preset value when the error between adjacent number of prediction deviation values is greater than the preset value.

[0060] The work recognition model can be a classification model. The classification model can be a random forest. The sample signal can be a historical pressure signal or a historical current signal. The work recognition model can be trained multiple times by the number of first historical peaks, the number of second historical peaks, the first historical characteristic value, and the second historical characteristic value of all historical time windows. For each training of the work recognition model, the controller can output multiple predicted work modes corresponding to the sample signal by the work recognition model. When the actual work mode corresponding to the sample signal is obtained, the controller can compare the predicted work mode with the actual work to determine multiple prediction deviation values of the work recognition model, so that the training effect of the model can be evaluated according to the multiple prediction deviation values.

[0061] When the error between the adjacent number of prediction deviation values is less than or equal to a preset value, the controller can determine that the work recognition model training is completed. That is, after continuous multiple training, the error of the prediction deviation value corresponding to each training does not exceed the preset value, at this time, the prediction accuracy of the work recognition model can have reached a maximum value, and the processor can determine that the work recognition model training is completed. When the error between the adjacent number of prediction deviation values is greater than the preset value, the processor can input the number of first historical peaks, the number of second historical peaks, the first historical characteristic value, and the second historical characteristic value of all historical time windows to the work recognition model again to train the work recognition model. That is, after continuous multiple training, the error of the prediction deviation value corresponding to each training exceeds the preset value, at this time, the prediction accuracy of the work recognition model can not have reached a maximum value, and the processor can train the work recognition model again until the training number of the work recognition model reaches a preset training number or the error between the adjacent number of prediction deviation values is less than or equal to the preset value, so as to improve the training accuracy of the work recognition model. The preset training number can be customized according to actual conditions.

[0062] In one embodiment, the loss value in the recognition model is used to represent the difference between the label value of the training sample and the output value of the model. When the loss value no longer decreases or the iteration number is greater than a preset value, the model training is completed. The grid search is used for parameter optimization. The parameters are adjusted in steps in turn. Through loop traversal, the parameters with the highest accuracy on the validation set are found from all the parameters. The model corresponding to the set of parameters is the final work recognition model.

[0063] In one embodiment, as Figure 3As shown, a flowchart for training a work recognition model is provided. The controller can input features X into a digging action recognition model to output a prediction Z by the digging action recognition model. The features X can include a number of first historical peak values, a number of second historical peak values, a first historical feature value, and a second historical feature value. The digging action recognition model can refer to a work recognition model. The prediction Z can refer to a predicted work mode of the engineering vehicle. The label Y can refer to an actual work mode corresponding to the features X. In the case that the prediction Z is output by the digging action recognition model, the controller can compare the label Y with the prediction Z to determine a prediction deviation value of the work digging action recognition model until the prediction deviation value does not significantly decrease in continuous multiple times of training, or the model training reaches a preset number of iterations.

[0064] In one embodiment, the method further comprises: obtaining work signals of the engineering vehicle in an actual work process as prediction signals, the prediction signals including a first work signal corresponding to the first main pump and a second work signal corresponding to the second main pump; determining a number of first work peak values and a first work feature value according to the first work signal respectively; determining a number of second work peak values and a second work feature value according to the second work signal respectively; inputting the number of first work peak values, the number of second work peak values, the first work feature value, and the second work feature value into the work recognition model to predict a work mode of the engineering vehicle in the actual work process.

[0065] The controller can obtain work signals of the engineering vehicle in an actual work process as prediction signals. The prediction signals can be voltage signals or current signals. The prediction signals can include a first work signal corresponding to a first main pump of the engineering vehicle and a second work signal corresponding to a second main pump. The controller can determine a number of first work peak values and a first work feature value according to the first work signal respectively. The controller can determine a number of second work peak values and a second work feature value according to the second work signal respectively. The work feature value can include a coefficient of variation and a variance. The work peak value can refer to an envelope peak value. The controller can input the number of first work peak values, the number of second work peak values, the first work feature value, and the second work feature value into the work recognition model to predict a work mode of the engineering vehicle in the actual work process by the work recognition model. The work mode of the engineering vehicle can include a digging work, a grading work, and a breaking work. By the above technical solution, the number of peak values and the feature values of all preset time windows are input into the work recognition model to predict the work mode of the engineering vehicle from the local change trend and the overall change trend of the work signals, which greatly improves the accuracy and reliability of determining the work mode, has low required artificial cost and time cost, and effectively improves the work quality and work efficiency of the engineering vehicle.

[0066] Figures 1-2 A flowchart of a method for determining a work mode in an embodiment is shown. It should be understood that although the steps in the flowchart are shown in a certain order, the steps are not necessarily executed in the order shown by the arrows. Unless otherwise specified herein, the steps are not necessarily executed in a strict order, and the steps can be executed in other orders. Moreover, Figures 1-2 Unless otherwise specified herein, the steps in the flowchart are not necessarily executed in the order shown by the arrows. Unless otherwise specified herein, the steps are not necessarily executed in a strict order, and the steps can be executed in other orders. Moreover, Figures 1-2 At least a part of the steps in the flowchart can include multiple sub-steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of the sub-steps or stages is not necessarily sequential, but can be executed in rotation or alternation with at least a part of other steps or sub-steps or stages of other steps.

[0067] In an embodiment, the controller is configured to run a program, wherein the program is configured to execute the method for determining a work mode when the program is running.

[0068] In an embodiment, as shown in Figure 4 A device for determining a work mode is provided, which includes a data acquisition device 401 and the controller 402 described above. The data acquisition device 401 can be configured to acquire work signals generated by the construction vehicle when working.

[0069] In an embodiment, a construction vehicle is provided, which includes at least one main pump and the device for determining a work mode described above.

[0070] In an embodiment, a storage medium is provided, which stores a program that is executed by a controller to implement the method for determining a work mode described above.

[0071] In an embodiment, a computer device is provided, which can be a server, and the internal structure diagram of the computer device can be as shown in Figure 5As shown in the figure. The computer device includes a processor A01, a network interface A02, a memory (not shown in the figure) and a database (not shown in the figure) connected by a system bus. Among them, the processor A01 of the computer device is used to provide computing and control capabilities. The memory of the computer device includes an internal memory A03 and a non-volatile storage medium A04. The non-volatile storage medium A04 stores an operating system B01, a computer program B02 and a database (not shown in the figure). The internal memory A03 provides an environment for the operating system B01 and the computer program B02 in the non-volatile storage medium A04 to run. The database of the computer device is used to store job signals and other data. The network interface A02 of the computer device is used to communicate with external terminals through network connection. The computer program B02 is executed by the processor A01 to implement a method for determining a job mode.

[0072] Those skilled in the art can understand that, Figure 5 The structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the computer device to which the scheme of the present application is applied. The specific computer device can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0073] The embodiment of the present application provides a device, the device includes a processor, a memory and a program stored in the memory and executable on the processor, and the processor implements the following steps when executing the program: obtaining a job signal of an engineering vehicle when the engineering vehicle works in a preset time period, the preset time period including a plurality of preset time windows; determining a peak value number of an envelope peak value corresponding to each preset time window and a feature value corresponding to each preset time window according to the job signal; inputting the peak value number and the feature value of all the preset time windows into a job recognition model to output a job mode corresponding to the engineering vehicle when the engineering vehicle works in the preset time period through the job recognition model.

[0074] In one embodiment, the engineering vehicle includes a first main pump and a second main pump, and the job signal includes a first job signal corresponding to the first main pump and a second job signal corresponding to the second main pump; determining the peak value number of the envelope peak value corresponding to each preset time window according to the job signal includes: filtering the first job signal and the second job signal respectively; performing differential processing on the filtered first job signal and the filtered second job signal respectively to determine a first envelope peak value corresponding to the filtered first job signal and a second envelope peak value corresponding to the filtered second job signal respectively; and determining the peak value number of the first envelope peak value and the peak value number of the second envelope peak value in each preset time window.

[0075] In one embodiment, the preset time period includes a plurality of time points, the work signal includes work parameter values corresponding to each time point, and the filtered first work signal and the filtered second work signal are subjected to differential processing to determine a first envelope peak value corresponding to the filtered first work signal and a second envelope peak value corresponding to the filtered second work signal, respectively. The method includes: obtaining a work parameter value corresponding to any one of the time points in the filtered first work signal or the filtered second work signal; determining a first parameter difference between a work parameter value of a current time point and a work parameter value of a previous time point, and a second parameter difference between the work parameter value of the current time point and a work parameter value of a next time point; and determining the work parameter value of the current time point as an envelope peak value corresponding to the work signal in a case where the first parameter difference is greater than a preset threshold and the second parameter difference is less than the preset threshold.

[0076] In one embodiment, the engineering vehicle includes a first main pump and a second main pump, the work signal includes a first work signal corresponding to the first main pump and a second work signal corresponding to the second main pump, the preset time period includes a plurality of preset time windows, each preset time window includes a plurality of time points, the work signal includes work parameter values corresponding to each time point, and the characteristic value includes a coefficient of variation and a variance. The method includes: determining, for the first work signal or the second work signal, a parameter mean of the work parameter values corresponding to each time point in the work signal; determining, according to the parameter mean and the work parameter value corresponding to each time point, a variance and a standard deviation of the work parameter values of each preset time window; and determining, according to the parameter mean and the standard deviation corresponding to each preset time window, a coefficient of variation corresponding to each preset time window.

[0077] In one embodiment, the method further includes a training step of the work recognition model. The training step includes: obtaining sample signals corresponding to a plurality of historical work modes of the engineering vehicle in a historical work process, and taking the sample signals in each historical work mode as training samples, wherein the engineering vehicle includes at least a first main pump and a second main pump, the sample signals include a first sample signal corresponding to the first main pump and a second sample signal corresponding to the second main pump, and a label of the training sample is a corresponding historical work mode; determining, according to the first sample signal, a number of first historical peak values corresponding to a historical time window and a first historical characteristic value in each historical work mode; determining, according to the second sample signal, a number of second historical peak values corresponding to the historical time window and a second historical characteristic value in each historical work mode; and inputting the number of first historical peak values, the number of second historical peak values, the first historical characteristic value, and the second historical characteristic value of all historical time windows into the work recognition model to train the work recognition model.

[0078] In one embodiment, the method further comprises: outputting, by the work recognition model, a plurality of predicted work modes corresponding to the sample signal for each training of the work recognition model; comparing each predicted work mode with an actual work mode to determine a plurality of prediction deviation values of the work recognition model; determining that the training of the work recognition model is completed in a case that an error between adjacent number of prediction deviation values is less than or equal to a preset value; in a case that the error between adjacent number of prediction deviation values is greater than the preset value, inputting the number of first historical envelope peaks, the number of second historical envelope peaks, the first historical feature value and the second historical feature value of all historical time windows to the work recognition model again to train the work recognition model until a preset training number of the work recognition model is reached or the error between adjacent number of prediction deviation values is less than or equal to the preset value.

[0079] In one embodiment, the engineering vehicle is an excavator, and the work modes of the engineering vehicle include any one of a digging work, a grading work and a breaking work.

[0080] The application also provides a computer program product adapted to execute a program for initializing the following method steps when executed on a data processing device: obtaining work signals of an engineering vehicle during a preset time period, the preset time period including a plurality of preset time windows; determining a peak number of envelope peaks corresponding to each preset time window and a feature value corresponding to each preset time window according to the work signals; inputting the peak numbers and the feature values of all preset time windows to a work recognition model to output a work mode corresponding to the engineering vehicle during the preset time period by the work recognition model.

[0081] In one embodiment, the engineering vehicle includes a first main pump and a second main pump, and the work signals include a first work signal corresponding to the first main pump and a second work signal corresponding to the second main pump; the determining of the peak number of envelope peaks corresponding to each preset time window according to the work signals includes: filtering the first work signal and the second work signal respectively; differentiating the filtered first work signal and the second work signal respectively to determine a first envelope peak corresponding to the filtered first work signal and a second envelope peak corresponding to the filtered second work signal respectively; and determining the peak number of the first envelope peaks and the peak number of the second envelope peaks in each preset time window.

[0082] In one embodiment, the preset time period includes a plurality of time points, the work signal includes work parameter values corresponding to each time point, and the filtered first work signal and the filtered second work signal are subjected to differential processing to determine a first envelope peak value corresponding to the filtered first work signal and a second envelope peak value corresponding to the filtered second work signal, respectively. The method includes: obtaining a work parameter value corresponding to any one of the time points in the filtered first work signal or the filtered second work signal; determining a first parameter difference between a work parameter value of a current time point and a work parameter value of a previous time point, and a second parameter difference between the work parameter value of the current time point and a work parameter value of a next time point; and determining the work parameter value of the current time point as the envelope peak value corresponding to the work signal in a case where the first parameter difference is greater than a preset threshold and the second parameter difference is less than the preset threshold.

[0083] In one embodiment, the engineering vehicle includes a first main pump and a second main pump, the work signal includes a first work signal corresponding to the first main pump and a second work signal corresponding to the second main pump, the preset time period includes a plurality of preset time windows, each preset time window includes a plurality of time points, the work signal includes work parameter values corresponding to each time point, and the characteristic value includes a coefficient of variation and a variance. The method includes: determining, for the first work signal or the second work signal, a parameter mean of the work parameter values corresponding to each time point in the work signal; determining, according to the parameter mean and the work parameter value corresponding to each time point, a variance and a standard deviation of the work parameter values of each preset time window; and determining, according to the parameter mean and the standard deviation corresponding to each preset time window, a coefficient of variation corresponding to each preset time window.

[0084] In one embodiment, the method further includes a training step of the work recognition model. The training step includes: obtaining sample signals corresponding to a plurality of historical work modes of the engineering vehicle in a historical work process, and taking the sample signals in each historical work mode as training samples, wherein the engineering vehicle includes at least a first main pump and a second main pump, the sample signals include a first sample signal corresponding to the first main pump and a second sample signal corresponding to the second main pump, and a label of the training sample is a corresponding historical work mode; determining, according to the first sample signal, a number of first historical peak values corresponding to a historical time window and a first historical characteristic value in each historical work mode; determining, according to the second sample signal, a number of second historical peak values corresponding to the historical time window and a second historical characteristic value in each historical work mode; and inputting the number of first historical peak values, the number of second historical peak values, the first historical characteristic value, and the second historical characteristic value of all historical time windows into the work recognition model to train the work recognition model.

[0085] In an embodiment, the method further comprises: outputting, by the work recognition model, a plurality of predicted work modes corresponding to the sample signal for each training of the work recognition model; comparing each predicted work mode with an actual work mode to determine a plurality of prediction deviation values of the work recognition model; determining that the training of the work recognition model is completed in a case where an error between adjacent number of prediction deviation values is less than or equal to a preset value; and inputting, in a case where the error between adjacent number of prediction deviation values is greater than the preset value, the number of first historical peak values, the number of second historical peak values, the first historical feature value, and the second historical feature value of all historical time windows to the work recognition model again to train the work recognition model until a preset training number of the work recognition model is reached or the error between adjacent number of prediction deviation values is less than or equal to the preset value.

[0086] In an embodiment, the engineering vehicle is an excavator, and the work modes of the engineering vehicle include any one of a digging work, a grading work, and a breaking work.

[0087] Those skilled in the art will understand that embodiments of the present application can be provided as methods, systems, or computer program products. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer-readable program code.

[0088] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as combinations of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions, which are executed via the processor of the computer or other programmable data processing apparatus, generate a means for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 The flow or flows and / or blocks in a flowchart and / or block diagram Figure 1 The means for carrying out the function specified by a flow or flows and / or blocks in a flowchart and / or block diagram.

[0089] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction means, which implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The flow or flows and / or blocks in a flowchart and / or block diagram Figure 1the function(s) specified in the block or blocks.

[0090] These computer program instructions can also be loaded into computer or other programmable data processing devices to cause a series of operational steps to be performed on the computer or other programmable devices to produce a computer implemented process such that the instructions which execute on the computer or other programmable devices provide steps for implementing the flowchart block(s) or flowchart flow(s) and / or portions thereof. Figure 1 the flowchart block(s) or flowchart flow(s) and / or portions thereof. Figure 1 the function(s) specified in the block or blocks.

[0091] In one typical configuration, the computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0092] The memory can include non-persistent memory and / or volatile memory, such as random access memory (RAM) about which the computer stores information such as computer program instructions. Memory is an example of computer readable media.

[0093] Computer readable media includes permanent and non-permanent, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital versatile discs (DVD) or other optical storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to computing devices. According to the definition herein, computer readable media does not include transitory media, such as modulated data signals and carrier waves.

[0094] It should also be noted that the terms "comprising", "including", or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article or apparatus that comprises a list of elements does not include only those elements recited, but can also include other elements not expressly listed or inherent to such process, method, article or apparatus. Without further limitation, an element preceded by "comprises a" does not, without more constraints, foreclose the existence of additional identical elements in the process, method, article or apparatus that includes the element.

[0095] The above merely provides an example of the present application, and is not intended to limit the present application. For those skilled in the art, the present application can have various modifications and changes. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall fall into the scope of claims of the present application.

Claims

1. A method for determining a mode of operation, characterized by, The method is applied to an engineering vehicle comprising a first main pump and a second main pump, and comprises: obtaining work signals of the engineering vehicle during work in a preset time period, the preset time period comprising a plurality of preset time windows, the work signals comprising a first work signal corresponding to the first main pump and a second work signal corresponding to the second main pump, each preset time window comprising a plurality of time points, and the work signals comprising work parameter values corresponding to each time point; determining, according to the work signals, a peak number of envelope peaks corresponding to each preset time window and a characteristic value corresponding to each preset time window, the characteristic value comprising a coefficient of variation and a variance; inputting the peak number and the characteristic value of all preset time windows into a work recognition model to output, by the work recognition model, a work mode corresponding to work of the engineering vehicle in the preset time period; wherein determining, according to the work signals, the characteristic value corresponding to each preset time window comprises: for the first work signal or the second work signal, determining a parameter mean of the work parameter values corresponding to each time point in the work signal; determining, according to the parameter mean and the work parameter value corresponding to each time point, a variance and a standard deviation of the work parameter value of each preset time window; determining, according to the parameter mean and the standard deviation corresponding to each preset time window, a coefficient of variation corresponding to each preset time window.

2. The method for determining a mode of operation of claim 1, wherein, determining, according to the work signals, the peak number of envelope peaks corresponding to each preset time window comprises: filtering the first work signal and the second work signal respectively; differentially processing the filtered first work signal and the filtered second work signal respectively to determine a first envelope peak corresponding to the filtered first work signal and a second envelope peak corresponding to the filtered second work signal respectively; determining the peak number of the first envelope peak and the peak number of the second envelope peak in each preset time window.

3. The method for determining a mode of operation of claim 2, wherein, differentially processing the filtered first work signal and the filtered second work signal to determine the first envelope peak corresponding to the filtered first work signal and the second envelope peak corresponding to the filtered second work signal respectively comprises: obtaining a work parameter value corresponding to any one time point in the filtered first work signal or the filtered second work signal; determining a first parameter difference between the work parameter value of the current time point and the work parameter value of the previous time point, and a second parameter difference between the work parameter value of the current time point and the work parameter value of the next time point; in a case where the first parameter difference is greater than a preset threshold and the second parameter difference is less than the preset threshold, determining the work parameter value of the current time point as an envelope peak corresponding to the work signal.

4. The method for determining a mode of operation of claim 1, wherein, The method further comprises a training step of the work recognition model, and the training step comprises: Obtaining sample signals of an engineering vehicle in a plurality of historical operation modes during historical operation processes, and taking the sample signals in each historical operation mode as training samples, wherein the engineering vehicle comprises at least a first main pump and a second main pump, the sample signals comprise first sample signals corresponding to the first main pump and second sample signals corresponding to the second main pump, and a label of the training sample is a corresponding historical operation mode; determining a number of first historical peak values corresponding to a historical time window and a first historical characteristic value in each historical operation mode according to the first sample signals; determining a number of second historical peak values corresponding to a historical time window and a second historical characteristic value in each historical operation mode according to the second sample signals; inputting the number of first historical peak values, the number of second historical peak values, the first historical characteristic value and the second historical characteristic value of all historical time windows into the operation recognition model to train the operation recognition model.

5. The method for determining a mode of operation of claim 4, wherein, The method further comprises: outputting, by the operation recognition model, a plurality of predicted operation modes corresponding to the sample signals for each training of the operation recognition model; comparing each predicted operation mode with an actual operation mode to determine a plurality of prediction deviation values of the operation recognition model; determining that the training of the operation recognition model is completed when an error between adjacent number of prediction deviation values is less than or equal to a preset value; in a case where the error between adjacent number of prediction deviation values is greater than the preset value, inputting again the number of first historical peak values, the number of second historical peak values, the first historical characteristic value and the second historical characteristic value of all historical time windows into the operation recognition model to train the operation recognition model until a training number of the operation recognition model reaches a preset training number or the error between adjacent number of prediction deviation values is less than or equal to the preset value.

6. The method for determining a mode of operation according to any one of claims 1 to 5, characterized in that The engineering vehicle is an excavator, and the operation modes of the engineering vehicle include any one of a digging operation, a grading operation and a breaking operation.

7. A controller characterized by comprising: The device is configured to perform the method for determining an operation mode according to any one of claims 1 to 6.

8. An apparatus for determining a mode of operation, characterized by The device comprises: a data acquisition device configured to acquire operation signals generated by the engineering vehicle during operation; and the controller according to claim 7.

9. An engineering vehicle characterized by, The engineering vehicle comprises: at least one main pump; and the device for determining an operation mode according to claim 8. The engineering vehicle comprises: at least one main pump; and the device for determining an operation mode according to claim 8.

Images