Electric mine truck power system energy efficiency optimization method based on deep learning

Abstract

The invention relates to the field of computer technology application, in particular to an electric mine truck power system energy efficiency optimization method based on deep learning, and the method comprises the steps: obtaining to-be-processed information; based on the driving resource influence parameter information of the moving object and a driving resource prediction model, obtaining a corresponding driving resource estimated value as an initial driving resource estimated value of the moving object; based on user information of a moving object and a target driving resource estimated value, correcting the initial driving resource estimated value to obtain a corresponding correction result, and taking the correction result as the target driving resource estimated value; based on the target driving resource estimated value, generating a driving resource mapping relation graph of the moving object and performing visual display; and in response to the received control instruction generation information, generating a control instruction acting on a power system of the moving object based on the driving resource mapping relation graph, and outputting the generated control instruction. According to the invention, the energy efficiency of the power system of the electric mine truck can be optimized.

Description

Technical Field

[0001] The present invention relates to the field of computer technology applications, and particularly to an energy efficiency optimization method for the power system of an electric mining truck based on deep learning. Background Art

[0002] With the accelerating penetration of the electrification process in the transportation field, special transportation equipment with the characteristics of fixed operation paths is undergoing a systematic power transformation. Among them, the electrification replacement of mining trucks (mining trucks) has become a core measure for the green upgrading of mines. However, electric mining trucks face severe energy consumption management challenges under the complex working conditions of open-pit mines: the unstructured terrain leads to dynamic changes in slopes (-15% to +25%), and heavy-load transportation causes continuous high-power demands (peak value ≥ 800 kW). Under this background, developing an intelligent energy distribution strategy based on working condition perception to achieve two-dimensional spatio-temporal optimization scheduling of the limited power of the power battery has become a decisive factor in improving transportation efficiency. Summary of the Invention

[0003] For the above technical problems, the technical solution adopted by the present invention is as follows: An embodiment of the present invention provides an energy efficiency optimization method for the power system of an electric mining truck based on deep learning. The method is executed by a computer program, and the method includes the following steps: S100, obtaining information to be processed; the information to be processed includes information on the driving resource impact parameters of a moving object and user information, and the information on the driving resource impact parameters includes terrain data of a target path and attribute parameters of the moving object.

[0004] S200, based on the information on the driving resource impact parameters of the moving object and a driving resource prediction model, obtaining a corresponding predicted value of the driving resource as the initial predicted value of the driving resource of the moving object.

[0005] S300, based on the user information of the moving object, correcting the initial predicted value of the driving resource to obtain a corresponding correction result as the target predicted value of the driving resource.

[0006] S400, based on the target predicted value of the driving resource, generating a driving resource mapping relationship diagram of the moving object and performing visual display; wherein, the driving resource mapping relationship diagram includes the mapping relationship between terrain data and driving resources.

[0007] S500, in response to receiving a control instruction generation message, generating a control instruction for the power system of the moving object based on the driving resource mapping relationship diagram and outputting the generated control instruction.

[0008] The present invention has at least the following beneficial effects: The energy efficiency optimization method for the power system of an electric mining truck based on deep learning provided by the embodiments of the present invention can, based on the terrain data of the target path and the attribute parameters of the moving object, obtain an initial estimated value of the driving resources required for the moving object to travel the target path through deep learning, correct the initial estimated value of the driving resources based on user information, and generate control instructions for the power system of the moving object based on the received control instruction generation information. In actual application scenarios, it can improve the prediction accuracy and eliminate the energy consumption deviation caused by operation differences among individuals, and thus can effectively optimize the energy efficiency of the power system of the electric mining truck.

[0009] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present invention, nor is it used to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0011] Figure 1 It is a flowchart of the energy efficiency optimization method for the power system of an electric mining truck based on deep learning provided by the embodiments of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0012] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0013] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs. The terms used in the description of the present invention in this specification are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0014] It should be noted that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts depict the steps as sequential processes, many of the steps can be implemented in parallel, concurrently, or simultaneously. In addition, the order of the steps can be rearranged. The process can be terminated when its operations are completed, but there can also be additional steps not included in the drawings. The process can correspond to a method, function, procedure, subroutine, subprogram, etc.

[0015] An embodiment of the present invention provides an energy efficiency optimization method for the power system of an electric mining truck based on deep learning, and the method is executed by a computer program. As Figure 1 shown, the method includes the following steps: S100, obtaining information to be processed; the information to be processed includes information on the driving resource impact parameters of the moving object and user information, and the driving resource impact parameter information includes the terrain data of the target path and the attribute parameters of the moving object.

[0016] In the embodiment of the present invention, the moving object is an electric mining truck. The driving resource refers to the energy consumption value, that is, the power consumption. The target path is the road that the moving object needs to pass through.

[0017] In the embodiment of the present invention, it aims to predict the energy consumption value required to pass through the target path based on the terrain data of the target path of the moving object and its own attribute parameters through simplified feature engineering. Therefore, in the embodiment of the present application, only the influencing factors with a greater impact on the energy consumption value are selected for analysis.

[0018] In the embodiment of the present invention, the terrain data of the target path may include the path length, road surface damage information, uphill ramp information, and road surface material. In a schematic embodiment, the terrain data can be data obtained by on-site measurement. In a preferred embodiment, the terrain data can be obtained based on the monitoring data of vibration sensors and the collected high-precision map. Among them, the monitoring data based on vibration sensors can obtain the pothole information of the target path, and other terrain data of the target path can be obtained based on the high-precision map.

[0019] In the embodiment of the present invention, the attribute parameters of the moving object may include the load, remaining power, and remaining battery life.

[0020] In the embodiment of the present invention, the uphill ramp is a ramp that requires the moving object to climb. The road surface material of the target path can be the same road surface material or different, and can be determined based on the actual situation.

[0021] S200, obtaining the corresponding predicted value of the driving resource based on the driving resource impact parameter information of the moving object and the driving resource prediction model.

[0022] In an embodiment of the present invention, the driving resource prediction model may be a trained deep learning model, and the deep learning model may be an existing deep learning model. For example, it may be a hybrid network model of a convolutional neural network and a long short-term memory artificial neural network.

[0023] In an embodiment of the present invention, the driving resource prediction model may be obtained by training based on a training data set. The training data set may be a collected historical data set and an enhanced data set obtained by enhancing the historical data set. Each piece of data includes information on the driving resource impact parameters of the corresponding moving object and the actual energy consumption value. During the training process, the encoded features corresponding to the driving resource impact parameter information are used as input information, and the actual energy consumption value is used as a label to train the original deep learning model.

[0024] Those skilled in the art know that any method of training the original deep learning model based on a training data set to obtain a trained deep learning model belongs to the protection scope of the present invention. The data enhancement method may be an existing data enhancement method, such as generating terrain data based on a GAN network.

[0025] Further, S200 may specifically include: S201, generating a path state feature vector and an attribute state feature vector of the moving object based on the driving resource impact parameter information.

[0026] S202, inputting the path state feature vector and the attribute state feature vector into the driving resource prediction model to obtain a corresponding predicted driving resource value.

[0027] Among them, the path state feature vector includes a path length feature, a road surface damage feature, an uphill ramp feature, and a road surface material feature. The path length feature is determined based on the path length and the set path length level determination information. The road surface damage feature is determined based on the road surface damage characterization value and the set road surface damage level determination information. The uphill ramp feature is determined based on the slope feature index and the set uphill ramp level determination information. The road surface material feature is determined based on the road surface material characterization value and the set road surface material level determination information.

[0028] In an embodiment of the present invention, the set path length level determination information may be obtained based on existing data. For example, it may be obtained based on the path length information in the training data set. Specifically, the path length may be divided into different levels based on the path length information in the training data set, and each level corresponds to a path length range. The specific division method may be implemented using existing methods.

[0029] The determined information of the road surface damage level can be obtained based on the road surface damage information in the training dataset. Specifically, the road surface damage degree can be divided into different levels based on the road surface damage information in the training dataset, and each level corresponds to a range of road surface damage characterization values. The specific division method can be implemented using existing methods.

[0030] The determined information of the uphill grade can be obtained based on the uphill information in the training dataset. Specifically, the uphill can be divided into different grades based on the uphill information in the training dataset, and each grade corresponds to a range of slope characteristic indices. The specific division method can be implemented using existing methods.

[0031] The determined information of the road surface material grade can be obtained based on the road surface material information in the training dataset. Specifically, the uphill road surface materials can be divided into different grades based on the road surface material information in the training dataset, and each grade corresponds to a range of road surface material characterization values.

[0032] In an embodiment of the present invention, the attribute status feature vector may include a load feature, a remaining power feature, and a battery remaining life feature. Among them, the load feature is determined based on the load and the determined information of the set load grade, the remaining power feature is determined based on the remaining power and the determined information of the set remaining power grade, and the battery remaining life feature is determined based on the battery remaining life and the determined information of the set battery remaining life grade.

[0033] In an embodiment of the present invention, the determined information of the set load grade, the determined information of the set remaining power grade, and the determined information of the set battery remaining life grade can all be determined based on the load information, the remaining power information, and the battery remaining life in the training dataset. Specifically, the load can be divided into different grades based on the load information in the training dataset, and each grade corresponds to a range of loads. The remaining power can be divided into different grades based on the remaining power information in the training dataset, and each grade corresponds to a range of remaining power. The battery remaining life can be divided into different grades based on the battery remaining life information in the training dataset, and each grade corresponds to a range of battery remaining life.

[0034] In an embodiment of the present invention, different levels of different features can be distinguished using different identification values.

[0035] Further, in an embodiment of the present invention, the road surface damage information may include pothole information, and the pothole information includes pothole depth and edge angle; the road surface damage characterization values of the moving object include a first pothole characterization value and a second pothole characterization value, where the first characterization value is used to characterize the steepness of the pothole edge, and the second characterization value is used to characterize the road surface roughness.

[0036] Specifically, the first pothole characterization value satisfies the following conditions: .

[0037] Among them, CV1 is the first pit characterization value, d i is the depth of the i-th pit, θ i is the edge angle of the i-th pit, f1() is the first function expression, and the value range of i is from 1 to m1, where m1 is the number of pits; f1(d i , θ i ) = d i × tanθ i .

[0038] In the embodiment of the present invention, the edge angle of the pit is equal to the included angle between the edge of the pit and the vertical direction.

[0039] Furthermore, the second pit characterization value satisfies the following conditions: .

[0040] Among them, CV2 is the second pit characterization value, f2() is the second function expression, N j is the number of pits within the j-th unit path length in the target path, the value range of j is from 1 to n, where n is the number of unit path lengths, Avgd j is the average depth of the pits within the j-th unit path length, wp j is the pit weight corresponding to the j-th unit path length, f2(N j , Avgd j ) = N j × Avgd j .

[0041] In the embodiment of the present invention, the unit path length can be set according to actual needs, for example, it can be 100 meters.

[0042] Furthermore, in the embodiment of the present invention, , N b is the number of pits within the b-th unit path length in the target path, the value range of b is from 1 to n, Avgd b is the average depth of the pits within the b-th unit path length.

[0043] Furthermore, in the embodiment of the present invention, the uphill ramp information may include the uphill ramp length and the slope angle.

[0044] Those skilled in the art know that the road surface damage characteristics may include two characteristic values, that is, two characteristic values determined based on the first pit characterization value and the second pit characterization value respectively.

[0045] Among them, the slope characteristic index of the moving object includes the first slope characteristic index and the second slope characteristic index. Among them, the first slope characteristic index characterizes the elevation change degree of the road surface, and the second slope characteristic index characterizes the flatness degree of the road surface.

[0046] Further, the first slope feature index satisfies the following condition: .

[0047] Wherein, RV1 is the first slope feature index, f3() is the third function expression, L r is the length of the r-th uphill lane in the uphill lane information, the value of r ranges from 1 to m2, and m2 is the number of uphill lanes in the uphill lane information, α r is the slope angle of the r-th uphill lane; f3(L r , α r ) = L r ×Sinα r .

[0048] Further, the second slope feature index satisfies the following condition: .

[0049] Wherein, f4() is the fourth function expression, Avgα is the average slope angle obtained based on the uphill lane information, and Maxα is the maximum slope angle in the uphill lane information.

[0050] In the embodiment of the present invention, f4(Avgα, Maxα) = w1×Avgα + w2×Maxα, where w1 is the weight corresponding to the average slope angle, w2 is the weight corresponding to the maximum slope angle, w1 and w2 can be experimental values, and w1 + w2 = 1. In a schematic embodiment, w1 can be greater than w2.

[0051] Those skilled in the art know that the uphill lane feature can include two feature values, namely, two feature values determined based on the first slope feature index and the second slope feature index respectively.

[0052] Further, in the embodiment of the present invention, the road surface material characterization value can satisfy the following condition: .

[0053] Wherein, MV is the road surface material characterization value, LM s is the resistance coefficient of the s-th road surface material in the target path, the value of s ranges from 1 to q, and q is the number of types of road surface materials included in the target path, W s is the weight of the s-th road surface material, which can be an experimental value. Generally, the greater the resistance coefficient of the road surface material, the greater the weight.

[0054] S300. Based on the user information of the moving object, correct the initial estimated value of the driving resource to obtain the corresponding correction result as the target estimated value of the driving resource.

[0055] In an embodiment of the present invention, the user information corresponding to the moving object is the ID of the user who performs the control operation on the moving object, that is, the ID of the driver. The ID of the user can be the identity identifier of the user, for example, it can be the driver's license number of the user.

[0056] Further, S300 may specifically include: Based on the user information, obtain the correction coefficient corresponding to the user information.

[0057] Use the correction coefficient to correct the initial predicted value of the driving resources of the moving object to obtain the target predicted value of the driving resources.

[0058] Further, the obtaining the correction coefficient corresponding to the user information based on the user information specifically includes: If the user information does not exist in the preset correction coefficient reference information table, determine that the correction coefficient corresponding to the user information is the reference correction coefficient; if the user information exists in the preset correction coefficient reference information table, use the correction coefficient in the preset correction coefficient reference information table as the correction coefficient corresponding to the user information.

[0059] In an embodiment of the present invention, the preset correction coefficient reference information stores the correction coefficient corresponding to each user. The correction coefficient of each user is determined based on the driving behavior habits of the user. Users with different driving behavior habits have different correction coefficients. The correction coefficient of each user can be determined based on the actual situation or historical operation behavior.

[0060] In an embodiment of the present invention, the reference correction coefficient can be a coefficient greater than 1 and can be an empirical value.

[0061] In an embodiment of the present invention, the target predicted value of the driving resources can be equal to the product of the correction coefficient and the initial predicted value of the driving resources.

[0062] In an embodiment of the present invention, by predicting the initial predicted value of the driving resources based on the user's driving behavior habits, it is possible to eliminate the energy consumption deviation caused by the operation differences among different people, making the obtained predicted value of the driving resources more accurate.

[0063] S400, generate a driving resource mapping relationship diagram of the moving object based on the target predicted value of the driving resources and perform visual display; wherein, the driving resource mapping relationship diagram includes the mapping relationship between the terrain data and the driving resources.

[0064] In an embodiment of the present invention, the driving resource mapping relationship diagram can be obtained based on the existing method, as long as it can display the available driving resources corresponding to each characteristic section on the target path.

[0065] In an embodiment of the present invention, the characteristic road sections include flat road sections, uphill ramps, downhill ramps, and potholes. In the driving resource mapping relationship diagram, the available driving resources corresponding to each flat road section, uphill ramp, downhill ramp, and pothole will be displayed.

[0066] Those skilled in the art know that the present invention knows that a pothole should belong to a region. However, for the convenience of description, the present invention approximately regards the pothole as a micro road section, but this will not affect the essence of the solution of the present invention.

[0067] In a schematic embodiment of the present invention, the available driving resources of each characteristic road section can be obtained based on a trained driving resource allocation model, and specifically can be obtained through the following steps: S401: Input the path state feature vector, attribute state feature vector, and target driving resource pre-estimation value of the moving object into the trained driving resource allocation model to obtain the available driving resources of the uphill road section and the available driving resources of the pothole section.

[0068] Among them, the trained driving resource allocation model can be a deep learning model, which can be trained based on a historical data set. During the training process, the input of the model is the path state feature vector, attribute state feature vector, and the actual driving resources with added noise of the moving object, and the output is the available driving resources of the uphill road section and the available driving resources of the pothole section. Among them, the added noise can follow the prediction error distribution, such as the Gaussian distribution N(0,σ²).

[0069] Those skilled in the art know that the specific training method of the driving resource allocation model can belong to the prior art.

[0070] S402: Use the remaining driving resources obtained by subtracting the available driving resources of the uphill road section and the available driving resources of the pothole section from the target driving resource pre-estimation value as the available driving resources of the flat road section and the downhill ramp.

[0071] S403: Based on the available driving resources of the uphill road section, determine the available driving resources corresponding to each uphill ramp in the target path; based on the available driving resources of the pothole section, determine the available driving resources corresponding to each pothole in the target path; and based on the available driving resources of the flat road section and the downhill ramp, determine the available driving resources corresponding to each flat road section and each uphill ramp in the target path.

[0072] Specifically, the available driving resources corresponding to each uphill ramp can meet the following conditions: UR r = (f3(L r , α r ) / RV1) × TUR.

[0073] Among them, UR ris the available driving resource value of the r-th uphill ramp in the target path, and TUR is the available driving resource of the uphill road section.

[0074] The available driving resources corresponding to each pothole can satisfy the following conditions: PR i = (f1(d i , θ i )) / CV1) × TPR.

[0075] where PR i is the available driving resource of the i-th pothole in the target path, and TPR is the available driving resource of the pothole section.

[0076] In the embodiments of the present invention, for simplicity of calculation, information such as resource savings or braking requirements on downhill ramps is not considered, and flat sections and downhill ramps are regarded as the same section. Specifically, the available driving resources of each flat section or downhill ramp can satisfy the following conditions: FR = WF × (LF / TLF) × TFR.

[0077] where FR is the available driving resource of each flat section or each downhill ramp, LF is the path length of each flat section or each downhill ramp, TLF is the total path length of all flat sections and the downhill ramps in the target path, TFR is the available driving resource of flat sections and downhill ramps, WF is the weight corresponding to each flat section or each downhill ramp, all flat sections have the same weight, all downhill ramps have the same weight, and the specific weight can be determined based on actual situations. For example, the weight of flat sections can be greater than the weight of downhill ramps.

[0078] S500, in response to receiving a control instruction to generate information, based on the driving resource mapping diagram, generate a control instruction for the power system of the moving object, and output the generated control instruction.

[0079] In the embodiments of the present invention, control instructions for controlling the power system of the moving object can be generated based on the available driving resources of each characteristic road section. The control instructions can be set according to actual needs as long as it can ensure that the energy consumption of each characteristic road section is less than or equal to the corresponding available driving resources. For example, it can include torque distribution instructions, power limit instructions, shift strategy instructions, etc. acting on each characteristic road section.

[0080] In the embodiments of the present invention, the control instruction generation information can be generated by user input, or generated based on an automatic trigger instruction set by the control system of the moving object, that is, once it is detected that the driving resource mapping diagram is generated, the control instruction generation information is automatically triggered to be generated. The control instruction will be sent to the power system of the moving object to guide the power output of the power system. It should be noted that in the embodiments of the present invention, the control instruction indicates a reference information, rather than an instruction used to forcibly restrict the power system to output according to the control instruction. In actual application scenarios, the user may change the output of the power system based on their own driving behavior habits, that is, the driver's final decision-making power is reserved.

[0081] Furthermore, in the embodiments of the present invention, the following steps are further included: S600, when receiving the monitoring data obtained by monitoring the driving resources of the moving object during driving, update the current driving resource mapping diagram based on the received monitoring data, and generate a control instruction for the power system of the moving object based on the updated driving resource mapping diagram. The initial value of the current driving resource mapping diagram is the initial driving resource mapping diagram.

[0082] In the embodiments of the present invention, the monitoring data can be obtained through an energy consumption monitoring system on the moving object.

[0083] Furthermore, S600 specifically includes: S601, based on the received monitoring data, determine whether the current characteristic section passed by the moving object is the last section of the target path. If so, calculate the driving resources actually used by the moving object to pass the target path, and execute S606; if not, execute S602.

[0084] S602, obtain the used driving resources used by the moving object to pass the current characteristic section as the current used driving resources, and compare the current used driving resources with the current available driving resources corresponding to the current characteristic section. If the used driving resources are greater than the corresponding current available driving resources, execute S603; otherwise, execute S601.

[0085] S603, update the current available driving resources based on the current used driving resources, that is, subtract the current used driving resources from the current available driving resources to obtain the updated available driving resources, and execute S604; where the initial value of the current available driving resources is the target driving resource estimated value.

[0086] S604. Use the updated available driving resources as the current available driving resources, determine the current feature vector of the moving object based on the terrain data and attribute parameters of the current remaining driving path of the moving object, and update the current driving resource mapping graph based on the current feature vector and the current available driving resources, then execute S605.

[0087] In the embodiment of the present invention, the current feature vector includes the current path state feature vector and the current attribute state feature vector. The acquisition method of the current feature vector can be obtained with reference to the implementation content of the foregoing S200, that is, the current path state feature vector is obtained based on the terrain data included in the current remaining driving path, and the current attribute state feature vector is obtained based on the current attribute parameters of the moving object. The current available driving resources can be obtained with reference to the implementation content of the foregoing S400, and the specific implementation of updating the current driving resource mapping graph also refers to the specific implementation content of S400. The initial value of the current driving resource mapping graph is the driving resource mapping graph generated by S400.

[0088] In the embodiment of the present invention, the update of the driving resource mapping graph can be carried out in a manner of edge computing plus cloud collaboration.

[0089] S605. Generate a control instruction for the power system acting on the moving object based on the current driving resource mapping graph, and then execute S601.

[0090] S606. If the difference between the driving resources actually used by the moving object through the target path and the corresponding target driving resource estimate is greater than 0 and the difference is greater than the set value, set the current operation identifier of the user corresponding to the moving object as the first identifier; otherwise, that is, if the driving resources actually used by the moving object through the target path are less than or equal to the corresponding target driving resource estimate, or if the driving resources actually used by the moving object through the target path are greater than the corresponding target driving resource estimate but the difference is less than or equal to the set value, set the current operation identifier of the user corresponding to the moving object as the second identifier, store the current operation identifier in the corresponding operation identifier record set, and exit the current control program, that is, end the power system energy consumption optimization process of the current target path.

[0091] In the embodiment of the present invention, the set value can be set according to actual needs. The initial value of the operation identifier record set for each user can be empty. The first identifier and the second identifier can be represented by different numbers. Those skilled in the art know that the operation identifier record set is used to record the difference information between the driving resources actually used by the user each time passing through a certain target path and the corresponding target driving resource estimate.

[0092] Further, in the embodiments of the present invention, the correction coefficient of each user in the preset correction coefficient reference information table can be updated based on the corresponding operation identification record set, specifically based on the last Q operation identifications in the operation identification record set. Further, the correction coefficient of each user can be updated through the following steps: S1. Obtain the quantity of the first identification among the last Q operation identifications in the operation identification record set corresponding to this user. If the quantity of the first identification is greater than the set quantity, it indicates that the operation behavior habit of this user has changed and the current correction coefficient needs to be updated, then execute S2; otherwise, do not update the current correction coefficient of the user.

[0093] In the embodiments of the present invention, Q can be an empirical value. The initial value of the current correction coefficient of the user is the reference correction coefficient. The set quantity can be an empirical value. For example, it can be equal to c×Q, where c can be a number greater than 0.5, for example, it can be set to 0.8.

[0094] S2. Update the current correction coefficient of the user in the following manner: k = kc + △k.

[0095] Wherein, k is the updated correction coefficient of the user, kc is the current correction coefficient of the user, and △k is the preset correction coefficient change value, which can be an empirical value.

[0096] Theoretically, the driving behavior habit of a user is generally fixed, but the situation of change cannot be excluded. For example, the target path of the moving object operated by the user has changed, resulting in a great change in the terrain data, and / or the own attributes of the moving object have changed, leading to a certain change in the driving behavior habit of the user.

[0097] In the embodiments of the present invention, the correction coefficient of each user can be updated based on the corresponding current operation identification record set, which can make the correction coefficient of the user more accurate.

[0098] Those skilled in the art know that all the data that needs to be processed in the embodiments of the present invention can be obtained from the corresponding devices through the computer interface of the computer device.

[0099] The embodiments of the present invention further provide an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are configured to execute the method described in the embodiments of the present invention.

[0100] The embodiments of the present invention further provide a computer-readable storage medium, storing computer-executable instructions, and the computer instructions are used to execute the method described in the embodiments of the present invention.

[0101] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in the present invention can be executed in parallel, sequentially, or in different orders, as long as the desired results of the technical solution disclosed in the present invention can be achieved, and no limitations are imposed herein.

[0102] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub - combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. An energy efficiency optimization method for the power system of an electric mining truck based on deep learning, characterized in that, The method is executed by a computer program, and the method includes the following steps: S100. Obtain the information to be processed; the information to be processed includes the driving resource influence parameter information of the moving object and the user information, and the driving resource influence parameter information includes the terrain data of the target path and the attribute parameters of the moving object; the terrain data of the target path includes the path length, road surface damage information, uphill ramp information, and road surface material; the attribute parameters of the moving object include the load, remaining power, and remaining battery life; S200. Based on the driving resource influence parameter information of the moving object and the driving resource prediction model, obtain the corresponding predicted driving resource value as the initial predicted driving resource value of the moving object; S300. Based on the user information of the moving object, correct the initial predicted driving resource value to obtain the corresponding correction result as the target predicted driving resource value; S400. Based on the target predicted driving resource value, generate a driving resource mapping relationship diagram of the moving object and perform visual display; wherein, the driving resource mapping relationship diagram includes the mapping relationship between the terrain data and the driving resource; S500. In response to receiving the control instruction generation information, based on the driving resource mapping relationship diagram, generate a control instruction for the power system of the moving object and output the generated control instruction.

2. The method according to claim 1, wherein S200 specifically includes: S201. Based on the driving resource influence parameter information, generate a path state feature vector and an attribute state feature vector of the moving object; S202. Input the path state feature vector and the attribute state feature vector into the driving resource prediction model to obtain the corresponding predicted driving resource value; Among them, the path state feature vector includes a path length feature, a road surface damage feature, an uphill ramp feature, and a road surface material feature. The path length feature is determined based on the path length and the set path length level determination information. The road surface damage feature is determined based on the road surface damage characterization value and the set road surface damage level determination information. The uphill ramp feature is determined based on the slope feature index and the set uphill ramp level determination information. The road surface material feature is determined based on the road surface material characterization value and the set road surface material level determination information; The attribute state feature vector includes a load feature, a remaining power feature, and a remaining battery life feature. The load feature is determined based on the load and the set load level determination information. The remaining power feature is determined based on the remaining power and the set remaining power level determination information. The remaining battery life feature is determined based on the remaining battery life and the set remaining battery life level determination information.

3. The method according to claim 1, characterized in that The road surface damage information includes pothole information, and the pothole information includes the pothole depth and the edge angle; the road surface damage characterization value of the moving object includes a first pothole characterization value and a second pothole characterization value; The first pit characterization value satisfies the following conditions: ; Among them, CV1 is the first pit characterization value, d i is the depth of the i-th pit, θ i is the edge angle of the i-th pit, f1() is the first function expression, and the value range of i is from 1 to m1, where m1 is the number of pits; The second pothole characterization value satisfies the following conditions: ； where CV2 is the second pit characterization value, f2() is the second function expression, N j is the number of pits within the j-th unit path length in the target path, where j ranges from 1 to n, and n is the number of unit path lengths, Avgd j is the average depth of the pits within the j-th unit path length, wp j is the pit weight corresponding to the j-th unit path length.

4. The method according to claim 3, wherein f1(d i , θ i ) = d i × tan θ i .

5. The method according to claim 3, wherein f2(N j , Avgd j ) = N j × Avgd j .

6. The method according to claim 1, characterized in that, The uphill ramp information includes the uphill ramp length and the slope angle; the slope feature index of the moving object includes a first slope feature index and a second slope feature index; The first slope characteristic index satisfies the following conditions: ; Among them, RV1 is the first slope feature index, f3( ) is the third function expression, and L r is the length of the r-th uphill lane in the uphill lane information, where r ranges from 1 to m2, and m2 is the number of uphill lanes in the uphill lane information, and α r is the slope angle of the r-th uphill lane; The second slope characteristic index satisfies the following conditions: ; Among them, f4( ) is the fourth function expression, Avgα is the average slope angle obtained based on the uphill ramp information, and Maxα is the maximum slope angle in the uphill ramp information.

7. The method according to claim 6, wherein f3(L r , α r ) = L r × Sinα r .

8. The method according to claim 6, characterized in that f4(Avgα, Maxα) = w1 × Avgα + w2 × Maxα, where w1 is the weight corresponding to the average slope angle and w2 is the weight corresponding to the maximum slope angle.

9. The method according to claim 1, characterized in that S300 specifically includes: Obtaining a correction coefficient corresponding to the user information based on the user information; Using the correction coefficient to correct the initial estimated value of the driving resources of the moving object to obtain the target estimated value of the driving resources.

10. The method according to claim 1, wherein It further includes the following steps: S600, when receiving the monitoring data obtained by monitoring the driving resources of the moving object during the driving process, updating the current driving resource mapping diagram based on the received monitoring data, and generating a control instruction for the power system of the moving object based on the updated driving resource mapping diagram. The initial value of the current driving resource mapping diagram is the initial driving resource mapping diagram.

Images