Vehicle control system

Abstract

The objective is to provide a vehicle control system that can achieve both drivetrain protection and off-road capability using a pre-trained model. [Solution] A vehicle control device comprising: a prediction unit that inputs vehicle driving data into a first trained model trained by machine learning and predicts feature quantities; an estimation unit that inputs the feature quantities into a second trained model trained by machine learning and estimates whether or not the vehicle can travel on the road surface; a setting unit that determines a limit value to limit the driving force or braking force of the vehicle according to the estimation of the estimation unit; and a control unit that controls the limit value according to the estimation and the result of whether or not the vehicle can travel on the road surface.

Description

Technical Field

[0006] ,

[0004] , , ,

[0005] , , , , ,

[0001] The present invention relates to a control device for a vehicle.

Background Art

[0002] Techniques have been developed to predict driving operations using a trained model obtained by training a neural network. The trained model predicts driving operations to avoid excessive loads on a transmission or the like (for example, Patent Document ･1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In a vehicle, it is required to ensure both protection of the drive system and running performance. Therefore, an object of the present invention is to provide a control device for a vehicle that can achieve both protection of the drive system and running performance using a trained model.

Means for Solving the Problems

[0005] The above object can be achieved by a vehicle control device including a prediction unit that inputs vehicle running data into a first trained model learned by machine learning to predict a feature amount, an estimation unit that inputs the feature amount into a second trained model learned by machine learning to estimate whether the vehicle can break through a road surface, a setting unit that determines a limit value for limiting the driving force or braking force of the vehicle according to the estimation of the estimation unit, and a control unit that controls the limit value according to the estimation and the result of whether the vehicle can break through or not.

Effects of the Invention

[0006] Using a pre-trained model, it is possible to provide a vehicle control system that can achieve both drivetrain protection and off-road capability. [Brief explanation of the drawing]

[0007] [Figure 1] Figure 1 is a schematic diagram of the vehicle according to this embodiment. [Figure 2] Figure 2 is a flowchart illustrating the process in the embodiment. [Figure 3] Figure 3 is a flowchart illustrating the process in the embodiment. [Modes for carrying out the invention]

[0008] The control device of the vehicle according to this embodiment will be described below with reference to the drawings. Figure 1 is a schematic diagram of the vehicle 1 according to this embodiment. Vehicle 1 is a four-wheel drive vehicle. Vehicle 1 includes an internal combustion engine 10, an automatic transmission 12, a transfer case 13, four drive wheels 16, two propeller shafts 20, two differential gears 22, a drive shaft 24, an accelerator pedal 26, an accelerator position sensor 28, a wheel speed sensor 30, a brake mechanism 32, and an ECU (Electronic Control Unit) 40.

[0009] The internal combustion engine 10 is, for example, a gasoline engine or a diesel engine. The automatic transmission 12 is a stepped transmission or a continuously variable transmission. The automatic transmission 12 is connected to the internal combustion engine 10. The transfer case 13 is connected to the automatic transmission 12.

[0010] Two propeller shafts 20 are connected to the transfer case 13. One of the two propeller shafts 20 extends forward of the vehicle 1. The other propeller shaft 20 extends rearward of the vehicle 1. Each of the two propeller shafts 20 is connected to a differential gear 22. A drive shaft 24 is connected to the differential gear 22 and the drive wheels 16.

[0011] The vehicle 1 moves when the driver of the vehicle 1 operates the accelerator pedal 26. The internal combustion engine 10 is the power source of the vehicle 1, and generates driving force by burning fuel such as gasoline. The driving force output by the internal combustion engine 10 is transmitted to the drive wheels 16 through the automatic transmission 12, transfer case 13, propeller shaft 20, and drive shaft 24. The drive wheels 16 rotate due to the driving force, and the vehicle 1 moves.

[0012] The greater the amount the accelerator pedal 26 is pressed (accelerator opening), the greater the torque (engine torque) output by the internal combustion engine 10. The smaller the accelerator opening, the less torque there is. The accelerator opening sensor 28 detects the accelerator opening.

[0013] The number of wheel speed sensors 30 and brake mechanisms 32 is four, the same as the number of drive wheels 16. Each of the four drive wheels 16 is provided with a wheel speed sensor 30 and a brake mechanism 32. The wheel speed sensor 30 detects the rotational speed (wheel speed) of the drive wheel 16. The brake mechanism 32 is, for example, a hydraulic brake, which applies braking force to the drive wheel 16, slowing down its rotation.

[0014] The ECU40 is the control unit of the vehicle 1 and is equipped with a processing unit such as a CPU (Central Processing Unit), and storage devices such as RAM (Random Access Memory) and ROM (Read Only Memory). The ECU40 performs various controls by executing programs stored in the ROM and storage devices. The ECU40 obtains the accelerator opening from the accelerator opening sensor 28. The ECU40 obtains the wheel speed of the four drive wheels 16 from the four wheel speed sensors 30.

[0015] The ECU 40 controls four brake mechanisms 32 to increase or decrease braking force according to the amount of brake pedal depression (not shown). The ECU 40 also controls the torque of the internal combustion engine 10 according to the accelerator pedal opening.

[0016] Depending on the road surface and driving conditions, the torque applied to the drivetrain, including the propeller shaft 20, differential gear 22, and drive shaft 24, changes. For example, the ECU 40 performs traction control (TRC) to adjust the engine torque and the braking force of the brake mechanism 32 to prevent the drive wheels 16 from slipping. Torque is applied to the drivetrain according to the braking force. Also, if there is a difference in rotational speed between the left and right drive wheels 16, the differential gear 22 generates torque to limit the differential. Applying excessive torque to the drivetrain may cause damage to the drivetrain. The higher the torque output by the internal combustion engine 10 and the braking force of the brake mechanism 32, the greater the torque applied to the drivetrain. The lower the engine torque and braking force, the smaller the torque of the drivetrain, but the less drivability of the vehicle 1. In this embodiment, the torque of the drivetrain is controlled and the drivability is improved by controlling the limit values ​​of the engine torque or braking force.

[0017] The ECU40's memory stores AI (Artificial Intelligence) models, such as the first and second trained models. These models are created using machine learning, with experimental and simulation data serving as training data. The first trained model is created, for example, using LSTM (Long Short-Term Memory). The second trained model is created, for example, using Random Forest. The first trained model predicts features by inputting driving data of vehicle 1. These features include wheel speed, vehicle speed, and torque during future driving of vehicle 1. The second trained model estimates whether vehicle 1 can traverse the road surface by inputting the predicted features.

[0018] The ECU 40 functions as a prediction unit 42 that inputs driving data into a first pre-trained model and predicts feature quantities. The ECU 40 functions as an estimation unit 44 that inputs the feature quantities into a second pre-trained model and estimates whether the vehicle 1 can traverse the road surface. The ECU 40 functions as a setting unit 46 that sets a limit value for restricting at least one of the engine torque and the braking force according to the estimation. The ECU 40 functions as a control unit 48 that controls the limit value according to the result of whether the vehicle 1 can traverse the road surface.

[0019] FIG. 2 and FIG. 3 are flowcharts illustrating the processing in the embodiment. The ECU 40 acquires data (step S10). The data includes driving data and road surface data. The driving data includes the accelerator opening degree, engine torque, rotational speed of the internal combustion engine 10, gear position of the automatic transmission 12, wheel speed, hydraulic pressure of the brake mechanism 32, acceleration of the vehicle 1, information around the vehicle 1, and the like. The road surface data is data regarding the type of road surface, such as a paved road or a muddy road, for example. The road surface data can be acquired from an in-vehicle camera or can also be estimated from the driving data.

[0020] Based on the road surface data, the ECU 40 selects a first pre-trained model suitable for the road surface from the stored machine learning pre-trained models. The ECU 40 inputs the driving data into the selected first pre-trained model. The first pre-trained model predicts the feature quantities (step S12). The ECU 40 inputs the feature quantities into a second pre-trained model and estimates whether the vehicle 1 can traverse the road surface (can traverse or get stuck) (step S14).

[0021] If an affirmative determination (Yes, estimated as traversable) is made in step S14, the ECU 40 selects a value L1 corresponding to the estimation of traversability as the limit value for the engine torque (step S16). The ECU 40 determines whether the vehicle 1 has actually traversed the road surface (step S18). If an affirmative determination (traversable) is made, the process ends.

[0022] If a negative determination (No, impassable) is made in step S18, the ECU 40 selects the value L2 at the time of impassable (stack) as the limit value of the engine torque (step S20). The ECU 40 predicts the maximum value Tr of the torque applied to the drive system from the engine torque, the gear ratio, the braking force, the ratio of the left and right wheel speeds, the inertia, and the like. The ECU 40 determines whether or not the maximum torque Tr is less than the threshold value Tth (step S22).

[0023] In the case of an affirmative determination, the ECU 40 determines whether or not the engine torque has reached the limit value L2 (step S24). If a negative determination is made in either of steps S22 and S24, the process ends. If affirmative determinations are made in both of steps S22 and S24, the ECU 40 changes the limit value L2 at the time of stack to a value larger than the value selected in step S20 (step S26). The ECU 40 sets the relearning flag of the second learned model (step S28). Relearning of the second learned model that was incorrect in the estimation of passability is performed. After step S28, the process ends.

[0024] If a negative determination (No, estimated as stack) is made in step S14, the ECU 40 selects the value L2 corresponding to the estimation (impassable) as the limit value for the engine torque (step S30). The ECU 40 determines whether or not the vehicle 1 has actually become stuck (step S32). If an affirmative determination (stack) is made, the process ends.

[0025] If the determination in step S32 is negative (No, drivable), the ECU 40 selects the value L1 for drivable conditions as the engine torque limit (step S34). The ECU 40 determines whether the maximum torque Tr applied to the drivetrain is greater than or equal to the threshold Tth (step S36). If the determination is positive, the ECU 40 determines whether the engine torque has reached the limit value L1 (step S38). If the determination in either step S36 or S38 is negative, the process ends. If the determination in both steps S36 and S38 is positive, the ECU 40 changes the limit value L1 for drivable conditions to a value smaller than the value selected in step S24 (step S40). The ECU 40 sets the retraining flag for the second learned model (step S42). After step S42, the process ends.

[0026] According to the embodiment, features are predicted by inputting driving data into a first trained model. By inputting the features into a second trained model, it is estimated whether or not vehicle 1 can traverse the road surface. If the driving result is the same as the estimation, the limit value does not change. If the driving result differs from the estimation, the ECU 40 changes the limit value.

[0027] Specifically, if vehicle 1 gets stuck despite being estimated to be able to traverse the terrain, the engine torque limit L2 at the time of getting stuck is increased. This improves traversability. If vehicle 1 successfully traverses the terrain despite being estimated to be stuck, the torque limit L1 at the time of traversal is decreased. This reduces the torque applied to the drivetrain, protecting it. This makes it possible to achieve both drivetrain protection and traversability.

[0028] In the example above, the limit value of the engine torque is controlled, but for example, one or both of the engine torque and brake force may be controlled. Increasing the limit values ​​for engine torque and brake force improves off-road capability (step S26 in Figure 2). Decreasing the limit values ​​also improves off-road capability (step S40 in Figure 3). If vehicle 1 has a motor generator (MG) as a power source, the ECU 40 sets a limit value for motor torque and controls the limit value. For example, a limit value may be set for the rotational speed of the internal combustion engine 10.

[0029] By turning on the retraining flag (steps S28 and S42), the second trained model is retrained. For example, data from when the pass / fail judgment was incorrect is input, and the model is retrained to be able to correctly determine the pass / fail judgment. Retraining improves the accuracy of the pass / fail judgment.

[0030] Although preferred embodiments of the present invention have been described in detail above, the present invention is not limited to these specific embodiments, and various modifications and changes are possible within the scope of the gist of the invention as described in the claims. [Explanation of symbols]

[0031] 1 Vehicle, 10 Internal combustion engine, 12 Automatic transmission, 13 Transfer case, 16 Drive wheels, 20 Propeller shaft, 22 Differential gear, 24 Drive shaft, 26 Accelerator pedal, 28 Accelerator position sensor, 30 Wheel speed sensor, 32 Brake mechanism, 40 ECU, 42 Prediction unit, 44 Estimation unit, 46 Setting unit, 48 Control unit

Claims

[Claim 1] A prediction unit that inputs vehicle driving data into a first pre-trained model trained by machine learning and predicts features, An estimation unit that inputs the aforementioned features into a second pre-trained model trained by machine learning and estimates whether or not the vehicle can traverse the road surface, A setting unit that determines a limit value to limit the driving force or braking force of the vehicle in accordance with the estimation of the estimation unit, A vehicle control device comprising: a control unit that controls the limit value according to the estimation and the result of whether the vehicle can travel the route;

Images