Predictive Control Methods and Systems Based on Road Adhesion Coefficient

Abstract

This application relates to the field of vehicle control technology and discloses a predictive control method and system based on road surface adhesion coefficient. In this method, a cloud server acquires real-time information uploaded by vehicles on the road. The cloud server can combine the real-time information with historical information from each grid to form a road surface adhesion coefficient matrix. The cloud server sends the road surface adhesion coefficient matrix to the vehicle. Based on the road surface adhesion coefficient matrix, the vehicle's current position, and navigation path, the vehicle determines the predicted road surface adhesion coefficient value in its current direction of travel. Based on the predicted road surface adhesion coefficient value, the vehicle calculates the upper limit of its regenerative braking capacity and / or the upper limit of its driving torque, and then controls the vehicle accordingly. The beneficial effect is improved vehicle control efficiency on low-adhesion roads.

Description

Technical Field

[0001] This application relates to the field of vehicle control technology, and in particular to a predictive control method and system based on road surface adhesion coefficient. Background Technology

[0002] With the booming development of the new energy vehicle industry, vehicle safety and driving experience under complex road conditions are receiving increasing attention.

[0003] Currently, traditional vehicles rely heavily on data from their own sensors to calculate the coefficient of friction. Vehicle controllers typically need to acquire this sensor data, calculate the coefficient of friction based on it, and then control the vehicle accordingly.

[0004] However, this traditional method has a certain time delay. When dealing with anomalies such as those at elevated road expansion joints, it's easy for the vehicle to have already moved out of the danger zone by the time the system intervenes. This results in low control efficiency. Summary of the Invention

[0005] This application provides a predictive control method and system based on the road surface adhesion coefficient, which solves the technical problem of low control efficiency of vehicles on low-adhesion roads in the prior art, and achieves the technical effect of improving the control efficiency of vehicles on low-adhesion roads.

[0006] To achieve the above objectives, the main technical solutions adopted in this application include:

[0007] In a first aspect, embodiments of this application provide a predictive control method based on road surface adhesion coefficient, applied to vehicles, the method comprising:

[0008] The road surface adhesion coefficient matrix of the road where the vehicle is currently located is obtained from the cloud server; and based on the road surface adhesion coefficient matrix, the current vehicle position and navigation path, the predicted value of the road surface adhesion coefficient of the current vehicle in the driving direction is determined; wherein, the road surface adhesion coefficient matrix is ​​calculated by the cloud server by fusing the real-time road surface adhesion coefficient uploaded by the vehicle on the road and the historical road surface adhesion coefficient of the road.

[0009] Based on the preset control coefficient, the vehicle parameters, and the predicted value of the road surface adhesion coefficient, the upper limit of the regenerative braking capacity and / or the upper limit of the driving torque of the vehicle are calculated; based on the current state of the vehicle, and the upper limit of the regenerative braking capacity and / or the upper limit of the driving torque of the vehicle, the vehicle is controlled to perform braking.

[0010] In this embodiment, by obtaining the predicted value of the road surface adhesion coefficient in the vehicle's driving direction from the cloud server, and calculating the upper limit of regenerative braking and driving torque based on multiple parameters to control the vehicle's braking, a method of predicting the road surface adhesion coefficient in advance based on cloud-vehicle collaboration is achieved. This allows the vehicle to perform feedforward control based on the predicted road surface adhesion coefficient value, thereby improving the vehicle's control efficiency based on the road surface adhesion coefficient, enhancing vehicle safety, and improving the user experience.

[0011] Secondly, embodiments of this application provide a predictive control method based on road surface adhesion coefficient, applied to a cloud server, the method comprising:

[0012] The system acquires real-time information uploaded by vehicles on the road, including the real-time road surface adhesion coefficient calculated by the vehicle based on sensor monitoring information and the vehicle's GPS information.

[0013] Based on the GPS information of each vehicle, the real-time information of each vehicle is mapped to a road grid; the road grid is obtained by the cloud server dividing the roads within a preset area based on a preset grid side length;

[0014] By integrating the real-time information and historical information of each grid, the predicted value and confidence level of the road surface adhesion coefficient of each grid are obtained.

[0015] In this embodiment, by acquiring real-time information uploaded by the vehicle and writing the real-time road surface adhesion coefficient in the real-time information into the corresponding grid based on the GPS information in the real-time information, and then calculating the predicted value and confidence level of the road surface adhesion coefficient of the grid based on the real-time information and historical information of the grid, the method of accurately predicting the predicted value and confidence level of the road surface adhesion coefficient corresponding to each grid on the road is achieved, which provides a basis for the feedforward control of the vehicle.

[0016] Thirdly, embodiments of this application provide a predictive control device based on the road surface adhesion coefficient, applied to a vehicle, comprising:

[0017] The acquisition module is used to obtain the road surface adhesion coefficient matrix of the road where the vehicle is currently located from the cloud server; and based on the road surface adhesion coefficient matrix, the current vehicle position and navigation path, determine the predicted value of the road surface adhesion coefficient of the current vehicle in the driving direction; wherein, the road surface adhesion coefficient matrix is ​​calculated by the cloud server by fusing the real-time road surface adhesion coefficient uploaded by the vehicle on the road and the historical road surface adhesion coefficient of the road.

[0018] The control module is used to calculate the upper limit of the vehicle's regenerative braking capacity and / or the upper limit of the driving torque based on preset control coefficients, vehicle parameters, and predicted road adhesion coefficient; and to control the vehicle to perform braking based on the vehicle's current state and the upper limit of the vehicle's regenerative braking capacity and / or the upper limit of the driving torque.

[0019] Fourthly, embodiments of this application provide a predictive control device based on the road surface adhesion coefficient, applied to a cloud server, comprising:

[0020] The acquisition module is used to acquire real-time information uploaded by vehicles on the road. The real-time information includes the real-time road surface adhesion coefficient calculated by the vehicle based on sensor monitoring information and the vehicle's GPS information. Based on the GPS information of each vehicle, the real-time information of each vehicle is mapped to the road grid. The road grid is obtained by the cloud server by dividing the road within a preset area based on the preset grid side length.

[0021] The fusion module is used to fuse real-time information and historical information from each grid to obtain the predicted value and confidence level of the road surface adhesion coefficient for each grid.

[0022] Fifthly, embodiments of this application provide a vehicle including a controller. The controller includes a memory and a processor, which are communicatively connected. The memory stores computer instructions, and the processor executes the computer instructions to perform the method described in any of the above embodiments.

[0023] Sixthly, embodiments of this application provide a cloud server, the cloud server comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to perform the method described in any of the above embodiments.

[0024] Seventhly, embodiments of this application provide a predictive control system based on road surface adhesion coefficient, including a vehicle and a cloud server.

[0025] Eighthly, embodiments of this application provide a computer-readable storage medium storing computer instructions, the computer instructions being configured to cause a computer to perform the method described in any of the first aspects above, or any of the methods described in the second aspect.

[0026] Ninthly, embodiments of this application provide a computer program product, including computer instructions, which are used to cause a computer to perform the method described in any of the first aspects above, or any of the methods described in the second aspect. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the specific embodiments of this application or the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0028] Figure 1 A flowchart illustrating a predictive control method based on road surface adhesion coefficient provided in this application embodiment;

[0029] Figure 2 A flowchart illustrating a predictive control method for a vehicle provided in an embodiment of this application;

[0030] Figure 3 A flowchart of a vehicle safety mode control provided in this application embodiment;

[0031] Figure 4 A flowchart illustrating a predictive control method based on road surface adhesion coefficient provided in this application embodiment;

[0032] Figure 5 A flowchart for confidence calculation provided in this application embodiment;

[0033] Figure 6 A schematic diagram of signaling interaction for a predictive control method based on road surface adhesion coefficient provided in an embodiment of this application;

[0034] Figure 7 A schematic diagram of signaling interaction for a predictive control method based on road surface adhesion coefficient provided in an embodiment of this application;

[0035] Figure 8 A structural diagram of a predictive control device based on road surface adhesion coefficient provided in an embodiment of this application;

[0036] Figure 9 A structural diagram of a predictive control device based on road surface adhesion coefficient provided in an embodiment of this application;

[0037] Figure 10 A schematic diagram of the structure of a vehicle provided in an embodiment of this application;

[0038] Figure 11 This is a schematic diagram of the structure of a cloud server provided in an embodiment of this application. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0040] Traditional vehicles rely solely on their own wheel speed sensors to calculate the coefficient of friction (μ), resulting in significant response lag. In real-world driving scenarios, as vehicles sequentially traverse the same road surface, their perception information is isolated and not shared. When a vehicle ahead detects a low-friction surface, such as ice or a slippery section, following vehicles must reach the same location to perceive the change in road conditions.

[0041] In daily driving, elevated expansion joints are a common driving condition. Because traditional vehicles are slow to detect changes in road surface adhesion coefficient and have a short slip distance on low-traction surfaces, by the time a vehicle reaches a low-traction area and detects the reduced adhesion coefficient, it has often already left that area. Applying control measures at this point can cause a sudden change in vehicle power, leading to abrupt acceleration. This sudden change not only causes significant discomfort to the driver and passengers but can also easily trigger driving panic.

[0042] In summary, current vehicles generally rely on a single source of perception data, consisting solely of real-time feedback data from their own sensors (such as wheel speed sensors and IMUs). Due to a lack of ability to predict road conditions ahead, ABS (Anti-lock Braking System) / TCS (Traction Control System) typically only activates when the wheels begin to slip. Furthermore, this response delay is significant, reaching or exceeding 150ms.

[0043] Taking elevated road expansion joint surfaces (which have a low coefficient of adhesion, i.e., low μ surfaces) as an example, according to current common solutions, when the system control intervenes, the vehicle has often already moved away from the danger zone. Moreover, when traditional feedback control intervenes, it causes a sudden change in drive or regenerative torque, which can cause discomfort to the driver and passengers, and even trigger panic.

[0044] To address the aforementioned issues, this application focuses on resolving the problems of slow response and poor user experience of new energy vehicles in complex road conditions under traditional sensing methods. Based on V2X data fusion sensing technology, a feedforward control strategy that combines energy recovery with drive anti-skid is proposed.

[0045] During the operation of new energy vehicles, especially in complex road conditions such as expansion joints on elevated bridges, icy and snowy roads, and slippery curves, the slow response of energy recovery and drive anti-skid control is a prominent issue. To address this, this application achieves predictive control of vehicles through vehicle-group data collaboration, solving the following two major challenges:

[0046] Firstly, traditional regenerative braking control methods cannot predict road conditions ahead, forcing vehicles to adopt conservative recovery strategies, such as setting a fixed deceleration of 0.3g. This strategy, when driving on high-friction surfaces (where the road surface has a high coefficient of friction), results in insufficient energy recovery potential and energy waste.

[0047] Secondly, even with a recovery deceleration limit of 0.3g, the vehicle may still exhibit instability when it reaches low-adhesion areas (such as when the road surface adhesion coefficient μ < 0.3g). To ensure driving stability, the system will quickly exit the energy recovery mode. However, due to the limitation of the electronic braking response speed, friction braking cannot replenish braking force in time, resulting in a loss of braking deceleration and causing panic among the driver.

[0048] Specifically, this application proposes a predictive control method based on the road surface adhesion coefficient. This method dynamically adjusts the vehicle's energy recovery limit based on the road surface adhesion coefficient μ estimated from the vehicle ahead. For example, if the road surface adhesion coefficient μ is greater than or equal to 0.8, it can be determined that the road surface ahead is a high-adhesion road surface, and the vehicle can allow a deceleration of 0.6g for energy recovery. Conversely, if the road surface adhesion coefficient μ is less than or equal to 0.3, it can be determined that the road surface ahead is a low-adhesion road surface, and the vehicle can limit the deceleration of energy recovery to 0.2g.

[0049] Furthermore, this application can also employ a feedforward continuous control strategy. That is, by using data from the leading vehicle, a certain distance in advance, such as 200m, the recovery torque of energy recovery is adjusted to a reasonable level, avoiding a "one-size-fits-all" approach to recovery values ​​that would waste recovery capacity.

[0050] Furthermore, traditional methods, which primarily use wheel speed sensors to identify the road surface adhesion coefficient, suffer from slow recognition and slow control response. For example, when accelerating on an elevated road in rainy weather, if the vehicle encounters an expansion joint with an adhesion coefficient μ that may be less than 0.3, the TCS (Traction Control System) will be triggered abruptly to reduce torque due to detected slippage. However, because the expansion joint is short and the slippage detection speed is slow, by the time the TCS triggers and executes the torque reduction, the vehicle has already moved out of the expansion joint. In this case, the TCS not only fails to provide control but also causes a sudden decrease in acceleration.

[0051] To address this situation, this application employs feedforward dynamic control by obtaining the estimated adhesion coefficient μ from the preceding vehicle. This allows the vehicle to proactively reduce torque in advance for road sections where the adhesion coefficient μ suddenly decreases ahead, effectively preventing the aforementioned situation from occurring, improving vehicle smoothness, and enhancing the user experience.

[0052] As can be seen, compared to traditional vehicles that need to acquire data from their own sensors for feedback control, resulting in a processing and control delay of more than 150ms, this application combines data from the leading vehicle, historical data, and data from the vehicle's own sensors to predict road conditions ahead in advance through feedforward control and pre-plan the processing, thereby reducing the delay, improving the smoothness of vehicle driving, and enhancing the user experience.

[0053] According to an embodiment of this application, a predictive control method based on road surface adhesion coefficient is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed on a computer device via a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here. The computer device can be a mobile terminal, a personal computer, a server, etc.

[0054] Figure 1 A flowchart of a predictive control method based on road surface adhesion coefficient provided for embodiments of this application is shown below. Figure 1 As shown, taking the vehicle as the execution subject, the vehicle can be equipped with two parts: a regenerative braking module and a traction control module. The regenerative braking module calculates the upper limit of regenerative braking capacity based on the predicted road surface adhesion coefficient, and then optimizes its control based on this upper limit. The traction control module calculates the upper limit of driving torque based on the predicted road surface adhesion coefficient, and then optimizes its control based on this upper limit. This optimization control of both the regenerative braking module and the traction control module based on the predicted road surface adhesion coefficient is essentially a vehicle-based feedforward control system. It addresses the problems of slow response or inflexible "one-size-fits-all" limit methods used in traditional solutions. The process includes the following steps:

[0055] S101. Obtain the road surface adhesion coefficient matrix of the road where the vehicle is currently located from the cloud server. Based on the road surface adhesion coefficient matrix, the current vehicle position, and the navigation path, determine the predicted value of the road surface adhesion coefficient for the current vehicle in the driving direction. The road surface adhesion coefficient matrix is ​​calculated by the cloud server by fusing real-time road surface adhesion coefficients uploaded by vehicles on the road and historical road surface adhesion coefficients.

[0056] For example, a vehicle can obtain the road surface adhesion coefficient matrix of the road where it is currently located from a cloud server via communication. Based on this road surface adhesion coefficient matrix, combined with the current vehicle position and navigation path, the vehicle can accurately determine the predicted value of the road surface adhesion coefficient in the current driving direction.

[0057] In one implementation, the road surface adhesion coefficient matrix is ​​a matrix composed of the predicted road surface adhesion coefficient values ​​corresponding to each grid after dividing the road in front of the vehicle into grids.

[0058] In one implementation, the vehicle can obtain the road surface adhesion coefficient matrix generated by the cloud server after aggregating the data uploaded by various vehicles on the road.

[0059] In one implementation, the cloud server can calculate the road surface adhesion coefficient matrix by combining the road surface adhesion coefficient data uploaded in real time by vehicles on the road with the historical road surface adhesion coefficient data of the road.

[0060] In one implementation, the wireless communication between the vehicle and the cloud server can be V2X communication.

[0061] In one implementation, the vehicle can obtain the road surface adhesion coefficient matrix by receiving a broadcast sent by a cloud server.

[0062] In another implementation, the vehicle can obtain the predicted value of the road adhesion coefficient in the driving direction directly from the cloud server by sending its location to the cloud server.

[0063] S102. Based on the preset control coefficients, vehicle parameters, and predicted road surface adhesion coefficient, calculate the upper limit of the vehicle's regenerative braking capacity and / or the upper limit of its driving torque; based on the vehicle's current state and the upper limit of its regenerative braking capacity and / or the upper limit of its driving torque, control the vehicle to perform braking.

[0064] For example, the vehicle calculates the upper limit of its regenerative braking capacity and / or the upper limit of its driving torque based on preset control coefficients, vehicle parameters, and the predicted value of the road adhesion coefficient obtained in step S101. Then, considering the vehicle's current state, the vehicle is controlled to perform a corresponding energy recovery operation based on this upper limit of regenerative braking capacity. And / or, considering the vehicle's current state, the vehicle is controlled to perform a corresponding torque reduction operation based on the upper limit of driving torque.

[0065] In one implementation, the preset control coefficient is set based on a combination of factors such as vehicle type and driving environment, and is used to adjust the weights and proportions in the calculation process.

[0066] In one implementation, the vehicle parameters include basic parameters such as the vertical force on the shaft where the drive motor is located and the wheel rolling radius. These parameters are the basis for calculating the upper limit of regenerative braking capacity and the upper limit of drive torque.

[0067] In one implementation, the predicted road surface adhesion coefficient reflects the road surface adhesion coefficient in the current vehicle's travel direction. This predicted road surface adhesion coefficient is calculated by a cloud server based on real-time vehicle-uploaded information and historical information.

[0068] In one implementation, the vehicle can calculate the upper limit of the regenerative braking capacity by multiplying the control coefficient of the regenerative braking capacity, the vertical force on the shaft where the drive motor is located, and the predicted value of the road adhesion coefficient. The formula can be:

[0069]

[0070] in, The calculated upper limit of regenerative braking capacity. This is the control coefficient corresponding to the regenerative braking capability. This refers to the vertical force on the shaft where the drive motor is located. It can be obtained through calculations by other modules. This is the predicted value of the road surface adhesion coefficient.

[0071] In one implementation, the vehicle can calculate the upper limit of the drive torque by multiplying the control coefficient of the drive torque, the vertical force on the shaft where the drive motor is located, the predicted value of the road adhesion coefficient, and the wheel rolling radius. The formula can be:

[0072]

[0073] in, Upper limit of drive torque. This is the control coefficient corresponding to the driving torque. This refers to the vertical force on the shaft where the drive motor is located. This is the predicted value of the road surface adhesion coefficient. This is the rolling radius of the wheel.

[0074] In this embodiment, by obtaining the predicted value of the road surface adhesion coefficient in the vehicle's driving direction from the cloud server, and calculating the upper limit of regenerative braking and driving torque based on multiple parameters to control the vehicle's braking, a method of predicting the road surface adhesion coefficient in advance based on cloud-vehicle collaboration is achieved. This allows the vehicle to perform feedforward control based on the predicted road surface adhesion coefficient value, thereby improving the vehicle's control efficiency based on the road surface adhesion coefficient, enhancing vehicle safety, and improving the user experience.

[0075] In one example, the specific execution process in step S102 above can be as follows: Figure 2As shown. The vehicle receives the road surface adhesion coefficient. After obtaining the predicted values, the corresponding optimized controller can be executed in both torque-driven and regenerative braking scenarios.

[0076] Regarding regenerative braking, the vehicle's feedforward control based on the predicted road adhesion coefficient may include: calculating an upper limit for regenerative braking capacity; and limiting the regenerative torque based on this upper limit to prevent fishtailing.

[0077] In terms of torque drive, the vehicle's feedforward control based on the predicted road adhesion coefficient may include: calculating the upper limit of the drive torque and limiting the drive torque based on the upper limit of the drive torque to prevent slippage.

[0078] The process can be illustrated in the following two examples.

[0079] In one example, in step S102 above, controlling the vehicle to perform braking based on the vehicle's current state and the vehicle's regenerative braking capacity upper limit includes:

[0080] S1021. The smaller of the vehicle's current braking capacity and the upper limit of regenerative braking capacity is used as the target regenerative braking force.

[0081] For example, when a vehicle performs braking control, it can first compare the vehicle's current braking capacity with the upper limit of regenerative braking capacity. The vehicle can then select the smaller value between the current braking capacity and the upper limit of regenerative braking capacity as the target regenerative braking force.

[0082] In one implementation, the vehicle's current braking capacity refers to the actual braking capacity that the vehicle can provide in its current state. This braking capacity is affected by various factors such as the state of the vehicle's braking system and the friction between the tires and the road surface.

[0083] In one implementation, the upper limit of regenerative braking capacity is the maximum capacity that the vehicle can achieve in terms of regenerative braking, which is calculated in advance based on vehicle parameters, road surface adhesion coefficient prediction, etc.

[0084] In one implementation, if the current braking capacity is less than the upper limit of the regenerative braking capacity, it indicates that the braking capacity currently provided by the vehicle is within a safe range and will not exceed the maximum capacity value, thus preventing slippage. Therefore, the vehicle can use this current braking capacity as the target regenerative braking force.

[0085] In one implementation, if the current braking capacity is greater than or equal to the upper limit of the regenerative braking capacity, it indicates that the braking capacity currently provided by the vehicle exceeds the safe range. If the vehicle continues to use this braking capacity, skidding may occur. Therefore, the vehicle can use this upper limit of the regenerative braking capacity as the target regenerative braking force.

[0086] In one implementation method, the formula for calculating the target regenerative braking force can be:

[0087]

[0088] in, To regenerate braking force for the target. This represents the vehicle's current braking capacity. The calculated upper limit of regenerative braking capacity.

[0089] S1022. Calculate the difference between the current braking capacity and the target regenerative braking force to obtain the target friction braking force.

[0090] For example, after determining the target regenerative braking force, the vehicle further calculates the target frictional braking force. Specifically, the vehicle can use the difference between the vehicle's current braking capacity and the target regenerative braking force as the target frictional braking force.

[0091] In one implementation, the target frictional braking force is the braking force that the frictional braking system needs to provide in order to supplement the braking demand that regenerative braking cannot meet.

[0092] In one implementation, the formula for calculating the target friction braking force can be:

[0093]

[0094] in, It is the friction braking force. This represents the vehicle's current braking capacity. To regenerate braking force for the target.

[0095] In another implementation, the vehicle can also establish a braking demand model, input parameters such as current braking capacity and target regenerative braking force into the model, and obtain the target friction braking force through the model's calculation rules, thereby improving the calculation accuracy of braking capacity.

[0096] S1023. Control the vehicle to perform regenerative braking based on the target regenerative braking force and the target friction braking force.

[0097] For example, the vehicle can control the vehicle to perform regenerative braking based on the calculated target regenerative braking force and target friction braking force.

[0098] In one implementation, the vehicle can send control commands to the regenerative braking system and the friction braking system respectively based on the calculated target regenerative braking force and target friction braking force, so that the regenerative braking system brakes according to the target regenerative braking force, and the friction braking system brakes according to the target friction braking force, and the two work together to achieve vehicle braking.

[0099] In one implementation, the vehicle can use open-loop control, directly generating commands based on the calculated target regenerative braking force and target friction braking force, and then sending and executing them.

[0100] In another implementation, the vehicle can also employ closed-loop control, monitoring its braking status in real time during braking. The vehicle can adjust its braking based on the monitoring results to ensure that the braking state meets the target regenerative braking force and the target frictional braking force, thereby ensuring optimal braking performance.

[0101] In this example, the target regenerative braking force is determined by comparing the smaller value between the current braking capacity and the upper limit of the regenerative braking capacity, and the target frictional braking force is obtained by calculating the difference. Then, based on the two, the vehicle is controlled to perform regenerative braking, so as to achieve the effect of reasonable distribution of regenerative braking and frictional braking during vehicle braking, and improve braking efficiency and stability.

[0102] In one example, in step S102 above, based on the current state of the vehicle and the upper limit of the vehicle's driving torque, controlling the vehicle to perform braking includes:

[0103] S1024. Based on the pedal opening or automatic driving command, the total required torque is obtained by analysis.

[0104] For example, during vehicle operation, the required driving force needs to be determined based on the driver's actions or instructions from the autonomous driving system. Typically, the vehicle can obtain pedal opening information through sensors or receive autonomous driving instructions from the autonomous driving system. The vehicle can then analyze and process the pedal opening or autonomous driving instructions to obtain the total required torque.

[0105] In one implementation, pedal opening refers to the degree to which the driver depresses the accelerator or brake pedal. Pedal opening directly reflects the driver's demand for vehicle power or braking.

[0106] In one implementation, the vehicle can acquire the pedal opening in real time using an angle sensor or displacement sensor mounted on the pedal. The vehicle can then analyze the data based on a pre-defined relationship between pedal opening and torque to obtain the total required torque.

[0107] In one implementation, the autonomous driving command is a command issued by the autonomous driving system to control the vehicle's movement after making a comprehensive judgment based on road conditions, navigation information, and other factors.

[0108] In one implementation, the vehicle can parse the torque value indicated in the autonomous driving command to determine the total required torque.

[0109] S1025, The smaller of the upper limit of the driving torque and the total required torque is taken as the actual output torque.

[0110] For example, after obtaining the total required torque, the vehicle needs to determine the actual torque it can output based on its own driving capabilities. The vehicle can compare the previously calculated upper limit of the driving torque with the total required torque and select the smaller value as the actual output torque.

[0111] In one implementation, the upper limit of the driving torque is the maximum torque value that the vehicle can provide in terms of driving, calculated in advance based on its own parameters, road surface adhesion coefficient, and other factors. Driving torque ensures the safety and stability of the vehicle during operation.

[0112] In one implementation, the formula for calculating the actual output torque can be:

[0113]

[0114] in, This represents the actual output torque. Upper limit of drive torque. This represents the total required torque.

[0115] S1026. Adjust the drive torque of the vehicle according to the actual output torque.

[0116] For example, the vehicle adjusts its drive torque based on the calculated actual output torque to achieve the driving state desired by the driver or the autonomous driving system.

[0117] In one implementation, the vehicle uses the actual output torque as the control target, and changes the vehicle's driving torque by adjusting the engine's output power, the electric motor's torque, etc., to achieve the required actual output torque.

[0118] In one implementation, when the vehicle is a gasoline-powered vehicle, the vehicle can change the engine's output power by adjusting parameters such as the engine's throttle opening and fuel injection quantity, thereby adjusting the drive torque.

[0119] In one implementation, when the vehicle is an electric vehicle, the vehicle can adjust the driving torque by controlling the magnitude and direction of the electric motor's current to change the motor's torque.

[0120] In one implementation, the vehicle can also use closed-loop control to monitor changes in drive torque in real time during the adjustment process, and dynamically optimize the adjustment strategy based on the monitoring results to ensure that the drive torque can accurately and quickly meet the requirements of the actual output torque.

[0121] In this example, the total required torque is analyzed based on the pedal opening or autonomous driving command, and the smaller value between the upper limit of the driving torque and the total required torque is selected as the actual output torque to control the vehicle to adjust the driving torque. This achieves the effect of accurately matching the vehicle's driving torque to the demand and ensuring driving power and stability.

[0122] In one example, the process of controlling the vehicle to perform braking based on the vehicle's upper limit of regenerative braking capacity and / or upper limit of drive torque also includes:

[0123] S103. If, during the adjustment of the regenerative braking force of the vehicle based on the target regenerative braking force, the first change per unit time is greater than a preset first change threshold, then the regenerative braking force at the current moment is adjusted according to the first change threshold and the regenerative braking force at the previous moment; and / or, if, during the adjustment of the output torque of the vehicle based on the actual output torque, the second change per unit time is greater than a preset second change threshold, then the output torque at the current moment is adjusted according to the second change threshold and the output torque at the previous moment.

[0124] For example, during the process of adjusting the vehicle's regenerative braking force or torque based on the target regenerative braking force and / or the actual output torque, in order to ensure the smooth operation of the vehicle, the vehicle will continuously monitor the changes in the regenerative braking force or output torque per unit time and record the first change in the regenerative braking force or the second change in the output torque.

[0125] The vehicle will compare the first change with a preset first change threshold. Alternatively, the vehicle will compare the second change with a preset second change threshold. If the first change is detected to be greater than the preset first change threshold, or the second change is detected to be greater than the preset second change threshold, then the vehicle will activate the adjustment mechanism.

[0126] In one implementation, the vehicle determines its current regenerative braking force based on a pre-set first change threshold and the regenerative braking force recorded at the previous moment. For example, the controller can use the sum of the regenerative braking force recorded at the previous moment and the first change threshold as the current regenerative braking force. And / or, the vehicle determines its current output torque based on a pre-set second change threshold and the output torque recorded at the previous moment. For example, the controller can use the sum of the output torque recorded at the previous moment and the second change threshold as the current output torque.

[0127] In another implementation, the vehicle determines the regenerative braking force and output torque at the current moment based on a preset maximum decay value. For example, the maximum decay value of the regenerative braking force per second can be 600 Nm.

[0128] In one implementation, when the predicted value of the road surface adhesion coefficient drops sharply, the variable may exceed the change threshold.

[0129] In this example, by adjusting the corresponding value at the current moment based on the threshold and the value at the previous moment when the change in the target regenerative braking force and / or the actual output torque per unit time exceeds a preset threshold, the effect of smooth transition of vehicle power output and improved driving comfort and stability is achieved.

[0130] In one example, the road adhesion coefficient matrix also includes a confidence level for each road adhesion coefficient. The vehicle can then be further controlled based on this confidence level. This process may include:

[0131] S104. If the confidence level of the predicted road surface adhesion coefficient for the current vehicle in the driving direction is less than or equal to a preset confidence threshold, then switch to single-vehicle control mode. Alternatively, if the current vehicle does not obtain the road surface adhesion coefficient matrix sent by the cloud server within a preset time period, then switch to single-vehicle control mode. The single-vehicle control mode indicates that the current vehicle calculates the real-time road surface adhesion coefficient based on sensor monitoring information and then performs regenerative braking control and torque control based on the real-time road surface adhesion coefficient.

[0132] For example, during vehicle operation, the reliability of the road surface adhesion coefficient prediction value obtained based on cloud data and the communication status with the cloud are evaluated in real time. When the vehicle determines that the confidence level of the road surface adhesion coefficient prediction value in the current driving direction is less than or equal to a preset confidence threshold, it means that the road surface adhesion coefficient prediction value provided by the cloud has low reliability, and the vehicle will switch to single-vehicle control mode.

[0133] In one implementation, while obtaining the predicted value of the road adhesion coefficient in the driving direction from the road adhesion coefficient matrix, the vehicle can also obtain the confidence level of the predicted value of the road adhesion coefficient from the same matrix.

[0134] One approach is to use confidence level as a quantitative evaluation index for the accuracy and reliability of predicted road surface adhesion coefficients. This confidence level comprehensively considers factors such as data sources, calculation methods, and comparisons with historical data.

[0135] In one implementation, if the confidence level of the predicted road surface adhesion coefficient is less than or equal to the confidence threshold, then the predicted road surface adhesion coefficient is considered unreliable. In this case, to improve the safety of normal vehicle operation, the vehicle can discard the predicted road surface adhesion coefficient and switch to single-vehicle control mode.

[0136] One implementation involves setting a confidence threshold, which is a pre-defined standard value for the vehicle, to determine whether the predicted road adhesion coefficient is reliable.

[0137] In one implementation, the vehicle can dynamically adjust the confidence threshold based on its historical operating conditions. For example, if the vehicle experiences anomalies after multiple re-adjustments of braking force and output torque based on the confidence threshold, the controller can raise the confidence threshold. Otherwise, if the vehicle does not experience anomalies after multiple re-adjustments of braking force and output torque based on the confidence threshold, the controller may not adjust the confidence threshold.

[0138] For example, if the vehicle fails to obtain the road surface adhesion coefficient matrix sent by the cloud server within a preset time period, it indicates that there may be a problem with the communication between the vehicle and the cloud or that the cloud data is not updated in time. In this case, the vehicle will also switch to single-vehicle control mode.

[0139] In one implementation, the execution flow of the time-based security mode setting can be as follows: Figure 3 As shown, when a V2X signal is detected, the vehicle can perform feedforward control normally based on the road adhesion coefficient matrix obtained from V2X. Furthermore, the vehicle can switch to traditional single-vehicle mode when a V2X signal is detected as lost.

[0140] In one implementation, the cloud server can periodically update the road surface adhesion coefficient matrix and then broadcast it. If the cloud server malfunctions, it may be unable to send the road surface adhesion coefficient matrix correctly.

[0141] In one implementation, the vehicle can receive the road surface adhesion coefficient matrix via broadcast. If the vehicle-side receiving equipment malfunctions, it may be unable to receive the road surface adhesion coefficient matrix properly.

[0142] In one implementation, the preset duration is a time range pre-set by the vehicle based on factors such as the stability of network communication and the frequency of data updates. For example, the preset duration could be 10ms.

[0143] In one implementation, the single-vehicle control mode is an autonomous control method for vehicles when they cannot rely on cloud data. In this mode, the vehicle calculates the real-time road adhesion coefficient based on information monitored by various sensors on its own.

[0144] Optionally, the sensor may include a wheel speed sensor, an acceleration sensor, a gyroscope, etc.

[0145] Optionally, the vehicle can use these sensors to monitor its driving status and the interaction between the tires and the road surface in real time.

[0146] Optionally, the vehicle can use specific algorithm models to calculate the real-time road surface adhesion coefficient based on the data collected by these sensors. Then, the vehicle can perform regenerative braking control and torque control based on the calculated real-time road surface adhesion coefficient.

[0147] Optionally, in non-single-vehicle control mode, the vehicle can also calculate the real-time road surface adhesion coefficient based on the data collected by these sensors. The vehicle can then upload this real-time road surface adhesion coefficient to a cloud server.

[0148] In this example, when the confidence level of the predicted road surface adhesion coefficient is not up to standard or the cloud-based road surface adhesion coefficient matrix is ​​not obtained on time, the system switches to single-vehicle control mode. Based on the sensor, the system calculates the real-time road surface adhesion coefficient and performs regenerative braking and torque control accordingly. This achieves the autonomous and precise control of the vehicle's braking and driving under complex road conditions, ensuring driving safety and stability.

[0149] Figure 4 A flowchart illustrating a predictive control method based on road surface adhesion coefficient provided in this application embodiment is shown below. Figures 1 to 3 Based on the illustrated embodiments, as Figure 4 As shown, with the cloud server as the execution subject, the process includes the following steps:

[0150] S201. Obtain real-time information uploaded by vehicles on the road, including the real-time road surface adhesion coefficient calculated by the vehicle based on sensor monitoring information and the vehicle's GPS information.

[0151] For example, in order to achieve a comprehensive and accurate understanding of the road surface adhesion coefficient, the cloud server can obtain real-time information uploaded by vehicles on the road.

[0152] In one implementation, the real-time information may include the real-time road surface adhesion coefficient of the vehicles' location and the vehicles' GPS information. Based on the GPS information, the vehicles can determine the location corresponding to the real-time road surface adhesion coefficient.

[0153] In one implementation, the sensors in the vehicle may include wheel speed sensors, acceleration sensors, etc. Based on the information collected by these sensors, the vehicle can calculate the real-time road surface adhesion coefficient using a specific algorithm formula.

[0154] In one implementation, the cloud server can receive information uploaded by the vehicle in real time by establishing a stable wireless communication connection with the vehicle, such as a 4G or 5G network or a dedicated vehicle-to-everything (V2X) communication protocol.

[0155] In another implementation, the vehicle can broadcast this real-time information. The cloud server can then receive the broadcast and obtain the information.

[0156] In one implementation, the cloud server can be a road testing device or an edge device. Each cloud server can be fixedly associated with a road in a specific area. The cloud server can establish a communication connection with a vehicle after it enters the area, and can disconnect the communication connection with the vehicle after it leaves the area.

[0157] S202. Based on the GPS information of each vehicle, the real-time information of each vehicle is mapped to the road grid. The road grid is obtained by the cloud server by dividing the roads within a preset area based on the preset grid side length.

[0158] For example, the cloud server can divide the roads in its corresponding area into a grid. This grid can have a fixed side length. After obtaining real-time information uploaded by each vehicle, the cloud server can determine the grid corresponding to the GPS information included in the real-time information. The cloud server can then write this real-time information into the grid.

[0159] In one implementation, the road grid is created by a cloud server that divides a preset area into roads based on a preset grid side length for more detailed management of road information. Since each cloud server can be fixedly associated with the roads in a specific area, this grid can be continuously used after being divided during initial setup.

[0160] In one implementation, the preset area can be a specific area in a city, or a section of highway, etc.

[0161] In one implementation, the preset grid side length can be adjusted according to actual needs and computational accuracy. The smaller the side length, the more refined the road division, but the computational load will also increase accordingly. Therefore, technicians can determine the final grid side length by comprehensively considering factors such as the computing power of the cloud server, the accuracy requirements of vehicle driving, and the variation range of the road surface adhesion coefficient.

[0162] In one implementation, the cloud server can spatially divide the road based on the road's geographic information and according to a preset grid side length.

[0163] In one implementation, the real-time road surface adhesion coefficient can be written into the grid from real-time information.

[0164] S203. By integrating real-time information and historical information from each grid, the predicted value and confidence level of the road surface adhesion coefficient for each grid are obtained.

[0165] For example, the cloud server can periodically update and calculate the predicted road adhesion coefficient and confidence level for each grid. The cloud server can complete the calculation of the predicted road adhesion coefficient and confidence level for that grid based on the real-time information and historical information recorded in that grid within that period.

[0166] In one implementation, historical information refers to information already existing in the grid. For example, this historical information may include the actual pavement adhesion coefficient corresponding to the grid, collected and stored before the current processing cycle. Alternatively, it may include previously calculated pavement adhesion coefficient predictions and confidence levels for the grid.

[0167] In one implementation, the cloud server can fuse real-time and historical information using a weighted average method. Alternatively, the cloud server can also fuse features of the real-time and historical information using machine learning algorithms.

[0168] In one implementation, the confidence level is a quantitative assessment of the accuracy and reliability of the predicted road adhesion coefficient. This confidence level reflects the degree to which the predicted value closely approximates the actual value.

[0169] In this embodiment, by acquiring real-time information uploaded by the vehicle and writing the real-time road surface adhesion coefficient in the real-time information into the corresponding grid based on the GPS information in the real-time information, and then calculating the predicted value and confidence level of the road surface adhesion coefficient of the grid based on the real-time information and historical information of the grid, the method of accurately predicting the predicted value and confidence level of the road surface adhesion coefficient corresponding to each grid on the road is achieved, which provides a basis for the feedforward control of the vehicle.

[0170] In one example, the historical information includes the grid's predicted road adhesion coefficient and confidence level at the previous time step.

[0171] In step S203 above, the real-time information and historical information of each grid are fused to obtain the predicted value and confidence level of the road surface adhesion coefficient for each grid, including:

[0172] S2031. Based on the real-time pavement adhesion coefficient, road condition type, and rainfall from all real-time information in the grid, the confidence level of the grid at the current moment is calculated. Among them, the road condition type indicates the dryness of the road surface.

[0173] For example, after obtaining the real-time road surface adhesion coefficient from all the real-time information written by a grid in the current calculation cycle, the cloud server can combine the road state type of the road where the grid is located and the rainfall of the environment where the road is located to calculate the confidence level of the grid at the current moment through a specific algorithm model.

[0174] In one implementation, the road condition type is a classification information used to indicate the type and dryness of the road surface. For example, it may include dry asphalt, wet asphalt, snow asphalt, ice asphalt, dry cement, wet cement, snow cement, ice cement, and metal pavement.

[0175] Optionally, the cloud server can identify the type and dryness of the road surface based on the image information uploaded by the vehicle through image recognition methods.

[0176] Optionally, the cloud server can also identify the type and dryness of the road surface based on road monitoring images.

[0177] In one implementation, rainfall data is obtained in real time from meteorological monitoring equipment or vehicle rain sensors. A cloud server can obtain this rainfall data from the vehicle. Alternatively, the cloud device can obtain rainfall data uploaded by meteorological monitoring equipment from the network.

[0178] In one implementation, the cloud server can use a preset empirical formula to calculate the confidence level based on the real-time road surface adhesion coefficient, road condition type, and rainfall.

[0179] In another implementation, the cloud server can also use machine learning algorithms to predict the confidence level by taking the real-time road surface adhesion coefficient, road condition type and rainfall as input features.

[0180] S2032. Based on the preset first weight, second weight, the mean of the real-time road surface adhesion coefficient in all real-time information in the grid, the confidence level of the grid at the current moment, and the predicted value and confidence level of the road surface adhesion coefficient in the historical information of the grid, the predicted value of the road surface adhesion coefficient of the grid at the current moment is calculated.

[0181] For example, the cloud server may have a first weight and a second weight preset. The first weight may correspond to real-time information, and the second weight may correspond to historical information. The cloud server can also calculate the mean of the real-time road surface adhesion coefficients written to the grid within the current period. The cloud server can use the first weight combined with the mean in the real-time information and the confidence level calculated in step S2031 above, and use the second weight combined with the predicted road surface adhesion coefficients and confidence levels in the historical information, to obtain the predicted coefficient value of the grid at the current moment using a specific calculation method.

[0182] In one implementation, the first weight and the second weight are parameters pre-set by the cloud server to balance real-time and historical information. The values ​​of the first weight and the second weight can be adjusted according to actual needs and a large amount of experimental data. Optionally, the sum of the first weight and the second weight can be 1.

[0183] In one implementation, the average of all real-time road surface adhesion coefficients in the grid is calculated by averaging the real-time road surface adhesion coefficients uploaded by all vehicles within the grid. This average value reflects the overall situation of road surface adhesion coefficients within the grid in the current period.

[0184] In one implementation, the predicted values ​​and confidence levels of the road surface adhesion coefficient in the historical information of the grid are the predicted values ​​and confidence levels of the road surface adhesion coefficient calculated by the cloud server in the previous period.

[0185] In one implementation, the cloud server can calculate the predicted value of the road surface adhesion coefficient in the current period by using a weighted summation method.

[0186] In this example, confidence levels are calculated by integrating real-time information from the grid, road conditions, and rainfall. Then, by combining preset first and second weights, the mean and confidence levels of the road surface adhesion coefficient in the real-time information and the predicted and confidence levels of the road surface adhesion coefficient in the historical information are fused to obtain the predicted value of the road surface adhesion coefficient in the current period. This method achieves the effect of accurately predicting the grid road surface adhesion coefficient and improves the reliability and accuracy of road condition assessment.

[0187] In one example, in step S2031 above, based on the real-time pavement adhesion coefficient, road condition type, and rainfall from all real-time information in the grid, the confidence level of the grid at the current moment is calculated, including:

[0188] Step 11: Calculate the variance value based on the real-time road surface adhesion coefficient in all real-time information in the grid.

[0189] For example, the cloud server can first obtain all real-time information written into the grid within the current period. Then, the cloud server can calculate the variance of the real-time road surface adhesion coefficient contained in all the real-time information in the grid to obtain the variance value.

[0190] In one implementation, the formula for calculating the variance can be:

[0191]

[0192] in, This is the variance of the real-time road surface adhesion coefficient within a calculated grid. Let be the i-th real-time road surface adhesion coefficient. This is the average of all real-time road surface adhesion coefficients written to this grid during the current period. This represents the number of real-time road surface adhesion coefficients written to this grid during the current period.

[0193] Step 12: Normalize the variance values ​​to obtain the variance consistency score.

[0194] For example, after the cloud server obtains the variance value, it needs to be normalized to obtain the variance consistency score in order to facilitate subsequent comprehensive analysis and comparison.

[0195] In one implementation, the cloud server can map the variance value to a specific range using a preset formula, thereby normalizing the variance value.

[0196] In another implementation, the cloud server can normalize the difference based on an exponent. The formula can be:

[0197]

[0198] in, The variance consistency score. This is the conversion factor. The variance value for one grid.

[0199] Step 13: Perform a fusion calculation on the mapping values ​​of road condition type and rainfall to obtain the environmental consistency score.

[0200] For example, in addition to considering the dispersion of the real-time pavement adhesion coefficient, the cloud server also comprehensively considers the impact of road environmental factors on the pavement adhesion coefficient. The cloud server can map road condition type and rainfall separately to obtain their respective mapping values. Then, the cloud server can fuse these two mapping values ​​to obtain an environmental consistency score.

[0201] In one implementation, the road condition type indicates the dryness of the road surface. For example, it could be dry, wet, waterlogged, snow-covered, icy, etc. Furthermore, the dryness of the road surface can also include the type of road surface, such as asphalt, cement, or metal.

[0202] In one implementation, the cloud server can store mapping tables between the road condition type and rainfall. The cloud server can then calculate the mapping value between the road condition type and rainfall based on these mapping tables.

[0203] In one implementation, the cloud server can use a weighted summation method to fuse the mapping values ​​of the road state type and rainfall to obtain an environmental consistency score.

[0204] Step 14: Calculate the weighted sum of variance consistency score and environmental consistency score based on the preset third and fourth weights to obtain the confidence level of the grid at the current moment.

[0205] For example, the cloud server can also be configured with a third weight and a fourth weight. The third weight can correspond to the variance consistency score, and the fourth weight can correspond to the environmental consistency score. Based on these third and fourth weights, the cloud server can perform a weighted sum of the variance consistency score and the environmental consistency score to obtain the grid's confidence level at the current moment.

[0206] In one implementation, the third and fourth weights are parameters pre-set by the cloud server based on actual needs and experience, used to adjust the weight of variance consistency score and environmental consistency score in the final confidence calculation.

[0207] In one implementation, the cloud server can use an expert experience method, inviting experts in relevant fields to provide suggested weight values ​​based on actual road conditions and data characteristics.

[0208] In another implementation, the cloud server can adopt a data-driven approach, using a large amount of experimental data and machine learning algorithms, such as genetic algorithms and particle swarm optimization algorithms, to automatically search for the optimal weight combination, thereby improving the accuracy of confidence calculation.

[0209] In one implementation, the formula for calculating the degree of stability can be:

[0210]

[0211] in, , where is the confidence level. This represents the environmental consistency score. The variance consistency score. It is the third weight. It is the fourth weight. And. The and The sum is 1. For example, It is 0.8. It is 0.2.

[0212] In this example, the variance of the real-time road surface adhesion coefficient within the grid is calculated, and then the variance consistency score is calculated based on this variance. The environmental consistency score is obtained by fusing the mapping value between road conditions and rainfall. The weighted sum of the variance consistency score and the environmental consistency score is calculated to obtain the final confidence level. This method achieves the effect of accurately quantifying the confidence level of the grid at the current moment and improves the accuracy of road condition assessment.

[0213] In one example, in step 13 above, the mapping values ​​of road condition type and rainfall are fused to obtain an environmental consistency score, including:

[0214] Step 131: Identify the road condition type based on the road surface image.

[0215] For example, to comprehensively assess the impact of road conditions on vehicle operation, the cloud server can acquire road surface images taken by the vehicle while obtaining real-time information from the vehicle. That is, the real-time information may include road surface images. Based on these images, the cloud server can identify the road type. Furthermore, based on these images, the cloud server can also identify the dryness of the road surface. Based on the road type and the dryness level, the cloud server can determine the road condition type.

[0216] In one implementation, the cloud server can extract and analyze features such as color, texture, and shape in the road surface image based on computer vision algorithms to obtain the road state type.

[0217] In one implementation, the road condition type may include dry asphalt, dry cement, wet asphalt, wet cement, snow-covered road surface, icy road surface, metal road surface, etc.

[0218] In one implementation, the cloud server can obtain road surface images taken by the vehicle's forward-facing camera for recognition in the process.

[0219] In one implementation, when the cloud server identifies multiple road state types in a grid based on the road surface image, the cloud server can select the identification result of the most numerous road state types as the final road state type.

[0220] Step 132: Determine the road surface score based on the road condition type and the preset first mapping table. The first mapping table indicates the mapping relationship between the road condition type and the road surface score.

[0221] For example, the cloud server stores a preset first mapping table. This first mapping table is pre-defined by the cloud server and is used to indicate the mapping relationship between different road condition types and road surface scores. After determining the road condition type, the cloud server can determine the corresponding road surface score by looking up the first mapping table.

[0222] For example, the first mapping table can be as shown in Table 1.

[0223] Table 1

[0224]

[0225] In one implementation, as shown in Table 1, dry road surfaces typically provide a better road adhesion coefficient, which is beneficial for vehicle driving. Therefore, dry road surface conditions correspond to higher road surface scores. Wet road surfaces increase the risk of vehicle skidding; therefore, waterlogged road surfaces correspond to lower road surface scores. Metallic road surfaces, snow-covered road surfaces, and icy road surfaces further increase the risk of vehicle skidding, thus resulting in even lower road surface scores.

[0226] In one implementation, the cloud server can refer to a large amount of real-world road test data and expert experience to quantify and set road scores based on the degree of impact of different road conditions on vehicle driving safety and comfort.

[0227] Step 133: Determine the weather score based on rainfall and a preset second mapping table. The second mapping table indicates the mapping relationship between rainfall and weather score.

[0228] For example, the cloud server stores a preset second mapping table. This second mapping table is pre-defined by the cloud server and is used to indicate the mapping relationship between different rainfall amounts and weather scores. The cloud server obtains the weather score based on the acquired rainfall amount by looking up the second mapping table.

[0229] In one implementation, the second mapping table can be as shown in Table 2.

[0230] Table 2

[0231]

[0232] In one implementation, heavier rainfall typically has a greater impact on vehicle braking and increases the risk of skidding. Therefore, as shown in Table 2, when rainfall is less than 2, the impact of rainfall can be ignored, and the weather score is set to 1. When rainfall is greater than 10, it can be considered a rainstorm. At this point, a rainstorm will cause severe water accumulation on the road surface, greatly affecting the braking and handling performance of the vehicle. Therefore, a minimum weather score of 0.7 can be set.

[0233] In one implementation, the cloud server can combine meteorological knowledge with actual road conditions, taking into account factors such as the amount and duration of rainfall and their impact on the road environment to set a weather score.

[0234] In one implementation, the cloud server can obtain the rainfall amount through a rain sensor installed on the vehicle. Optionally, if the cloud server obtains multiple rainfall amounts within a period, it can calculate the average of these multiple rainfall amounts as the final rainfall amount.

[0235] Step 134: Using the preset fifth and sixth weights, calculate the weighted sum of the road surface score and the weather score to obtain the environmental consistency score.

[0236] For example, the cloud server can also be configured with a fifth weight and a sixth weight. The fifth weight corresponds to the road surface score, and the sixth weight corresponds to the weather score. The cloud server can calculate a weighted sum of the road surface score and the weather score based on these fifth and sixth weights to obtain the environmental consistency score.

[0237] In one implementation, the formula for calculating the environmental consistency score can be:

[0238]

[0239] in, Score for environmental consistency. Points are awarded for the road surface. Score the weather. It is the fifth weight. It is the sixth weight.

[0240] In one implementation, the fifth and sixth weights are parameters pre-set by the cloud server based on actual needs and experience, used to adjust the weight of road surface score and weather score in the final environmental consistency score calculation.

[0241] For example, if road condition type has a greater impact on environmental consistency, the fifth weight can be set larger. Conversely, if rainfall has a more critical impact, the sixth weight can be set larger.

[0242] In one implementation, the cloud server can use an expert evaluation method, where experts in the relevant field provide suggested weight values ​​based on actual road conditions and data characteristics.

[0243] In another implementation, the cloud server can use data analysis to collect a large amount of actual road environment data, analyze the correlation between road surface score and weather score and environmental consistency, and then determine reasonable weight values.

[0244] In this example, by identifying road condition types from road surface images, mapping road scores based on road condition types, mapping weather scores based on rainfall, and then weighting and fusing the road scores and weather scores to obtain the final environmental consistency score, a comprehensive quantification of the road environment is achieved, thereby improving data accuracy.

[0245] In one example, the confidence level can be calculated as follows: Figure 5 As shown. The cloud server can perform the following steps:

[0246] A401, the cloud server receives real-time information sent by the vehicle.

[0247] A402. The cloud server performs a data validity check on the real-time information. If the real-time information data is valid, proceed to A403. If the real-time information data is invalid, proceed to A404.

[0248] In one implementation, the data validity check may include whether the data packet is complete and whether the parsing is correct.

[0249] A403. The cloud server performs spatiotemporal consistency checks on real-time information.

[0250] In one implementation, a spatiotemporal consistency check can be used to check whether the location of the real-time information is within the road grid of the cloud server, and whether the time difference between the real-time information and the current time is less than a time threshold.

[0251] A404. If the real-time information data is invalid, the cloud server can directly set the confidence level of the grid to 0.

[0252] A405. The cloud server queries the device credit score of the vehicle that sent this real-time information.

[0253] For example, the cloud server can retrieve the vehicle's credit score from a preset credit table based on the vehicle information included in the real-time information. If the vehicle's equipment credit score is low, it indicates a high probability that the data collected by the vehicle or the calculated real-time road adhesion coefficient is abnormal.

[0254] A406, Obtain environment parameters from the cloud server.

[0255] For example, a cloud server can acquire environmental information such as road images and rainfall, and generate corresponding environmental parameters such as road scores and weather scores.

[0256] A407, weighted comprehensive calculation is performed on the cloud server.

[0257] A408, Output confidence level.

[0258] In this example, the cloud server performs data validity and spatiotemporal consistency checks on the received real-time vehicle information. Combined with vehicle equipment credit scores and environmental parameters, a weighted comprehensive calculation is performed to achieve accurate output of grid confidence, thereby ensuring the reliability of road information.

[0259] In one example, in step S2032 above, the predicted coefficient value of the grid at the current moment is calculated based on the preset first weight, second weight, the mean of the real-time road surface adhesion coefficient in all real-time information in the grid, the confidence level of the grid at the current moment, and the predicted values ​​and confidence levels of the road surface adhesion coefficient in the historical information of the grid, including:

[0260] Step 21: Calculate the first weight, the product of the mean of the real-time road surface adhesion coefficient in all real-time information in the grid and the confidence level of the grid at the current time, to obtain the first parameter.

[0261] For example, when the cloud server calculates the predicted value of the road surface adhesion coefficient at the current moment of the grid, it first calculates the first parameter based on the real-time road surface adhesion coefficient and confidence level in the real-time information, combined with the first weight.

[0262] In one implementation, the cloud server can obtain the first parameter by calculating the product of the first weight, the mean of the real-time road surface adhesion coefficients in all real-time information in the grid, and the confidence level of the grid at the current moment.

[0263] Step 22: Calculate the second weight, and obtain the second parameter by multiplying the predicted value of the road surface adhesion coefficient and the confidence level in the historical information of the grid.

[0264] For example, in order to make full use of the grid's historical information, the cloud server can calculate the second parameter based on the second weight, as well as the predicted value and confidence level of the road surface adhesion coefficient in the historical information.

[0265] In one implementation, the cloud server can calculate the second weight, as well as the product of the predicted road surface adhesion coefficient and the confidence level in the historical information of the grid, to obtain the second parameter.

[0266] Step 23: Use the sum of the first parameter and the second parameter as the predicted value of the road surface adhesion coefficient of the grid at the current moment.

[0267] For example, after the cloud server obtains the first and second parameters, it sums them up to achieve the fusion of real-time and historical information and obtain the final predicted value of the road surface adhesion coefficient.

[0268] In one implementation, the calculation formula can be written as:

[0269]

[0270] in, This represents the predicted value of the road surface adhesion coefficient for the grid within the current period. This represents the average real-time road surface adhesion coefficient for that grid in the real-time information. This represents the confidence level of the grid in the real-time information. This is the predicted value of the road surface adhesion coefficient for this grid in historical information. This represents the confidence level of the grid in historical information.

[0271] in, This is the first weight. The value of this first weight can be between 0 and 1. For example, it can be 0.7. This is the second weight. The value of this second weight can be between 0 and 1. For example, it can be 0.3. and The sum can be 1.

[0272] In this example, by combining the product of the average real-time road surface adhesion coefficient and the confidence level in the real-time information under the first weight, and the product of the predicted road surface adhesion coefficient and the confidence level in the historical information under the second weight, the effect of accurately predicting the predicted value of the road surface adhesion coefficient of the grid in the current period is achieved.

[0273] In one example, the real-time information also includes a timestamp. Based on this timestamp, the steps taken by the cloud server for processing may include:

[0274] S204. Calculate the time difference between the current time and the timestamp.

[0275] For example, after obtaining real-time information uploaded by the vehicle, the cloud server can retrieve a timestamp from that real-time information. The cloud server can then calculate the time difference between that timestamp and the current moment.

[0276] In one implementation, the current time is the system time when the cloud server is processing data.

[0277] In one implementation, the timestamp is the time when the real-time information was sent.

[0278] In one implementation, the time difference can indicate whether the real-time information received by the cloud server meets the real-time requirements.

[0279] S205. If the time difference is greater than the time threshold, then discard the real-time information.

[0280] For example, after obtaining the time difference, the cloud server compares it with a preset time threshold to filter the real-time information. Specifically, if the time difference is greater than the time threshold, the cloud server can determine that the real-time information was sent too far in time and cannot be used. In this case, the cloud server can directly discard the real-time information.

[0281] In one implementation, the time threshold can be determined based on the update cycle of the predicted road adhesion coefficient and the confidence level. For example, if the update cycle is once every 1 second, then the time threshold can be 1 second.

[0282] S206. If the time difference is less than or equal to the time threshold, then the real-time road surface adhesion coefficient in the real-time information is written into the grid corresponding to the GPS information based on the GPS information in the real-time information.

[0283] For example, if the time difference is less than or equal to the time threshold, it indicates that the real-time information still has a certain timeliness, and the cloud server will further process the information. The cloud server can write the real-time road surface adhesion coefficient from the real-time information into the grid corresponding to the GPS information based on the GPS information in the real-time information.

[0284] In one implementation, the GPS information is data recording the vehicle's location at the time the real-time information was generated. This GPS information may include longitude, latitude, and other information. This GPS information can accurately determine the vehicle's location in geographic space.

[0285] In one implementation, a grid is a geographical region divided by a cloud server for convenient management and analysis of road information. Each grid has its specific geographical range and identifier. Optionally, each grid can correspond to a certain range of longitude and latitude.

[0286] In one implementation, regarding the writing method, the cloud server can determine the grid to which the coordinates belong based on the latitude and longitude coordinates in the GPS information and the longitude and latitude range corresponding to the grid. Then, the cloud server can write the real-time road surface adhesion coefficient from this real-time information to the corresponding data storage location of that grid.

[0287] In this example, by calculating the time difference between the current moment and the timestamp in the real-time information, and comparing this time difference with a time threshold, the real-time information is filtered, thus improving the timeliness of the real-time information used.

[0288] Figure 6 This application provides a schematic diagram of the signaling interaction for a predictive control method based on road surface adhesion coefficient, as illustrated in the embodiments of this application. Figures 1 to 5 Based on the illustrated embodiments, as Figure 6 As shown, the execution entities include a cloud server and vehicles. To distinguish the currently executing vehicle, this embodiment can further classify the vehicle into a lead vehicle and a following vehicle.

[0289] The lead vehicle is the vehicle located in front of the following vehicle. The following vehicle can be located behind the lead vehicle, within a preset distance range. For example, the following vehicle may be within 100-200 meters behind the lead vehicle. Optionally, when there are multiple vehicles within the preset distance range in front of the following vehicle, the cloud server can identify these multiple vehicles as the lead vehicle.

[0290] After obtaining real-time information uploaded by the lead vehicle, the cloud server determines the corresponding following vehicle based on the lead vehicle's position. The cloud server can then send a road surface adhesion coefficient heatmap generated based on the real-time information uploaded by the lead vehicle to the following vehicle. The following vehicle can then perform drive anti-skid feedforward control and / or regenerative braking feedforward control based on this road surface adhesion coefficient heatmap, thereby improving vehicle driving safety and braking stability.

[0291] The signaling interaction process includes the following steps:

[0292] S301, the lead vehicle uploads real-time information to the cloud server. This real-time information may include the real-time road surface adhesion coefficient.

[0293] For example, the lead vehicle uploads real-time road surface adhesion coefficient, GPS coordinates, and timestamps.

[0294] S302 The cloud server integrates the real-time information received by each grid in the current period, as well as the historical information of that grid, to obtain the latest predicted value of the road surface adhesion coefficient in the current period.

[0295] For example, the cloud server calculates and generates a predicted value for the road surface adhesion coefficient based on the real-time uploaded road surface adhesion coefficient, GPS coordinates, timestamp, and historical data.

[0296] S303, the cloud server sends the predicted value of the road surface adhesion coefficient to the following vehicle.

[0297] S304. After obtaining the predicted value of the road surface adhesion coefficient, the following vehicle can calculate the feedforward control quantity based on the predicted value of the road surface adhesion coefficient. The feedforward control quantity may include regenerative braking capability and / or drive torque.

[0298] For example, the following vehicle calculates a feedforward control quantity based on the predicted road adhesion coefficient. The following vehicle can then generate a torque limiting command based on this feedforward control quantity. The following vehicle can then send this torque limiting command to the actuator.

[0299] S305, the following vehicle can use the actuators inside the vehicle to perform the regenerative braking capability and / or drive torque.

[0300] In this embodiment, by uploading real-time information from the lead vehicle, fusing multi-source information into a cloud server to predict the road surface adhesion coefficient and then distributing it, the following vehicle calculates and executes the feedforward control quantity based on this information. This achieves the effect of improving driving safety and stability by optimizing driving control based on accurate road surface information.

[0301] Furthermore, in this process, the lead vehicle's data upload time in 5G NR low-latency mode is typically less than or equal to 50ms. The cloud server can be an edge server, and its data processing time is typically less than or equal to 30ms. The following vehicle uses the CAN FD bus to transmit torque limiting commands, and the execution time of these commands is typically less than or equal to 20ms. Comprehensive calculations show that, using the method of this application, the total time for torque control of the following vehicle based on the lead vehicle's data is less than 100ms. This time is better than the currently commonly required 150ms.

[0302] Figure 7 This application provides a schematic diagram of the signaling interaction for a predictive control method based on road surface adhesion coefficient, as illustrated in the embodiments of this application. Figures 1 to 6 Based on the illustrated embodiments, as Figure 7 As shown, the process includes the following steps:

[0303] S401, cloud server obtains the road adhesion coefficient of the vehicle in front. Relevant data.

[0304] For example, the lead vehicle can use V2X communication to transmit the road surface adhesion coefficient. The relevant data is uploaded to the cloud server. The cloud server can run a preset fusion algorithm for road surface adhesion coefficient to calculate and generate a heat map of the road surface adhesion coefficient.

[0305] In one implementation, the road surface adhesion coefficient The relevant data may include road surface adhesion coefficient μ, GPS coordinates, timestamps, etc.

[0306] In one implementation, the heat map of the road surface adhesion coefficient is the road surface adhesion coefficient matrix.

[0307] In one implementation, the lead vehicle is the vehicle that leads the way.

[0308] S402, the cloud server can send the heat map of the road surface adhesion coefficient to the following vehicles.

[0309] In one implementation, the following vehicle is located behind the lead vehicle. In this application, the logic for calculating the data needs to consider the relationship between vehicles. However, in actual data transmission, it is usually sent via broadcast, without considering the relationship between vehicles. Vehicles can directly obtain usable data from the acquired road surface adhesion coefficient matrix for processing. Furthermore, the cloud server can directly obtain the road surface adhesion coefficients of all vehicles. Relevant data.

[0310] S403, following vehicles can achieve drive anti-skid feedforward control and regenerative braking feedforward control based on the data in the road surface adhesion coefficient matrix.

[0311] In this embodiment, the relevant data of the road surface adhesion coefficient of the preceding vehicle is obtained through the cloud server and a heat map is generated and sent to the following vehicle. The following vehicle then implements drive anti-skid and regenerative braking feedforward control based on the heat map data, thereby improving the vehicle's driving safety and braking stability.

[0312] Based on the above embodiments, an example of calculating the predicted road surface adhesion coefficient value obtained by vehicle group collaborative calculation using the above method on the cloud server may include the following steps:

[0313] S501. Obtain real-time vehicle group data transmitted by the lead vehicle via V2X broadcast. This real-time vehicle group data may include the road surface adhesion coefficient μ, GPS coordinates, timestamp, and confidence factor.

[0314] S502, a cloud server, integrates historical data and real-time vehicle group data to generate a heat map of the road surface adhesion coefficient μ. Optionally, the resolution of this heat map is less than or equal to 10m.

[0315] The specific calculation process for the heatmap of the road surface adhesion coefficient μ may include:

[0316] S5021, the cloud server can obtain real-time vehicle group data. This real-time vehicle group data may include the road surface adhesion coefficient μ, GPS coordinates, timestamp, and confidence level uploaded by the lead vehicle. The road surface adhesion coefficient μ can be denoted as... This confidence level can be denoted as... .

[0317] S5022, The cloud server can access its own stored historical data. This historical data may include the historical pavement adhesion coefficient μ, historical confidence level, and weather compensation factor for the same road segment stored on the cloud server. The historical pavement adhesion coefficient μ can be denoted as... This historical confidence level can be denoted as... .

[0318] S5023, the cloud server can acquire environmental data. This environmental data may include real-time rainfall. This real-time rainfall can be obtained through a meteorological service. Alternatively, the real-time rainfall can be uploaded by the lead vehicle. The lead vehicle can detect this rainfall using its onboard solar and rain sensors.

[0319] S5024. The cloud server can divide the road into a 10m × 10m grid, with each grid associated with a matrix of road surface adhesion coefficient μ. Optionally, the number of grids can be adjusted according to accuracy requirements, road width, and other information. The cloud server can map the road surface adhesion coefficient μ to the corresponding grid based on the GPS coordinates uploaded by the lead vehicle.

[0320] The S5025 and cloud server can use a weighted fusion algorithm to update the road surface adhesion coefficient μ within each grid. The final formula for calculating the road surface adhesion coefficient μ can be:

[0321] Optionally, the allocation of weights for real-time and historical data can be adaptively adjusted based on factors such as weather. For example, the weights can be appropriately increased on rainy days. To 0.9, corresponding adjustment. Up to 0.1.

[0322] S5026, the cloud server can generate heat map matrices.

[0323] For example, a three-dimensional heat map matrix is ​​formed based on the already generated gridded two-dimensional GPS coordinate information and the weighted matrix of the road surface adhesion coefficient μ.

[0324] S5027, the cloud server can obtain the predicted value of the road surface adhesion coefficient μ at a given location by looking up a table based on the generated heat map matrix and GPS coordinates.

[0325] In this embodiment, by fusing real-time vehicle group data uploaded by the lead vehicle with historical data from the cloud and using a weighted fusion algorithm to generate a high-resolution heat map of the road surface adhesion coefficient, the effect of accurately predicting the road surface adhesion coefficient at various locations on the road can be achieved.

[0326] In one implementation, the lead vehicle, cloud server, and following vehicles constitute a vehicle group data interaction system.

[0327] The lead vehicle is equipped with a feature extractor for the road surface adhesion coefficient μ. The lead vehicle can upload its estimated road surface adhesion coefficient μ to a cloud server. The lead vehicle can also simultaneously upload its vehicle coordinates and timestamp.

[0328] The cloud server stores the spatiotemporal matrix of the road surface adhesion coefficient μ. Furthermore, the cloud server can use a weather compensation factor to correct the road surface adhesion coefficient μ.

[0329] Optionally, the weather compensation factor can be obtained by looking up a table after obtaining the rainfall amount. Specifically, the correspondence between the rainfall amount and the weather compensation factor can be shown in Table 1.

[0330] The following vehicle can incorporate a built-in feedforward controller. This controller calculates the driving torque and regenerative braking force based on the road surface adhesion coefficient μ, which contains spatiotemporal information, sent from a cloud server, thus achieving feedforward control. This feedforward control ensures that the following vehicle's response delay for different road surface adhesion coefficients is less than 150ms.

[0331] In one example, the following vehicle may also be equipped with a safety redundancy design. This safety redundancy design can switch to traditional single-vehicle control mode when either of the following conditions is met: first, the V2X signal loss duration is greater than 200ms; second, the cloud server data confidence level is less than 0.7.

[0332] Figure 8 A structural diagram of a predictive control device based on road surface adhesion coefficient provided in this application embodiment is shown below. Figure 8 As shown, the predictive control device 800 based on the road surface adhesion coefficient is applied to a vehicle and includes:

[0333] The acquisition module 801 is used to obtain the road surface adhesion coefficient matrix of the road where the current vehicle is located from the cloud server; and based on the road surface adhesion coefficient matrix, the current vehicle position and navigation path, determine the predicted value of the road surface adhesion coefficient of the current vehicle in the driving direction; wherein, the road surface adhesion coefficient matrix is ​​calculated by the cloud server by fusing the real-time road surface adhesion coefficient uploaded by the vehicle on the road and the historical road surface adhesion coefficient of the road.

[0334] The control module 802 is used to calculate the upper limit of the vehicle's regenerative braking capacity and / or the upper limit of the driving torque based on the preset control coefficient, the vehicle parameters, and the predicted value of the road surface adhesion coefficient; and to control the vehicle to perform braking based on the current state of the vehicle and the upper limit of the vehicle's regenerative braking capacity and / or the upper limit of the driving torque.

[0335] In one example, control module 802 is used for:

[0336] The smaller of the vehicle's current braking capacity and the upper limit of regenerative braking capacity is used as the target regenerative braking force;

[0337] The difference between the current braking capacity and the target regenerative braking force is calculated to obtain the target friction braking force;

[0338] The vehicle is controlled to perform regenerative braking based on the target regenerative braking force and the target friction braking force.

[0339] In one example, control module 802 is used for:

[0340] The total required torque is obtained by analyzing the pedal opening or automatic driving command.

[0341] The smaller of the upper limit of the driving torque and the total required torque is taken as the actual output torque;

[0342] The vehicle's drive torque is adjusted based on the actual output torque.

[0343] In one example, control module 802 is used for:

[0344] If, during the process of adjusting the regenerative braking force of the vehicle based on the target regenerative braking force, the first change per unit time is greater than the preset first change threshold, then the regenerative braking force at the current moment is adjusted according to the first change threshold and the regenerative braking force at the previous moment.

[0345] And / or,

[0346] If, during the adjustment of the vehicle's output torque based on the actual output torque, the second change per unit time is greater than the preset second change threshold, then the output torque at the current moment is adjusted according to the second change threshold and the output torque at the previous moment.

[0347] In one example, the road surface adhesion coefficient matrix also includes a confidence level for each road surface adhesion coefficient; control module 802 is used for:

[0348] If the confidence level of the predicted road adhesion coefficient of the current vehicle in the direction of travel is less than or equal to the preset confidence threshold, then switch to single-vehicle control mode.

[0349] Alternatively, if the current vehicle does not obtain the road surface adhesion coefficient matrix sent by the cloud server within a preset time period, it will switch to single-vehicle control mode.

[0350] Among them, the single-vehicle control mode represents the current vehicle calculating the real-time road surface adhesion coefficient based on sensor monitoring information, and then performing regenerative braking control and torque control based on the real-time road surface adhesion coefficient.

[0351] Further functional descriptions of the above modules and units are the same as those in the corresponding embodiments described above, and will not be repeated here.

[0352] Figure 9 A structural diagram of a predictive control device based on road surface adhesion coefficient provided in this application embodiment is shown below. Figure 9 As shown, the predictive control device 900 based on the road surface adhesion coefficient is applied to a cloud server and includes:

[0353] The acquisition module 901 is used to acquire real-time information uploaded by vehicles on the road. The real-time information includes the real-time road surface adhesion coefficient calculated by the vehicle based on sensor monitoring information and the vehicle's GPS information. Based on the GPS information of each vehicle, the real-time information of each vehicle is mapped to the road grid. The road grid is obtained by the cloud server dividing the road within a preset area based on the preset grid side length.

[0354] The fusion module 902 is used to fuse real-time information and historical information from each grid to obtain the predicted value and confidence level of the road surface adhesion coefficient for each grid.

[0355] In one example, the historical information includes the predicted value and confidence level of the road surface adhesion coefficient of the grid at the previous time step; the fusion module 902 is used for:

[0356] Based on the real-time pavement adhesion coefficient, road condition type, and rainfall from all real-time information in the grid, the confidence level of the grid at the current moment is calculated; among which, the road condition type indicates the dryness of the road surface.

[0357] Based on the preset first weight, second weight, the mean of the real-time road surface adhesion coefficient in all real-time information in the grid, the confidence level of the grid at the current moment, and the predicted value and confidence level of the road surface adhesion coefficient in the historical information of the grid, the predicted value of the road surface adhesion coefficient of the grid at the current moment is calculated.

[0358] In one example, the fusion module 902 is used for:

[0359] The variance value is calculated based on the real-time road surface adhesion coefficient from all real-time information in the grid.

[0360] The variance is normalized to obtain the variance consistency score;

[0361] The mapping values ​​of road condition type and rainfall are fused and calculated to obtain the environmental consistency score;

[0362] Based on the preset third and fourth weights, the weighted sum of the variance consistency score and the environmental consistency score is calculated to obtain the confidence level of the grid at the current moment.

[0363] In one example, the fusion module 902 is used for:

[0364] The road condition type is determined based on road surface image recognition.

[0365] Based on the road condition type and a preset first mapping table, the road surface score is determined; the first mapping table indicates the mapping relationship between the road condition type and the road surface score.

[0366] Based on rainfall and a pre-set second mapping table, a weather score is determined; the second mapping table indicates the mapping relationship between rainfall and weather score.

[0367] Using the preset fifth and sixth weights, the weighted sum of the road surface score and the weather score is calculated to obtain the environmental consistency score.

[0368] In one example, based on preset first weights, second weights, the mean of real-time road surface adhesion coefficients in all real-time information of the grid, the grid's confidence level at the current moment, and the predicted road surface adhesion coefficients and confidence levels in the grid's historical information, the predicted road surface adhesion coefficient of the grid at the current moment is calculated, including:

[0369] The first parameter is obtained by multiplying the first weight, the mean of the real-time road surface adhesion coefficient in all real-time information in the grid, and the confidence level of the grid at the current time.

[0370] The second weight is calculated, and the second parameter is obtained by multiplying the predicted value of the road surface adhesion coefficient and the confidence level in the historical information of the grid.

[0371] The sum of the first and second parameters is used as the predicted value of the road surface adhesion coefficient of the grid at the current moment.

[0372] In one example, the real-time information also includes a timestamp; the fusion module 902 is used for:

[0373] Calculate the time difference between the current time and the timestamp;

[0374] If the time difference is greater than the time threshold, the real-time information is discarded.

[0375] If the time difference is less than or equal to the time threshold, the real-time road surface adhesion coefficient in the real-time information is written into the grid corresponding to the GPS information based on the GPS information in the real-time information.

[0376] Further functional descriptions of the above modules and units are the same as those in the corresponding embodiments described above, and will not be repeated here.

[0377] In this embodiment, the predictive control device 800 and the predictive control device 900 based on the road surface adhesion coefficient are presented in the form of functional units. Here, a unit refers to an application-specific integrated circuit (ASIC), a processor and memory that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0378] Figure 10 A structural diagram of a vehicle provided in this application embodiment, such as Figure 10As shown, the vehicle 1000 includes: one or more processors 1001, a memory 1002, and interfaces for connecting the various components, including high-speed interfaces and low-speed interfaces. The processor 1001 may be a central processing unit, a network processor, or a combination thereof.

[0379] The memory 1002 stores instructions executable by at least one processor 1001 to cause at least one processor 1001 to perform the method shown in the above embodiments.

[0380] The memory 1002 may include a program storage area and a data storage area, wherein the program storage area may store the operating system and application programs required for at least one function; and the data storage area may store data created based on the use of the computer device.

[0381] The memory 1002 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 1002 may also include a combination of the above types of memory.

[0382] The computer device also includes a communication interface 1003 for communicating with other devices or communication networks.

[0383] In one implementation, the processor 1001, memory 1002, and communication interface 1003 can be located in a controller of the vehicle.

[0384] Figure 11 A structural diagram of a cloud server provided in an embodiment of this application is shown below. Figure 11 As shown, the vehicle 1100 includes: one or more processors 1101, a memory 1102, and interfaces for connecting the various components, including high-speed interfaces and low-speed interfaces. The processor 1101 may be a central processing unit, a network processor, or a combination thereof.

[0385] The memory 1102 stores instructions executable by at least one processor 1101 to cause at least one processor 1101 to perform the method shown in the above embodiments.

[0386] The memory 1102 may include a program storage area and a data storage area, wherein the program storage area may store the operating system and application programs required for at least one function; and the data storage area may store data created based on the use of the computer device.

[0387] The memory 1102 may include volatile memory, such as random access memory; the memory may also include non-volatile memory, such as flash memory, hard disk or solid-state drive; the memory 1102 may also include a combination of the above types of memory.

[0388] The computer device also includes a communication interface 1103 for communicating with other devices or communication networks.

[0389] This application also provides a computer-readable storage medium in which the methods described in this application can be implemented in hardware or firmware, or implemented as recordable on a storage medium, or implemented as computer code originally stored on a remote storage medium or a non-transitory machine-readable storage medium and to be stored on a local storage medium after being downloaded over a network, so that the methods described herein can be stored on such software processing on a storage medium using a general-purpose computer, a special-purpose processor, or programmable or special-purpose hardware.

[0390] This application provides a computer program product including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the method of any embodiment of this application.

[0391] Although embodiments of this application have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of this application, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A predictive control method based on road surface adhesion coefficient, characterized in that, Applied to vehicles, the method includes: The road surface adhesion coefficient matrix of the road where the vehicle is currently located is obtained from the cloud server; and based on the road surface adhesion coefficient matrix, the current vehicle position and navigation path, the predicted value of the road surface adhesion coefficient of the current vehicle in the driving direction is determined; wherein, the road surface adhesion coefficient matrix is ​​calculated by the cloud server by fusing the real-time road surface adhesion coefficient uploaded by the vehicle on the road and the historical road surface adhesion coefficient of the road. Based on the preset control coefficient, the vehicle parameters, and the predicted value of the road adhesion coefficient, the upper limit of the regenerative braking capacity and / or the upper limit of the driving torque of the vehicle are calculated; based on the current state of the vehicle, and the upper limit of the regenerative braking capacity and / or the upper limit of the driving torque of the vehicle, the vehicle is controlled to perform braking. The road surface adhesion coefficient matrix is ​​a matrix composed of the predicted road surface adhesion coefficient and the confidence level of each grid after dividing the road in front of the current vehicle into grids. The calculation of the predicted road adhesion coefficient and confidence level for each grid includes: The system acquires real-time information uploaded by each vehicle on the road, including the real-time road surface adhesion coefficient calculated by the vehicle based on sensor monitoring information and the vehicle's GPS information; and maps the real-time information of each vehicle to the grid based on the GPS information of each vehicle. Based on the real-time road surface adhesion coefficient from all real-time information in the grid, the variance value is calculated; the variance value is normalized to obtain the variance consistency score; the mapping values ​​of road surface dryness and rainfall in the grid are fused to obtain the environmental consistency score; according to the preset third weight and fourth weight, the weighted sum of the variance consistency score and the environmental consistency score is calculated to obtain the confidence level of the grid at the current time. Based on the preset first weight, second weight, the mean of the real-time road surface adhesion coefficient in all real-time information in the grid, the confidence level of the grid at the current moment, and the predicted value and confidence level of the road surface adhesion coefficient in the historical information of the grid, the predicted value of the road surface adhesion coefficient of the grid at the current moment is calculated.

2. The method according to claim 1, characterized in that, Based on the current state of the vehicle and the upper limit of the vehicle's regenerative braking capability, controlling the vehicle to perform braking includes: The smaller of the vehicle's current braking capacity and the upper limit of the regenerative braking capacity is used as the target regenerative braking force; The target friction braking force is obtained by calculating the difference between the current braking capacity and the target regenerative braking force. The vehicle is controlled to perform regenerative braking based on the target regenerative braking force and the target friction braking force.

3. The method according to claim 1, characterized in that, Based on the current state of the vehicle and the upper limit of the vehicle's drive torque, controlling the vehicle to perform braking includes: The total required torque is obtained by analyzing the pedal opening or automatic driving command. The smaller of the upper limit of the driving torque and the total required torque is taken as the actual output torque; The vehicle's drive torque is adjusted based on the actual output torque.

4. The method according to any one of claims 1-3, characterized in that, The method further includes: If, during the adjustment of the regenerative braking force of the vehicle based on the target regenerative braking force, the first change per unit time is greater than the preset first change threshold, then the regenerative braking force at the current moment is adjusted according to the first change threshold and the regenerative braking force at the previous moment. And / or, If, during the adjustment of the vehicle's output torque based on the actual output torque, the second change per unit time is greater than a preset second change threshold, then the output torque at the current moment is adjusted according to the second change threshold and the output torque at the previous moment.

5. The method according to any one of claims 1-3, characterized in that, The road surface adhesion coefficient matrix also includes the confidence level of each road surface adhesion coefficient; the method further includes: If the confidence level corresponding to the predicted value of the road surface adhesion coefficient of the current vehicle in the driving direction is less than or equal to the preset confidence threshold, then switch to single vehicle control mode. Alternatively, if the current vehicle does not obtain the road surface adhesion coefficient matrix sent by the cloud server within a preset time period, it will switch to single-vehicle control mode. The single-vehicle control mode refers to the current vehicle calculating the real-time road surface adhesion coefficient based on sensor monitoring information, and then performing regenerative braking control and torque control based on the real-time road surface adhesion coefficient.

6. A predictive control method based on road surface adhesion coefficient, characterized in that, Applied to a cloud server, the method includes: The system acquires real-time information uploaded by vehicles on the road, including the real-time road surface adhesion coefficient calculated by the vehicle based on sensor monitoring information and the vehicle's GPS information. Based on the GPS information of each vehicle, the real-time information of each vehicle is mapped to a road grid; the road grid is obtained by the cloud server dividing the roads within a preset area based on a preset grid side length; The variance value is calculated based on the real-time road surface adhesion coefficient in all the real-time information in the grid. The variance values ​​are normalized to obtain the variance consistency score; The mapping values ​​of road surface dryness and rainfall are fused and calculated to obtain an environmental consistency score; Based on the preset third and fourth weights, the weighted sum of the variance consistency score and the environment consistency score is calculated to obtain the confidence level of the grid at the current moment. Based on the preset first weight, second weight, the mean of the real-time road surface adhesion coefficient in all the real-time information in the grid, the confidence level of the grid at the current time, and the predicted value and confidence level of the road surface adhesion coefficient in the historical information of the grid, the predicted value of the road surface adhesion coefficient of the grid at the current time is calculated. The road adhesion coefficient matrix is ​​formed by combining the predicted road adhesion coefficient values ​​and confidence levels of each grid at the current time.

7. The method according to claim 6, characterized in that, The environmental consistency score is obtained by fusing the mapping values ​​of the road surface dryness and the rainfall, including: The dryness of the road surface is determined based on road surface image recognition. Based on the road surface dryness and a preset first mapping table, a road surface score is determined; the first mapping table indicates the mapping relationship between the road surface dryness and the road surface score. A weather score is determined based on the rainfall and a preset second mapping table; the second mapping table indicates the mapping relationship between the rainfall and the weather score. Using preset fifth and sixth weights, the weighted sum of the road surface score and the weather score is calculated to obtain the environmental consistency score.

8. The method according to claim 6, characterized in that, Based on preset first weights, second weights, the mean of the real-time road surface adhesion coefficients in all real-time information in the grid, the confidence level of the grid at the current moment, and the predicted road surface adhesion coefficients and confidence levels in the historical information of the grid, the predicted coefficient value of the grid at the current moment is calculated, including: The first parameter is obtained by multiplying the first weight, the mean of the real-time road surface adhesion coefficient in all the real-time information in the grid, and the confidence level of the grid at the current time. The second weight is calculated, and the second parameter is obtained by multiplying the predicted value of the road surface adhesion coefficient and the confidence level in the historical information of the grid. The sum of the first parameter and the second parameter is used as the predicted value of the road surface adhesion coefficient of the grid at the current moment.

9. The method according to any one of claims 6-8, characterized in that, The real-time information also includes a timestamp; the method includes: Calculate the time difference between the current time and the timestamp; If the time difference is greater than the time threshold, the real-time information is discarded. If the time difference is less than or equal to the time threshold, then the real-time road surface adhesion coefficient in the real-time information is written into the grid corresponding to the GPS information based on the GPS information in the real-time information.

10. A vehicle, characterized in that, include: A memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, the processor executing the computer instructions to perform the method of any one of claims 1 to 5.

11. A cloud server, characterized in that, include: A memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, the processor executing the computer instructions to perform the method of any one of claims 6 to 9.

Images