Device for predicting road surface state

Abstract

To provide a device for predicting a road surface state that is capable of accurately predicting a future state of a road surface.SOLUTION: A device for predicting a road surface state comprises: an acquisition section for acquiring friction force information related to friction force in between the wheels of a vehicle and a road surface; and a prediction section for predicting a future state of the road surface on the basis of the friction force information acquired by the acquisition section, by using a prediction model which has been learned mechanically according to supervised learning, and in which the friction force information is comprised in an explanatory variable and a future state of the road surface is an objective variable.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to a road surface condition prediction device. [Background technology]

[0002] There is a technique for predicting the future state of a road surface based on the behavior of a vehicle (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2018-024278 A Summary of the Invention [Problem to be solved by the invention]

[0004] The future state of the road surface changes depending on the frictional force between the wheels of the vehicle and the road surface. Since the above technology does not take such frictional force into consideration, there is a risk that the future state of the road surface cannot be predicted with high accuracy.

[0005] SUMMARY OF THE PRESENT DISCLOSURE An object of the present invention is to provide a road surface condition prediction device capable of predicting the future condition of a road surface with high accuracy. [Means for solving the problem]

[0006] The above objective can be achieved by a road surface condition prediction device including an acquisition unit that acquires friction force information regarding the friction force between the wheels of a vehicle and a road surface, and a prediction unit that predicts the future state of the road surface based on the friction force information acquired by the acquisition unit using a machine-learned prediction model through supervised learning that includes the friction force information as an explanatory variable and has the future state of the road surface as a target variable. Effect of the Invention

[0007] According to the present invention, it is possible to provide a road surface condition prediction device that can accurately predict the future condition of a road surface. [Brief description of the drawings]

[0008] [Figure 1] FIG. 1 is a schematic diagram of a vehicle. [Diagram 2] FIG. 2A is a flowchart illustrating the road surface prediction control, and FIG. 2B is a flowchart illustrating the resonance countermeasure control. [Diagram 3] FIG. 3A is a timing chart illustrating the countermeasure processing, and FIG. 3B is a timing chart illustrating a modified example of the countermeasure processing. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] [Vehicle outline] 1 is a schematic configuration diagram of a vehicle 1. The vehicle 1 is an electric vehicle equipped with a motor 2. The vehicle 1 is equipped with the motor 2, a propeller shaft 3, a differential gear 4, drive shafts 51 and 52, drive wheels 61 and 62, a PCU (Power Control Unit) 7, a battery 8, and an ECU (Electric Control Unit) 10.

[0010] The motor 2 functions as a prime mover that outputs torque when supplied with electric power. The motor 2 also functions as a generator that generates electric power when braking the vehicle 1. The electric power generated by the motor 2 is supplied to a battery 8 via a PCU 7.

[0011] The motor 2 is connected to drive wheels 61 and 62 via the propeller shaft 3, the differential gear 4, and the drive shafts 51 and 52. The torque of the motor 2 is transmitted to the drive wheels 61 and 62, causing the vehicle 1 to run.

[0012] The ECU 10 includes a central processing unit (CPU), a random access memory (RAM), a read only memory (ROM), a storage device, etc., and performs various controls by executing programs stored in the ROM and the storage device. The ECU 10 is an example of a road surface condition prediction device, and functionally realizes an acquisition unit and a prediction unit, which will be described in detail later.

[0013] The ECU 10 can communicate with a server 100 that stores a dynamic map 101 via a network NW. The ECU 10 can obtain map data of the dynamic map 101 from the server 100 and display the dynamic map 101 together with the current position of the vehicle 1 on a display 96 provided in the interior of the vehicle 1.

[0014] The ECU 10 is electrically connected to an ignition switch 90, wheel rotation speed sensors 91 and 92, an axle rotation speed sensor 93, an accelerator opening sensor 94, a vehicle speed sensor 95, a display 96, and a GPS (Global Positioning System) module 97. The ignition switch 90 detects the on / off state of the ignition. The wheel rotation speed sensors 91 and 92 detect the rotation speeds of the driving wheels 61 and 62, respectively (hereinafter referred to as wheel speeds). The axle rotation speed sensor 93 detects the rotation speed of the propeller shaft 3. The accelerator opening sensor 94 detects the accelerator opening operated by the driver. The vehicle speed sensor 95 detects the vehicle speed of the vehicle 1. The display 96 displays the current position of the vehicle 1 and a map of the surroundings of the vehicle 1. The GPS module 97 detects the current position of the vehicle 1.

[0015] In such a vehicle 1, resonance may occur while the vehicle 1 is traveling. The resonance of the vehicle 1 mainly occurs while the vehicle 1 is traveling on an undulating road. It is considered that the resonance while traveling on an undulating road occurs because the frequency at which the drive wheels 61 and the like repeatedly spin and contact the ground due to the forced fluctuation from the undulating road overlaps with the unsprung resonance frequency and torsional resonance frequency of the vehicle 1. In this embodiment, the ECU 10 predicts the future state of the road surface using a prediction model described below. Furthermore, the ECU 10 predicts whether or not resonance will occur before passing over the road surface using the prediction result, and executes countermeasure processing against the resonance when it is predicted that the vehicle 1 will resonate. The prediction model is stored in advance in the ROM of the ECU 10.

[0016] [Prediction model] The prediction model is a prediction model that has been machine-learned by supervised learning. The explanatory variables of the prediction model include friction force information regarding the friction force between the wheels of the vehicle and the road surface. The objective variable of the prediction model is the road surface level indicating the future state of the road surface. The road surface level will be described later. As the supervised learning, for example, any one or a combination of logistic regression, random forest, support vector machine, k-nearest neighbor method, and neural network may be used.

[0017] The friction force information included in the explanatory variables of the prediction model includes information on the vehicle's driving force, the ground load of the wheels, information on the friction coefficient, and information on the wheel's driving force.

[0018] The information on the driving force of the vehicle is the driving force when the wheels of the vehicle are not slipping on the road surface and are in a gripping state. In this case, the driving force of the vehicle is approximately equal to the friction force.

[0019] The information on the wheel ground load is the wheel ground load when the vehicle wheels are in a slipping state relative to the road surface. This is because the greater the wheel ground load, the greater the frictional force. The information on the wheel ground load includes, for example, the vehicle weight and running at unsprung resonance. The condition for unsprung resonance to occur is when the vehicle is running near the unsprung resonance vehicle speed. The unsprung resonance vehicle speed is calculated from the pitch of the wavy road and the vehicle's inherent unsprung resonance frequency. The grip state and the slip state can be distinguished based on the difference between the vehicle speed and the wheel speed.

[0020] The information on the friction coefficient includes, for example, the type of wheels, such as summer tires, studless tires, high-grip tires, etc.

[0021] The information on the driving force of the wheels is, for example, the accelerator opening, the torque of the engine or motor which is the driving power source, and the occurrence condition of torsional resonance. The occurrence condition of torsional resonance is when the vehicle is traveling near the torsional resonance vehicle speed. The torsional resonance vehicle speed is the pitch of the wavy road multiplied by the inherent torsional resonance frequency of the vehicle.

[0022] As described above, the explanatory variables of the prediction model include friction information related to the friction between the wheels and the road surface. For example, the future road surface state will be different for a road surface with a relatively high friction force and a road surface with a low friction force. Therefore, by using a prediction model that includes friction information as an explanatory variable, it is possible to accurately predict the future road surface state.

[0023] The explanatory variables of the prediction model may include the frequency of travel on the road surface for each vehicle speed range, the frequency of travel on the road surface for each torque range, or the frequency of travel combining the vehicle speed range and the torque range. The explanatory variables may include the weather. The explanatory variables may also include the frequency of travel combining the torque, vehicle speed, wheel type, and weather. The explanatory variables may also include the depth of the unevenness of the road surface. The depth of the unevenness of the road surface may be estimated, for example, from an image captured by an on-board camera. The explanatory variables may also include the manner in which the wheels rise. In this case, the difference between the vehicle speed and the wheel speed may be used, for example, as the manner in which the wheels rise. The explanatory variables may also include the rotational fluctuation of the propeller shaft 3.

[0024] The road surface level, which is the objective variable of the prediction model, is calculated as, for example, one of road surface levels A to X. For example, road surface level A indicates the flattest road surface state, and road surface level X indicates the roughest road surface state. The road surface level indicates the road surface state at each predetermined time lapse from the present. For example, it may be the road surface state for each day from the present, or the road surface state for each week.

[0025] FIG. 2A is a flowchart illustrating road surface prediction control. This control is repeatedly executed while the ignition is on. The ECU 10 acquires the explanatory variables described above while the vehicle 1 is traveling (step S1). As described above, the explanatory variables include the driving force of the vehicle 1, the torque of the motor 2, the accelerator opening, whether the vehicle speed of the vehicle 1 is near the unsprung resonance vehicle speed or the torsional resonance vehicle speed (hereinafter referred to as the resonance vehicle speed), the type of wheels of the vehicle 1, the weight of the vehicle 1, and the like. Step S1 is an example of a process executed by the acquisition unit.

[0026] The driving force of the vehicle 1 is calculated by the ECU 10 based on, for example, the rotation speed and torque of the motor 2, the differential ratio, the radius of the wheels, etc. The rotation speed and torque of the motor 2 may be calculated by the ECU 10 based on, for example, the power supplied to the motor 2, or may be acquired by the ECU 10 based on the detection values ​​of sensors. The accelerator opening and vehicle speed are acquired by the ECU 10 based on the detection values ​​of an accelerator opening sensor 94 and a vehicle speed sensor 95. The type of wheels of the vehicle 1 needs to be input to the ECU 10 in advance. The weight of the vehicle 1 is stored in the ECU 10 in advance.

[0027] The unsprung resonance frequency and the torsional resonance frequency (hereinafter referred to as resonance frequency) used to judge whether the vehicle speed is near the resonant vehicle speed must be input to the ECU 10 in advance. The ECU 10 acquires the pitch of the road surface used in the above judgment from the dynamic map 101. The dynamic map 101 is added with information on the road surface pitch based on information from other vehicles. When resonance occurs when another vehicle is traveling at a specified point, the road surface pitch calculated based on the vehicle speed at that time and the resonance frequency of that vehicle is reflected in the dynamic map 101. Note that if the road surface pitch at the traveling point of the vehicle 1 is not included in the dynamic map 101, the information on whether the vehicle speed is near the resonant vehicle speed is excluded from the explanatory variables.

[0028] Next, the ECU 10 uses the above-mentioned prediction model to predict the road surface level at the position of the vehicle 1 from which the explanatory variables have been acquired (step S2). As described above, the explanatory variables of the prediction model include frictional force information, so that the road surface level can be predicted with high accuracy. Step S2 is an example of a process executed by the prediction unit. The ECU 10 reflects the predicted road surface level at the position of the vehicle 1 in the dynamic map 101 (step S3). In detail, the ECU 10 transmits the position information of the vehicle 1 and the road surface level to the server 100. The position information of the vehicle 1 can be detected by the GPS module 97. The server 100 reflects the received position information and road surface level in the dynamic map 101. The server 100 updates the road surface level at predetermined time intervals and reflects it in the dynamic map 101.

[0029] FIG. 2B is a flowchart illustrating the resonance countermeasure control. This control is repeatedly executed while the ignition is on. The ECU 10 predicts whether or not resonance will occur in the vehicle 1 based on the road surface level at the planned travel location of the vehicle 1 acquired from the dynamic map 101 (step S11). For example, when the road surface level is included in N to X, the ECU 10 predicts that resonance will occur in the vehicle 1 when the vehicle 1 travels through the planned travel location. If the result of step S11 is No, the ECU 10 turns off the resonance prediction flag (step S12), and this control ends. If the result of step S11 is Yes, the ECU 10 turns on the resonance prediction flag (step S13), and executes a countermeasure process described later (step S14).

[0030] [Countermeasures] FIG. 3A is a timing chart illustrating the countermeasure processing. FIG. 3A shows the transition of the resonance prediction flag and the maximum torque of the motor 2. When the resonance prediction flag is switched on (time t1), the maximum torque of the motor 2 is immediately limited. This makes it possible to avoid the occurrence of resonance in the vehicle 1. This countermeasure processing is suitable for vehicles in which the maximum torque of the motor 2 is relatively large. This is because even if the maximum torque of the motor 2 is immediately limited in this way, the motor 2 has a large reserve capacity, so there is little impact on drivability.

[0031] FIG. 3B is a timing chart showing a modified example of the countermeasure processing. FIG. 3B shows the transition of the resonance prediction flag, the integrated value of the rotation fluctuation amount of the propeller shaft 3, and the maximum torque of the motor 2. The integrated value of the rotation fluctuation amount of the propeller shaft 3 is calculated by the ECU 10. When the resonance prediction flag is switched on (time t1), the threshold value for the integrated value for limiting the maximum torque of the motor 2 is switched from value α to a smaller value β. As a result, when the integrated value becomes equal to or greater than value β (time t2), the maximum torque of the motor 2 is limited. This countermeasure processing is suitable for vehicles in which the maximum torque of the motor 2 is relatively small. This is because a vehicle with a relatively small maximum torque has a small driving force, and therefore, if the maximum torque of the motor 2 is limited, the impact on drivability is large.

[0032] In the above embodiment, the vehicle 1 is an electric vehicle, but the present invention is not limited to this. For example, the vehicle may be an engine vehicle equipped with an engine as a power source for running. The vehicle may also be a hybrid vehicle equipped with an engine and a motor as power sources for running.

[0033] In the above embodiment, the motor 2 and the drive wheels 61 and 62 are directly connected, but this is not limited to this, and the contents of the above embodiment can also be applied to a mechanism in which they are connected via a clutch, transmission, etc.

[0034] Although the embodiments of the present invention have been described in detail above, the present invention is not limited to such specific embodiments, and various modifications and variations are possible within the scope of the gist of the present invention described in the claims. [Explanation of symbols]

[0035] 1 vehicle 10 ECU (Road condition prediction device, acquisition unit, prediction unit)

Claims

[Claim 1] an acquisition unit that acquires friction force information regarding a friction force between a wheel of a vehicle and a road surface; and a prediction unit that predicts the future state of the road surface based on the friction force information acquired by the acquisition unit, using a machine-learned prediction model through supervised learning that includes the friction force information as an explanatory variable and has the future state of the road surface as a target variable.

Images