CN117930804A - Fault processing equipment and processing method for mining truck - Google Patents

Abstract

The invention discloses fault processing equipment for a mining truck, which comprises a complete vehicle fault controller, a central gateway, a CAN data recorder, a fault display instrument and a cloud server, wherein a VCU controller is adopted, a CAN communication network is utilized, all faults of the mining truck, such as hard line faults, CAN line faults, operation logic faults and the like, are integrated and standardized and are sent to visual hardware, such as a display screen and the like, and are uploaded to the cloud in real time.

Description

Fault processing equipment and processing method for mining truck

Technical Field

The invention relates to the technical field of CAN communication, in particular to fault processing equipment and a fault processing method for a mining truck.

Background

Mining trucks are indispensable equipment in mine production and mainly used for material transportation inside mines. Because the mining truck is subjected to a severe working environment and challenges during the use process, faults often occur, and for the faults, the faults need to be timely checked and repaired so as to ensure the normal operation of the mining truck and even mine production. In the development of mining trucks, the processing of faults, for example, the transmission to a display screen via a CAN network for visualization and subsequent optimization of the processing, is thus a necessary development. The existing fault processing scheme has the defects of imperfect performance, poor linkage and completion degree, insufficient equipment independence, high cost depending on mining sites and the like.

The technical scheme adopts the VCU controller and utilizes the CAN communication network, and based on the self-defined communication protocol, all faults such as hard line faults, CAN line faults, operation logic faults and the like of the mining truck are integrated and standardized and transmitted to visual hardware such as a display screen and the like and uploaded to the cloud in real time, and the fault type and the possibility CAN be rapidly positioned through the processing of a machine learning algorithm, so that equipment CAN learn and identify potential fault modes from fault data and make an improvement scheme.

Disclosure of Invention

The invention aims to provide fault processing equipment for a mining truck, which aims to solve the problems of poor linkage among all parts, high cost due to dependence on mine support, imperfect design and the like in the whole vehicle fault processing process of the existing mine vehicle fault processing scheme in the background technology.

In order to achieve the above purpose, the present invention provides the following technical solutions: a fault handling apparatus for a mining truck, comprising: the system comprises a whole vehicle fault controller, a central gateway, a CAN data recorder, a fault display instrument and a cloud server, wherein the whole vehicle fault controller realizes fault detection and whole vehicle communication signal processing and response functions, carries out logic judgment and processing by collecting feedback signals of all equipment controllers at the lower layer of a control logic structure, and controls feedback actions of the equipment to faults through all the equipment controllers at the lower layer; the central gateway comprises a central processor, a communication interface and one or more electronic control units, wherein the central processor is used for executing control logic, data processing and communication tasks, and the communication interface is used for realizing communication between the central processor and various subsystems and external devices in the vehicle; electronic controls are used to perform various functions and coordinate communications within the vehicle; the fault display instrument is used for displaying relevant information of the vehicle to a driver; the cloud server is used for storing related data generated in the running process of the vehicle, processing the related data through a fault processing algorithm preset in the cloud server, and giving related vehicle control commands.

As a preferable technical scheme, the control content of the whole vehicle fault controller comprises vehicle power failure control, vehicle steering failure control, vehicle braking failure control and vehicle communication network failure management.

As a preferable technical scheme, the fault display instrument comprises a digital display screen, a multifunctional information display screen and a digital instrument panel meter, and is connected with the vehicle navigation system, wherein the digital display screen is used for displaying the speed, the engine rotating speed, the fuel level, the mileage and time and faults, the multifunctional information display screen is used for displaying navigation maps, media information, vehicle setting and warning information, and the digital instrument panel meter is used for displaying the speed, the rotating speed, the oil quantity, the mileage or other information in a digital or other multimedia mode.

The fault processing method for the mining truck, which uses the fault processing equipment, further comprises the following steps:

Step S1: classifying faults; the failure of electrical systems or electronic equipment on the vehicle, the related hardware related to the transmission of basic signals of wires, cables, sensors and control units of the vehicle is defined as hard-line failure; the vehicle CAN line fault comprises a CAN line fault defined as a fault in a controller area network system; a fault generated by the operation of the vehicle according to the correct control logic is defined as a logic fault;

Step S2: processing and transmitting faults and forming a fault information table: firstly, coding each specific fault to form a fault code table, when the fault is generated, reporting no fault or waiting time exceeding a fault message transmission period, accumulating the highest fault level in all faults sent out by each period of an internal set of a processor for the time, forming a message by accumulated fault times, fault codes and other information from the power-on of the vehicle to the time of reporting the fault, and sequencing and sending according to the fault level; accumulating the messages after the fault messages and various sensor information enter the gateway, collecting fault codes and corresponding times in all the messages when no message is transmitted or the waiting time exceeds the fault message transmission period, recording the fault reporting time, and then looking up a table to obtain corresponding fault grades to form a fault information table;

Step S3: after all relevant data such as a fault information table and the like are stored in a database, accurately reported faults and relevant data are collected and recorded; and processing the related data by using convolutional neural network to filter and obtain effective fault data

Step S4: after the effective data of the real-time running of the vehicle is obtained, the vehicle data is subjected to supervised learning by using a pair of multi-mode Support Vector Machines (SVM), and the predictive maintenance of the mine car can be realized after a period of time under the support of a large amount of effective data and a deep learning model.

As a preferred technical solution, in step S3, a depth residual contraction network is used to filter to obtain effective data, and soft thresholding is used to eliminate noise-related features; embedding soft thresholding as a non-linear layer into a residual building block, specifically comprising:

1) A residual shrinkage construction module for constructing a shared threshold value among channels in a conventional depth residual network;

2) In this inter-channel shared threshold residual contraction building block, global averaging is applied over the absolute values of the feature map to obtain a one-dimensional vector;

3) Inputting the one-dimensional vector into a two-layer fully connected network to obtain a scale parameter;

The Sigmoid function normalizes the above-mentioned scaling parameter between zero and one, and takes as a threshold the average value of the absolute value of this scaling parameter multiplied by the feature map.

Compared with the prior art, the invention has the beneficial effects that:

(1) The VCU controller is adopted, a CAN communication network is utilized, all faults such as hard line faults, CAN line faults, operation logic faults and the like of the mining truck are integrated and standardized based on a self-defined communication protocol, the faults are sent to visual hardware such as a display screen and the like and uploaded to a cloud end in real time, the fault types and the possibility CAN be rapidly positioned through the processing of a machine learning algorithm, and the equipment CAN learn and identify potential fault modes from fault data and make an improvement scheme.

(2) The adopted algorithm is based on a very mature deep neural network, only a residual shrinkage construction module of the sharing threshold value among channels is added, and soft thresholding is used for eliminating the characteristics related to noise; the soft thresholding is used as a nonlinear layer to be embedded into a residual error construction module, and the obtained depth residual error shrinkage network is used for processing effective data, so that on one hand, the stability of an algorithm is ensured, and on the other hand, the pertinence and the efficiency of algorithm processing are improved.

(3) Based on the development and operation flow of the mining truck, three specific faults are proposed: 1. hard wire failure 2.Can wire failure 3. Logic failure. The first two are failure-based transmission sources, and the third means that VCU control logic develops possible failures that need to be defined. All three faults are transmitted into the controller through a CAN1 line, and then transmitted to the cloud server database through a CAN2 line for processing or remote control, and simultaneously transmitted to the display screen and the fault instrument. And the data CAN be received by the CAN data recorder in the transmission process, so that fault messages and related messages CAN be detected in real time during debugging conveniently, and the faults CAN be solved in real time. By classifying faults, the difficulty of fault processing is reduced, and the fault processing efficiency is improved.

Drawings

FIG. 1 is a schematic diagram of a fault handling apparatus for a mining truck according to the present invention;

FIG. 2 is a diagram illustrating a failure message format according to the present invention;

FIG. 3 is an example of a fault information table obtained by gateway processing according to the present invention;

FIG. 4 is a graph showing the effect of faults on the instrument of the present invention;

FIG. 5 is a CAN data logger of the present invention recording data;

FIG. 6 is a depth residual network basic structure of the present invention;

FIG. 7 is a basic structure of a depth residual contraction network used in the present invention;

FIG. 8 is a core of a depth residual contraction network used in the present invention;

FIG. 9 is a visual result of classifying fault data using a Gaussian function in accordance with the present invention;

FIG. 10 is a confusion matrix function (confusion matrix) for predictive failure classification in accordance with the present invention, where precision and duplicate rates are obtained using the predicted value (predict), the true value (real).

Detailed Description

The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

Referring to fig. 1, the present invention provides a technical solution: the system comprises a whole vehicle fault controller, a central gateway, a CAN data recorder, a fault display instrument and a cloud server, wherein the whole vehicle fault controller realizes fault detection and whole vehicle communication signal processing and response functions, carries out logic judgment and processing by collecting feedback signals of all equipment controllers at the lower layer of a control logic structure, and controls feedback actions of the equipment to faults through all the equipment controllers at the lower layer; the central gateway comprises a central processor, a communication interface and one or more electronic control units, wherein the central processor is used for executing control logic, data processing and communication tasks, and the communication interface is used for realizing communication between the central processor and various subsystems and external devices in the vehicle; electronic controls are used to perform various functions and coordinate communications within the vehicle; the fault display instrument is used for displaying relevant information of the vehicle to a driver; the cloud server is used for storing related data generated in the running process of the vehicle, processing the related data through a fault processing algorithm preset in the cloud server, and giving related vehicle control commands.

The scheme simultaneously provides a corresponding fault processing method for the mining truck, and the reverse method is theoretically based on the development and operation flow of the mining truck and provides three specific faults: 1. hard wire failure 2.Can wire failure 3. Logic failure. The first two are failure-based transmission sources, and the third means that VCU control logic develops possible failures that need to be defined. All three faults are transmitted into the controller through a CAN1 line, and then transmitted to the cloud server database through a CAN2 line for processing or remote control, and simultaneously transmitted to the display screen and the fault instrument. And the data CAN be received by the CAN data recorder in the transmission process, so that fault messages and related messages CAN be detected in real time during debugging conveniently, and the faults CAN be solved in real time. (please refer to FIG. 1)

The specific fault generation and treatment flow is as follows:

First, hard-line faults are defined as problems encountered by electrical systems or electronic equipment on the vehicle, which typically involve the associated hardware of the basic signaling of the vehicle's wires, cables, sensors, control units, etc. These faults may cause the vehicle to degrade or cease to function entirely. For example: battery failure, wire breakage or short circuit, electrical connection problems, etc. If the fault occurs, the fault is automatically generated in the operation of the mine card, the fault is transmitted into the whole vehicle fault controller through the CAN1 by each component controller, the whole vehicle fault controller responds to corresponding countermeasures, and the fault is sent to the CAN2 according to protocol arrangement (refer to figure 2).

Second, CAN line failure vehicle CAN line failure generally refers to a problem in the controller area network (Controller Area Network, CAN) system. The CAN bus is a communication system commonly used in modern vehicles that allows data communication between various Electronic Control Units (ECU) in the vehicle. CAN bus is commonly used to transmit information for engine control, brake systems, air bag systems, vehicle sensors, etc. CAN line faults are defined as self-faults of various components of the vehicle, such as the motor, the engine, and the like. For example: TCU failure, ECU failure, etc. If the fault occurs, the fault is automatically generated in the operation of the mine card, the fault is transmitted into the whole vehicle fault controller through the CAN1 by each component controller, the whole vehicle fault controller responds to corresponding countermeasures, and the fault is sent to the CAN2 according to protocol arrangement (refer to figure 2).

Again, a logic fault is defined as a fault in the vehicle operating logic that does not operate in accordance with the correct control logic. If the fault is detected by the whole vehicle fault controller, the whole vehicle fault controller autonomously judges the states detected by the controllers, responds to corresponding countermeasures and sends out the corresponding countermeasures to CAN2 according to the self-defined protocol arrangement (refer to figure 2).

However, more faults are greatly affected by various factors such as vehicle running conditions, vehicle running environments, driving habits of drivers and the like, so that the current corresponding countermeasures of the controller are not necessarily completely suitable for the current fault conditions, and due to the large weight and heavy load of the mining truck, the current corresponding strategies of the whole-vehicle fault controller can only prevent the occurrence of worse conditions when more faults occur, rather than judging and analyzing the fault reasons and perfectly solving the fault points through the whole-vehicle fault controller.

Processing and transmitting faults: firstly, coding each specific fault to form a fault code table, when the fault is generated, reporting the fault (waiting time exceeds the fault message sending period), accumulating the highest fault level in all faults sent out by each period of the processor internal set for the time, accumulating the number of times of faults, fault codes and other information from the time of powering up the vehicle to the time of reporting the fault to form a message (refer to fig. 2), and sequencing and sending the message according to the fault level. From CAN2 into the meter display (see fig. 4), gateway, CAN data logger (see fig. 5). In the gateway, after fault messages and various sensor information enter, the messages are accumulated, no messages are transmitted (waiting time exceeds the fault message sending period), fault codes and corresponding times in all the messages are collected, fault reporting time is recorded, and then corresponding fault grades are obtained through table lookup, so that a fault information table is formed. (please refer to FIG. 3)

After all relevant data such as the fault information table and the like are stored in the database, accurately reported faults and relevant data are collected and recorded. The effective data is then filtered using a depth residual contraction network. Describing a model:

First, a depth residual network is described (see fig. 6). The residual construction module is the basic component. The residual construction module comprises two batch normalization, two rectification linear units, two convolution layers and an identity path. The identity path is the key to making the depth residual network superior to convolutional neural networks. The gradient of the cross entropy loss function, in a common convolutional neural network, is counter-propagating layer by layer. When using an identity path, the gradient can flow back to the previous layer more efficiently, so that the parameters can be updated more efficiently. Fig. 6-b and 6-c show two residual construction modules capable of outputting feature maps of different sizes. Here, the reason for reducing the size of the output feature map is that the operation amount of the subsequent layer can be reduced; the reason for increasing the number of channels is to facilitate the integration of different features into a strongly discriminative feature. Fig. 6-d shows the overall framework of the depth residual network, including an input layer, a convolution layer, a number of residual building blocks, a batch normalization, a ReLU activation function, a global averaging pool, and a fully connected output layer.

The depth residual contraction network (see fig. 7) used in the present scheme is optimized above the depth residual network, and soft thresholding is used to eliminate noise-related features. Soft thresholding is embedded as a non-linear layer into the residual construction block. More importantly, the threshold is automatically learned in the residual construction module.

In fig. 7-a, the residual contraction building block named "inter-channel shared threshold", unlike the residual building block in fig. 6-a, there is a special block to estimate the threshold required for soft thresholding (fig. 8). In this particular module, global averaging is applied over the absolute values of the feature map to obtain a one-dimensional vector. This one-dimensional vector is then input into a two-layer fully connected network to obtain a scaling parameter. The Sigmoid function scales this scale parameter between zero and one. This scaling parameter is then multiplied by the average of the absolute values of the feature map as a threshold. In this way, the threshold can be controlled within a suitable range without bringing the output characteristics to zero. The simplified architecture of the proposed inter-channel shared-threshold depth residual contraction network is shown in fig. 7-b, which is similar to the classical depth residual network in fig. 6-d. The only difference is that the present invention replaces the common residual construction module with the residual contraction module (RSBU-CS) of the inter-channel shared threshold. A number RSBU-CS are stacked so that the noise-related features are gradually cut down. Another advantage is that the threshold is automatically learned rather than manually set by an expert, so that expertise in the signal processing domain is not required when implementing an inter-channel shared threshold depth residual contraction network.

After obtaining effective data of real-time running of the vehicle, a pair of multimode Support Vector Machines (SVM) are utilized to conduct supervised learning on the vehicle data. The model thus trained can be used to derive a probability of whether the input data belongs to a certain known fault type. Faults are classified into 3 general categories, for example as follows:

1. The training data total 300. The goal is to divide into 3 classes, and a one-to-many approach requires training 3 SVM models (each model is a two-class, so a division of positive and negative samples is required). Positive samples are all from this category, negative samples are randomly selected from the other two categories, and the number of negative samples is guaranteed to be the same as the number of positive samples, thus obtaining training samples X. Setting the label to +1, -1, thus obtaining a sample label Y; other parameters are not default to each class of training SVM model.

2. The test sample does not need to divide positive and negative samples for each class, but only needs to know the test data and the sample label. Each test sample obtains a score in 3 SVM models, and the maximum score is used for judging which class the sample belongs to.

3. Classification is performed using gaussian kernel functions and the learning effect is visualized as in fig. 9.

4. Test set data 30. And then utilizing confusion matrix confusion matrix functions in MATLAB (figure 10), obtaining precision by means of real labels and predictive labels, recall (recall) and comprehensive evaluation index (F-measure). Referring to fig. 10, the precision of class 0 is a/(a+b+g), and the recall is a/(a+b+c); the precision of the category 1 is e/(b+e+h), and the recall is e/(d+e+f); the precision of category 1 is i/(c+f+i), and the recall is i/(g+h+i).

5. Finally, the effect of the learning model is obtained.

FIG. 9 is a visual result of classifying fault data using a Gaussian function in accordance with the present invention;

FIG. 10 is a confusion matrix function (confusion matrix) for predictive failure classification in accordance with the present invention, where precision and duplicate rates are obtained using the predicted value (predict), the true value (real).

From the above results, it can be seen that with the support of a large amount of available data and deep learning models, predictive maintenance of the mine car can be achieved after a period of time (vehicle maintenance teams can plan and perform maintenance tasks based on predictive analysis).

After the processing of the server is completed, if suggestions such as a standby control strategy which can be adjusted and optimized in real time exist, the suggestions are sent to the gateway and then are transmitted to the display screen for display. Then, the CAN data record is manually used for manual detection and observation, and the machine learning result is referred to so as to optimize the control strategy and hardware.

The technical scheme adopts the VCU controller and utilizes the CAN communication network, and based on the self-defined communication protocol, all faults such as hard line faults, CAN line faults, operation logic faults and the like of the mining truck are integrated and standardized and transmitted to visual hardware such as a display screen and the like and uploaded to the cloud in real time, and the fault type and the possibility CAN be rapidly positioned through the processing of a machine learning algorithm, so that equipment CAN learn and identify potential fault modes from fault data and make an improvement scheme. Meanwhile, the adopted algorithm is based on a very mature deep neural network, only a residual shrinkage construction module of the inter-channel sharing threshold value is added, and soft thresholding is used for eliminating the noise-related characteristics; the soft thresholding is used as a nonlinear layer to be embedded into a residual error construction module, and the obtained depth residual error shrinkage network is used for processing effective data, so that on one hand, the stability of an algorithm is ensured, and on the other hand, the pertinence and the efficiency of algorithm processing are improved.

And, moreover, the method comprises the steps of. The scheme is based on the development and operation flow of the mining truck, and three specific faults are proposed: 1. hard wire failure 2.Can wire failure 3. Logic failure. The first two are failure-based transmission sources, and the third means that VCU control logic develops possible failures that need to be defined. All three faults are transmitted into the controller through a CAN1 line, and then transmitted to the cloud server database through a CAN2 line for processing or remote control, and simultaneously transmitted to the display screen and the fault instrument. And the data CAN be received by the CAN data recorder in the transmission process, so that fault messages and related messages CAN be detected in real time during debugging conveniently, and the faults CAN be solved in real time. By classifying faults, the difficulty of fault processing is reduced, and the fault processing efficiency is improved.

The relevant hardware configuration is as follows:

the hardware of the complete vehicle fault controller has the following electrical functions:

1. The VCU needs to have the power supply reverse connection protection function, and is not damaged after lasting 120 seconds;

2. Surge and overvoltage protection of a power supply;

Esd protection (antistatic protection);

4. Overvoltage and overcurrent and over-temperature protection of the power device;

5. to the ground, short circuit and open circuit protection and diagnosis are carried out on a power supply;

6. Ground wire loss protection;

7. All sensors have a default state at the time of failure.

8. The VCU has a low power consumption mode, and the power consumption of the VCU is less than 5mA. And after detecting the wake-up signal, exiting the low power consumption mode.

9. In the range of 9-32V power supply voltage, all functions of VCU can be normally realized.

10. In the range of 6.5-9V power supply voltage, VCU functions are limited, but the following functions can work normally:

a. The microprocessor will normally operate

Normal CAN communication

11. The controller is not damaged when the power supply voltage of the power supply is 16-32V.

The central gateway is configured as follows:

1. Control Unit (Control Unit): the vehicle central gateway is comprised of one or more electronic control units that are responsible for performing various functions and coordinating communications within the vehicle.

2. Processor (Processor): the central gateway is typically equipped with a high-performance processor (embedded microprocessor or microcontroller) for performing control logic, data processing and communication tasks.

3. Memory (Memory): the central gateway includes a memory for storing configuration files, an operating system, software programs, and log data.

4. Communication interface (Communication Interfaces): the vehicle center gateway has a plurality of communication interfaces for communicating with various subsystems and external devices within the vehicle. These interfaces mainly comprise a CAN bus, a LIN bus.

5. Data bus (Data bus): the vehicle central gateway is connected via different data buses to other control units and sensors/actuators in the vehicle in order to transmit and receive data. The CAN bus (Controller Area Network) is typically one of the most common data buses.

6. Power Supply (Power Supply): the central gateway requires a power supply, typically provided by the battery system of the vehicle. It also includes a power management circuit to ensure that the necessary power is maintained when the vehicle is turned off.

7. Safety function (Security Features): since the central gateway is involved in the security and data protection of the vehicle interior system, it is typically provided with security functions such as data encryption and authentication.

8. Diagnostic interface (Diagnostics Interface): the central gateway is typically provided with a diagnostic interface so that a vehicle maintenance and diagnostic engineer can communicate with it to detect and solve problems.

9. Software (Software): the software run by the central gateway includes an operating system, control logic, communication protocol stacks, and interface software with other vehicle systems.

The cloud server is configured as follows:

1. Central Processing Unit (CPU): the CPU is the brain of the server, performing the computational tasks. The CPU of the cloud server may have different core numbers and clock speeds to accommodate different workload demands.

2. Memory (RAM): memory is used to store executing applications and data and is critical to performance. Cloud servers typically provide different capacities of RAM to accommodate different workloads.

3. And (3) storing: cloud servers typically include two types of storage:

and (3) local storage: typically Solid State Disks (SSDs) or mechanical hard disks (HDDs) for the installation of operating systems and application programs.

Network attached storage: typically network attached storage volumes (e.g., cloud disks) for persistent data storage and backup.

4. Network interface: cloud servers typically have multiple network interfaces that allow connection to different virtual networks, as well as public and private networks.

Gpu (graphics processing unit): for workloads requiring graphics processing, machine learning, deep learning, etc., cloud servers may be equipped with GPUs, providing additional computing power.

6. Operating system: the cloud server may run various operating systems, including LinuX, windows, etc., depending on the user's needs.

7. Data center location: cloud service providers typically provide servers at multiple data centers, and users may choose to deploy their servers at specific geographic locations to meet performance, regulatory compliance, and data storage requirements.

8. Security hardware: cloud servers are typically equipped with hardware security functions, such as encryption modules, hardware firewalls, and security certificates, to protect data and system security.

9. Scalability and elasticity: the cloud server allows expansion or contraction as needed to accommodate changes in traffic and workload.

10. Monitoring and management tool: cloud service providers typically provide monitoring and management tools to help users monitor server performance, perform automation tasks, and manage resources.

The whole vehicle fault display instrument is configured as follows:

1. digital display screen: the instrument is provided with a plurality of digital display screens for displaying information such as vehicle speed, engine speed, fuel level, mileage, time, faults and the like. These displays may be liquid crystal displays providing digital and graphic displays to provide more information.

2. Multifunctional information display: the instrument is equipped with a multi-function information display that can display navigation maps, media information, vehicle settings, warning information, etc.

3. Digitizer dial: the vehicle adopts a full-digital instrument panel to replace the traditional pointer instrument. These dashboards can display different information as required, not just fixed dials.

4. Navigation system: the meter may be integrated with a navigation system providing real-time navigation instructions and map display.

5. And (3) media control: the meter has a media control function, allowing the driver to control multimedia devices such as music, radio, CD player, etc.

6. Bluetooth: the meter supports bluetooth functionality allowing the driver to connect to a cell phone or other device for hands-free conversation and music playback.

7. Driving assistance system: the meter may integrate driving assistance systems such as blind spot monitoring, lane keeping assistance, adaptive cruise control, etc.

8. Vehicle diagnostics and warnings: the instrument directly obtains the specific fault code and the corresponding occurrence time and frequency through J1939 protocol analysis and displays the specific fault code and the corresponding occurrence time and frequency on the instrument, such as engine fault, low tire pressure, insufficient fuel and the like.

9. Touch screen control: the meter employs touch screen technology, allowing the driver to control and set via the touch screen.

And (3) voice recognition: the meter is equipped with a voice recognition system that allows the driver to control functions and access information using voice commands.

Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.

Claims (5)

1. A fault handling apparatus for a mining truck, comprising: the system comprises a whole vehicle fault controller, a central gateway, a CAN data recorder, a fault display instrument and a cloud server, wherein the whole vehicle fault controller realizes fault detection and whole vehicle communication signal processing and response functions, logic judgment and processing are carried out by collecting feedback signals of all equipment controllers at the lower layer of a control logic structure, and feedback actions of the equipment to faults are controlled by all the equipment controllers at the lower layer; the central gateway comprises a central processor, a communication interface and one or more electronic control units, wherein the central processor is used for executing control logic, data processing and communication tasks, and the communication interface is used for realizing communication between the central processor and various subsystems and external equipment in the vehicle; the electronic control is used to perform various functions and coordinate communications within the vehicle; the fault display instrument is used for displaying relevant information of the vehicle to a driver; the cloud server is used for storing related data generated in the running process of the vehicle, processing the related data through a fault processing algorithm preset in the cloud server, and giving related vehicle control commands.

2. The fault handling apparatus for a mining truck according to claim 1, wherein the control content of the complete vehicle fault controller includes vehicle power failure control, vehicle steering failure control, vehicle brake failure control, vehicle communication network failure management.

3. The fault handling apparatus for a mining truck according to claim 1, wherein the fault display instrument comprises a digital display screen for displaying vehicle speed, engine speed, fuel level, mileage, time, faults, a multifunction information display screen for displaying navigation map, media information, vehicle settings, warning information, and a digital dashboard for displaying vehicle speed, fuel quantity, mileage or other information in digital or other multimedia form, and is connected to a car navigation system.

4. A fault handling method for a mining truck using a fault handling apparatus as claimed in any one of claims 1 to 3, further comprising the steps of:

Step S1: classifying faults; the failure of electrical systems or electronic equipment on the vehicle, the related hardware related to the transmission of basic signals of wires, cables, sensors and control units of the vehicle is defined as hard-line failure; the vehicle CAN line fault comprises a CAN line fault defined as a fault in a controller area network system; a fault generated by the operation of the vehicle according to the correct control logic is defined as a logic fault;

Step S2: processing and transmitting faults and forming a fault information table: firstly, coding each specific fault to form a fault code table, when the fault is generated, reporting no fault or waiting time exceeding a fault message transmission period, accumulating the highest fault level in all faults sent out by each period of an internal set of a processor for the time, forming a message by accumulated fault times, fault codes and other information from the power-on of the vehicle to the time of reporting the fault, and sequencing and sending according to the fault level; accumulating the messages after the fault messages and various sensor information enter the gateway, collecting fault codes and corresponding times in all the messages when no message is transmitted or the waiting time exceeds the fault message transmission period, recording the fault reporting time, and then looking up a table to obtain corresponding fault grades to form a fault information table;

step S3: after all relevant data such as a fault information table and the like are stored in a database, accurately reported faults and relevant data are collected and recorded; and processing the related data by using a convolutional neural network to filter to obtain effective fault data;

step S4: after the effective data of the real-time running of the vehicle is obtained, the vehicle data is subjected to supervised learning by using a pair of multi-mode support vector machines, and the predictive maintenance of the mine car can be realized after a period of deep learning time under the support of a large amount of effective data and a convolutional neural network.

5. The fault handling method for a mining truck according to claim 4, wherein in step S3, a depth residual shrinkage network is used to filter out valid data, and soft thresholding is used to eliminate noise-related features; embedding soft thresholding as a non-linear layer into a residual building block, specifically comprising:

1) A residual shrinkage construction module for constructing a shared threshold value among channels in a conventional depth residual network;

2) In this inter-channel shared threshold residual contraction building block, global averaging is applied over the absolute values of the feature map to obtain a one-dimensional vector;

3) Inputting the one-dimensional vector into a two-layer fully connected network to obtain a scale parameter;

4) The Sigmoid function normalizes the above-mentioned scaling parameter between zero and one, and takes as a threshold the average value of the absolute value of this scaling parameter multiplied by the feature map.