Working condition recognition method and device for engineering machinery, storage medium and processor

Abstract

This invention relates to the field of construction machinery technology, and discloses a method, device, storage medium, and processor for identifying the operating conditions of construction machinery. The identification method of this invention includes: acquiring basic data and main pump pressure signals of the construction machinery; identifying the basic actions of the construction machinery using a pre-set basic action identification model based on the basic data, obtaining basic action identification results; and identifying the operating condition stages of the construction machinery using a pre-set operating condition stage identification model based on the basic action identification results and the main pump pressure signal, obtaining operating condition stage identification results. This achieves a multi-level operating condition identification method, reducing the complexity and computational requirements of the operating condition identification model, and improving the accuracy and stability of the operating condition identification results. Simultaneously, it can cover more basic actions and operating condition stages, broadening the coverage of the identification model, and can be used in various scenarios, improving the environmental adaptability of the operating condition identification method.

Description

Technical Field

[0001] This invention relates to the field of engineering machinery technology, specifically to a method for identifying the working conditions of engineering machinery, a device for identifying the working conditions of engineering machinery, a machine-readable storage medium, and a processor. Background Technology

[0002] With rising labor and material costs and increasing demands for quality and schedule during project construction, intelligent construction, characterized by data-driven project decision-making based on on-site equipment and personnel operational information, has become an important direction for future development. For construction machinery responsible for large-scale lifting of heavy objects, accurate analysis and feedback of operational efficiency-related information can not only significantly improve the intelligence level of individual machines but also form a crucial foundation for achieving intelligent construction.

[0003] The statistics of the operating efficiency of construction machinery include lifting weight sensing and lifting speed sensing. In order to accurately count the operating efficiency of construction machinery, it is necessary to count the lifting speed of each working condition, the duration of each working condition, and compare the lifting speed of each working condition. All of these require the identification of the working condition of the construction machinery.

[0004] Currently, the existing technology for identifying the working conditions of construction machinery often uses a visual information-based method. However, this method is affected by light, and the basic actions of the actuators of construction machinery are similar in various working conditions, which leads to inaccurate identification of the working conditions based on visual information and a narrow coverage. Summary of the Invention

[0005] The purpose of this invention is to provide a method for identifying the working conditions of construction machinery, a device for identifying the working conditions of construction machinery, a machine-readable storage medium, and a processor. This method for identifying the working conditions of construction machinery can realize multi-level working condition identification, can cover more basic actions and working condition stages, and can improve the accuracy and stability of the working condition identification results.

[0006] To achieve the above objectives, the first aspect of this application provides a method for identifying the operating conditions of engineering machinery, comprising:

[0007] Acquire basic data and main pump pressure signals from the construction machinery;

[0008] Based on the aforementioned basic data, a pre-set basic motion recognition model is used to identify the basic motions of the engineering machinery, thereby obtaining the basic motion recognition results.

[0009] Based on the basic action recognition results and the main pump pressure signal, a preset working condition stage recognition model is used to identify the working condition stage of the engineering machinery, and the working condition stage recognition results are obtained.

[0010] The data types of the basic data include at least one of the following: actuator pilot control signal, actuator displacement, actuator tilt angle, and hydraulic pump flow rate.

[0011] In this embodiment of the application, the construction process of the basic action recognition model includes:

[0012] Obtain first sample data, which includes first sample basic data under each basic action and basic action labels corresponding to the first sample basic data;

[0013] The first sample basic data for each basic action is input into the first neural network to obtain the predicted basic action.

[0014] The parameters of the first neural network are adjusted according to the predicted basic action and the basic action label corresponding to the basic data of the first sample in the first sample data to obtain the basic action recognition model.

[0015] In this embodiment of the application, the construction process of the working condition stage identification model includes:

[0016] Acquire second sample data, which includes basic actions and main pump pressure signals under each operating condition stage, as well as operating condition stage labels corresponding to the basic actions;

[0017] The basic actions and main pump pressure signals under each working condition stage are respectively input into the second neural network to obtain the predicted working condition stage;

[0018] The parameters of the second neural network are adjusted based on the predicted operating condition stage and the labels of each operating condition stage in the second sample data to obtain the operating condition stage recognition model.

[0019] In this embodiment of the application, the data type of the basic data is the actuator pilot control signal, and the actuator pilot control signal includes multiple actuator pilot control signal data within a first preset time range from the current time;

[0020] The step of identifying the basic movements of the construction machinery using a pre-set basic movement recognition model based on the basic data, and obtaining the basic movement recognition results, includes:

[0021] A1: Perform feature extraction on the pilot control signal of the actuator to obtain a feature vector;

[0022] A2: Normalize the feature vector to obtain the normalized feature vector;

[0023] A3: Input the normalized feature vector into the preset basic action recognition model to perform basic action recognition and obtain the basic action recognition result;

[0024] A4: Determine whether the basic motion recognition process has ended;

[0025] A5: Output the basic action recognition result after the basic action recognition process is completed;

[0026] A6: If the basic action recognition process is not completed, remove the actuator pilot control signal data that is furthest from the current time in the actuator pilot control signal, and obtain the actuator pilot control signal data of the next time to update the actuator pilot control signal, so as to obtain a new actuator pilot control signal and return to execute A1.

[0027] In this embodiment of the application, the step of identifying the working stage of the construction machinery based on the basic action recognition result and the main pump pressure signal using a preset working stage recognition model to obtain the working stage recognition result includes:

[0028] B1: Acquire multiple basic action recognition results and the main pump pressure signal corresponding to each basic action recognition result within a second preset time range from the current time, and combine the multiple basic action recognition results and the main pump pressure signal corresponding to each basic action recognition result into a working condition stage recognition feature vector;

[0029] B2: Normalize the operating condition stage identification feature vector to obtain the normalized operating condition stage identification feature vector.

[0030] B3: Input the normalized working condition stage identification feature vector into the preset working condition stage identification model to identify the working condition stage and obtain the working condition stage identification result.

[0031] B4: Determine whether the working condition stage identification process has ended;

[0032] B5: Output the working condition stage identification result after the working condition stage identification process is completed;

[0033] B6: If the working condition stage identification process is not completed, remove the basic action identification result and the corresponding main pump pressure signal that are farthest from the current time in the working condition stage identification feature vector, and obtain the basic action identification result and the corresponding main pump pressure signal of the next time moment to update the working condition stage identification feature vector, obtain a new working condition stage identification feature vector, and return to execute B2.

[0034] In this embodiment of the application, after executing B3, the following is also included:

[0035] Obtain historical operating condition stage identification results;

[0036] Based on the historical operating condition stage identification results and the preset intelligent verification rule library, the operating condition stage identification results are verified and corrected to obtain the corrected operating condition stage identification results.

[0037] In this embodiment of the application, the historical working condition stage identification result includes the working condition stage identification result of the previous moment, and multiple working condition stage identification results within a third preset time range from the previous moment;

[0038] The step of verifying and correcting the operating condition stage identification results based on the historical operating condition stage identification results and a pre-set intelligent verification rule base to obtain corrected operating condition stage identification results includes:

[0039] Based on the identification results of multiple working condition stages within the third preset time range from the previous moment, the actual working condition stage at the current moment is determined;

[0040] Based on the current actual working condition stage and the working condition stage identification result of the previous time, the corresponding working condition stage identification result correction rule is matched in the preset intelligent verification rule library.

[0041] Based on the correction rules for the identified operating conditions, the identified operating conditions are verified and corrected to obtain the corrected identified operating conditions.

[0042] In this embodiment of the application, after obtaining the working condition stage identification result, the method further includes:

[0043] Based on the basic motion recognition results and the working condition stage recognition results, the hoisting operation speed information is calculated;

[0044] Obtain lifting weight statistics;

[0045] Based on the hoisting operation speed information and the hoisting weight statistics, the operation efficiency information of the construction machinery is obtained.

[0046] A second aspect of this application provides a working condition identification device for engineering machinery, comprising:

[0047] The acquisition module is used to acquire basic data and main pump pressure signals of the construction machinery; wherein, the data type of the basic data includes at least one of the following: actuator pilot control signal, actuator displacement, actuator tilt angle, and hydraulic pump flow rate;

[0048] The motion recognition module is used to recognize the basic motions of the construction machinery based on the basic data and a preset basic motion recognition model, so as to obtain the basic motion recognition results.

[0049] The working condition identification module is used to identify the working condition stage of the construction machinery based on the basic action identification result and the main pump pressure signal, using a preset working condition stage identification model, and obtain the working condition stage identification result.

[0050] A third aspect of this application provides a processor configured to perform the above-described engineering machinery operating condition identification method.

[0051] A fourth aspect of this application provides a machine-readable storage medium storing instructions that, when executed by a processor, configure the processor to perform the above-described engineering machinery operating condition identification method.

[0052] The above technical solution first acquires the basic data and main pump pressure signal of the construction machinery. Then, based on the basic data and main pump pressure signal, it sequentially performs basic action recognition and working condition stage recognition using a basic action recognition model and a working condition stage recognition model, thereby obtaining the working condition stage recognition result. By using the basic data and main pump pressure signal, the working condition recognition of the construction machinery is completed in a multi-step manner, from basic action to working condition stage, thus realizing a multi-level working condition recognition method. Compared with single-step working condition recognition methods, this reduces the complexity of the working condition recognition model and the computational requirements. The multi-level working condition recognition method can improve the accuracy and stability of the working condition recognition results. Simultaneously, the basic action recognition model and working condition stage recognition model can cover more basic actions and working condition stages, broadening the coverage of the recognition model and making it applicable to various scenarios, thus improving the environmental adaptability of the working condition recognition method.

[0053] Other features and advantages of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0054] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the following detailed description to explain the invention, but do not constitute a limitation thereof. In the drawings:

[0055] Figure 1 The illustration shows a flowchart of a method for identifying the working conditions of engineering machinery according to an embodiment of this application;

[0056] Figure 2 This schematic diagram illustrates a structural block diagram of an engineering machinery operating condition identification device according to an embodiment of this application;

[0057] Figure 3 The diagram illustrates the internal structure of a computer device according to an embodiment of this application.

[0058] Explanation of reference numerals in the attached figures

[0059] 410 - Acquisition Module; 420 - Action Recognition Module; 430 - Operating Condition Recognition Module; A01 - Processor; A02 - Network Interface; A03 - Internal Memory; A04 - Display Screen; A05 - Input Device; A06 - Non-Volatile Storage Medium; B01 - Operating System; B02 - Computer Program. Detailed Implementation

[0060] 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. It should be understood that the specific embodiments described herein are only for illustration and explanation of the embodiments of this application and are not intended to limit the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0061] It should be noted that if the embodiments of this application involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of the components in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicators will also change accordingly.

[0062] Furthermore, if the embodiments of this application involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed in this application.

[0063] Please refer to Figure 1 , Figure 1 The illustration shows a flowchart of a method for identifying the operating conditions of engineering machinery according to an embodiment of this application. Figure 1As shown in one embodiment of this application, a method for identifying the working conditions of construction machinery is provided. It should be noted that this method can be applied to construction cranes, earthmoving machinery, pump trucks, aerial work platforms, etc. Specifically, for cranes, this method can be applied to work scenarios involving winching, luffing, and slewing movements; for earthmoving machinery, it can be applied to earthmoving operations such as excavation, shoveling, pushing, or leveling of soil and gravel, including actions such as bucket movement, boom movement, forearm movement, and slewing, with working conditions including excavation, leveling, slope repair, and breaking. For ease of explanation, this embodiment primarily uses cranes as an example, specifically an example of a crane performing a lifting operation. The method for identifying the working conditions of construction machinery includes the following steps:

[0064] Step 210: Obtain the basic data and main pump pressure signal of the construction machinery; wherein, the data type of the basic data includes at least one of the following: actuator pilot control signal, actuator displacement, actuator tilt angle, and hydraulic pump flow rate.

[0065] Step 220: Based on the basic data, use a preset basic motion recognition model to identify the basic motions of the engineering machinery and obtain the basic motion recognition results;

[0066] Step 230: Based on the basic action recognition results and the main pump pressure signal, the working condition stage of the engineering machinery is identified using a preset working condition stage recognition model to obtain the working condition stage recognition results.

[0067] The above technical solution first acquires the basic data and main pump pressure signal of the construction machinery. Then, based on the basic data and main pump pressure signal, it sequentially performs basic action recognition and working condition stage recognition using a basic action recognition model and a working condition stage recognition model, thereby obtaining the working condition stage recognition result. By using the basic data and main pump pressure signal, the working condition recognition of the construction machinery is completed in a multi-step manner, from basic action to working condition stage, thus realizing a multi-level working condition recognition method. Compared with single-step working condition recognition methods, this reduces the complexity of the working condition recognition model and the computational requirements. The multi-level working condition recognition method can improve the accuracy and stability of the working condition recognition results. Simultaneously, the basic action recognition model and working condition stage recognition model can cover more basic actions and working condition stages, broadening the coverage of the recognition model and making it applicable to various scenarios, thus improving the environmental adaptability of the working condition recognition method.

[0068] In the above implementation process, the basic motion recognition model and the working condition stage recognition model are used to identify the basic motion recognition results and the working condition stage recognition results respectively. The output of the basic motion recognition model and the main pump pressure signal are used as the input of the working condition stage recognition model. This realizes a multi-level working condition recognition method, reduces the amount of calculation, and is conducive to obtaining the working condition recognition results quickly.

[0069] In this embodiment, the basic data of the construction machinery can be any one of the following: actuator pilot control signal, actuator displacement, actuator tilt angle, and hydraulic pump flow rate. The actuator pilot control signal is relatively easy to obtain, as it can be obtained from the central control unit on the construction machinery. The main pump pressure signal can be acquired by a pressure sensor on the construction machinery. For ease of explanation, this embodiment mainly uses the actuator pilot control signal as the basic data. The actuator pilot control signal can be a current or voltage signal output from the gantry controller to the multi-way valve, a multi-way valve pilot pressure signal, etc.

[0070] It should be noted that, in this embodiment, the basic data and main pump pressure signal can be the basic data and main pump pressure signal collected at the current moment, or the basic data and main pump pressure signal within a certain time range.

[0071] Accordingly, step 220 can use the basic data to identify the basic movements of the engineering machinery using a preset basic movement recognition model to obtain the basic movement recognition result.

[0072] Taking the lifting operation of engineering cranes as an example, the complete lifting operation process can be divided into five stages based on the phased tasks of transferring the suspended load in space: lifting preparation, lifting, unloading, empty return, and idle. The lifting preparation stage mainly involves adjusting the crane's posture to load the load; the lifting stage mainly involves the coordinated actions of the crane's actuators to transfer the suspended load in space; the unloading stage involves unloading the load; the empty return stage is the process of the crane's hook moving to the next work point; and the idle stage is when the actuators stop moving to assist workers in loading, unloading, maintaining the stability of the crane and the suspended load, and determining the work position. Typically, engineering cranes complete the operation in the sequence of lifting preparation—lifting—unloading—empty return, with occasional idle periods. It should be noted that the stage can also be determined based on other operational situations of the crane, which will not be elaborated upon here.

[0073] Accordingly, taking the lifting operation of engineering cranes as an example, the basic actuator actions corresponding to the five working conditions are shown in Table 1. Table 1 is a table showing the correspondence between the working conditions of engineering cranes and the actions of the actuators. In the table, 1 represents that the actuator has action, and 0 represents that the actuator has no action. As can be seen from the table, except for the idling condition, the operation and sequence of the actuators are determined by the working environment and the operator's habits. For example, in the lifting and transporting condition, the engineering crane can perform one or more of the following actions: single winch action, single luffing action, single slewing action, winch + luffing combined action, winch + slewing combined action, luffing + slewing combined action, and winch + luffing + slewing combined action. It should be noted that after determining the working conditions based on other operating situations, different basic actions can also be determined, which can be determined according to the actual situation, and will not be elaborated here.

[0074] Table 1 Correspondence between working conditions and actuator actions of engineering lifting machinery

[0075] idler 0 0 0 / hoisting preparation 1 or 0 1 or 0 1 or 0 small area hoisting 1 or 0 1 or 0 1 or 0 large area Unloading 1 or 0 1 or 0 1 or 0 small area Return without load 1 or 0 1 or 0 1 or 0 large area

[0076] Prior to step 220, a basic motion recognition model can be pre-built and trained using a neural network. This basic motion recognition model can be pre-installed in the engineering machinery. In some embodiments, the construction process of the pre-installed basic motion recognition model includes the following steps:

[0077] First, the first sample data is obtained, which includes the first sample basic data under each basic action, and the basic action label corresponding to the first sample basic data.

[0078] In this embodiment, the first sample data includes multiple sets of data, each set including first sample basic data and corresponding basic action labels. Taking the lifting operation completed by engineering cranes as an example, the first sample basic data can be the actuator pilot control signal, or the actuator pilot control signal waveforms corresponding to various basic actions of the engineering cranes can be used as the basic action classification markers, with different basic actions corresponding to different actuator pilot control signal waveforms. Among them, the basic actions include a total of 8 types: single winch action, single luffing action, single slewing action, winch + luffing combined action, winch + slewing combined action, luffing + slewing combined action, winch + luffing + slewing combined action, and idle.

[0079] Then, the first sample basic data under each basic action is input into the first neural network to obtain the predicted basic action;

[0080] Finally, the parameters of the first neural network are adjusted according to the predicted basic action and the basic action label corresponding to the basic data of the first sample in the first sample data to obtain the basic action recognition model.

[0081] In this embodiment, the first neural network can be a linear neural network, a feedback neural network, a multilayer feedforward neural network (BP neural network), etc., among which the BP neural network has strong nonlinear mapping ability and flexible network structure. The number of intermediate layers and the number of neurons in each layer can be arbitrarily set according to specific circumstances, and it has strong generalization ability and fault tolerance. Using the BP neural network can obtain a more stable and reliable basic action recognition model. Taking the lifting operation of engineering cranes as an example, a basic action recognition model for engineering cranes can be established based on the BP neural network. The segmentation labels of all basic actions are used as model inputs. The basic actions are predicted by the model. The predicted basic actions and their corresponding labels are input into a preset loss function to obtain the corresponding loss value. The model parameters are adjusted according to the loss value so that the predicted basic actions are the same as the basic action labels corresponding to the basic data of the first sample, so that the model has a sufficient recognition accuracy. Finally, the basic action recognition model is trained.

[0082] Accordingly, prior to step 230, the operating condition stage identification model can also be pre-built, and the construction process of the operating condition stage identification model includes:

[0083] First, acquire the second sample data, which includes the basic actions and main pump pressure signals under each operating condition stage, as well as the operating condition stage labels corresponding to the basic actions.

[0084] In this embodiment, the second sample data includes multiple sets of data. Each set of data includes a basic action and main pump pressure signal under a certain working condition stage, as well as the corresponding working condition stage label. Taking the completion of hoisting operations by engineering cranes as an example, the basic action and main pump pressure signal waveform characteristics corresponding to each working condition stage of the engineering cranes can be used as the dividing criteria for working condition stages. Among them, the working condition stages include hoisting preparation, hoisting, unloading, no-load return, and idle, totaling 5 types.

[0085] Then, the basic actions and main pump pressure signals under each working condition stage are input into the second neural network to obtain the predicted working condition stage.

[0086] Finally, the parameters of the second neural network are adjusted according to the predicted operating condition stage and the labels of each operating condition stage in the second sample data to obtain the operating condition stage recognition model.

[0087] In this embodiment, the second neural network can employ deep learning networks such as recurrent neural networks (RNNs) and long short-term memory (LSTM) networks. LSTM effectively captures semantic relationships between long sequences, mitigating the vanishing or exploding gradient phenomenon. Its main feature lies in its gated control, including forget gates, input gates, cell state gates, and output gates. This acts as a "processor" that determines the usefulness of information. LSTM can better handle time-series tasks, solves the long-term dependency problem of traditional RNNs, and alleviates the vanishing gradient problem caused by backpropagation during training, thus making the resulting working condition stage recognition model more stable and reliable.

[0088] Taking the lifting operation of engineering cranes as an example, a working condition stage recognition model can be established based on an LSTM deep learning network. The LSTM model mainly includes an input layer, an LSTM layer, a fully connected layer, and an output layer. The waveform features of all working condition stages can be used as the model input. The working condition stages are predicted by the model. The predicted working condition stages and their corresponding labels are input into a preset loss function to obtain the corresponding loss values. The model parameters are then adjusted based on the loss values ​​to ensure that the predicted working condition stages are the same as the working condition stage labels corresponding to the basic actions, so that the model has sufficient recognition accuracy. Finally, the working condition stage recognition model is trained.

[0089] In the above implementation process, the use of neural network training to obtain the working condition stage recognition model and the basic action recognition model helps to improve the reliability of the working condition recognition results. Using a BP neural network model to train the basic action recognition model reduces the cost of working condition recognition and makes it easier to integrate into construction machinery; using an LSTM neural network model to train the working condition stage recognition model results in a more stable and reliable model, thereby improving the accuracy and reliability of working condition recognition. Using the actuator pilot control signal and main pump pressure as data sources, employing BP neural network models and LSTM deep learning models, and providing system sampling to reduce data sampling frequency, these multiple measures ensure low-cost and low-computing-power requirements for working condition recognition in construction machinery, making it easier to integrate into construction machinery.

[0090] In some embodiments, since the actuator pilot control signal is relatively easy to obtain in construction machinery, taking the lifting operation of construction crane as an example, in order to facilitate working condition identification, the data type of the basic data can be the actuator pilot control signal. Accordingly, the actuator pilot control signal can be a real-time actuator pilot control signal, and the basic actions performed by the construction crane can be identified in real time based on the actuator pilot control signal; the actuator pilot control signal can also be multiple actuator pilot control signal data within a certain time range, and then the basic actions currently performed by the construction crane can be identified based on the actuator pilot control signal.

[0091] In one embodiment, the basic actions performed by the construction machinery can be identified in real time based on the pilot control signal of the actuator. The step of identifying the basic actions of the construction machinery using a preset basic action recognition model based on the basic data to obtain the basic action recognition result includes the following steps:

[0092] Step A1: Extract features from the pilot control signal of the actuator to obtain a feature vector;

[0093] In this embodiment, mean filtering is first used to reduce noise and transient interference in the actuator pilot control signal; then, based on the computing power of the host controller, system sampling is used to reduce the sampling frequency of the actuator pilot control signal; then, the time-domain feature values ​​of the reduced-frequency pilot control signal data are extracted to form a feature vector.

[0094] Taking the lifting operation of engineering cranes as an example, the pilot control signal data of the luffing cylinder, winch motor, and slewing motor of the engineering cranes within a certain time range at the current moment is collected through the CAN bus. After extraction and frequency reduction, the time-domain feature values ​​obtained include: the average value of the amplitude increase pilot signal, the average value of the amplitude decrease pilot signal, the average value of the winch rise pilot signal, the average value of the winch fall pilot signal, the average value of the clockwise slewing pilot signal, and the average value of the counterclockwise slewing pilot signal. The feature vector formed can be expressed as: X = [x1, x2, x3, x4, x5, x6], where X is the feature vector constructed from the time-domain feature values; x1 is the average value of the amplitude increase pilot signal, x2 is the average value of the amplitude decrease pilot signal; x3 is the average value of the winch rise pilot signal, x4 is the average value of the winch fall pilot signal; x5 is the average value of the clockwise slewing pilot signal, and x6 is the average value of the counterclockwise slewing pilot signal.

[0095] Step A2: Normalize the feature vector to obtain the normalized feature vector;

[0096] In this embodiment, the normalization process described above can be achieved by substituting the time-domain eigenvalues ​​in the eigenvector into the normalization formula to calculate the normalized eigenvalues, thereby obtaining the normalized eigenvector. The normalization formula can be:

[0097] Where, x new x is the normalized time-domain eigenvalue; x is the unnormalized time-domain eigenvalue; x max x represents the maximum value of the time-domain feature value corresponding to the category in the feature vector; min This represents the minimum value of the time-domain eigenvalue corresponding to the category x in the eigenvector. In practical applications, other normalization methods such as nonlinear normalization can also be used.

[0098] Step A3: Input the normalized feature vector into the preset basic action recognition model to perform basic action recognition and obtain the basic action recognition result;

[0099] Taking the lifting operation of engineering cranes as an example, the normalized feature vector is input into the basic motion recognition model. The basic motion performed by the engineering crane is determined based on the sequence number with the highest output probability value in the model output. For example, when the sequence number with the highest probability value is 1-8, the basic motions of the engineering crane at the current moment are respectively: idle, single winch motion, single luffing motion, single slewing motion, winch + luffing combined motion, winch + slewing combined motion, luffing + slewing combined motion, and winch + luffing + slewing combined motion. It should be noted that the above sequence numbers can be set in advance when training the basic motion recognition model, and the sequence number with the highest output probability value can be set as the output result.

[0100] In one embodiment, the obtained basic action recognition results may be stored in hardware devices such as processors and caches.

[0101] In one embodiment, the obtained basic motion recognition results can also be output from the aforementioned hardware device to an output device in real time. The output device can be a memory, display device, terminal, communication module, etc., so that the user can obtain the basic motion recognition results in a timely manner. For example, the basic motion recognition results stored in the processor can be output to the display device.

[0102] To further improve the accuracy of basic motion recognition, the basic motion recognition process can be performed cyclically until an end signal is received, at which point the basic motion recognition result is output. The actuator pilot control signal includes multiple actuator pilot control signal data within a first preset time range from the current time. Specifically, after completing step A3, step A4 can be executed: determining whether the basic motion recognition process has ended. In this embodiment, determining whether the basic motion recognition process has ended can be done by checking for an end signal. This end signal can be a switch that is activated by the operator according to the actual situation to obtain the end signal. Alternatively, it can be determined whether the basic motion recognition process has ended by checking for a power outage.

[0103] Step A5: After the basic action recognition process is completed, output the basic action recognition result;

[0104] In this real-time example, the basic motion recognition results can be output from the aforementioned hardware devices to an output device, which can be a memory, display device, terminal, communication module, etc., so that users can obtain the basic motion recognition results in a timely manner. For example, the basic motion recognition results stored in the processor can be output to the display device.

[0105] Step A6: If the basic action recognition process is not completed, remove the actuator pilot control signal data that is furthest from the current time in the actuator pilot control signal, and obtain the actuator pilot control signal data of the next time to update the actuator pilot control signal, so as to obtain a new actuator pilot control signal and return to execute A1.

[0106] In this embodiment, since the actuator pilot control signal includes actuator pilot control signal data within a certain time range, when the basic action recognition process is not finished, as time changes, the actuator pilot control signal data furthest from the current time can be removed, and then the actuator pilot control signal data of the next time moment can be added to form a new actuator pilot control signal. Then, steps A1-A4 are repeated until the recognition is finished, thereby ensuring the accuracy and real-time performance of the basic action recognition result.

[0107] Accordingly, in order to make the obtained operating condition stage identification results more accurate, the basic action identification results and the main pump pressure signal within a certain time range can be used as data sources to identify the operating condition stage.

[0108] In one embodiment, the basic motion recognition result and the main pump pressure signal are real-time basic motion recognition results and main pump pressure signals. The step of identifying the operating stage of the construction machinery using a preset operating stage recognition model based on the basic motion recognition result and the main pump pressure signal to obtain the operating stage recognition result includes the following steps:

[0109] Step B1: Obtain multiple basic action recognition results and the main pump pressure signal corresponding to each basic action recognition result within a second preset time range from the current time, and combine the multiple basic action recognition results and the main pump pressure signal corresponding to each basic action recognition result into a working condition stage recognition feature vector.

[0110] In this embodiment, the output results of the basic motion recognition model of the construction machinery are extracted within a certain time range from the current time. Simultaneously, the main pump pressure signal data within the same time range from the current time is collected via the CAN bus to obtain multiple basic motion recognition results and the corresponding main pump pressure signals within a second preset time range from the current time. Then, the multiple basic motion recognition results and the corresponding main pump pressure signals are preprocessed to obtain the operating condition stage recognition feature vector. The preprocessing includes first using mean filtering to reduce noise and transient interference in the main pump pressure signal; then, based on the computing power of the host controller, using system sampling to reduce the sampling frequency of the basic motion recognition results and the main pump pressure signal; and finally, combining the frequency-reduced basic motion recognition results and the main pump pressure signal to form the operating condition stage recognition feature vector.

[0111] For example, taking the lifting operation of engineering cranes as an example, the basic motion recognition results within 0.5 seconds of the current time are extracted from the basic motion recognition model. At the same time, the main pump pressure signal data within 0.5 seconds of the current time are collected. After preprocessing, these basic motion recognition results and the main pump pressure signal are combined into a vector to obtain the working condition stage recognition feature vector.

[0112] Step B2: Normalize the operating condition stage identification feature vector to obtain the normalized operating condition stage identification feature vector.

[0113] In this embodiment, the normalization process is the same as that in step A2, and will not be described again here. The obtained normalized working condition stage identification feature vector can be expressed as:

[0114] Where, x 1,n The normalized main pump pressure signal; x 2,n This represents the normalized basic action recognition result; n is the amount of data.

[0115] Step B3: Input the normalized working condition stage identification feature vector into the preset working condition stage identification model to identify the working condition stage and obtain the working condition stage identification result.

[0116] Taking the lifting operation of engineering cranes as an example, the normalized working condition stage identification feature vector is input into the working condition stage identification model. The working condition stage of the engineering crane is determined based on the sequence number with the highest output probability value in the model output results. When the sequence number with the highest probability value is 1-5, the working condition stage of the engineering crane at the current moment is respectively idle, lifting preparation, lifting, unloading, and empty return. It should be noted that the above sequence number can be set in advance when training the working condition stage identification model, and the sequence number with the highest output probability value can be set as the output result.

[0117] In one embodiment, the obtained operating condition stage identification result may be stored in hardware devices such as processors and caches.

[0118] In one embodiment, the obtained operating condition stage identification results can also be output from the aforementioned hardware device to an output device in real time. The output device can be a memory, display device, terminal, communication module, etc., so that users can obtain the operating condition stage identification results in a timely manner. For example, the operating condition stage identification results stored in the processor can be output to the display device.

[0119] To further improve the accuracy of operating condition stage identification, the identification process can be repeated until a termination signal is received, at which point the identification result is output. Specifically, after step B3, step B4 can be executed: determining whether the operating condition stage identification process has ended. In this embodiment, determining whether the basic action identification process has ended can be done by checking for a termination signal. This termination signal can be a switch that is activated by the operator based on the actual situation. Alternatively, it can be determined whether the power has been cut off to determine whether the basic action identification process has ended.

[0120] Step B5: After the working condition stage identification process is completed, output the working condition stage identification result;

[0121] In this real-time example, the obtained operating condition stage identification results can be output from the aforementioned hardware devices to an output device, which can be a memory, display device, terminal, communication module, etc., so that users can obtain the operating condition stage identification results in a timely manner. For example, the operating condition stage identification results stored in the processor can be output to the display device.

[0122] Step B6: If the working condition stage identification process is not completed, remove the basic action identification result and the corresponding main pump pressure signal that are farthest from the current time in the working condition stage identification feature vector, and obtain the basic action identification result and the corresponding main pump pressure signal at the next time to update the working condition stage identification feature vector, obtain a new working condition stage identification feature vector, and return to execute B2.

[0123] In this embodiment, since the basic action recognition result and the main pump pressure signal include the basic action recognition result and the main pump pressure signal identified within a certain time range, when the working condition stage recognition process has not ended, as time changes, the basic action recognition result and the main pump pressure signal furthest from the current time can be removed, and then the basic action recognition result and the main pump pressure signal of the next time moment can be added to form a new working condition stage recognition feature vector. Then, steps B2-B4 are repeated until the recognition ends, thereby ensuring the accuracy and real-time performance of the working condition stage recognition result.

[0124] In some embodiments, taking the lifting operation of engineering cranes as an example, considering that the load and basic actions are basically the same in other working conditions besides the lifting phase, errors can easily occur in the identification of working conditions. Therefore, in order to improve the accuracy of the working condition identification results, the identification results can be further checked and corrected to obtain more accurate working condition identification results. Specifically, after executing B3, the following steps may be included:

[0125] First, obtain the historical operating condition stage identification results; in this embodiment, the historical operating condition stage identification results can be the operating condition stage identification results of any one or more previous times.

[0126] Then, based on the historical operating condition stage identification results and the preset intelligent verification rule library, the operating condition stage identification results are verified and corrected to obtain the corrected operating condition stage identification results.

[0127] In this embodiment, the preset intelligent verification rule base is an intelligent verification rule base built on expert knowledge. It contains verification rules for various situations. For example, if the identification result of the historical working condition stage is G1 and the identification result of the working condition stage is G2, the identification result of the working condition stage will be corrected to G1.

[0128] In the above implementation process, by combining the historical operating condition stage identification results and the pre-set intelligent verification rule library, the operating condition stage identification results can be further verified and corrected, making the obtained operating condition stage identification results more accurate.

[0129] In some embodiments, to further improve the accuracy of the verification and correction, the current working condition stage identification result can be verified and corrected based on information such as the current and previous identification results, so that it meets the actual operation specifications, action sequence, system safety protection measures, etc. Specifically, the historical working condition stage identification results include the working condition stage identification result of the previous moment, as well as multiple working condition stage identification results within a third preset time range from the previous moment;

[0130] The process of verifying and correcting the operating condition stage identification results based on the historical operating condition stage identification results and a pre-set intelligent verification rule base to obtain corrected operating condition stage identification results includes the following steps:

[0131] First, based on the identification results of multiple working condition stages within the third preset time range from the previous moment, the actual working condition stage at the current moment is determined;

[0132] Then, based on the current actual working condition stage and the working condition stage identification result of the previous time, the corresponding working condition stage identification result correction rule is matched in the preset intelligent verification rule library.

[0133] Finally, based on the working condition stage identification result correction rules, the working condition stage identification result is checked and corrected to obtain the corrected working condition stage identification result.

[0134] In this embodiment, an intelligent verification rule base can be established based on expert knowledge. Based on information such as the current and previous identification results, the identification results of the current operating condition stage are verified and corrected to meet actual operating specifications, action sequences, and system safety protection measures. In specific implementation, the verification rules in the intelligent verification rule base can be set in the form of IF-THEN to formulate a series of rules related to the correction of the operating condition stage identification results. For example, IF represents the conditions that the corrected identification results must meet. These conditions include three types: the previous operating condition stage identification result (PS), the current operating condition stage identification result (CS) before correction, and the current actual operating condition stage (RCS) determined based on historical identification results. The conditions can be one or more of these three types. For example, the previous operating condition stage identification result PS may inevitably produce momentary misidentifications due to signal fluctuations, waveform similarities, and model errors. Therefore, the current actual operating condition stage RCS can be determined based on historical operating condition stage identification results. Then, the current operating condition stage identification result CS is corrected based on the previous operating condition stage identification result PS and the current actual operating condition stage RCS to improve the correction effect of momentary operating condition stage errors on the original current operating condition stage identification result CS. THEN represents the corrected operating condition stage identification result, i.e., the corrected current operating condition stage identification result CS. This corrected operating condition stage identification result can be pre-set based on expert experience. For example, if the previous working condition stage identification result was "lifting preparation", and the current actual working condition stage is determined to be "lifting" based on the historical working condition stage identification results, and the current working condition stage identification result is "unloading", then we can match the rule "IF" in the intelligent verification rule base, which means that the previous working condition stage identification result was "lifting preparation" and the current actual working condition stage is "lifting". If we get "THEN" in the rule, then we can correct the current working condition stage identification result to "lifting preparation".

[0135] It is understandable that if the working condition stage identification result in the corresponding working condition stage identification result correction rule is the same as the identified working condition stage identification result, then the identification is correct and no correction is required.

[0136] In the above implementation process, after the identification results of the working condition stage are obtained, they are further combined with the historical working condition stage identification results and the preset intelligent verification rule library for verification and correction. This enables the verification of the working condition stage identification results, ensuring the reliability of the working condition stage identification results. At the same time, it can also correct the working condition stage identification results that may have identification errors, thereby further improving the accuracy of the working condition stage identification results.

[0137] The above technical solution first acquires the basic data and main pump pressure signal of the construction machinery. Then, based on the basic data and main pump pressure signal, it sequentially performs basic action recognition and working condition stage recognition using a basic action recognition model and a working condition stage recognition model, thereby obtaining the working condition stage recognition result. By using the basic data and main pump pressure signal, the working condition recognition of the construction machinery is completed in a multi-step manner, from basic action to working condition stage, thus realizing a multi-level working condition recognition method. Compared with single-step working condition recognition methods, this reduces the complexity of the working condition recognition model and the computational requirements. The multi-level working condition recognition method can improve the accuracy and stability of the working condition recognition results. Simultaneously, the basic action recognition model and working condition stage recognition model can cover more basic actions and working condition stages, broadening the coverage of the recognition model and making it applicable to various scenarios, thus improving the environmental adaptability of the working condition recognition method. By employing a basic motion recognition model and a working condition stage recognition model to identify the basic motion and working condition stage recognition results respectively, and using the output of the basic motion recognition model along with the main pump pressure signal as the input of the working condition stage recognition model, a multi-level working condition recognition method is achieved. This reduces the computational load and facilitates the rapid acquisition of working condition recognition results. The inputs to this working condition recognition method are all signal sources already configured in the host machine. The selection of neural networks and signal frequency reduction techniques result in low cost and low computational requirements, making it more suitable for the configuration of engineering machinery host machines.

[0138] In some embodiments, after obtaining the operating condition stage identification result, the following steps are further included:

[0139] First, based on the basic motion recognition results and the working condition stage recognition results, the hoisting operation speed information is calculated;

[0140] In this embodiment, after obtaining the working condition stage identification results, the hoisting operation speed can be statistically analyzed based on the basic action identification results and the working condition stage identification results. The hoisting operation speed statistics include the number of hoisting operations, the average time per hoisting operation, the average duration of each working condition stage in a single hoisting operation, and the average duration of single / compound actions (hoisting, luffing, and slewing) during a single hoisting operation. The number of hoisting operations is calculated as follows: if the previous working condition stage was an empty return, and the current time is not an empty return, then the number of hoisting operations N increases by 1. The other indicators can be calculated using the following formula:

[0141] T = ∑T i ;

[0142]

[0143]

[0144] Where T is the average time per hoisting operation; T i n is the average duration of each working stage in a single lifting operation;i Δt represents the number of times the identification result appears in each working condition stage; Δt represents the time interval for identifying working condition stages; TT represents the number of times the identification result appears in each working condition stage. j The average duration of single / compound actions of hoisting, luffing, and slewing in a single hoisting operation; nj is the number of times the identification results of single / compound actions of hoisting, luffing, and slewing occur; among all the above parameters, i = 1, 2, ..., 5 represent the idle, hoisting preparation, hoisting, unloading, and no-load return working conditions, respectively, and j = 1, 2, 3 represent single / compound actions including hoisting, single / compound actions including luffing, and single / compound actions including slewing, respectively.

[0145] By analyzing the speed of hoisting operations, the project manager can gain insights into the operational status and develop appropriate construction strategies. For example, if the preparation phase of a single hoisting operation is prolonged, it may indicate an inadequate personnel allocation. Similarly, if the duration of the slewing single / compound motion is significantly longer than that of the luffing and winch movements, it may suggest an improper placement of the load. In such cases, the project manager will need to develop a corresponding construction strategy.

[0146] Then, the lifting weight statistics are obtained; in this embodiment, the lifting weight statistics include the total lifting weight and the average single lifting weight. The aforementioned lifting weight can be obtained using existing technologies such as lifting weight sensing based on outrigger reaction force, lifting weight sensing based on tension sensors, and lifting weight calculation based on torque limiters.

[0147] Finally, based on the hoisting operation speed information and the hoisting weight statistics, the engineering machinery operation efficiency information is obtained.

[0148] In this embodiment, since the efficiency statistics of construction machinery operation include two aspects, hoisting operation speed and hoisting weight, the operation information of the construction machinery hoisting process is divided into two levels: basic actions of the execution structure and working condition stage, according to the completed phased tasks. Based on the response characteristics of hydraulic system parameters, machine learning methods are used to identify the basic actions and working condition stages in sequence to complete the statistics of operation speed information. This allows for more comprehensive information statistics, which is beneficial for project managers to make reasonable project decisions.

[0149] Figure 1 This is a flowchart illustrating the method for identifying the working conditions of construction machinery in this embodiment. It should be understood that, although... Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0150] In one embodiment, such as Figure 2 As shown, Figure 2 The schematic diagram illustrates a structural block diagram of a construction machinery operating condition identification device according to an embodiment of this application. The device includes an acquisition module 410, an action recognition module 420, and an operating condition identification module 430, wherein:

[0151] The acquisition module 410 is used to acquire basic data and main pump pressure signal of the construction machinery; wherein, the data type of the basic data includes at least one of the following: actuator pilot control signal, actuator displacement, actuator tilt angle, and hydraulic pump flow rate;

[0152] The motion recognition module 420 is used to recognize the basic motions of the engineering machinery based on the basic data and a preset basic motion recognition model to obtain the basic motion recognition result.

[0153] The working condition identification module 430 is used to identify the working condition stage of the construction machinery based on the basic action identification result and the main pump pressure signal, using a preset working condition stage identification model, and obtain the working condition stage identification result.

[0154] The engineering machinery working condition identification device includes a processor and a memory. The acquisition module 410, action identification module 420 and working condition identification module 430 are all stored in the memory as program units. The processor executes the program modules stored in the memory to realize the corresponding functions.

[0155] The processor contains a kernel, which retrieves the corresponding program units from memory. One or more kernels can be configured, and multi-level operating condition identification can be achieved by adjusting kernel parameters.

[0156] The memory may include non-permanent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0157] This application provides a storage medium on which a program is stored. When the program is executed by a processor, it implements the above-described method for identifying the working conditions of engineering machinery.

[0158] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 3 As shown in the figure, the computer device includes a processor A01, a network interface A02, a display screen A04, an input device A05, and a memory (not shown) connected via a system bus. The processor A01 provides computing and control capabilities. The memory includes internal memory A03 and a non-volatile storage medium A06. The non-volatile storage medium A06 stores an operating system B01 and a computer program B02. The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 stored in the non-volatile storage medium A06. The network interface A02 is used for communication with external terminals via a network connection. When the computer program is executed by the processor A01, it implements a method for identifying the working conditions of engineering machinery. The display screen A04 can be a liquid crystal display (LCD) or an e-ink display. The input device A05 can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0159] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0160] In one embodiment, the engineering machinery operating condition identification device provided in this application can be implemented as a computer program, and the computer program can be implemented as follows: Figure 3 The device runs on the computer shown. The computer's memory can store the various program modules that make up the engineering machinery operating condition identification device, for example, Figure 2 The acquisition module 410, action recognition module 420, and working condition recognition module 430 are shown. The computer program composed of these modules causes the processor to execute the steps in the engineering machinery working condition recognition methods of the various embodiments of this application described in this specification.

[0161] Figure 3 The computer device shown can be used as follows Figure 2 The acquisition module 410 in the engineering machinery condition identification device shown executes step 210. The computer equipment can execute step 220 through the action identification module 420 and step 230 through the condition identification module 430.

[0162] This application provides a device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it performs the following steps:

[0163] Acquire basic data and main pump pressure signals from the construction machinery;

[0164] Based on the aforementioned basic data, a pre-set basic motion recognition model is used to identify the basic motions of the engineering machinery, thereby obtaining the basic motion recognition results.

[0165] Based on the basic action recognition results and the main pump pressure signal, a preset working condition stage recognition model is used to identify the working condition stage of the engineering machinery, and the working condition stage recognition results are obtained.

[0166] The data types of the basic data include at least one of the following: actuator pilot control signal, actuator displacement, actuator tilt angle, and hydraulic pump flow rate.

[0167] In one embodiment, the construction process of the basic action recognition model includes:

[0168] Obtain first sample data, which includes first sample basic data under each basic action and basic action labels corresponding to the first sample basic data;

[0169] The first sample basic data for each basic action is input into the first neural network to obtain the predicted basic action.

[0170] The parameters of the first neural network are adjusted according to the predicted basic action and the basic action label corresponding to the basic data of the first sample in the first sample data to obtain the basic action recognition model.

[0171] In one embodiment, the construction process of the working condition stage identification model includes:

[0172] Acquire second sample data, which includes basic actions and main pump pressure signals under each operating condition stage, as well as operating condition stage labels corresponding to the basic actions;

[0173] The basic actions and main pump pressure signals under each working condition stage are respectively input into the second neural network to obtain the predicted working condition stage;

[0174] The parameters of the second neural network are adjusted based on the predicted operating condition stage and the labels of each operating condition stage in the second sample data to obtain the operating condition stage recognition model.

[0175] In one embodiment, the data type of the basic data is an actuator pilot control signal, and the actuator pilot control signal includes multiple actuator pilot control signal data within a first preset time range from the current time;

[0176] The step of identifying the basic movements of the construction machinery using a pre-set basic movement recognition model based on the basic data, and obtaining the basic movement recognition results, includes:

[0177] A1: Perform feature extraction on the pilot control signal of the actuator to obtain a feature vector;

[0178] A2: Normalize the feature vector to obtain the normalized feature vector;

[0179] A3: Input the normalized feature vector into the preset basic action recognition model to perform basic action recognition and obtain the basic action recognition result;

[0180] A4: Determine whether the basic motion recognition process has ended;

[0181] A5: Output the basic action recognition result after the basic action recognition process is completed;

[0182] A6: If the basic action recognition process is not completed, remove the actuator pilot control signal data that is furthest from the current time in the actuator pilot control signal, and obtain the actuator pilot control signal data of the next time to update the actuator pilot control signal, so as to obtain a new actuator pilot control signal and return to execute A1.

[0183] In one embodiment, the step of identifying the operating stage of the construction machinery using a preset operating stage identification model based on the basic motion identification result and the main pump pressure signal, to obtain the operating stage identification result, includes:

[0184] B1: Acquire multiple basic action recognition results and the main pump pressure signal corresponding to each basic action recognition result within a second preset time range from the current time, and combine the multiple basic action recognition results and the main pump pressure signal corresponding to each basic action recognition result into a working condition stage recognition feature vector;

[0185] B2: Normalize the operating condition stage identification feature vector to obtain the normalized operating condition stage identification feature vector.

[0186] B3: Input the normalized working condition stage identification feature vector into the preset working condition stage identification model to identify the working condition stage and obtain the working condition stage identification result.

[0187] B4: Determine whether the working condition stage identification process has ended;

[0188] B5: Output the working condition stage identification result after the working condition stage identification process is completed;

[0189] B6: If the working condition stage identification process is not completed, remove the basic action identification result and the corresponding main pump pressure signal that are farthest from the current time in the working condition stage identification feature vector, and obtain the basic action identification result and the corresponding main pump pressure signal of the next time moment to update the working condition stage identification feature vector, obtain a new working condition stage identification feature vector, and return to execute B2.

[0190] In one embodiment, after executing B3, the following is also included:

[0191] Obtain historical operating condition stage identification results;

[0192] Based on the historical operating condition stage identification results and the preset intelligent verification rule library, the operating condition stage identification results are verified and corrected to obtain the corrected operating condition stage identification results.

[0193] In one embodiment, the historical working condition stage identification result includes the working condition stage identification result of the previous moment, and multiple working condition stage identification results within a third preset time range from the previous moment.

[0194] The step of verifying and correcting the operating condition stage identification results based on the historical operating condition stage identification results and a pre-set intelligent verification rule base to obtain corrected operating condition stage identification results includes:

[0195] Based on the identification results of multiple working condition stages within the third preset time range from the previous moment, the actual working condition stage at the current moment is determined;

[0196] Based on the current actual working condition stage and the working condition stage identification result of the previous time, the corresponding working condition stage identification result correction rule is matched in the preset intelligent verification rule library.

[0197] Based on the correction rules for the identified operating conditions, the identified operating conditions are verified and corrected to obtain the corrected identified operating conditions.

[0198] In one embodiment, after obtaining the operating condition stage identification result, the method further includes:

[0199] Based on the basic motion recognition results and the working condition stage recognition results, the hoisting operation speed information is calculated;

[0200] Obtain lifting weight statistics;

[0201] Based on the hoisting operation speed information and the hoisting weight statistics, the operation efficiency information of the construction machinery is obtained.

[0202] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0203] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations 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, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0204] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0205] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0206] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0207] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0208] Computer-readable media include both permanent and non-permanent, removable and non-removable media, which can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, 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 technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0209] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0210] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method of work condition recognition of a working machine, characterized in that, include: Acquire basic data and main pump pressure signals from the construction machinery; Based on the aforementioned basic data, a pre-set basic motion recognition model is used to identify the basic motions of the engineering machinery, thereby obtaining the basic motion recognition results. Based on the basic motion recognition results and the main pump pressure signal, a preset working condition stage recognition model is used to identify the working condition stage of the construction machinery to obtain the working condition stage recognition result; the basic motion recognition result output by the basic motion recognition model and the main pump pressure signal are used as the input of the working condition stage recognition model. The basic data includes at least one of the following: actuator pilot control signal, actuator displacement, actuator tilt angle, and hydraulic pump flow rate; the basic action refers to the actuator action.

2. The working condition recognition method of a working machine according to claim 1, characterized by, The construction process of the basic action recognition model includes: Obtain first sample data, which includes first sample basic data under each basic action and basic action labels corresponding to the first sample basic data; The first sample basic data for each basic action is input into the first neural network to obtain the predicted basic action. The parameters of the first neural network are adjusted according to the predicted basic action and the basic action label corresponding to the basic data of the first sample in the first sample data to obtain the basic action recognition model.

3. The working condition recognition method of a working machine according to claim 1, characterized by, The construction process of the working condition stage identification model includes: Acquire second sample data, which includes basic actions and main pump pressure signals under each operating condition stage, as well as operating condition stage labels corresponding to the basic actions; The basic actions and main pump pressure signals under each working condition stage are respectively input into the second neural network to obtain the predicted working condition stage; The parameters of the second neural network are adjusted based on the predicted operating condition stage and the labels of each operating condition stage in the second sample data to obtain the operating condition stage recognition model.

4. The working-conditions identification method of a working machine according to claim 1, characterized by, The data type of the basic data is the actuator pilot control signal, which includes multiple actuator pilot control signal data within a first preset time range from the current time. The step of identifying the basic movements of the construction machinery using a pre-set basic movement recognition model based on the basic data, and obtaining the basic movement recognition results, includes: A1: Perform feature extraction on the pilot control signal of the actuator to obtain a feature vector; A2: Normalize the feature vector to obtain the normalized feature vector; A3: Input the normalized feature vector into the preset basic action recognition model to perform basic action recognition and obtain the basic action recognition result; A4: Determine whether the basic motion recognition process has ended; A5: Output the basic action recognition result after the basic action recognition process is completed; A6: If the basic action recognition process is not completed, remove the actuator pilot control signal data that is furthest from the current time in the actuator pilot control signal, and obtain the actuator pilot control signal data of the next time to update the actuator pilot control signal, so as to obtain a new actuator pilot control signal and return to execute A1.

5. The working-conditions identification method of a working machine according to claim 1, characterized by, The step of identifying the operating stages of the construction machinery based on the basic action recognition results and the main pump pressure signal using a preset operating stage recognition model, and obtaining operating stage recognition results, includes: B1: Acquire multiple basic action recognition results and the main pump pressure signal corresponding to each basic action recognition result within a second preset time range from the current time, and combine the multiple basic action recognition results and the main pump pressure signal corresponding to each basic action recognition result into a working condition stage recognition feature vector; B2: Normalize the operating condition stage identification feature vector to obtain the normalized operating condition stage identification feature vector. B3: Input the normalized working condition stage identification feature vector into the preset working condition stage identification model to identify the working condition stage and obtain the working condition stage identification result. B4: Determine whether the working condition stage identification process has ended; B5: Output the working condition stage identification result after the working condition stage identification process is completed; B6: If the working condition stage identification process is not completed, remove the basic action identification result and the corresponding main pump pressure signal that are farthest from the current time in the working condition stage identification feature vector, and obtain the basic action identification result and the corresponding main pump pressure signal of the next time moment to update the working condition stage identification feature vector, obtain a new working condition stage identification feature vector, and return to execute B2.

6. The working condition recognition method of a working machine according to claim 5, characterized by, After executing B3, the following is also included: Obtain historical operating condition stage identification results; Based on the historical operating condition stage identification results and the preset intelligent verification rule library, the operating condition stage identification results are verified and corrected to obtain the corrected operating condition stage identification results.

7. The working condition recognition method of a working machine according to claim 6, characterized by, The historical operating condition stage identification results include the operating condition stage identification results of the previous moment, as well as multiple operating condition stage identification results within a third preset time range from the previous moment. The step of verifying and correcting the operating condition stage identification results based on the historical operating condition stage identification results and a pre-set intelligent verification rule base to obtain corrected operating condition stage identification results includes: Based on the identification results of multiple working condition stages within the third preset time range from the previous moment, the actual working condition stage at the current moment is determined; Based on the current actual working condition stage and the working condition stage identification result of the previous time, the corresponding working condition stage identification result correction rule is matched in the preset intelligent verification rule library. Based on the correction rules for the identified operating conditions, the identified operating conditions are verified and corrected to obtain the corrected identified operating conditions.

8. The working condition recognition method of a working machine according to claim 1, characterized by, After obtaining the operating condition stage identification results, the following is also included: Based on the basic motion recognition results and the working condition stage recognition results, the hoisting operation speed information is calculated; Obtain lifting weight statistics; Based on the hoisting operation speed information and the hoisting weight statistics, the operation efficiency information of the construction machinery is obtained.

9. A working condition recognition device for a working machine, characterized in that include: The acquisition module is used to acquire basic data and main pump pressure signals of the construction machinery; wherein, the data type of the basic data includes at least one of the following: actuator pilot control signal, actuator displacement, actuator tilt angle, and hydraulic pump flow rate; The motion recognition module is used to identify the basic motions of the construction machinery based on the basic data and a preset basic motion recognition model, and to obtain the basic motion recognition result. The basic motions refer to the motions of the actuators. The working condition identification module is used to identify the working condition stage of the construction machinery based on the basic action identification result and the main pump pressure signal using a preset working condition stage identification model, and obtain the working condition stage identification result; the basic action identification result output by the basic action identification model and the main pump pressure signal are used as the input of the working condition stage identification model.

10. A processor, comprising: It is configured to perform the engineering machinery operating condition identification method according to any one of claims 1 to 8.

11. A machine-readable storage medium having instructions stored thereon, the instructions comprising: When executed by a processor, this instruction causes the processor to be configured to perform the engineering machinery operating condition identification method according to any one of claims 1 to 8.

Images