Unmanned driving remote control collaborative decision-making method and system

Abstract

The present application provides an unmanned driving remote control collaborative decision-making method and system, which relates to the field of unmanned driving remote control collaborative decision-making technology, and can improve the technical problem of poor accident prediction ability of unmanned driving remote control collaborative decision-making system in related technologies. The method includes: constructing a prediction model based on fault factors, the prediction model is used to predict the number of faults of vehicles in the mining area; and determining key partitions based on the number of faults in each partition; pushing the video screen corresponding to the key partition to the remote operator; obtaining the route and current status information of each vehicle in the key partition, the current status information includes the current speed and current position; judging the intersection time of the two vehicles based on the route and current status information; when the intersection time of the two vehicles is within the intersection threshold, and the current speed of one of the vehicles is abnormal, judging that the collision possibility of the two vehicles reaches the collision threshold, and switching the video screen of the key partition to the video screen of the vehicle.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of collaborative decision-making for unmanned driving remote control, and in particular, to a method and system for collaborative decision-making for unmanned driving remote control. Background Art

[0002] Existing unmanned remote control collaborative decision-making systems have shown obvious limitations in practical applications. These systems can often only passively respond to events that have already occurred, but cannot actively predict and prevent accidents. They lack effective accident prediction capabilities and cannot take measures to avoid or mitigate the impact of accidents before they occur. Summary of the invention

[0003] The embodiments of the present application provide a method and system for collaborative decision-making of unmanned driving remote control, which are used to improve the technical problem in related technologies that the unmanned driving remote control collaborative decision-making system lacks effective accident prediction capabilities.

[0004] To achieve the above objectives, the embodiments of the present application adopt the following technical solutions:

[0005] In a first aspect, an embodiment of the present application provides an unmanned driving remote control collaborative decision-making method, which is applied to vehicle control in a mining area, wherein the mining area includes multiple partitions, and each partition has a corresponding video screen. The method includes: obtaining a fault factor of a vehicle fault; constructing a prediction model based on the fault factor, wherein the prediction model is used to predict the number of faults of vehicles in the mining area; predicting the number of faults in each partition based on the prediction model, and determining a key partition based on the number of faults in each partition; pushing the video screen corresponding to the key partition to a remote operator; obtaining the route and current status information of each vehicle in the key partition, wherein the current status information includes a current speed and a current position; judging a meeting time of the two vehicles based on the route and the current status information; when the meeting time of the two vehicles is within a meeting threshold and the current speed of one of the vehicles is abnormal, judging that the possibility of collision between the two vehicles reaches a collision threshold, and switching the video screen of the key partition to the video screen of the vehicle.

[0006] In a possible implementation of the first aspect, the remote operator obtains video images through a monitoring device, and pushing the video images corresponding to the key partitions to the remote operator includes: sorting the multiple key partitions based on the number of faults corresponding to the key partitions, and pushing the key partitions in sequence according to the rankings to the middle area of ​​the display page of the monitoring device.

[0007] In a possible implementation of the first aspect, switching the video picture of the key partition to the video picture of the vehicle includes: pushing the video picture of the vehicle to a display page corresponding to the key partition.

[0008] In a possible implementation of the first aspect, the speed of the vehicles when they intersect is a first speed, and when a difference between a current speed of the vehicles when they intersect and the first speed is greater than a speed threshold, it is determined that the current speed of the vehicle is abnormal.

[0009] In a possible implementation manner of the first aspect, the prediction model is expressed as:

[0010] ;

[0011] in, is the number of failures at time t, is the number of vehicles that have not failed at time t, m represents the total number of vehicles, is the first failure factor, The second fault factor.

[0012] In a possible implementation of the first aspect, the fault factor includes a first fault factor and a second fault factor, wherein the first fault factor is used to represent an initial rate of vehicle failure, and the second fault factor is used to represent a rate at which vehicle failure increases with usage time and environmental influences.

[0013] In a possible implementation of the first aspect, predicting the number of failures in each partition based on the prediction model includes: acquiring historical failure data, the failure data including the number of failures in multiple different time periods; obtaining an intermediate value of the failure factor based on the historical failure data, current actual failure data and the prediction model; and predicting the number of failures in each partition based on the prediction model corresponding to the intermediate value of the failure factor.

[0014] In a possible implementation of the first aspect, an intermediate value of the fault factor is obtained based on the historical fault data, the current actual fault data and the prediction model, including: constructing a data model based on the historical fault data; acquiring first predicted fault data based on the data model; obtaining second predicted fault data based on the current actual fault data and the first predicted fault data; and obtaining the intermediate value of the fault factor based on the data of the trend node in the second predicted fault data and the prediction model.

[0015] In a possible implementation of the first aspect, the second predicted fault data is obtained based on the current actual fault data and the first predicted fault data, including: obtaining the average deviation between characteristic data in the first predicted fault data and the current actual fault data, the characteristic data being the data in the first predicted fault data corresponding to the current actual fault data; and correcting the first predicted fault data according to the average deviation to obtain the second predicted fault data.

[0016] In the second aspect, the present application also provides an unmanned driving remote control collaborative decision-making system, which is used to execute the method described in the first aspect, and the system includes: an acquisition module, which is used to acquire fault data; a model construction module, which is coupled to the acquisition module, and the model construction module is used to construct a prediction model and a data model; and a processing module, which is coupled to the acquisition module and is also coupled to the model construction module, and the processing module is used to predict the number of faults based on the fault data and the prediction model.

[0017] Beneficial effects: The present application first constructs a prediction model through fault factors, and predicts the number of faults in each partition through the prediction model to obtain the key partitions that need special attention, and then pushes the video images of the key partitions to the remote operator, and judges the collision possibility of the vehicle based on the vehicle's route and current status information in the key partitions. Finally, the video images of the vehicles whose collision possibility exceeds the collision threshold are pushed to the remote operator, so that the remote operator can manage the vehicles in the mining area more accurately, intervene before the collision may occur, reduce the probability of vehicle collision in the mining area, and improve the technical problem of the lack of effective accident prediction capability in the driving remote control collaborative decision-making system in the related technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic diagram of a method provided for some embodiments of the present application;

[0019] Figure 2 for Figure 1 Schematic diagram of the process of step S300;

[0020] Figure 3 for Figure 2 Schematic diagram of the process of step S320;

[0021] Figure 4 A schematic diagram of the structure of a system provided for some embodiments of the present application. DETAILED DESCRIPTION

[0022] The technical solutions in the embodiments of the present application will be described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments.

[0023] Existing unmanned remote control collaborative decision-making systems have shown obvious limitations in practical applications. These systems can often only passively respond to events that have already occurred, but cannot actively predict and prevent accidents. They lack effective accident prediction capabilities and cannot take measures to avoid or mitigate the impact of accidents before they occur.

[0024] Therefore, the accident prediction capability of the unmanned driving remote control collaborative decision-making method in the prior art is relatively poor. The embodiment of the present application provides an unmanned driving remote control collaborative decision-making method and system, which is used to improve the technical problem that the unmanned driving remote control collaborative decision-making method in the related art lacks effective accident prediction capability.

[0025] The method is applied to vehicle control in a mining area. The vehicles include various movable vehicles and movable devices such as unmanned operation robots. In the example of this application, the mining area vehicles are in automatic driving mode by default. When encountering special circumstances or active takeover, the manual driving mode is triggered and controlled by a remote operator. Based on this, each vehicle is also equipped with a corresponding video acquisition device, and each vehicle has its corresponding video screen.

[0026] The mining area includes a plurality of mining area partitions, each of which is provided with a corresponding video monitoring device, so that each of the partitions has one or more corresponding video images. Exemplarily, when the partition size and the monitoring area of ​​the monitoring device are appropriate, each partition can have a total video image. Alternatively, each partition has a plurality of monitoring devices, each monitoring device is responsible for monitoring a portion of the area within the partition, and when the images of the corresponding partition need to be monitored, the video images of the plurality of monitoring devices can be played in turn or in parallel.

[0027] The remote operator obtains the video screen through the monitoring device. That is, the video monitoring equipment in the mining area transmits the video screen to the monitoring device of the remote operator, and the monitoring device presents the video screen on the display page for the remote operator to view. The display page can have multiple display areas, and each display area plays a video screen of a partition. For example, the display page is divided into 9 display areas (the first to the ninth display areas, arranged in a nine-square grid), and each display area plays a video screen of a partition.

[0028] like Figures 1 to 3 As shown, the method includes:

[0029] S100: Obtain the failure factor of the vehicle failure.

[0030] The propensity of a vehicle to fail is significantly affected by its inherent characteristics and external factors, which together determine the potential vulnerability of the vehicle to failure stress. The inherent characteristics of the vehicle, such as manufacturing quality, intensity of use, maintenance history, vehicle age, adaptability to adverse operating conditions, load type, etc., are all key factors affecting its failure.

[0031] For example, a high initial failure rate for vehicles can quickly increase the number of vehicle failures in a mine. Vehicles that are used intensively may fail more frequently. Vehicles with a poor maintenance history may be more sensitive to specific failures. In addition, the adaptability of vehicles to adverse operating conditions and resistance to previous failure cases also affect their potential risk of failure.

[0032] The above factors are all internal conditions of each vehicle. Therefore, the failure factors of the present application may include a first failure factor, which is used to characterize the initial rate of failure of the vehicle when there is no external influence. This reflects the initial rate of failure of the vehicle based on its own conditions and usage history, without being affected by other factors. The larger the value of the first failure factor, the higher the potential risk of failure of the mining vehicle.

[0033] Vehicle failure is also closely related to external environmental factors. External environmental factors such as road conditions, climate change, vehicle usage frequency, etc., all affect the vehicle's failure behavior and the frequency of failure in different ways. These factors together constitute the external conditions for vehicle failure.

[0034] For example, vehicles may break down more often when road conditions are bad. Climate change may affect the failure behavior of vehicles, making them more likely to break down in bad weather conditions. Frequently used vehicles may break down earlier due to accumulated fatigue.

[0035] As the vehicle is used, the failure rate may increase due to factors such as wear, aging, or insufficient maintenance of the vehicle. Therefore, the failure factor in the present application may also include a second failure factor, which is used to characterize the failure rate of the vehicle that increases with the use time and environmental influences. This rate reflects the increase in the risk of failure caused by natural aging and environmental pressure during the use of the vehicle. The larger the value of the second failure factor, the more significant the increase in the vehicle failure rate with time and environmental influences.

[0036] S200: construct a prediction model based on the failure factors, wherein the prediction model is used to predict the number of failures of mining vehicles. The prediction model can be expressed as:

[0037] ;

[0038] in, is the number of failures at time t, is the number of vehicles that have not failed at time t, m represents the total number of vehicles, is the first failure factor, The second fault factor.

[0039] In the above prediction model, the term where the first failure factor is located represents the failure rate caused by the vehicle's own factors, which is proportional to the number of vehicles that have not yet failed. It simulates vehicles that may fail due to their own conditions and usage history without being affected by other factors. When the number of non-failed vehicles is large (i.e., there are few failed vehicles and many non-failed vehicles), this part contributes more to the failure rate.

[0040] The term for the second failure factor represents the failure rate caused by external factors, which is proportional to the product of the number of vehicles that have failed and the proportion of vehicles that have not failed. It simulates the risk of failure that increases with natural aging and environmental pressures in the fleet, or the remaining non-failed vehicles due to the same influences (such as the same usage pattern, the same lack of maintenance, etc.). When there are many failed vehicles, this part contributes more to the failure rate, but as the number of failures approaches the total number, this term will decrease, that is, its growth rate will slow down, thereby limiting the growth of the failure rate.

[0041] This prediction model describes the dynamic changes in the number of vehicle failures over time. The prediction model regards the vehicle failure process as a dynamic system that is jointly affected by its inherent characteristics and external environmental factors. It describes the failure of the vehicle in a specific state within a certain period of time. The first failure factor reflects the initial failure rate of the vehicle itself, while the second failure factor reflects the impact of time and environmental factors on the failure.

[0042] The prediction model in this application takes into account the number limit of vehicle failures by introducing the concept of the proportion of non-faulty vehicles to the total number of vehicles. This proportion changes dynamically as the failure process progresses, so that the model can capture the nonlinear characteristics of vehicle failures, increase the accuracy of the prediction, and thus improve the monitoring effect of vehicle failures.

[0043] In addition, the parameters of the prediction model are designed to be estimable and can be quantified through past data, ensuring the feasibility and effectiveness of the model in practical applications.

[0044] S300 , predicting the number of faults in each partition based on the prediction model, and determining a key partition based on the number of faults in each partition.

[0045] The number of faults in each partition in the mine area is predicted according to the prediction model. When the number of faults in a partition is greater than a set value, such as 5, the partition is judged to have a high possibility of collision or danger, and is defined as a key partition that requires the remote operator to pay close attention to. It can be understood that the set value of the number of faults can also be determined based on the average number of faults in the mine area in the past.

[0046] Exemplarily, predicting the number of failures of each partition based on the prediction model includes:

[0047] S310: Acquire historical fault data, where the fault data includes the number of faults in multiple different time periods.

[0048] Obtain the fault data of the mine area in previous years, which is defined as historical fault data. For example, the fault data of the mine area in any year may include the total number of vehicles, the total number of faults, the number of faults in different time periods, etc. This step can be obtained through the storage module in the management control system of the mine area or from the fault record table of the mine area personnel. It can be understood that the number of vehicles and environment of each partition are different. In order to improve the prediction accuracy of each partition, the historical fault data of the corresponding partition can be obtained first when selecting data.

[0049] S320: Obtain an intermediate value of the fault factor according to the historical fault data, the current actual fault data and the prediction model. Exemplarily, S320 includes:

[0050] S321. Construct a data model according to the historical fault data.

[0051] Exemplarily, the data model may be composed of a plurality of models for prediction. For example, the data model may be composed of a weighted BP neural network model, a fuzzy neural network model, and an adaptive probabilistic neural network model. The weight of each model may be determined by the accuracy of its predicted results relative to the actual results. The higher the accuracy, the greater its weight in constituting the data model. It is understandable that the choice of model may be based on the needs of data prediction by those skilled in the art to select an existing model or build a new model. This application does not impose any restrictions on this, as long as the corresponding model can obtain predicted data based on historical data.

[0052] S32. Obtain first predicted fault data according to the data model.

[0053] The trained data model is used to predict the number of faults to obtain first predicted fault data, such as predicting the fault data from the first day to the thirtieth day of a month in a sub-area of ​​the mining area.

[0054] S323, obtaining second predicted fault data according to the current actual fault data and the first predicted fault data. Obtaining an average deviation between characteristic data in the first predicted fault data and the current actual fault data, wherein the characteristic data is data in the first predicted fault data corresponding to the current actual fault data.

[0055] Exemplarily, the current actual fault data includes data from the first to the tenth day of the mining area, and the data to be predicted includes data from the twentieth to the thirtieth day.

[0056] The data corresponding to the current actual fault data in the first predicted fault data is taken out, that is, the number of faults from the first to the tenth day in the first predicted fault data is analyzed with the number of faults from the first to the tenth day in the current actual fault data to obtain the average deviation between the two.

[0057] For example, the daily deviation value between the first predicted fault data and the current actual fault data from the first day to the tenth day is calculated, and all the deviation values ​​are summed up and divided by the number of days to obtain the average deviation for the period.

[0058] The first predicted fault data is corrected according to the average deviation, such as adding the average deviation to each data in the first fault data to obtain the second predicted fault data. This application does not limit the correction method of the first data according to the average deviation, which can be selected by technicians in this field according to their needs.

[0059] In this way, by calculating the deviation between the actual fault data and the first predicted fault data and correcting the predicted value accordingly, more accurate predicted fault data can be obtained. This method allows the model to adjust itself according to the latest actual data to improve the accuracy of future predictions. At the same time, the present application can obtain a large amount of fault data through the data model, so that the model training can be completed without a large amount of historical data.

[0060] S324: Obtain an intermediate value of the fault factor according to the data of the trend node in the second predicted fault data and the prediction model.

[0061] For example, the mine area starts monitoring vehicle failures on March 1, 2023, and records the number of vehicles that fail. The number of vehicles that fail on the first day is 5. In the following days, the number of vehicles that fail gradually increases. The number of vehicles that fail from March 1 to March 3, 2023 increases by 3, and the number of vehicles that fail from March 4 to March 5, 2023 increases by 6. From March 6 to March 10, 2023, the number of vehicles that fail increases by 7. Subsequently, the growth rate begins to slow down. From March 11 to March 13, 2023, the increase is 4, and from March 14 to March 15, 2023, the increase is 3. From March 16 to March 19, 2023, the number of vehicles that fail increases by only 1, and stabilizes after March 20, 2023.

[0062] In this example, the vehicle failure situation changes over time, where March 1 is the time point when vehicle failure monitoring starts. March 4 is the turning point of the initial increase in the number of vehicle failures. March 6 is the turning point between the initial increase and the stable period. March 11 is the turning point between the stable period and the slowdown period of the number of vehicle failures. After March 14 is the turning point between the slowdown period and the decline period. Therefore, in the entire monitoring period, there can be several trend nodes: the time point when monitoring starts, the turning point between the initial increase and the stable period, the turning point between the stable period and the slowdown period, and the turning point between the slowdown period and the decline period.

[0063] The number of faulty vehicles corresponding to each trend node in the second predicted fault data is obtained, substituted into the fault prediction model, and the prediction model is solved to obtain the corresponding intermediate value of the fault factor.

[0064] In the above example, by looking at the trend of the number of vehicle failures over time, several key turning points in the failure progression process can be identified. By analyzing these data points, a more accurate estimate can be made of the initial rate of vehicle failures and the rate of failures that increase over time and due to environmental influences.

[0065] S330: predict the number of failures in each partition according to the prediction model corresponding to the intermediate value of the failure factor. Substitute the obtained intermediate value of the failure factor into the prediction model, illustratively, substitute the determined first failure factor and second failure factor into the prediction model, and predict the number of failures in each partition from the 20th day to the 30th day according to the prediction model.

[0066] S400: Push the video screen corresponding to the key partition to the remote operator. Sort the multiple key partitions based on the number of faults corresponding to the key partitions, and push the key partitions in order to the middle area of ​​the display page of the monitoring device according to the ranking.

[0067] Exemplarily, the video picture of the key partition with the largest number of faults, i.e. the first ranked partition, is pushed to the middle display area of ​​the display partition, such as the fifth display area. The video pictures of the second to third ranked key partitions are pushed to the fourth and fifth display areas in sequence.

[0068] In this way, the partition with the largest number of faults can be given priority by the remote operator. It can be understood that the division of display partitions and the push of key partition video images can be set by those skilled in the art according to needs, as long as the key partition with the largest number of faults is pushed to the remote operator first.

[0069] S500: Acquire the travel route and current status information of each vehicle in the key partition, wherein the current status information includes current speed and current position.

[0070] Key zones have a higher probability of failure, and due to failures such as sensing failure, braking failure, acceleration failure, or movement failure, the faulty vehicle may be in the wrong place at the corresponding time, causing collision risks with other vehicles, especially when two vehicles meet. Therefore, it is necessary to give priority to the intersection of vehicles in key zones.

[0071] For example, each vehicle has a corresponding route in the mine, which can be preset or generated by the vehicle's control system based on the current task and environment. By communicating with each vehicle, the vehicle's route, current speed, and current location information can be obtained through the vehicle's control system. It is understandable that each vehicle can be provided with a communication module with communication function.

[0072] S600: Determine a meeting time of the two vehicles based on the travel route and the current state information.

[0073] When the vehicle's route, current position, and current speed are known, the vehicle's position in the next period of time can be determined, thereby obtaining the intersection time of the two vehicles. When the intersection time of the two vehicles is lower than the intersection threshold, such as 10s, attention should be paid to the intersection of the two vehicles. It is understandable that the intersection threshold can also be set to 1s, 5s, 8s, 20s, etc., which can be set by those skilled in the art according to needs.

[0074] S700: When the intersection time of the two vehicles is within the intersection threshold and the current speed of one of the vehicles is abnormal, it is determined that the collision possibility of the two vehicles reaches the collision threshold, and the video image of the key partition is switched to the video image of the vehicle.

[0075] For example, when the mining area vehicle meets, it travels at the first speed by default, and when it does not meet, it travels at the second speed, and the first speed is less than the second speed. Based on the road conditions and control errors in the mining area, the speed when meeting may be slightly lower or higher than the first speed, but the error should not be too large, and the error is defined as the speed threshold. When the difference between the speed when meeting and the first speed is greater than the speed threshold, it is determined that the vehicle may have a fault, such as a braking fault, an acceleration fault, or a moving fault.

[0076] When it is detected that the intersection time of the vehicles is within the intersection threshold, that is, the two vehicles will intersect in a short time, the current speed of the vehicles is detected. If the difference between the current speed of any vehicle and the first speed is greater than the speed threshold, it is determined that the vehicle may have a fault and the intersection of the two vehicles has a high risk of collision, that is, the possibility of collision reaches the collision threshold, and a remote operator is required to observe or take over.

[0077] Therefore, when the intersection time of the two vehicles is within the intersection threshold and the current speed of one of the vehicles is abnormal, the video screen of the key partition is switched to the video screen of the vehicle, that is, the video screen of the vehicle is pushed to the display area of ​​the display page corresponding to the key partition, so that the remote operator can determine whether intervention is needed.

[0078] In the second aspect, the present application also provides an unmanned driving remote control collaborative decision-making system, such as Figure 4 As shown, the system is used to perform the method as described above, and the system includes an acquisition module 100, a model construction module 200 and a processing module 300. The acquisition module 100 is used to acquire fault data. The model construction module 200 is coupled to the acquisition module 100, and the model construction module 200 is used to construct a prediction model and a data model. The processing module 300 is coupled to the acquisition module 100, and the processing module 300 is also coupled to the model construction module 200, and the processing module 300 is used to predict the number of faults according to the fault data and the prediction model.

[0079] Through the description of the above implementation methods, the technicians in the relevant field can clearly understand that the prediction method in the above embodiment can be implemented by means of software plus the necessary general hardware platform, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the relevant technology, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, disk, CD), including a number of instructions to enable a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods of various embodiments of the present invention.

[0080] The above contents are only specific implementation methods of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions within the technical scope disclosed in the present application shall be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. An unmanned remote control collaborative decision-making method, applied to vehicle control in a mining area, wherein the mining area includes a plurality of partitions, each of which has a corresponding video screen, characterized in that: include: Obtaining the failure factor of the vehicle failure; constructing a prediction model based on the failure factors, the prediction model being used to predict the number of failures of vehicles in a mining area; Predicting the number of failures in each partition based on the prediction model, and determining a key partition based on the number of failures in each partition; Pushing the video images corresponding to the key partitions to the remote operator; Acquire the travel route and current status information of each vehicle in the key zone, wherein the current status information includes current speed and current position; Determining a crossing time of the two vehicles based on the travel route and the current state information; When the intersection time of the two vehicles is within the intersection threshold and the current speed of one of the vehicles is abnormal, it is determined that the collision possibility of the two vehicles reaches the collision threshold, and the video screen of the key partition is switched to the video screen of the vehicle.

2. The unmanned driving remote control collaborative decision-making method according to claim 1 is characterized in that: The remote operator obtains video images through a monitoring device, and the video images corresponding to the key partitions are pushed to the remote operator, including: sorting the multiple key partitions based on the number of faults corresponding to the key partitions, and pushing the key partitions in sequence according to the ranking to the middle area of ​​the display page of the monitoring device.

3. The unmanned driving remote control collaborative decision-making method according to claim 2 is characterized in that: The switching the video picture of the key partition to the video picture of the vehicle includes: pushing the video picture of the vehicle to a display page corresponding to the key partition.

4. The unmanned driving remote control collaborative decision-making method according to claim 1 is characterized in that: The speed of the vehicles when they meet is a first speed. When the difference between the current speed of the vehicles when they meet and the first speed is greater than a speed threshold, it is determined that the current speed of the vehicle is abnormal.

5. The unmanned driving remote control collaborative decision-making method according to claim 1 is characterized in that: The expression of the prediction model is: ； in, is the number of failures at time t, is the number of vehicles that have not failed at time t, m represents the total number of vehicles, is the first failure factor, The second fault factor.

6. The unmanned driving remote control collaborative decision-making method according to claim 5 is characterized in that: The fault factors include a first fault factor and a second fault factor, wherein the first fault factor is used to represent an initial rate of vehicle faults, and the second fault factor is used to represent a rate at which vehicle faults increase with use time and environmental influences.

7. The unmanned driving remote control collaborative decision-making method according to claim 1 is characterized in that: Predicting the number of failures of each partition based on the prediction model includes: Acquire historical fault data, wherein the fault data includes the number of faults in multiple different time periods; Obtaining an intermediate value of the fault factor according to the historical fault data, the current actual fault data and the prediction model; The number of failures of each partition is predicted according to the prediction model corresponding to the intermediate value of the failure factor.

8. The unmanned driving remote control collaborative decision-making method according to claim 7 is characterized in that: According to the historical fault data, the current actual fault data and the prediction model, an intermediate value of the fault factor is obtained, including: constructing a data model according to the historical fault data; Acquire first predicted fault data according to the data model; Obtaining second predicted fault data according to current actual fault data and the first predicted fault data; The intermediate value of the fault factor is obtained according to the data of the trend node in the second predicted fault data and the prediction model.

9. The unmanned driving remote control collaborative decision-making method according to claim 8, characterized in that: The obtaining second predicted fault data according to the current actual fault data and the first predicted fault data includes: Obtaining an average deviation between characteristic data in the first predicted fault data and the current actual fault data, the characteristic data being data in the first predicted fault data corresponding to the current actual fault data; The first predicted fault data is corrected according to the average deviation to obtain the second predicted fault data.

10. An unmanned driving remote control collaborative decision-making system, characterized in that: The system is used to perform the method according to any one of claims 1 to 9, and the system comprises: An acquisition module, wherein the acquisition module is used to acquire fault data; A model construction module, coupled to the acquisition module, the model construction module is used to construct a prediction model and a data model; and a processing module, wherein the processing module is coupled to the acquisition module and is also coupled to the model construction module, and the processing module is used to predict the number of faults based on the fault data and the prediction model.

Images