A rental management method and system based on engineering machinery monitoring

Abstract

The present invention discloses a leasing management method and system based on engineering machinery monitoring, which belongs to the technical field of data identification and engineering machinery management. By obtaining the multi-dimensional monitoring parameters to be identified on the engineering machinery, and performing dimensional correction and overall correction on the multi-dimensional monitoring parameters to be identified, the enhanced final characteristic parameters are obtained. Finally, the final characteristic parameters are identified through the central computing node, and the operating status of the engineering machinery can be obtained more accurately, thereby realizing the leasing management of the engineering machinery. Different alarm information and the trajectory parameters of the engineering machinery can also be sent to the equipment terminal of the operation and maintenance personnel according to the operating status, thereby realizing intelligent monitoring and further management.

Description

Technical Field

[0001] The present invention belongs to the technical field of data identification and engineering machinery management, and in particular relates to a leasing management method and system based on engineering machinery monitoring. Background Art

[0002] Construction machinery is primarily used in various construction projects and typically operates in various mechanical industry environments. The operating status of construction machinery is related to information such as fuel consumption, the driver's working hours, and workload. In practical applications, construction machinery is often provided by leasing companies, providing clients with construction machinery. During the leasing process, effective monitoring of construction machinery is often lacking, hindering leasing companies' ability to effectively manage the machinery. Summary of the Invention

[0003] The present invention provides a leasing management method and system based on engineering machinery monitoring, which are used to solve the problem in the prior art that leasing companies are unable to effectively manage the leasing of engineering machinery.

[0004] In one aspect, the present invention provides a leasing management method based on engineering machinery monitoring, comprising:

[0005] Obtaining multi-dimensional monitoring parameters to be identified on the engineering machinery, and inputting the multi-dimensional monitoring parameters to be identified into the edge computing node of the corresponding dimension;

[0006] Inputting the corresponding multi-dimensional monitoring parameters to be identified into the sub-feature extraction model through the edge computing node to obtain the initial feature parameters of each dimension under the first constraint condition;

[0007] Inputting the multidimensional monitoring parameters to be identified into a central computing node, and inputting the multidimensional monitoring parameters to be identified into a comprehensive feature extraction model through the central computing node to obtain comprehensive feature parameters of all dimensions under the second constraint condition; the dimension of the comprehensive feature parameter is the same as the sum of the dimensions of all initial feature parameters;

[0008] Obtaining final characteristic parameters of the multi-dimensional monitoring parameters to be identified based on the initial characteristic parameters and the comprehensive characteristic parameters, and identifying the final characteristic parameters through a central computing node to obtain the operating status of the construction machinery;

[0009] According to the operating status of the construction machinery, different alarm information and the trajectory parameters of the construction machinery are sent to the equipment terminals of the operation and maintenance personnel through the central computing node to complete the rental management of the construction machinery.

[0010] Furthermore, the multi-dimensional monitoring parameters to be identified on the engineering machinery include: fuel consumption information, positioning information, start and stop records, operating time and driving speed at continuous time points.

[0011] Furthermore, the multi-dimensional monitoring parameters to be identified on the construction machinery are obtained, including:

[0012] The monitoring device on the engineering machinery samples the multi-dimensional monitoring data to be identified within a time period according to a preset sampling frequency to obtain a sampling data packet;

[0013] The central computing node and the edge computing node respectively obtain the sampling data packets stored in the monitoring equipment on the engineering machinery with the preset first delay condition to obtain the multi-dimensional monitoring parameters to be identified.

[0014] Furthermore, it also includes:

[0015] Acquiring anti-tampering parameters of the monitoring device according to a preset second time delay condition, wherein the anti-tampering parameters include at least a first sensing time sequence of the pressure sensor, a second sensing time sequence of the vibration sensor, and a third sensing time sequence of the distance sensor;

[0016] Obtaining an anti-dismantling feature matrix based on the first sensing time series, the second sensing time series, and the third sensing time series; identifying the anti-dismantling feature matrix using a pre-trained machine learning model at a central computing node to obtain a disassembly status of the monitoring device, where the disassembly status includes whether the monitoring device is disassembled or operating normally;

[0017] When the disassembly status indicates that the monitoring device is disassembled, a warning message indicating that the monitoring device is disassembled is sent to the device terminal of the operation and maintenance personnel through the central computing node, and the real-time positioning information or the last acquired positioning information in the monitoring device is transmitted to the device terminal of the operation and maintenance personnel to realize anti-disassembly monitoring of the monitoring device, thereby ensuring the accuracy of the multi-dimensional monitoring parameters to be identified.

[0018] Furthermore, the edge computing node inputs the corresponding multi-dimensional monitoring parameter to be identified into the sub-feature extraction model to obtain the initial feature parameters of each dimension under the first constraint condition, including:

[0019] Through different edge computing nodes, the fuel consumption information, positioning information, start-stop records, running time and driving speed at continuous time points are input into different sub-feature extraction models to obtain the initial feature parameters of the fuel consumption information, positioning information, start-stop records, running time and driving speed under the corresponding first constraint conditions.

[0020] Furthermore, the method for obtaining the first constraint condition includes:

[0021] Build a sub-feature extraction model using a machine learning model to obtain the initial first constraint condition;

[0022] The historical data of a certain dimension monitoring parameter and the corresponding manual correction data are used as the first training data, and the sub-feature extraction model is trained using the first training data until the training meets the pre-set first end condition and obtains the first constraint condition; the first constraint condition is provided by the trained sub-feature extraction model.

[0023] Furthermore, the multidimensional monitoring parameters to be identified are input into a central computing node, and the multidimensional monitoring parameters to be identified are input into a comprehensive feature extraction model through the central computing node to obtain comprehensive feature parameters of all dimensions under the second constraint condition, including:

[0024] Constructing a target characteristic parameter matrix based on the fuel consumption information, positioning information, start and stop records, operating time, and driving speed at the consecutive time points;

[0025] The target feature parameter matrix is ​​input into the comprehensive feature extraction model through the central computing node to obtain the comprehensive feature parameters of the target feature parameter matrix under the second constraint condition.

[0026] Furthermore, the method for obtaining the second constraint condition includes:

[0027] Build a comprehensive feature extraction model using a machine learning model to obtain the initial second constraint;

[0028] The historical target feature parameter matrix corresponding to the multi-dimensional monitoring parameter to be identified and the corresponding manual correction matrix are used as the second training data, and the comprehensive feature extraction model is trained using the second training data until the training meets the pre-set second end condition to obtain the second constraint condition; the second constraint condition is provided by the trained comprehensive feature extraction model.

[0029] Furthermore, according to the initial characteristic parameters and the comprehensive characteristic parameters, final characteristic parameters of the multi-dimensional monitoring parameters to be identified are obtained, and the final characteristic parameters are identified by the central computing node to obtain the operating status of the construction machinery, including:

[0030] The initial characteristic parameters and the comprehensive characteristic parameters are fused by data fusion to obtain the final characteristic parameters of the multi-dimensional monitoring parameters to be identified;

[0031] The operation status monitoring model pre-deployed on the central computing node is used to identify the final characteristic parameters to obtain the operation status of the construction machinery.

[0032] On the other hand, the present invention provides a rental management system based on engineering machinery monitoring, comprising a data acquisition module, a first feature acquisition module, a second feature acquisition module, an operation monitoring module, and a warning management module;

[0033] The data acquisition module is used to obtain the multi-dimensional monitoring parameters to be identified on the engineering machinery, and input the multi-dimensional monitoring parameters to be identified into the edge computing node of the corresponding dimension;

[0034] The first feature acquisition module is used to input the corresponding multi-dimensional monitoring parameter to be identified into the sub-feature extraction model through the edge computing node to obtain the initial feature parameters of each dimension under the first constraint condition;

[0035] The second feature acquisition module is used to input the multidimensional monitoring parameters to be identified into a central computing node, and input the multidimensional monitoring parameters to be identified into a comprehensive feature extraction model through the central computing node to obtain comprehensive feature parameters of all dimensions under the second constraint condition; the dimension of the comprehensive feature parameter is the same as the sum of the dimensions of all initial feature parameters;

[0036] The operation monitoring module is used to obtain the final characteristic parameters of the multi-dimensional monitoring parameters to be identified based on the initial characteristic parameters and the comprehensive characteristic parameters, and identify the final characteristic parameters through the central computing node to obtain the operating status of the construction machinery;

[0037] The warning management module is used to send different alarm information and trajectory parameters of the construction machinery to the equipment terminal of the operation and maintenance personnel through the central computing node according to the operating status of the construction machinery, thereby completing the rental management of the construction machinery.

[0038] The present invention provides a leasing management method and system based on engineering machinery monitoring, which obtains the multi-dimensional monitoring parameters to be identified on the engineering machinery, and performs dimensional correction and overall correction on the multi-dimensional monitoring parameters to be identified, thereby obtaining the enhanced final characteristic parameters, and finally identifying the final characteristic parameters through the central computing node, so as to more accurately obtain the operating status of the engineering machinery, thereby realizing the leasing management of the engineering machinery, and can also send different alarm information and the trajectory parameters of the engineering machinery to the equipment terminal of the operation and maintenance personnel according to the operating status, thereby realizing intelligent monitoring and further management. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0040] Figure 1 A flowchart of a rental management method based on engineering machinery monitoring provided by an embodiment of the present invention.

[0041] Figure 2 A schematic structural diagram of a rental management system based on engineering machinery monitoring provided by an embodiment of the present invention.

[0042] The above drawings illustrate specific embodiments of the present invention, which will be described in more detail below. These drawings and the accompanying description are not intended to limit the scope of the present invention in any way, but rather to illustrate the concept of the present invention to those skilled in the art by reference to specific embodiments. DETAILED DESCRIPTION

[0043] Exemplary embodiments will be described in detail herein, examples of which are illustrated in the accompanying drawings. In the following description, when referring to the drawings, like numbers in different figures represent like or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all possible embodiments consistent with the present invention. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present invention, as detailed in the appended claims.

[0044] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0045] Example 1

[0046] like Figure 1 As shown, the present invention provides a leasing management method based on engineering machinery monitoring, comprising:

[0047] S101: Acquire multi-dimensional monitoring parameters to be identified on engineering machinery, and input the multi-dimensional monitoring parameters to be identified into an edge computing node of a corresponding dimension.

[0048] Anti-disassembly monitoring equipment is installed on construction machinery. The monitoring equipment collects multi-dimensional monitoring data to be identified within the sampling period at a preset sampling frequency, and forms a data packet with the multi-dimensional monitoring data to be identified within each period. This allows the data packet to be pulled from the monitoring equipment at regular intervals according to the preset delay conditions to obtain the multi-dimensional monitoring parameters to be identified. The delay condition is used to indicate that the data packet acquisition operation is performed within a fixed time interval. It is worth noting that a sampling frequency that is too high will result in a larger data dimension, requiring a large amount of network resources and computing resources, while a sampling frequency that is too low will result in a decrease in data representativeness, resulting in a decrease in recognition effect. Therefore, the sampling frequency can be set according to the actual needs of the leasing company to achieve a balance between resource usage and the final effect.

[0049] S102: Inputting the corresponding multi-dimensional monitoring parameters to be identified into a sub-feature extraction model through the edge computing node to obtain initial feature parameters of each dimension under the first constraint condition.

[0050] The sub-feature extraction model can be constructed using a machine learning model, such as a BP (Back Propagation) neural network, an RBF (Radical Basis Function) neural network, or a convolutional neural network. It is worth noting that the above neural network is merely an example of this embodiment, and other neural networks can also be used to construct the sub-feature extraction model to achieve different extraction effects of initial feature parameters.

[0051] By training the sub-feature extraction model, the sub-feature extraction model can implicitly contain the first constraint condition, thereby implementing the first constraint on the monitoring parameters of a certain dimension, and finally obtaining the initial feature parameters of each dimension under the first constraint condition.

[0052] Optionally, the input corresponding to the sub-feature extraction model can be manually calibrated correction parameters. By learning the relationship between historical data and the correction parameters, data correction can be achieved to obtain more accurate initial feature parameters. It is also possible to set a warning level for a certain dimension of data (e.g., normal is level 0, abnormal is levels 1-4, and the higher the level, the worse the operating status of the construction machinery). By pre-learning the relationship between manually calibrated historical data of a certain dimension and the warning level, it is possible to identify the warning level of real-time monitoring data and ultimately obtain the initial feature parameters.

[0053] S103: Input the multidimensional monitoring parameters to be identified into a central computing node, and input the multidimensional monitoring parameters to be identified into a comprehensive feature extraction model through the central computing node to obtain comprehensive feature parameters of all dimensions under the second constraint condition. The dimensions of the comprehensive feature parameters are the same as the sum of the dimensions of all initial feature parameters.

[0054] The comprehensive feature extraction model identifies the multidimensional monitoring parameters to be identified, and multidimensional parameters are more suitable for constructing matrices. Therefore, a convolutional neural network can be used to construct the comprehensive feature extraction model. It is worth noting that the above neural network is only used as an example of this embodiment. Other neural networks can also be used to construct the comprehensive feature extraction model, and multidimensional parameters can also be constructed as vectors for identification to achieve different comprehensive feature parameter extraction effects.

[0055] By training the comprehensive feature extraction model, the comprehensive feature extraction model can implicitly include the second constraint condition, thereby implementing the second constraint on the multi-dimensional monitoring parameters to be identified, and finally obtaining the comprehensive feature parameters of all dimensions under the second constraint condition.

[0056] Optionally, when the sub-feature extraction model implements correction of a parameter in a certain dimension, the comprehensive feature extraction model should implement correction of parameters in all dimensions. When the sub-feature extraction model implements recognition of warning levels, the comprehensive feature extraction model should also implement recognition of warning levels. The sub-feature extraction model can realize the temporal correlation of a parameter in a certain dimension, and the comprehensive feature extraction model can realize the spatial correlation of multi-dimensional parameters, thereby more effectively extracting data features.

[0057] S104 , obtaining final characteristic parameters of the multi-dimensional monitoring parameters to be identified based on the initial characteristic parameters and the comprehensive characteristic parameters, and identifying the final characteristic parameters through a central computing node to obtain the operating status of the construction machinery.

[0058] After extracting the initial and comprehensive feature parameters, they need to be fused to obtain the final parameters to be identified. Finally, the parameters to be identified are identified to achieve precise monitoring of construction machinery. The extraction of temporal and spatial features provided by the embodiments of the present invention, followed by secondary identification, allows for better learning of potential connections and data features within the data. Compared to existing technologies that directly identify all monitoring parameters, this approach offers higher recognition accuracy and stronger data processing capabilities.

[0059] Optionally, a machine learning model can be set up to specifically identify the final feature parameters. After training this specially set machine learning model, a mapping relationship between the final feature parameters and the operating status of the construction machinery can be established, thereby realizing automatic monitoring of the construction machinery. Compared with the existing technology, this allows leasing companies to monitor the construction machinery in real time, making management more convenient. The operating status of the construction machinery can be excellent, good, fair, poor, and very poor. However, it should be noted that the above operating statuses are only examples in this embodiment and can also be set to other statuses.

[0060] S105. According to the operation status of the construction machinery, different alarm information and trajectory parameters of the construction machinery are sent to the equipment terminal of the operation and maintenance personnel through the central computing node to complete the rental management of the construction machinery.

[0061] Optionally, in addition to issuing alarm information, management strategies can also be preset. When different alarm information is sent to the equipment terminals of operation and maintenance personnel, preset management strategies can be provided to the operation and maintenance personnel, thereby further assisting the leasing company in the management of construction machinery.

[0062] The present invention provides a leasing management method and system based on engineering machinery monitoring, which obtains the multi-dimensional monitoring parameters to be identified on the engineering machinery, and performs dimensional correction and overall correction on the multi-dimensional monitoring parameters to be identified, thereby obtaining the enhanced final characteristic parameters, and finally identifying the final characteristic parameters through the central computing node, so as to more accurately obtain the operating status of the engineering machinery, thereby realizing the leasing management of the engineering machinery, and can also send different alarm information and the trajectory parameters of the engineering machinery to the equipment terminal of the operation and maintenance personnel according to the operating status, thereby realizing intelligent monitoring and further management.

[0063] Optionally, the leasing management method based on engineering machinery monitoring further includes:

[0064] Receive the return operation transmitted by the customer account and obtain the target customer account. It is worth noting that when the customer account does not have a rental order, the return operation is not allowed to be generated, thereby avoiding invalid data search.

[0065] Using the target customer account as the key value, search in the database to obtain the target rental order corresponding to the target customer account.

[0066] Based on the target rental order and pre-set rental rules, the rental amount for the target rental order is calculated, and a payment request equal to the calculated rental amount is initiated to the target customer account. Once the payment is successful, the location of the construction machinery corresponding to the target rental order is determined to be within a preset area. If so, the return is allowed; otherwise, the return is not allowed. It is worth noting that the preset area can be the internal area of ​​the rental company or a divided area. If the preset area is not within the rental company's internal area, the construction machinery can be equipped with a remotely controlled electromagnetic lock to achieve remote locking operation.

[0067] Optionally, to ensure customer safety, a reminder or confirmation message can be sent to the customer's account before remote locking or returning the vehicle to confirm that the customer is outside the vehicle. Multiple cameras can also be installed on the construction machinery to collect data about the surrounding environment. When a return operation is initiated, the surrounding data can be used to determine whether the construction machinery is in a safe location. If so, the return is allowed; otherwise, the return is not allowed.

[0068] In this embodiment, the multi-dimensional monitoring parameters to be identified on the engineering machinery include: fuel consumption information, positioning information, start and stop records, operating time, and driving speed at consecutive time points.

[0069] Each time point represents a sampling point. It is worth noting that in addition to the above-mentioned multi-dimensional monitoring parameters to be identified, other monitoring parameters can be set according to actual needs to achieve more accurate operation status identification.

[0070] In this embodiment, obtaining the multi-dimensional monitoring parameters to be identified on the construction machinery includes:

[0071] The monitoring equipment on the engineering machinery samples the multi-dimensional monitoring data to be identified within a time period according to a preset sampling frequency to obtain a sampling data packet.

[0072] The sampling frequency refers to the time interval for collecting data. Data is collected once every time interval, so that a sample data packet can be obtained.

[0073] The central computing node and the edge computing node acquire sampled data packets stored in the monitoring equipment on the engineering machinery using a preset first delay condition to obtain the multi-dimensional monitoring parameters to be identified. The first delay condition indicates that the data packet acquisition operation is performed within a fixed time interval.

[0074] In this embodiment, it also includes:

[0075] The anti-dismantling parameters of the monitoring device are obtained under a preset second time delay condition, wherein the anti-dismantling parameters include at least a first sensing time sequence of the pressure sensor, a second sensing time sequence of the vibration sensor, and a third sensing time sequence of the distance sensor.

[0076] According to the first sensing time series, the second sensing time series and the third sensing time series, an anti-dismantling feature matrix is ​​obtained, and the anti-dismantling feature matrix is ​​identified by a pre-trained machine learning model through a central computing node to obtain the disassembly status of the monitoring device, wherein the disassembly status includes whether the monitoring device is disassembled or the monitoring device is operating normally.

[0077] When the disassembly status indicates that the monitoring device is disassembled, a warning message indicating that the monitoring device is disassembled is sent to the device terminal of the operation and maintenance personnel through the central computing node, and the real-time positioning information or the last acquired positioning information in the monitoring device is transmitted to the device terminal of the operation and maintenance personnel to realize anti-disassembly monitoring of the monitoring device, thereby ensuring the accuracy of the multi-dimensional monitoring parameters to be identified.

[0078] By performing anti-dismantling identification, the accuracy of monitoring parameters of construction machinery can be effectively guaranteed to prevent customers from falsifying monitoring data, thereby achieving more accurate operating status monitoring.

[0079] In this embodiment, the edge computing node inputs the corresponding multi-dimensional monitoring parameters to be identified into the sub-feature extraction model to obtain the initial feature parameters of each dimension under the first constraint condition, including:

[0080] Through different edge computing nodes, the fuel consumption information, positioning information, start-stop records, running time and driving speed at continuous time points are input into different sub-feature extraction models to obtain the initial feature parameters of the fuel consumption information, positioning information, start-stop records, running time and driving speed under the corresponding first constraint conditions.

[0081] In this embodiment, the method for obtaining the first constraint condition includes:

[0082] A sub-feature extraction model is constructed using a machine learning model to obtain the initial first constraint condition.

[0083] The sub-feature extraction model is trained using historical data of a monitoring parameter of a certain dimension and corresponding manually corrected data as first training data until the training meets a pre-set first termination condition, thereby obtaining a first constraint condition. The first constraint condition is provided by the trained sub-feature extraction model.

[0084] In this embodiment, the multidimensional monitoring parameters to be identified are input into a central computing node, and the multidimensional monitoring parameters to be identified are input into a comprehensive feature extraction model through the central computing node to obtain comprehensive feature parameters of all dimensions under the second constraint condition, including:

[0085] A target characteristic parameter matrix is ​​constructed using the fuel consumption information, positioning information, start and stop records, operating time, and driving speed at the continuous time points.

[0086] The target feature parameter matrix is ​​input into the comprehensive feature extraction model through the central computing node to obtain the comprehensive feature parameters of the target feature parameter matrix under the second constraint condition.

[0087] In this embodiment, the method for obtaining the second constraint condition includes:

[0088] A comprehensive feature extraction model is constructed using a machine learning model to obtain the initial second constraint.

[0089] The comprehensive feature extraction model is trained using the historical target feature parameter matrix corresponding to the multi-dimensional monitoring parameter to be identified and the corresponding manual correction matrix as second training data until the training meets a pre-set second termination condition, thereby obtaining a second constraint condition. The second constraint condition is provided by the trained comprehensive feature extraction model.

[0090] Optionally, after obtaining the initial feature parameters and the comprehensive feature parameters, the two need to be fused, and the fusion method of the two is different depending on the data. More specifically, when the initial feature parameters and the comprehensive feature parameters are both revised data, the total dimension of all initial feature parameters is the same as the dimension of the comprehensive feature parameters, so the feature parameters of each dimension can be weighted and summed to obtain the final feature parameters of the multi-dimensional monitoring parameters to be identified. When the initial feature parameters and the comprehensive feature parameters are both warning levels, the two are combined into a new matrix to retain both spatial and temporal features. It is worth noting that even when the initial feature parameters and the comprehensive feature parameters are both revised data, the initial feature parameters and the comprehensive feature parameters can be combined into a new matrix.

[0091] In this embodiment, the final characteristic parameters of the multi-dimensional monitoring parameters to be identified are obtained based on the initial characteristic parameters and the comprehensive characteristic parameters, and the final characteristic parameters are identified by the central computing node to obtain the operating status of the construction machinery, including:

[0092] The initial characteristic parameters and the comprehensive characteristic parameters are fused in a data fusion manner to obtain the final characteristic parameters of the multi-dimensional monitoring parameters to be identified.

[0093] The operation status monitoring model pre-deployed on the central computing node is used to identify the final characteristic parameters to obtain the operation status of the construction machinery.

[0094] Optionally, an enhanced monitoring strategy can be set. When the operating status of the construction machinery is in an unreliable state, the enhanced monitoring strategy is called to update the operating status monitoring model and increase the data sampling frequency, thereby achieving more accurate data monitoring.

[0095] The final characteristic parameters of the multidimensional monitoring parameters to be identified are matrix parameters. Therefore, a convolutional neural network model can be used to construct an operation status monitoring model to realize the identification of the operation status of construction machinery.

[0096] In this embodiment, a training method for an operating status monitoring model is provided, specifically:

[0097] The running status monitoring model is initialized, and a hyperparameter group corresponding to the running status monitoring model is obtained, where each hyperparameter group includes all hyperparameters of the running status monitoring model.

[0098] Using historical data as input and manual labels as output, the objective function value of each hyperparameter group is obtained. The objective function value is set to the inverse or negative of the error function. When the derivative of the error function is not taken, the denominator is set to a small constant (such as 0.0001) to avoid the denominator being zero.

[0099] Guide the update of parameter individuals based on the optimal individual with the largest fitness, including:

[0100] X(t+1)=ω(t)*D'*e bl *cos(2πl)+X best (t)

[0101] ω(t)=ω i +(ω f -ω i )exp(-(αt / T max ) 2 )

[0102] D'=|Levy(λ)⊕X best (t)-X(t)|

[0103] Among them, X(t+1) represents the updated hyperparameter group, ω(t) represents the updated weight factor, D' represents the coefficient vector, e represents a natural constant, b represents a constant, l represents a random number between (-1,1), X best (t) represents the optimal hyperparameter group, X(t) represents the hyperparameter group to be updated, π represents pi, ω i represents the initial weight factor, ω f represents the final weight factor, α represents the convergence factor, α=2-(2 / (1+e -t / Tmax )), t represents the current update round number, T max Represents the maximum number of update rounds set, Levy(λ) represents a random search path that obeys the Levy distribution, and ⊕ represents point-to-point multiplication.

[0104] For the updated hyperparameter group, a secondary search is performed on each dimension of its parameters, and the search results are saved using a greedy algorithm to obtain the secondary updated hyperparameter group.

[0105] The secondary search can be:

[0106]

[0107] Among them, X(t+1) p represents the p-th dimension hyperparameter in the updated hyperparameter group, and the total dimension of the hyperparameter is D, κ p represents the updated X(t+1) p , π represents the ratio of circumference to circumference, and the step direction represents the direction in which the weight increases by the first step, step1.

[0108] When the fitness of a certain hyperparameter group increases for multiple consecutive times, a new individual is generated to replace the hyperparameter individual.

[0109] When the training reaches the preset number of times or the error function value is less than the set threshold, the hyperparameter group with the largest fitness value is obtained to obtain the target hyperparameters of the operation status monitoring model.

[0110] The above search algorithm takes into account both global search and local search, which can effectively improve the diversity of the training process while ensuring the training effect, improving the training accuracy, and avoiding falling into the local optimum.

[0111] Example 2

[0112] like Figure 2 As shown, the present invention provides a rental management system based on engineering machinery monitoring, including a data acquisition module 201, a first feature acquisition module 202, a second feature acquisition module 203, an operation monitoring module 204 and a warning management module 205.

[0113] The data acquisition module 201 is used to acquire the multi-dimensional monitoring parameters to be identified on the engineering machinery, and input the multi-dimensional monitoring parameters to be identified into the edge computing node of the corresponding dimension.

[0114] The first feature acquisition module 202 is used to input the corresponding multi-dimensional monitoring parameters to be identified into the sub-feature extraction model through the edge computing node to obtain the initial feature parameters of each dimension under the first constraint condition.

[0115] The second feature acquisition module 203 is configured to input the multidimensional monitoring parameters to be identified into a central computing node, which then inputs the multidimensional monitoring parameters to be identified into a comprehensive feature extraction model to obtain comprehensive feature parameters for all dimensions under the second constraint. The dimensions of the comprehensive feature parameters are the same as the sum of the dimensions of all initial feature parameters.

[0116] The operation monitoring module 204 is used to obtain the final characteristic parameters of the multi-dimensional monitoring parameters to be identified based on the initial characteristic parameters and the comprehensive characteristic parameters, and identify the final characteristic parameters through the central computing node to obtain the operation status of the construction machinery.

[0117] The warning management module 205 is used to send different alarm information and trajectory parameters of the construction machinery to the equipment terminal of the operation and maintenance personnel through the central computing node according to the operating status of the construction machinery, so as to complete the rental management of the construction machinery.

[0118] The rental management system based on engineering machinery monitoring described in this embodiment can execute the method and technical solution described in Example 1, and its beneficial effects and principles are similar, so they will not be repeated in this embodiment.

[0119] Those skilled in the art will readily appreciate other embodiments of the present invention after considering the specification and practicing the invention disclosed herein. The present invention is intended to cover any variations, uses, or adaptations of the present invention that follow the general principles of the present invention and include common knowledge or customary techniques in the art not disclosed herein. It should be understood that the present invention is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and variations can be made without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

Claims

1. A leasing management method based on engineering machinery monitoring, characterized in that: include: Obtaining multi-dimensional monitoring parameters to be identified on the engineering machinery, and inputting the multi-dimensional monitoring parameters to be identified into the edge computing node of the corresponding dimension; Inputting the corresponding multi-dimensional monitoring parameters to be identified into the sub-feature extraction model through the edge computing node to obtain the initial feature parameters of each dimension under the first constraint condition; Inputting the multidimensional monitoring parameters to be identified into a central computing node, and inputting the multidimensional monitoring parameters to be identified into a comprehensive feature extraction model through the central computing node to obtain comprehensive feature parameters of all dimensions under the second constraint condition; the dimension of the comprehensive feature parameter is the same as the sum of the dimensions of all initial feature parameters; Obtaining final characteristic parameters of the multi-dimensional monitoring parameters to be identified based on the initial characteristic parameters and the comprehensive characteristic parameters, and identifying the final characteristic parameters through a central computing node to obtain the operating status of the construction machinery; Based on the operating status of the construction machinery, different alarm information and the trajectory parameters of the construction machinery are sent to the equipment terminals of the operation and maintenance personnel through the central computing node to complete the rental management of the construction machinery; Obtain the multi-dimensional monitoring parameters to be identified on the construction machinery, including: The monitoring device on the engineering machinery samples the multi-dimensional monitoring data to be identified within a time period according to a preset sampling frequency to obtain a sampling data packet; The central computing node and the edge computing node respectively obtain the sample data packets stored in the monitoring equipment on the engineering machinery with a preset first delay condition to obtain the multi-dimensional monitoring parameters to be identified; The edge computing node inputs the corresponding multi-dimensional monitoring parameters to be identified into the sub-feature extraction model to obtain the initial feature parameters of each dimension under the first constraint condition, including: Inputting fuel consumption information, positioning information, start-stop records, operating time, and driving speed at consecutive time points into different sub-feature extraction models through different edge computing nodes to obtain initial feature parameters of fuel consumption information, positioning information, start-stop records, operating time, and driving speed under the corresponding first constraint condition; Methods for obtaining the first constraint condition include: Build a sub-feature extraction model using a machine learning model to obtain the initial first constraint condition; Using historical data of a certain dimensional monitoring parameter and corresponding manually corrected data as first training data, the sub-feature extraction model is trained using the first training data until the training meets a pre-set first end condition, thereby obtaining a first constraint condition; the first constraint condition is provided by the trained sub-feature extraction model; The multidimensional monitoring parameters to be identified are input into a central computing node, and the multidimensional monitoring parameters to be identified are input into a comprehensive feature extraction model through the central computing node to obtain comprehensive feature parameters of all dimensions under the second constraint condition, including: Constructing a target characteristic parameter matrix based on the fuel consumption information, positioning information, start and stop records, operating time, and driving speed at the consecutive time points; Inputting the target feature parameter matrix into the comprehensive feature extraction model through the central computing node to obtain comprehensive feature parameters of the target feature parameter matrix under the second constraint condition; Methods for obtaining the second constraint condition include: Build a comprehensive feature extraction model using a machine learning model to obtain the initial second constraint; The comprehensive feature extraction model is trained using the historical target feature parameter matrix corresponding to the multi-dimensional monitoring parameter to be identified and the corresponding manual correction matrix as second training data until the training meets a pre-set second end condition, thereby obtaining a second constraint condition; the second constraint condition is provided by the trained comprehensive feature extraction model; According to the initial characteristic parameters and the comprehensive characteristic parameters, final characteristic parameters of the multi-dimensional monitoring parameters to be identified are obtained, and the final characteristic parameters are identified by the central computing node to obtain the operating status of the construction machinery, including: The initial characteristic parameters and the comprehensive characteristic parameters are fused by data fusion to obtain the final characteristic parameters of the multi-dimensional monitoring parameters to be identified; The operation status monitoring model pre-deployed on the central computing node is used to identify the final characteristic parameters to obtain the operation status of the construction machinery.

2. The leasing management method based on engineering machinery monitoring according to claim 1 is characterized in that: The multi-dimensional monitoring parameters to be identified on the engineering machinery include: fuel consumption information, positioning information, start and stop records, operating time and driving speed at consecutive time points.

3. The leasing management method based on engineering machinery monitoring according to claim 1 is characterized in that: Also includes: Acquiring anti-tampering parameters of the monitoring device according to a preset second time delay condition, wherein the anti-tampering parameters include at least a first sensing time sequence of the pressure sensor, a second sensing time sequence of the vibration sensor, and a third sensing time sequence of the distance sensor; Obtaining an anti-dismantling feature matrix based on the first sensing time series, the second sensing time series, and the third sensing time series; identifying the anti-dismantling feature matrix using a pre-trained machine learning model at a central computing node to obtain a disassembly status of the monitoring device, where the disassembly status includes whether the monitoring device is disassembled or operating normally; When the disassembly status indicates that the monitoring device is disassembled, a warning message indicating that the monitoring device is disassembled is sent to the device terminal of the operation and maintenance personnel through the central computing node, and the real-time positioning information or the last acquired positioning information in the monitoring device is transmitted to the device terminal of the operation and maintenance personnel to realize anti-disassembly monitoring of the monitoring device, thereby ensuring the accuracy of the multi-dimensional monitoring parameters to be identified.

4. A leasing management system based on engineering machinery monitoring, which is capable of executing the leasing management method based on engineering machinery monitoring according to any one of claims 1 to 3, characterized in that: It includes a data acquisition module, a first feature acquisition module, a second feature acquisition module, an operation monitoring module and a warning management module; The data acquisition module is used to obtain the multi-dimensional monitoring parameters to be identified on the engineering machinery, and input the multi-dimensional monitoring parameters to be identified into the edge computing node of the corresponding dimension; The first feature acquisition module is used to input the corresponding multi-dimensional monitoring parameter to be identified into the sub-feature extraction model through the edge computing node to obtain the initial feature parameters of each dimension under the first constraint condition; The second feature acquisition module is used to input the multidimensional monitoring parameters to be identified into a central computing node, and input the multidimensional monitoring parameters to be identified into a comprehensive feature extraction model through the central computing node to obtain comprehensive feature parameters of all dimensions under the second constraint condition; the dimension of the comprehensive feature parameter is the same as the sum of the dimensions of all initial feature parameters; The operation monitoring module is used to obtain the final characteristic parameters of the multi-dimensional monitoring parameters to be identified based on the initial characteristic parameters and the comprehensive characteristic parameters, and identify the final characteristic parameters through the central computing node to obtain the operating status of the construction machinery; The warning management module is used to send different alarm information and trajectory parameters of the construction machinery to the equipment terminal of the operation and maintenance personnel through the central computing node according to the operating status of the construction machinery, thereby completing the rental management of the construction machinery.

Images