Method and apparatus for quantifying the work efficiency of industrial machinery.

Abstract

The present invention provides a method and apparatus for quantifying the work efficiency of a work machine, enabling continuous, real-time monitoring of the work process of the work machine and improving the accuracy of evaluating the work efficiency of the work machine. [Solution] The method includes receiving sensing data of the work environment from various sensors installed on the work machine, extracting operational characteristics of the work state of the work machine based on the sensing data and a pre-trained operational recognition model, wherein the pre-trained operational recognition model is a neural network based on machine learning for generating operational characteristics of the work state of the work machine, and quantifying the work indicators of the work machine based on the operational characteristics, wherein the work indicators are used to characterize the work efficiency of the work machine.

Description

Technical Field

[0001] This application relates to the technical field of computers, and particularly to a method and device for quantifying the working efficiency of a working machine.

Background Art

[0002] In the field of working machines, working machines (e.g., excavators) are widely applied as important working devices in industries such as construction, mining, and infrastructure construction. The working efficiency of a working machine is directly related to the progress and cost-effectiveness of a construction project.

[0003] In related technologies, the conventional method for evaluating the working efficiency of a working machine mainly depends on manual observation and empirical judgment. Since it depends on the personal experience and observation of an operator, the evaluation results are easily affected by subjective factors, and the consistency and accuracy of the evaluation are limited. In addition, manual observation cannot realize continuous real-time monitoring of the working process of the working machine, and the immediate evaluation of and response to the working efficiency are limited.

Summary of the Invention

Problems to be Solved by the Invention

[0004] Embodiments of this application provide a method and device for quantifying the working efficiency of a working machine. To basically understand some aspects of the disclosed embodiments, a brief summary is provided below. This summary part is not a general comment, nor does it determine key / important components or explain the protection scope of these embodiments. Its sole purpose is to represent some concepts in a simple form and serve as a preamble to the following detailed description.

Means for Solving the Problems

[0005] In a first aspect, embodiments of this application provide a method for quantifying the working efficiency of a working machine. The method includes: Receiving perception data on the working environment from various sensors installed on the working machine, Based on sensing data and a pre-trained motion recognition model, the operational characteristics of the working state of the work machine are extracted, and the pre-trained motion recognition model is a neural network based on machine learning for generating operational characteristics of the working state of the work machine. This includes quantifying work indicators of a work machine based on its operational characteristics, and ensuring that these work indicators characterize the work efficiency of the work machine.

[0006] One option is to extract the operational characteristics of the working state of the work machine based on sensing data and a pre-trained motion recognition model. By inputting sensing data into a pre-trained motion recognition model, the operational characteristics of the working state of the work machine can be extracted. Outputting operational characteristics corresponding to the sensing data, This includes using the motion characteristics corresponding to the sensing data as the motion characteristics of the working state of the work machine.

[0007] One option is to generate a pre-trained action recognition model. The process involves collecting historical sensing data about the work environment from various sensors installed on the work machine, and the historical sensing data including image frames, inertial measurement unit data, and machine state data. The steps include: obtaining model training samples by specifying the operational characteristics of the working state of the work machine in the image frame, IMU data, and machine state data; The steps include: introducing a predetermined loss function to the neural network to obtain an action recognition model; The process includes the step of machine learning an action recognition model based on model training samples to generate a pre-trained action recognition model.

[0008] As an option, the given loss function is the cross-entropy loss function, and the cross-entropy loss function is, JPEG0007852961000001.jpg25170

[0009] As an option, the operational characteristics include the operational type and the operational execution time. Quantifying the work indicators of a work machine based on its operational characteristics is When the operation type satisfies the predetermined operation completion conditions and a loading completion command is received as feedback from the work machine, the operation execution time is set as the operation completion time, and the predetermined operation completion conditions are that the work machine loads the object to be loaded onto the transport vehicle. Obtain the start time of the operation corresponding to the operation type, Calculate and cache the time difference between the end time and the start time of the operation, This includes calculating work cycles and the average value of work cycles based on cached time series data to obtain work indicators for the work machines.

[0010] As an option, calculating the work cycle and the average value of the work cycle based on the cached time series is This includes adding up cached time series to obtain the total time required for the work machine to complete an entire loading process, The formula for calculating the average value of the work cycle is: JPEG0007852961000002.jpg27170

[0011] As for options, the methods are: If the operation type satisfies predetermined operation termination conditions and has not received a loading completion command feedback from the work machine, the process further includes continuing to perform the step of receiving sensing data of the work environment from various sensors installed on the work machine.

[0012] As for options, the methods are: If the operation type satisfies a predetermined operation start condition, the operation execution time is set to the operation start time corresponding to the operation type, and the predetermined operation start condition further includes the work machine obtaining the object to be loaded.

[0013] As an option, the operational characteristics include the actual loading amount during the work operation. Quantifying the work indicators of a work machine based on its operational characteristics is To calculate the actual loading amount during the work operation and the ratio between the predetermined maximum loading amount and the actual loading amount, This includes converting ratios into percentage format to obtain operational metrics for the work machines.

[0014] In a second embodiment, the present invention provides a device for quantifying the work efficiency of a work machine, and the device is A receiving module for receiving sensing data about the work environment from various sensors installed on work machinery, Based on sensing data and a pre-trained motion recognition model, it is used to extract the motion characteristics of the working state of a work machine, and the pre-trained motion recognition model is an extraction module which is a neural network for generating motion characteristics of the working state of a work machine based on machine learning, The system comprises a quantification module used to quantify work indicators of a work machine based on its operational characteristics, wherein the work indicators are used to characterize the work efficiency of the work machine. [Effects of the Invention]

[0015] The technical proposal relating to the embodiment of this application may include the following beneficial effects.

[0016] In the embodiments of the present invention, in one aspect, sensing data of the work environment from various sensors installed on the work machine is received and processed, and said sensing data is received in real time, thereby enabling continuous real-time monitoring of the work process of the work machine. In another aspect, based on the sensing data and a pre-trained motion identification model, motion characteristics of the work state of the work machine are extracted, and said model is a neural network based on machine learning for generating motion characteristics of the work state of the work machine, thereby enabling said neural network to accurately identify motion characteristics of the work state of the work machine, thereby improving the accuracy of evaluating the work efficiency of the work machine.

[0017] It should be understood that the above general description and the following detailed description are merely illustrative and explanatory, and do not limit the present application.

Brief Description of the Drawings

[0018] The drawings herein are incorporated into the specification and constitute a part of this specification, showing embodiments consistent with the present application and used together with the specification to interpret the principles of the present application. [Figure 1] FIG. 1 is a flowchart of a method for quantifying the working efficiency of a working machine according to an embodiment of the present application. [Figure 2] FIG. 2 is a flowchart of a model training method according to the present application. [Figure 3] FIG. 3 is a structural schematic diagram of a device for quantifying the working efficiency of a working machine according to the present application. [Figure 4] FIG. 4 is a structural schematic diagram of an electronic device according to an embodiment of the present application.

Modes for Carrying Out the Invention

[0019] The following description and drawings fully disclose specific embodiments of the present application so that those skilled in the art can practice them.

[0020] Obviously, the described embodiments are part of the embodiments of the present application, not all of them. Based on the embodiments of the present application, any other embodiments obtained by those skilled in the art without creative efforts belong to the protection scope of the present application.

[0021] When the following description relates to the drawings, unless otherwise specified, the same numerals in different drawings indicate the same or similar elements. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present application, which are described in detail in the appended claims.

[0022] In the description of this application, it should be understood that terms such as "first," "second," etc., are for explanatory purposes only and should not be understood as indicating or suggesting relative importance. A person skilled in the art will be able to understand the specific meaning of the above terms in this application, taking into account the specific circumstances. Also, in the description of this application, unless otherwise specified, "plural" means two or more. "And / or" is used to describe the relationship between related objects and indicates that there may be three relationships. For example, "A and / or B" may indicate three situations: "A exists independently," "A and B exist simultaneously," and "B exists independently." The letter " / " generally indicates that the preceding and succeeding related objects are in an "or" relationship.

[0023] To solve the above-mentioned related technical problems, the present application provides a method and apparatus for quantifying the work efficiency of a work machine. In one embodiment of the present application, sensing data of the work environment from various sensors installed on the work machine is received and processed, and the sensing data is received in real time, thereby enabling continuous real-time monitoring of the work process of the work machine. In another embodiment, the operational characteristics of the work state of the work machine are extracted based on the sensing data and a pre-trained motion identification model, and the model is a neural network based on machine learning for generating operational characteristics of the work state of the work machine, thereby enabling the neural network to accurately identify operational characteristics of the work state of the work machine, thereby improving the accuracy of evaluating the work efficiency of the work machine. These will be explained in detail below with reference to exemplary embodiments.

[0024] The method for quantifying the work efficiency of a work machine according to an embodiment of the present invention will be described in detail below with reference to Figures 1 and 2. This method may be implemented by a computer program or executed in a work machine work efficiency quantification device based on a von Neumann architecture. The computer program may be integrated into an application or executed as an independent tool application.

[0025] Figure 1 is a flowchart of a method for quantifying the work efficiency of a work machine according to an embodiment of the present invention. As shown in Figure 1, the method of the embodiment of the present invention may include the following steps.

[0026] S101 Receives sensing data about the work environment from various sensors installed on the work machine. The work machine is a mechanical vehicle currently performing a work task, such as an excavator. The various sensors installed on the work machine are positioned by experts on the target work machine according to actual experience and may be placed in multiple different locations on the work machine. The various sensors include high-resolution cameras, inertial measurement units (IMUs), and machine condition sensors. The high-resolution cameras are for capturing high-resolution video images during the loading process of the excavator, and these images can provide visual information about the work of the excavator, including digging and loading operations. The inertial measurement units may measure the acceleration and angular velocity of the excavator in space, and may also measure the direction of the magnetic field, and the machine condition sensors are for monitoring the machine condition of the excavator, which includes, but is not limited to, important parameters such as hydraulic pressure, temperature, and engine speed.

[0027] By setting the camera resolution to 1920*1080 and the frame rate to 30fps, it is possible to accurately capture the loading motion of the work machinery.

[0028] In some embodiments, appropriate sensor types, such as high-resolution cameras, IMUs, and machine condition sensors, are selected according to the type of work machine and work needs. Sensors are placed in critical locations on the work machine to ensure comprehensive capture of work environment and machine condition information. Data acquisition modules are attached to the work machine, and these modules are connected to various sensor interfaces to receive sensor data in real time. A communication network is established, and the data acquisition modules transmit sensor data to the server side. This transmission means includes wired connections or wireless communication technologies, such as Wi-Fi, Bluetooth®, 4G / 5G, etc., and the server side receives sensing data about the work environment from various sensors installed on the work machine.

[0029] S102 is a neural network that extracts the operational characteristics of the working state of a work machine based on sensing data and a pre-trained motion recognition model, and the pre-trained motion recognition model generates the operational characteristics of the working state of a work machine based on machine learning. The neural network may also be a convolutional neural network.

[0030] In some embodiments, the process of extracting operational characteristics of the working state of a work machine based on sensing data and a pre-trained motion recognition model specifically includes: inputting sensing data into a pre-trained motion recognition model to extract operational characteristics of the working state of the work machine; outputting operational characteristics corresponding to the sensing data; and using the operational characteristics corresponding to the sensing data as operational characteristics of the working state of the work machine.

[0031] In embodiments of the present invention, generating a pre-trained motion recognition model includes the steps of: collecting historical sensing data of the work environment from various sensors installed on the work machine, the historical sensing data including image frames, inertial measurement unit data and machine state data; obtaining model training samples by specifying the motion characteristics of the work state of the work machine in the image frames, IMU data and machine state data; obtaining a motion recognition model by introducing a predetermined loss function to the neural network; and generating a pre-trained motion recognition model by machine learning the motion recognition model based on the model training samples.

[0032] Specifically, the process of machine learning an action recognition model based on model training samples to generate a pre-trained action recognition model includes inputting model training samples into the action recognition model, outputting the model's cross-entropy loss value, generating a pre-trained action recognition model if the cross-entropy loss value reaches a minimum and the number of training iterations reaches a predetermined threshold, and otherwise continuing to input model training samples into the action recognition model until the model loss value reaches a minimum.

[0033] Specifically, the model training process is performed on a single computer equipped with an NVIDIA GeForce RTX4090 graphics card, with 300 training iterations and a batch size of 64 training samples each time the model is trained.

[0034] Furthermore, the accuracy of motion recognition may be determined during the model training process, and the formula for calculating accuracy is: JPEG0007852961000003.jpg30170

[0035] Specifically, the given loss function is the cross-entropy loss function, and the cross-entropy loss function is, JPEG0007852961000004.jpg25170

[0036] S103 Quantifies the work indicators of a work machine based on its operational characteristics, and these work indicators are used to characterize the work efficiency of the work machine.

[0037] The operational characteristics include the operation type and the time of execution.

[0038] In some embodiments, the process of quantifying the work indicators of a work machine based on its operational characteristics specifically includes: setting the operation execution time as the operation end time when the operation type satisfies predetermined operation completion conditions and a loading completion command is received as feedback from the work machine, with the predetermined operation completion conditions being that the work machine loads the object to be loaded onto the transport vehicle; obtaining the operation start time corresponding to the operation type; calculating and caching the time difference between the operation end time and the operation start time; and obtaining the work indicators of the work machine by calculating the work cycle and the average value of the work cycle based on the cached time series.

[0039] Specifically, the process of calculating work cycles and the average value of work cycles based on cached time series specifically includes adding up the cached time series to obtain the total time required for the work machine to complete an entire loading process, The formula for calculating the average value of the work cycle is:

[0040] In some embodiments, if the operation type satisfies predetermined operation termination conditions and has not received a loading completion command fed back from the work machine, the step of receiving sensing data of the work environment from various sensors installed on the work machine is continued.

[0041] In some embodiments, if the operation type satisfies a predetermined operation start condition, the operation execution time is set to the operation start time corresponding to the operation type, and the predetermined operation start condition is that the work machine obtains the object to be loaded.

[0042] The operational characteristics include the actual loading volume during the work operation.

[0043] In some embodiments, the process of quantifying the work index of a work machine based on its operational characteristics specifically includes calculating the actual loading amount during the work operation and the ratio between a predetermined maximum loading amount and the actual loading amount, and converting the ratio into a percentage format to obtain the work index of the work machine.

[0044] Furthermore, the visualization interface allows operators and managers to see work indicators for the work machines in real time, thereby facilitating real-time display and exchange of data.

[0045] In the embodiments of the present invention, in one aspect, sensing data of the work environment from various sensors installed on the work machine is received and processed, and said sensing data is received in real time, thereby enabling continuous real-time monitoring of the work process of the work machine. In another aspect, based on the sensing data and a pre-trained motion identification model, motion characteristics of the work state of the work machine are extracted, and said model is a neural network based on machine learning for generating motion characteristics of the work state of the work machine, thereby enabling said neural network to accurately identify motion characteristics of the work state of the work machine, thereby improving the accuracy of evaluating the work efficiency of the work machine.

[0046] Figure 2 is a flowchart of the model training method according to the embodiment of the present application. As shown in Figure 2, the method according to the embodiment of the present application is Step S201 involves collecting historical sensing data of the working environment from various sensors installed on the work machine, and the historical sensing data includes image frames, inertial measurement unit data, and machine state data. Step S202 involves obtaining a model training sample by specifying the operational characteristics of the working state of the work machine in the image frame, IMU data, and machine state data. Step S203 involves introducing a predetermined loss function to the neural network to obtain an action recognition model. The process may also include step S204, which involves machine learning an action recognition model based on model training samples to generate a pre-trained action recognition model.

[0047] In the embodiments of the present invention, in one aspect, sensing data of the work environment from various sensors installed on the work machine is received and processed, and said sensing data is received in real time, thereby enabling continuous real-time monitoring of the work process of the work machine. In another aspect, based on the sensing data and a pre-trained motion identification model, motion characteristics of the work state of the work machine are extracted, and said model is a neural network based on machine learning for generating motion characteristics of the work state of the work machine, thereby enabling said neural network to accurately identify motion characteristics of the work state of the work machine, thereby improving the accuracy of evaluating the work efficiency of the work machine.

[0048] The following are apparatus embodiments of the present application, which may be used to carry out method embodiments of the present application. Details not disclosed in the apparatus embodiments of the present application may be referenced to the method embodiments of the present application.

[0049] Figure 3 shows a schematic diagram of the structure of a work efficiency quantification device for a work machine according to an exemplary embodiment of the present invention. The work efficiency quantification device for the work machine may be implemented as all or part of an electronic device by software, hardware, or a combination of both. The device 1 comprises a receiving module 10, an extraction module 20, and a quantification module 30.

[0050] The receiving module 10 is used to receive sensing data about the working environment from various sensors installed on the work machine. The extraction module 20 is used to extract the operational characteristics of the working state of the work machine based on sensing data and a pre-trained operational recognition model. The pre-trained operational recognition model is a neural network based on machine learning for generating operational characteristics of the working state of the work machine. The quantification module 30 is used to quantify the work indicators of a work machine based on its operational characteristics, and these work indicators are used to characterize the work efficiency of the work machine.

[0051] It should be explained that when the work efficiency quantification device for a work machine according to the above embodiment performs the work efficiency quantification method for a work machine, only the divisions of each functional module described above are explained as an example. In actual applications, the above functions can be completed by assigning them to different functional modules as needed, that is, the internal structure of the equipment can be divided into different functional modules to complete all or part of the functions described above. Furthermore, the embodiments of the work efficiency quantification device for a work machine and the work efficiency quantification method for a work machine according to the above embodiment belong to the same concept, and their specific implementation process may be described by referring to the method embodiment, so a detailed explanation is omitted here.

[0052] The numbering of the embodiments in the present application above is for illustrative purposes only and does not indicate any ranking of the embodiments.

[0053] In the embodiments of the present invention, in one aspect, sensing data of the work environment from various sensors installed on the work machine is received and processed, and said sensing data is received in real time, thereby enabling continuous real-time monitoring of the work process of the work machine. In another aspect, based on the sensing data and a pre-trained motion identification model, motion characteristics of the work state of the work machine are extracted, and said model is a neural network based on machine learning for generating motion characteristics of the work state of the work machine, thereby enabling said neural network to accurately identify motion characteristics of the work state of the work machine, thereby improving the accuracy of evaluating the work efficiency of the work machine.

[0054] This application further provides a computer-readable medium in which program instructions are stored, and when these program instructions are executed by a processor, a method for quantifying the work efficiency of a work machine according to each of the above-described embodiments of the method is realized.

[0055] The present invention further provides a computer program product including instructions, which, when executed on a computer, causes the computer to execute the methods for quantifying the work efficiency of the work machines of each of the above-described embodiment.

[0056] Figure 4 is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. As shown in Figure 4, the electronic device 1000 may include at least one processor 1001, at least one network interface 1004, a user interface 1003, a memory 1005, and at least one communication bus 1002.

[0057] The communication bus 1002 is used to enable communication between these components.

[0058] The user interface 1003 may include a display and a camera, and optionally, the user interface 1003 may further include a standard wired interface and a wireless interface.

[0059] The network interface 1004 may optionally include a standard wired interface or a wireless interface (e.g., a Wi-Fi interface).

[0060] The processor 1001 may include one or more processing cores. The processor 1001 connects various parts of the electronic device 1000 using various interfaces and circuits, and executes various functions and processing data of the electronic device 1000 by operating or executing instructions, programs, code sets or instruction sets stored in memory 1005, and by retrieving data stored in memory 1005. Optionally, the processor 1001 may be implemented in the form of at least one hardware component from among Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), and Programmable Logic Array (PLA). The processor 1001 may integrate one or more combinations from among a Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily processes operating devices, user interfaces, and application programs, the GPU is responsible for rendering and plotting content to be displayed on the display, and the modem handles wireless communication. As can be understood, the above modem may be implemented independently on a single chip, rather than being integrated into the processor 1001.

[0061] Memory 1005 may include random access memory (RAM) or read-only memory. Optionally, memory 1005 may include a non-transitory computer-readable storage medium. Memory 1005 may be used to store instructions, programs, code, code sets, or instruction sets. Memory 1005 may include a program storage area and a data storage area. The program storage area can store instructions for implementing an operating device, instructions for at least one function (e.g., touch function, audio playback function, image playback function, etc.), instructions for implementing each of the above method embodiments, etc., while the data storage area can store data related to each of the above method embodiments, etc. Memory 1005 may optionally be at least one storage device further separated from the processor 1001. As shown in Figure 4, memory 1005, which is a computer storage medium, may include an operating device, a network communication module, a user interface module, and an application program for quantifying the work efficiency of a work machine.

[0062] In the electronic device 1000 shown in Figure 4, the user interface 1003 is primarily used to provide an input interface to the user and to acquire data entered by the user. However, the processor 1001 may also be used to call an application program for quantifying the work efficiency of the work machine stored in memory 1005 and to perform the following operations, which include: Receiving sensing data about the work environment from various sensors installed on the work machine, Based on sensing data and a pre-trained motion recognition model, the operational characteristics of the working state of the work machine are extracted, and the pre-trained motion recognition model is a neural network based on machine learning for generating operational characteristics of the working state of the work machine. This includes quantifying work indicators of a work machine based on its operational characteristics, and ensuring that these work indicators characterize the work efficiency of the work machine.

[0063] In one embodiment, when the processor 1001 performs the step of extracting the operational characteristics of the working state of the work machine based on sensing data and a pre-trained motion identification model, specifically, By inputting the sensing data into a pre-trained motion recognition model, the motion characteristics of the working state of the work machine are extracted. Outputs the operational characteristics corresponding to the sensing data. The operation is performed to use the motion characteristics corresponding to the sensing data as the motion characteristics of the working state of the work machine.

[0064] In one embodiment, when the processor 1001 performs the step of generating a pre-trained behavioral identification model, specifically, The system collects historical sensing data about the work environment from various sensors installed on the work machine, and the historical sensing data includes image frames, inertial measurement unit data, and machine state data. Model training samples are obtained by clearly indicating the operational characteristics of the working state of the work machine in the image frame, IMU data, and machine state data. A predetermined loss function is introduced to the neural network to obtain an action recognition model. This process involves machine learning an action recognition model based on model training samples to generate a pre-trained action recognition model.

[0065] In one embodiment, when the processor 1001 performs the step of quantifying the work indicators of the work machine based on its operating characteristics, specifically, When the operation type satisfies the predetermined operation completion conditions and a loading completion command is received as feedback from the work machine, the operation execution time is set as the operation completion time, and the predetermined operation completion conditions are that the work machine loads the object to be loaded onto the transport vehicle. Get the start time of the operation corresponding to the operation type. The time difference between the end time of the operation and the start time of the operation is calculated and cached. This operation calculates the work cycle and the average value of the work cycle based on the cached time series, and then obtains the work indicators for the work machine.

[0066] In one embodiment, the processor 1001 further, If the operation type satisfies the predetermined operation termination conditions and no loading completion command has been received as feedback from the work machine, the operation continues to perform the step of receiving sensing data about the work environment from various sensors installed on the work machine.

[0067] In one embodiment, the processor 1001 further, If the operation type satisfies the predetermined start conditions, the operation execution time is set to the start time corresponding to the operation type, and the predetermined start conditions are that the work machine obtains the object to be loaded.

[0068] In one embodiment, when the processor 1001 quantifies the work indicators of the work machine based on its operating characteristics, specifically, The actual loading amount during the work operation and the ratio between the predetermined maximum loading amount and the actual loading amount are calculated. The operation involves converting the ratio into a percentage format to obtain the work indicator for the work machine.

[0069] In the embodiments of the present invention, in one aspect, sensing data of the work environment from various sensors installed on the work machine is received and processed, and said sensing data is received in real time, thereby enabling continuous real-time monitoring of the work process of the work machine. In another aspect, based on the sensing data and a pre-trained motion identification model, motion characteristics of the work state of the work machine are extracted, and said model is a neural network based on machine learning for generating motion characteristics of the work state of the work machine, thereby enabling said neural network to accurately identify motion characteristics of the work state of the work machine, thereby improving the accuracy of evaluating the work efficiency of the work machine.

[0070] As those skilled in the art will understand, the implementation of all or part of the processes in the above embodiments may be achieved by instructing the relevant hardware with a computer program, and the program for quantifying the work efficiency of the work machine may be stored on a computer-readable storage medium, and the program may include the processes of each embodiment of the above methods when executed. The storage medium for the program for quantifying the work efficiency of the work machine may be a magnetic disk, an optical disk, a read-only memory, or a random access memory, etc.

[0071] The above description is merely a preferred embodiment of the present application and, naturally, does not limit the scope of the claims of the present application. Therefore, equivalent modifications made based on the claims of the present application still fall within the scope of protection of the present application.

Claims

1. A method for quantifying the work efficiency of a work machine, wherein the method is: Receiving sensing data about the work environment from various sensors installed on the work machine, Based on the aforementioned sensing data and a pre-trained motion identification model, the operational characteristics of the working state of the work machine are extracted. This includes quantifying the work indicators of the work machine based on the aforementioned operating characteristics, The aforementioned pre-trained motion recognition model is a neural network based on machine learning for generating motion characteristics of the working state of the work machine, The aforementioned work indicators are for characterizing the work efficiency of the work machine, The aforementioned operational characteristics include the operational type and the operational execution time. Quantifying the work indicators of the work machine based on the aforementioned operating characteristics is When the operation type is set to load the object to be loaded onto the transport vehicle, and the predetermined operation completion conditions are met, and a loading completion command is received as feedback from the work machine, the operation execution time is set as the operation completion time. To obtain the start time of the operation corresponding to the aforementioned operation type, The time difference between the end time of the operation and the start time of the operation is calculated and cached, A method for quantifying the work efficiency of a work machine, characterized by comprising: calculating a work cycle and the average value of the work cycle based on a cached time series, and obtaining a work indicator for the work machine.

2. Extracting operational characteristics of the working state of the work machine based on the aforementioned sensing data and the aforementioned pre-trained motion identification model is: By inputting the aforementioned sensing data into the pre-trained motion identification model, the operational characteristics of the working state of the work machine corresponding to the sensing data are extracted. Outputting the operation characteristics corresponding to the sensing data, The method according to claim 1, characterized in that the operation characteristics corresponding to the sensing data are defined as operation characteristics of the working state of the work machine.

3. Generating the pre-trained motion recognition model is The steps include collecting historical sensing data of the working environment from various sensors installed on the aforementioned work machine, and the historical sensing data including image frames, inertial measurement unit data, and machine state data, The steps include: obtaining a model training sample by specifying the operational characteristics of the working state of the work machine in the image frame, inertial measurement unit data, and machine state data; The steps include: introducing a predetermined loss function to the neural network to obtain an action recognition model; The method according to 1 or 2, characterized by comprising the step of machine learning the motion recognition model based on the model training samples to generate the pre-trained motion recognition model.

4. The method according to the present invention, characterized in that the predetermined loss function is the cross-entropy loss function shown in the following formula 1. [Math 1] Here, L is the cross-entropy loss value, N is the number of samples in the model training sample, C is the number of categories in the operating feature, y i, c is the true category label of the operating feature, and p i, c is the predicted category label of the operating feature.

5. The aforementioned method, The method according to 1 or 2, further comprising the step of receiving sensing data of the working environment from various sensors installed on the working machine if the operation type satisfies the predetermined operation termination conditions and has not received a loading completion command fed back from the working machine.

6. The aforementioned method, The method according to 1 or 2, further comprising the condition that when the operation type satisfies a predetermined operation start condition, the operation execution time is set to the operation start time corresponding to the operation type, and the predetermined operation start condition is that the work machine obtains the object to be loaded.

7. The aforementioned operational characteristics include the actual loading amount during the work operation. Quantifying the work indicators of the work machine based on the aforementioned operating characteristics is The actual loading amount in the aforementioned work operation and the ratio between the predetermined maximum loading amount and the actual loading amount are calculated. The method according to 1 or 2, characterized in that it includes converting the ratio into a percentage format to obtain an operational indicator for the work machine.

8. A device for quantifying the work efficiency of a work machine, wherein the device is A receiving module for receiving sensing data about the work environment from various sensors installed on work machinery, An extraction module is used to extract the operational characteristics of the working state of the work machine based on the aforementioned sensing data and a pre-trained motion recognition model, wherein the pre-trained motion recognition model is a neural network based on machine learning for generating operational characteristics of the working state of the work machine. A quantification module is used to quantify the work indicators of the work machine based on the aforementioned operating characteristics, wherein the work indicators are used to characterize the work efficiency of the work machine, The aforementioned operational characteristics include the operational type and the operational execution time. When the quantification module quantifies the work indicators of the work machine based on the operating characteristics, When the operation type is defined as the work machine loading the object to be loaded onto the transport vehicle, and the predetermined operation completion conditions are met, and a loading completion command is received as feedback from the work machine, the operation execution time shall be set as the operation completion time. The start time of the operation corresponding to the aforementioned operation type is obtained, The time difference between the end time of the operation and the start time of the operation is calculated and cached. A device for quantifying the work efficiency of a work machine, characterized by calculating the work cycle and the average value of the work cycle based on cached time series data, and obtaining work indicators for the work machine.

Images