Vehicle behavior estimation method and vehicle behavior estimation device

Abstract

A controller comprising a vehicle behavior estimation unit for estimating sprung and unsprung state quantities of wheels. The vehicle behavior estimation unit: estimates the sprung and unsprung state quantities of front wheels and the sprung state quantity of one rear wheel by using a first AI estimation unit, which is artificial intelligence; and calculates the sprung and unsprung state quantities of the remaining rear wheel from the sprung and unsprung state quantities of the front wheels and the sprung state quantity of the one rear wheel. The vehicle behavior estimation unit estimates the sprung and unsprung state quantities of the rear wheels as needed by using a second AI estimation unit, which is artificial intelligence.

Description

[Technical Field]

[0001] The present disclosure relates to a vehicle behavior estimation method and a vehicle behavior estimation device for estimating state quantities of, for example, sprung and unsprung parts of a wheel. [Background technology]

[0002] Patent Document 1 discloses a damper control system that uses a machine learning algorithm to control damper characteristics with high responsiveness and robustness. This damper control system includes a damper control means that controls the characteristics of a damper used in a vehicle suspension, and a processing means that receives feedback data related to vehicle behavior measured on the vehicle, applies arithmetic processing identified by executing a machine learning algorithm to the feedback data, and outputs control variables obtained by the arithmetic processing to the damper control means. The damper control means controls the damper characteristics based on the control variables used internally in the damper control means, and replaces the internally used control variables with new control variables output by the processing means. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2021-17168 Summary of the Invention [Problem to be solved by the invention]

[0004] According to the above-mentioned conventional technology, control variables are obtained by computational processing (artificial intelligence) specified by executing a machine learning algorithm, and the damper characteristics are controlled based on these control variables. However, updating the control variables using such artificial intelligence at high frequencies results in an excessively high computational load, which is unrealistic from the perspective of computational resources.

[0005] An object of the present invention is to provide a vehicle behavior estimation method and a vehicle behavior estimation device that can reduce the computational load imposed by artificial intelligence. [Means for solving the problem]

[0006] One embodiment of the present invention is a vehicle behavior estimation method that uses artificial intelligence to estimate the state quantities of the sprung and unsprung masses of each wheel, in which the state quantities of the front sprung and unsprung masses and the sprung masses for one rear wheel are estimated using artificial intelligence, the remaining state quantities of the rear sprung and unsprung masses are calculated from the state quantities of the front sprung and unsprung masses and the sprung masses for one rear wheel, and the state quantities of the rear sprung and unsprung masses are estimated using artificial intelligence as necessary.

[0007] One embodiment of the present invention is a vehicle behavior estimation device that uses artificial intelligence to estimate the state quantities of the sprung and unsprung masses of each wheel. The state quantities of the front sprung and unsprung masses and the sprung masses for one rear wheel are estimated using artificial intelligence, and the remaining state quantities of the rear sprung and unsprung masses are calculated from the state quantities of the front sprung and unsprung masses and the sprung masses for one rear wheel, and the state quantities of the rear sprung and unsprung masses are estimated using artificial intelligence as needed.

[0008] According to one embodiment of the present invention, the computational load of artificial intelligence can be reduced. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is an overall configuration diagram showing a four-wheeled automobile to which a controller according to an embodiment of the present invention is applied; [Figure 2] FIG. 2 is a control block diagram showing a controller in FIG. [Figure 3] FIG. 2 is a block diagram showing a vehicle behavior estimating unit according to the first and third embodiments. [Figure 4] FIG. 2 is an explanatory diagram showing an example of a neural network that constitutes an FL state estimation unit. [Figure 5]FIG. 10 is a block diagram showing a vehicle behavior estimation unit according to a second embodiment. [Figure 6] FIG. 2 is an explanatory diagram showing the trajectory of a wheel during steady turning; DETAILED DESCRIPTION OF THE INVENTION

[0010] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A vehicle behavior estimation method and a vehicle behavior estimation device according to an embodiment of the present invention will be described in detail below with reference to the accompanying drawings, taking as an example a case where the method and device are applied to a four-wheeled vehicle.

[0011] 1 and 2 show a vehicle behavior control device 1. The vehicle behavior control device 1 is made up of a suspension device 5 that constitutes a damping force generating device, and a controller 11 that constitutes a vehicle control device. In FIG. 1, for example, left and right front wheels and left and right rear wheels (hereinafter collectively referred to as wheels 3) are provided on the underside of a vehicle body 2 that constitutes the body of the vehicle. The wheels 3 are made up of tires 4, and the tires 4 act as springs that absorb small irregularities in the road surface.

[0012] The suspension device 5 is installed between the vehicle body 2 and the vehicle wheel 3. The suspension device 5 is composed of a suspension spring 6 (hereinafter referred to as the spring 6) and an adjustable damping force shock absorber (hereinafter referred to as the variable damper 7) installed in parallel with the spring 6 between the vehicle body 2 and the wheel 3.

[0013] 2 shows a case where one set of suspension devices 5 is provided between the vehicle body 2 and the wheels 3. However, a total of four sets of suspension devices 5 are provided individually and independently between, for example, four wheels 3 and the vehicle body 2, and only one of these sets is shown schematically in FIG.

[0014] Here, the variable damper 7 of the suspension device 5 is configured using a damping force adjustable hydraulic shock absorber interposed between the vehicle body 2 and the wheel 3. This variable damper 7 is provided with a damping force variable actuator 8 consisting of a damping force adjustment valve or the like in order to continuously adjust the characteristics of the generated damping force (i.e., the damping force characteristics) from hard characteristics (hard characteristics) to soft characteristics (soft characteristics). Note that the damping force variable actuator 8 does not necessarily have to be configured to continuously adjust the damping force characteristics, and may be capable of adjusting the damping force in multiple stages, for example, two or more stages. Furthermore, the variable damper 7 may be of a pressure control type or a flow rate control type. The variable damper 7 may also be of a type that controls viscosity, such as a magnetorheological fluid or an electrorheological fluid.

[0015] The controller 11 constitutes a control device. The controller 11 is configured by, for example, a microcomputer as a control device that controls the damping characteristics of the variable damper 7. The controller 11 is connected to, for example, a CAN 10 (Controller Area Network), which is a line network required for data communication. The controller 11 acquires data related to the behavior of the vehicle (hereinafter referred to as behavior information) through the CAN 10. At this time, the behavior information includes, for example, longitudinal acceleration (longitudinal G), lateral acceleration (lateral G), steering angle, yaw rate, command current (FB current), wheel speed, etc. Therefore, this behavior information is input to the controller 11. The output side of the controller 11 is connected to the damping force variable actuator 8 of the variable damper 7.

[0016] The controller 11 also has a storage unit including a ROM, a RAM, a non-volatile memory, etc. The storage unit of the controller 11 stores various programs, information (vehicle information), data, etc. for controlling the variable damper 7. The controller 11 estimates the sprung velocity and relative velocity (piston velocity) as vehicle state quantities based on the behavior information. The controller 11 calculates the force to be generated by the variable damper 7 (force generating mechanism) of the suspension device 5 based on the estimated vehicle state quantities, and outputs a control signal (command current) to the variable damping force actuator 8 of the suspension device 5.

[0017] 2, the controller 11 includes a control unit 12, a weight parameter storage unit 13, and a data reading unit 14. The control unit 12 includes a vehicle behavior estimating unit 15 and a control value calculating unit 20.

[0018] The controller 11 acquires data (behavior information) relating to the vehicle behavior during driving from the communication data transmitted via the CAN 10 within the vehicle, calculates suspension control values ​​based on this behavior information and the weight parameters in the weight parameter memory unit 13, and controls the variable damper 7 of the suspension device 5.

[0019] In this embodiment, the CAN is used as an example of the in-vehicle network, but other in-vehicle networks may be used, such as CAN FD (CAN with Flexible Data Rate), FlexRay, or in-vehicle Ethernet.

[0020] Next, the damping force control of the suspension device 5 performed by the controller 11 will be described. The vehicle behavior estimator 15 in the control unit 12 is a vehicle behavior estimator. The vehicle behavior estimator 15 reads weight parameters stored in the weight parameter memory unit 13 via the data reader 14. Furthermore, the vehicle behavior estimator 15 estimates data related to suspension control required by the control value calculator 20, specifically, instantaneous values ​​related to suspension control, based on the behavior information and weight parameters received from the CAN 10. The control value calculator 20 calculates a suspension control value (e.g., a target damping force) for controlling the damping force of the variable damper 7 of the suspension device 5 based on the instantaneous value input from the vehicle behavior estimator 15. The control unit 20 outputs a command current as a control signal to the damping force variable actuator 8 of the variable damper 7 based on the suspension control value.

[0021] In this embodiment, the vehicle behavior estimation unit 15 estimates data (sprung speed, relative speed) related to suspension control for each of the four wheels by referring to behavior information, such as wheel speed, longitudinal acceleration, lateral acceleration, and yaw rate, among the data constantly transmitted to the CAN 10. The estimation results are then transferred to the downstream control value calculation unit 20. The vehicle behavior estimation unit 15 estimates sensor data (vehicle state quantities) related to suspension control for each of the four wheels. The vehicle behavior estimation unit 15 is configured, for example, with a neural network. The weight parameters used in the neural network are determined in advance for the vehicle by machine learning, and therefore the rigidity characteristics of the entire vehicle are also reflected, improving the accuracy of suspension control according to the vehicle behavior.

[0022] Weighting parameters and behavior information are input to the vehicle behavior estimation unit 15. Based on the weighting parameters and the behavior information, the vehicle behavior estimation unit 15 estimates instantaneous values, such as sensor data such as sprung speed and relative speed, and outputs the values.

[0023] As shown in FIG. 3, the vehicle behavior estimation unit 15 includes a first AI estimation unit 16, a second AI estimation unit 17, a non-AI estimation unit 18, and an AI estimation selection unit 19. The vehicle behavior estimation unit 15 estimates the sprung and unsprung velocities of the four wheels using the first AI estimation unit 16, the second AI estimation unit 17, and the non-AI estimation unit 18. The vehicle behavior estimation unit 15 acquires the sprung and relative velocities of the four wheels based on the estimated sprung and unsprung velocities of the four wheels. At this time, the relative velocities are calculated as the difference between the sprung and unsprung velocities. The vehicle behavior estimation unit 15 outputs the sprung and relative velocities of the four wheels as vehicle state quantities to a control value calculation unit 20 at a downstream stage.

[0024] The first AI estimation unit 16 is an artificial intelligence (AI). The first AI estimation unit 16 includes an FL state estimation unit 16A, an FR state estimation unit 16B, and an RL state estimation unit 16C. The FL state estimation unit 16A estimates the sprung velocity and unsprung velocity of the left front wheel based on behavior information. The FR state estimation unit 16B estimates the sprung velocity and unsprung velocity of the right front wheel based on behavior information. As a result, the first AI estimation unit 16 estimates the sprung velocity and unsprung velocity of the left front wheel and the right front wheel as front sprung and unsprung state quantities (hereinafter referred to as front sprung state quantities). In addition, the RL state estimation unit 16C estimates the sprung velocity of the left rear wheel based on behavior information. Therefore, the first AI estimation unit 16 estimates the state quantities of the front sprung and unsprung masses and the left rear sprung mass among the eight state quantities of the front sprung and unsprung masses and the rear sprung and unsprung masses.

[0025] Like the first AI estimation unit 16, the second AI estimation unit 17 is also an artificial intelligence. The second AI estimation unit 17 includes an RL state estimation unit 17A and an RR state estimation unit 17B. The RL state estimation unit 17A estimates the unsprung velocity of the left rear wheel based on behavior information. The RR state estimation unit 17B estimates the sprung velocity and unsprung velocity of the right rear wheel based on behavior information. As a result, the second AI estimation unit 17 estimates the unsprung velocity of the left rear wheel and the right rear wheel and the sprung velocity of the right rear wheel as rear sprung and unsprung state quantities (hereinafter referred to as rear sprung state quantities).

[0026] The FL state estimation unit 16A, the FR state estimation unit 16B, the RL state estimation unit 16C, the RL state estimation unit 17A, and the RR state estimation unit 17B are configured in the same manner as the neural network of the vehicle behavior estimation unit disclosed in, for example, Patent Publication No. 2022-191913.

[0027] As an example, a specific configuration of the FL state estimation unit 16A will be described with reference to Fig. 4. As shown in Fig. 4, the FL state estimation unit 16A is configured, for example, by a neural network. The neural network is configured as a three-layer hierarchical neural network in which elements of an input layer (number of elements i) 101, a hidden layer (number of elements j) 102, and an output layer (number of elements k) 103 are hierarchically connected. Each element of the input layer 101 and each element of the hidden layer 102 are connected by weights W1ij (i = 1 to I, j = 1 to J), and each element of the hidden layer 102 and each element of the output layer 103 are connected by weights W2jk (j = 1 to J, k = 1 to K). Information on these weights (hereinafter referred to as weight parameters) is expressed as a determinant of the weights W1ij and W2jk. The weight parameters are obtained in advance by machine learning and stored in the weight parameter storage unit 13. Although this example shows a neural network with a single, all-element-connected hidden layer 102, which is the simplest possible configuration, the present invention is not limited to this. For example, a neural network with two or more hidden layers 102 may also be used.

[0028] The neural network's machine learning is performed based on data on vehicle state quantities (sprung speed, unsprung speed) and behavior information data acquired in advance by a data acquisition vehicle. The data acquisition vehicle has the same specifications as the vehicle on which the first AI estimation unit 16 and second AI estimation unit 17 of the vehicle behavior estimation unit 15 are installed, and is equipped with various sensors (speed sensor, acceleration sensor, etc.) for acquiring vehicle state quantities. The neural network's machine learning adjusts weight parameters to learn the correlation between the vehicle state quantity data (sprung speed, unsprung speed) acquired by the data acquisition vehicle and the behavior information data. The weight parameters obtained as a result of learning are stored in the weight parameter memory unit 13.

[0029] In this embodiment, the first AI estimation unit 16 and the second AI estimation unit 17 perform machine learning to determine the correlation between vehicle state quantity data and behavior information data acquired by a data acquisition vehicle, but the present invention is not limited to this. For example, a vehicle model corresponding to the vehicle on which the first AI estimation unit 16 and the second AI estimation unit 17 are installed may be constructed, and the first AI estimation unit 16 and the second AI estimation unit 17 may perform machine learning based on data acquired by a simulation using this vehicle model.

[0030] As shown in Figure 4, the input layer 101 of the neural network is composed of a first input layer 101A to which time-series data of, for example, the wheel speed of the left front wheel is input, and a second input layer 101B to which time-series data of, for example, the longitudinal acceleration of the vehicle is input. The output layer 103 outputs instantaneous values ​​of the sprung and unsprung velocities of the suspension device 5, assumed to be attached to, for example, the left front wheel of the vehicle. The number of elements in the hidden layer 102 is generally determined by the number of elements in the input layer 101 and the output layer 103, and is set to a number that maximizes the accuracy of vehicle behavior estimation by the neural network. The number of elements in the output layer 103 is determined by the output specifications of the vehicle behavior estimation.

[0031] For ease of explanation, Fig. 4 shows the neural network of an FL state estimation unit 16A for controlling a suspension device 5 attached to the left front wheel of a vehicle as an example. The FR state estimation unit 16B, the RL state estimation unit 16C, the RL state estimation unit 17A, and the RR state estimation unit 17B are also configured with neural networks similar to those of the FL state estimation unit 16A. Therefore, the vehicle behavior estimation unit 15 includes the FL state estimation unit 16A, the FR state estimation unit 16B, the RL state estimation unit 16C, the RL state estimation unit 17A, and the RR state estimation unit 17B, which are configured with neural networks.

[0032] 4, the FL state estimation unit 16A outputs the sprung velocity and unsprung velocity of the left front wheel based on the wheel speed and longitudinal acceleration as behavior information. However, the present invention is not limited to this. The FL state estimation unit 16A may output the sprung velocity and unsprung velocity of the left front wheel based on the wheel speed, longitudinal acceleration, lateral acceleration, and yaw rate as behavior information. Furthermore, the FL state estimation unit 16A may output the sprung velocity and unsprung velocity of the left front wheel based on the wheel speed, longitudinal acceleration, lateral acceleration, yaw rate, steering angle, and command current supplied to the damping force variable actuator 8 of the left front wheel as behavior information. This also applies to the FR state estimation unit 16B, the RL state estimation unit 16C, the RL state estimation unit 17A, and the RR state estimation unit 17B.

[0033] As shown in FIG. 3, the non-AI estimation unit 18 calculates three state quantities for the rear wheels from the estimation results of the first AI estimation unit 16. In this way, the non-AI estimation unit 18 reduces the calculation load on the vehicle behavior estimation unit 15. The non-AI estimation unit 18 includes an RL state estimation unit 18A and an RR state estimation unit 18B. The RL state estimation unit 18A estimates the unsprung velocity of the left rear wheel based on the estimation results of the FL state estimation unit 16A and behavior information. The RR state estimation unit 18B estimates the unsprung velocity of the right rear wheel based on the estimation results of the FR state estimation unit 16B and behavior information. The RR state estimation unit 18B estimates the sprung velocity of the right rear wheel based on the estimation results of the FL state estimation unit 16A, the FR state estimation unit 16B, and the RL state estimation unit 16C and behavior information.

[0034] Specifically, the RL state estimation unit 18A calculates the unsprung velocity of the left rear wheel by performing a delay process for the wheelbase on the unsprung velocity of the left front wheel estimated by the FL state estimation unit 16A of the first AI estimation unit 16. The RR state estimation unit 18B calculates the unsprung velocity of the right rear wheel by performing a delay process for the wheelbase on the unsprung velocity of the right front wheel estimated by the FR state estimation unit 16B of the first AI estimation unit 16. Furthermore, the RR state estimation unit 18B calculates the sprung velocity of the right rear wheel, which is the remaining wheel, from the sprung velocities of the three wheels estimated by the FL state estimation unit 16A, the FR state estimation unit 16B, and the RL state estimation unit 16C by treating the vehicle body 2 as a rigid body. As a result, the non-AI estimation unit 18 calculates the unsprung velocities of the left rear wheel and the right rear wheel and the sprung velocity of the right rear wheel as rear spring vertical state quantities.

[0035] The AI ​​estimation selection unit 19 selects whether the unsprung velocity of the rear wheels and the like are to be calculated by the non-AI estimation unit 18 or estimated by the second AI estimation unit 17. The AI ​​estimation selection unit 19 selects either the non-AI estimation unit 18 or the second AI estimation unit 17 based on, for example, the steering angle included in the behavior information. For example, when the steering angle is smaller than a predetermined angle, the road surface trajectories of the front wheels and the rear wheels generally coincide. Therefore, when the steering angle is smaller than a predetermined angle, the AI ​​estimation selection unit 19 selects the non-AI estimation unit 18. At this time, the non-AI estimation unit 18 calculates the sprung velocity and unsprung velocity of the rear wheels based on the sprung velocity and unsprung velocity of the front wheels. On the other hand, when the steering angle is larger than a predetermined angle, the AI ​​estimation selection unit 19 selects the second AI estimation unit 17. At this time, the second AI estimation unit 17 estimates the sprung velocity and unsprung velocity of the rear wheels based on the behavior information.

[0036] Thus, vehicle behavior estimation unit 15 according to this embodiment estimates the state quantities of the front sprung and unsprung masses and the state quantities of the sprung masses for one rear wheel using first AI estimation unit 16, which is artificial intelligence; calculates the remaining rear sprung and unsprung masses from the state quantities of the front and the sprung masses for one rear wheel; and, if necessary, estimates the rear sprung and unsprung masses using second AI estimation unit 17, which is also artificial intelligence. In this way, vehicle behavior estimation unit 15 estimates the rear vehicle behavior (rear sprung masses) using a formula based on the results of estimation of the front vehicle behavior (front sprung masses) by the artificial intelligence (first AI estimation unit 16). As a result, in this embodiment, the computational load of estimation by artificial intelligence can be reduced compared to when all state quantities of the front and rear are estimated by artificial intelligence.

[0037] Furthermore, if the vehicle behavior estimation unit 15 determines that it is difficult to estimate the rear spring vertical state quantity from the front spring vertical state quantity due to steering or the like, the state quantities of the sprung and unsprung parts of all wheels are estimated by the first AI estimation unit 16 and the second AI estimation unit 17, which are artificial intelligences. This ensures a predetermined estimation accuracy for the state quantities of the sprung and unsprung parts of all four wheels, and ensures the performance of the suspension control.

[0038] Under normal circumstances, vehicle behavior estimation unit 15 uses artificial intelligence to estimate the state quantities of the front sprung and unsprung masses and the sprung mass of one rear wheel, and only when the trajectories of the front and rear wheels do not match does it estimate the state quantities of the rear sprung and unsprung masses using artificial intelligence. This reduces the computational load of estimation by artificial intelligence under normal circumstances. On the other hand, when the trajectories of the front and rear wheels do not match, the state quantities of the sprung and unsprung masses of all wheels are estimated by artificial intelligences, first AI estimation unit 16 and second AI estimator 17, thereby ensuring suspension control performance.

[0039] Of the eight state quantities for the front sprung and unsprung masses and the rear sprung and unsprung masses, the vehicle behavior estimation unit 15 estimates only the state quantities for the front sprung and unsprung masses and the state quantity for the left rear sprung mass using the first AI estimation unit 16, which is artificial intelligence, and calculates the remaining quantities from the state quantities for the front sprung masses and the state quantities for the sprung mass of one rear wheel. Therefore, the non-AI estimation unit 18 of the vehicle behavior estimation unit 15 can calculate the state quantity for the rear unsprung mass (unsprung velocity) based on the state quantity for the front unsprung mass (unsprung velocity). Furthermore, the non-AI estimation unit 18 of the vehicle behavior estimation unit 15 can calculate the state quantity for the sprung mass (sprung velocity) of the remaining one wheel (right rear wheel) from the state quantities for the sprung masses (sprung velocity) of the three wheels by regarding the vehicle body 2 as a rigid body.

[0040] If the vehicle behavior estimation unit 15 determines that the trajectories of the front and rear wheels are different, it estimates the state quantities of the sprung and unsprung masses for both the front and rear wheels using artificial intelligence. Therefore, even in a situation where the state quantities of the rear unsprung masses cannot be calculated from the state quantities of the front unsprung masses, the state quantities of the sprung and unsprung masses for all wheels can be estimated by the first AI estimation unit 16 and the second AI estimation unit 17, which are both artificial intelligences.

[0041] In the first embodiment, the first AI estimator 16 estimates the sprung velocity of the left rear wheel in addition to the sprung and unsprung velocities of the left front and right front wheels, but the present invention is not limited to this. The first AI estimator 16 may estimate the sprung velocity of the right rear wheel instead of the left rear wheel. In this case, the second AI estimator 17 and the non-AI estimator 18 estimate or calculate the sprung velocity of the left rear wheel, which is the remaining wheel, based on the sprung velocities of the three wheels estimated by the first AI estimator 16.

[0042] 2 and 5 show a second embodiment. The second embodiment is characterized in that the vehicle behavior estimation unit estimates or calculates the sprung acceleration and unsprung acceleration of the four wheels, and then acquires the sprung velocity and unsprung velocity of the four wheels. In the second embodiment, the same components as those in the first embodiment are denoted by the same reference numerals, and their description will be omitted.

[0043] The vehicle behavior estimating unit 21 according to the second embodiment is a vehicle behavior estimating device. The vehicle behavior estimating unit 21 reads weight parameters stored in the weight parameter storage unit 13 via the data reading unit 14. Furthermore, the vehicle behavior estimating unit 21 estimates data related to suspension control required by the control value calculating unit 20, specifically, instantaneous values ​​related to suspension control, based on the behavior information and weight parameters received from the CAN 10.

[0044] The vehicle behavior estimation unit 21 includes a first AI estimation unit 22, a second AI estimation unit 23, a non-AI estimation unit 24, an AI estimation selection unit 25, and an integrator 26. The vehicle behavior estimation unit 21 estimates the sprung acceleration and unsprung acceleration of the four wheels using the first AI estimation unit 22, the second AI estimation unit 23, and the non-AI estimation unit 24. The vehicle behavior estimation unit 21 integrates the estimated sprung acceleration and unsprung acceleration of the four wheels to obtain the sprung velocity and unsprung velocity of the four wheels. The vehicle behavior estimation unit 21 obtains the sprung velocity and relative velocity of the four wheels based on the sprung velocity and unsprung velocity of the four wheels. The vehicle behavior estimation unit 21 outputs the sprung velocity and relative velocity of the four wheels as vehicle state quantities to the control value calculation unit 20 at the subsequent stage.

[0045] The first AI estimation unit 22 is an artificial intelligence. The first AI estimation unit 22 includes an FL state estimation unit 22A, a FR state estimation unit 22B, and a RL state estimation unit 22C. The FL state estimation unit 22A estimates the sprung acceleration and unsprung acceleration of the left front wheel based on behavior information. The FR state estimation unit 22B estimates the sprung acceleration and unsprung acceleration of the right front wheel based on behavior information. As a result, the first AI estimation unit 22 estimates the sprung acceleration and unsprung acceleration of the left front wheel and the right front wheel as front sprung and unsprung state quantities. In addition, the RL state estimation unit 22C estimates the sprung acceleration of the left rear wheel based on behavior information. Therefore, the first AI estimation unit 22 estimates the front sprung and unsprung state quantities and the rear left sprung state quantity out of the eight state quantities of the front sprung and unsprung parts and the rear sprung and unsprung parts.

[0046] Like the first AI estimation unit 22, the second AI estimation unit 23 is also artificial intelligence. The second AI estimation unit 23 includes an RL state estimation unit 23A and an RR state estimation unit 23B. The RL state estimation unit 23A estimates the unsprung acceleration of the left rear wheel based on behavior information. The RR state estimation unit 23B estimates the sprung acceleration and unsprung acceleration of the right rear wheel based on behavior information. As a result, the second AI estimation unit 23 estimates the unsprung accelerations of the left rear wheel and right rear wheel and the sprung acceleration of the right rear wheel as rear spring vertical state quantities.

[0047] The FL state estimation unit 22A, the FR state estimation unit 22B, the RL state estimation unit 22C, the RL state estimation unit 23A, and the RR state estimation unit 23B according to the second embodiment are configured similarly to the FL state estimation unit 16A, the FR state estimation unit 16B, the RL state estimation unit 16C, the RL state estimation unit 17A, and the RR state estimation unit 17B according to the first embodiment. That is, the FL state estimation unit 22A, the FR state estimation unit 22B, the RL state estimation unit 22C, the RL state estimation unit 23A, and the RR state estimation unit 23B according to the second embodiment are configured by a neural network. This neural network receives weight parameters that have learned the correlation between behavior information and unsprung acceleration, etc. The FL state estimation unit 22A, the FR state estimation unit 22B, the RL state estimation unit 22C, the RL state estimation unit 23A, and the RR state estimation unit 23B estimate unsprung acceleration, etc. based on the weight parameters and the behavior information.

[0048] The non-AI estimation unit 24 calculates the three state quantities on the rear wheel side from the estimation results of the first AI estimation unit 22. In this way, the non-AI estimation unit 24 reduces the calculation load on the vehicle behavior estimation unit 21. The non-AI estimation unit 24 includes an RL state estimation unit 24A and an RR state estimation unit 24B. The RL state estimation unit 24A estimates the unsprung acceleration of the left rear wheel based on the estimation results of the FL state estimation unit 22A and behavior information. The RR state estimation unit 24B estimates the unsprung acceleration of the right rear wheel based on the estimation results of the FR state estimation unit 22B and behavior information. The RR state estimation unit 24B estimates the sprung acceleration of the right rear wheel based on the estimation results of the FL state estimation unit 22A, the FR state estimation unit 22B, and the RL state estimation unit 22C and behavior information.

[0049] Specifically, the RL state estimation unit 24A calculates the unsprung acceleration of the left rear wheel by performing a delay process for the wheelbase on the unsprung acceleration of the left front wheel estimated by the FL state estimation unit 22A of the first AI estimation unit 22. The RR state estimation unit 24B calculates the unsprung acceleration of the right rear wheel by performing a delay process for the wheelbase on the unsprung acceleration of the right front wheel estimated by the FR state estimation unit 22B of the first AI estimation unit 22. Furthermore, the RR state estimation unit 24B calculates the sprung acceleration of the right rear wheel, which is the remaining wheel, from the sprung accelerations of the three wheels estimated by the FL state estimation unit 22A, the FR state estimation unit 22B, and the RL state estimation unit 22C by treating the vehicle body 2 as a rigid body. As a result, the non-AI estimation unit 24 calculates the unsprung accelerations of the left rear wheel and the right rear wheel and the sprung acceleration of the right rear wheel as rear spring vertical state quantities.

[0050] The AI ​​estimation selection unit 25 selects whether the unsprung acceleration of the rear wheels and the like are calculated by the non-AI estimation unit 24 or estimated by the second AI estimation unit 23. The AI ​​estimation selection unit 25 according to the second embodiment is configured in a manner similar to the AI ​​estimation selection unit 19 according to the first embodiment.

[0051] For this reason, the AI ​​estimation selection unit 25 selects either the non-AI estimation unit 24 or the second AI estimation unit 23 based on, for example, the steering angle included in the behavior information. When the steering angle is smaller than a predetermined angle, the AI ​​estimation selection unit 19 selects the non-AI estimation unit 24. At this time, the non-AI estimation unit 24 calculates the sprung acceleration and unsprung acceleration of the rear wheels based on the sprung acceleration and unsprung acceleration of the front wheels. On the other hand, when the steering angle is larger than the predetermined angle, the AI ​​estimation selection unit 25 selects the second AI estimator 23. At this time, the second AI estimator 23 estimates the sprung acceleration and unsprung acceleration of the rear wheels based on the behavior information.

[0052] The integrator 26 integrates the sprung acceleration and unsprung acceleration of the four wheels estimated by the first AI estimation unit 22, the second AI estimation unit 23, and the non-AI estimation unit 24. In this way, the integrator 26 calculates and outputs the sprung velocities and unsprung velocities of the four wheels. The vehicle behavior estimation unit 21 acquires the sprung velocities and relative velocities of the four wheels based on the sprung velocities and unsprung velocities of the four wheels.

[0053] Thus, the second embodiment can also achieve substantially the same effects as the first embodiment. In the second embodiment, the first AI estimation unit 22, the second AI estimation unit 23, and the non-AI estimation unit 24 of the vehicle behavior estimation unit 21 estimate or calculate the sprung acceleration and the unsprung acceleration of the four wheels as the state quantities of the sprung and unsprung parts. Therefore, the vehicle behavior estimation unit 21 can obtain the sprung velocities and the unsprung velocities of the four wheels by integrating the sprung acceleration and the unsprung acceleration of the four wheels.

[0054] In the second embodiment, the first AI estimator 22 estimates the sprung acceleration of the left rear wheel in addition to the sprung and unsprung accelerations of the left front and right front wheels, but the present invention is not limited to this. The first AI estimator 22 may estimate the sprung acceleration of the right rear wheel instead of the left rear wheel. In this case, the second AI estimator 23 and the non-AI estimator 24 estimate or calculate the sprung acceleration of the left rear wheel, which is the remaining wheel, based on the sprung accelerations of the three wheels estimated by the first AI estimator 22.

[0055] Next, Figures 2, 3, and 6 show a third embodiment. The third embodiment is characterized in that the vehicle behavior estimation unit calculates the trajectories of the front and rear wheels based on the steering angle or lateral acceleration, and determines the deviation of the trajectories. In the third embodiment, the same components as those in the first embodiment are denoted by the same reference numerals, and their description will be omitted.

[0056] The vehicle behavior estimating unit 31 according to the third embodiment is a vehicle behavior estimating device. The vehicle behavior estimating unit 31 reads weight parameters stored in the weight parameter storage unit 13 via the data reading unit 14. Furthermore, the vehicle behavior estimating unit 31 estimates data related to suspension control required by the control value calculating unit 20, specifically, instantaneous values ​​related to suspension control, based on the behavior information and weight parameters received from the CAN 10.

[0057] The vehicle behavior estimation unit 31 includes a first AI estimation unit 16, a second AI estimation unit 17, a non-AI estimation unit 18, and an AI estimation selection unit 32. The vehicle behavior estimation unit 31 estimates the sprung and unsprung velocities of the four wheels using the first AI estimation unit 16, the second AI estimation unit 17, and the non-AI estimation unit 18. The vehicle behavior estimation unit 31 acquires the sprung and relative velocities of the four wheels based on the estimated sprung and unsprung velocities of the four wheels. The vehicle behavior estimation unit 31 outputs the sprung and relative velocities of the four wheels as vehicle state quantities to the control value calculation unit 20 at the subsequent stage.

[0058] The AI ​​estimation selection unit 32 selects whether the unsprung mass velocity of the rear wheels and the like is to be calculated by the non-AI estimation unit 18 or estimated by the second AI estimation unit 17. The AI ​​estimation selection unit 32 calculates the trajectories of the front and rear wheels based on the steering angle and lateral acceleration, and determines the deviation of the trajectories. Specifically, the AI ​​estimation selection unit 32 calculates the trajectories of the front and rear wheels based on the steering angle, lateral acceleration, and vehicle speed.

[0059] A specific method for calculating the deviation of the trajectory is shown below. For example, when it is assumed that the turn is a steady circular turn, the yaw rate r is expressed by the following formula (1).

[0060]

number

[0061] Here, δ is the steering angle, A is the stability factor, V is the vehicle speed, Lf is the longitudinal distance from the center of gravity to the front wheels, and Lr is the longitudinal distance from the center of gravity to the rear wheels. The sum of Lf and Lr is the wheelbase Lw. The sideslip angle β at this time is expressed by equation 2.

[0062]

number

[0063] Here, K is a proportionality constant called cornering power. When lateral acceleration a is used as an input, yaw rate r can be calculated using equation 3.

[0064]

number

[0065] The relationship between the center of gravity position and the turning center (Xcen, Ycen) is expressed by Equation 5 based on Equation 4.

[0066]

number

[0067]

number

[0068] The coordinates (Xfl, Yfl) and (Xrl, Yrl) of the front and rear wheels are expressed by the following equation (6) based on the size of the vehicle body 2.

[0069]

number

[0070] Based on these, the turning radius Rfl of the front wheels is expressed by the following equation (7).

[0071]

number

[0072] Similarly, the turning radius Rrl of the rear wheels is expressed by the following equation (8).

[0073]

number

[0074] If the difference between the turning radius Rfl of the front wheels and the turning radius Rrl of the rear wheels is large compared to the tire width, the AI ​​estimation selection unit 32 determines that it is difficult to estimate rear wheel information from the front wheel information. In this case, the AI ​​estimation selection unit 32 switches from the non-AI estimation unit 18 to the second AI estimator 17 to acquire the unsprung speed of the rear wheels, etc. In other words, if the difference between the turning radius Rfl of the front wheels and the turning radius Rrl of the rear wheels is large compared to the tire width, the second AI estimator 17, which is an artificial intelligence, estimates the unsprung speed of the rear wheels, etc., rather than the non-AI estimation unit 18 calculating the unsprung speed of the rear wheels using a formula.

[0075] As shown in Equations 1 to 8, the trajectories of the front and rear wheels can be calculated based on either the steering angle δ or the lateral acceleration a. Therefore, the vehicle behavior estimating unit 31 calculates the trajectories of the front and rear wheels based on the steering angle δ or the lateral acceleration a. However, the vehicle behavior estimating unit 31 may also calculate the trajectories of the front and rear wheels based on both the steering angle δ and the lateral acceleration a. For example, the vehicle behavior estimating unit 31 may compare the wheel trajectories calculated based on the steering angle δ with the wheel trajectories calculated based on the lateral acceleration a, and use the larger value (maximum value) to calculate the final wheel trajectories.

[0076] Thus, the third embodiment can achieve substantially the same effects as the first embodiment. In the third embodiment, the vehicle behavior estimation unit 31 calculates the trajectories of the front and rear wheels based on the steering angle δ or the lateral acceleration a, and determines the deviation of the trajectories. Therefore, when the calculated deviation of the trajectories is small compared to the tire width, the vehicle behavior estimation unit 31 selects the non-AI estimation unit 18 and can calculate the unsprung velocity of the rear wheels and the like using a formula. On the other hand, when the calculated deviation of the trajectory is large compared to the tire width, the vehicle behavior estimation unit 31 selects the second AI estimation unit 17, which is an artificial intelligence, and can estimate the unsprung velocity of the rear wheels and the like.

[0077] In the third embodiment, the first AI estimation unit 16, the second AI estimation unit 17, and the non-AI estimation unit 18 according to the first embodiment are used, but the first AI estimation unit 22, the second AI estimation unit 23, and the non-AI estimation unit 24 according to the second embodiment may also be used.

[0078] In the third embodiment, the AI ​​estimation selection unit 32 switches from the non-AI estimation unit 18 to the second AI estimation unit 17 when the deviation between the trajectories of the front and rear wheels is greater than a predetermined value, but the present invention is not limited to this. The AI ​​estimation selection unit 32 may also switch from the non-AI estimation unit 18 to the second AI estimation unit 17 when the vehicle is moving backward, for example. In other words, the AI ​​estimation selection unit 32 may select either the second AI estimation unit 17 or the non-AI estimation unit 18 depending on the traveling direction of the vehicle. This configuration can be applied to the first and second embodiments.

[0079] In the first embodiment, the controller 11 acquires vehicle behavior information including wheel speeds through the CAN 10, but the present invention is not limited to this. For example, the controller 11 may acquire detected values ​​of various sensors directly from the various sensors. Alternatively, the controller 11 may acquire behavior information from another controller, etc. This configuration can be applied to the second and third embodiments.

[0080] In the above-described embodiments, the force generating mechanism is a variable damper 7 made up of a semi-active damper. The present invention is not limited to this, and an active damper (either an electric actuator or a hydraulic actuator) may be used as the force generating mechanism. In the above-described embodiments, the force generating mechanism that generates an adjustable force between the vehicle body 2 side and the wheel 3 side is configured by a variable damper 7 made up of a damping force adjustable hydraulic shock absorber. The present invention is not limited to this, and the force generating mechanism may be configured by an air suspension, a stabilizer (kinesus), an electromagnetic suspension, or the like, in addition to a hydraulic shock absorber.

[0081] In the above embodiments, the vehicle behavior control device is used for a four-wheeled automobile. However, the present invention is not limited to this and can also be applied to, for example, a work vehicle, a transport vehicle such as a truck or a bus.

[0082] The present invention is not limited to the above-described embodiments and includes various modifications. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and are not necessarily limited to those including all of the described configurations. Furthermore, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment, or to add the configuration of another embodiment to the configuration of one embodiment. Furthermore, it is possible to add, delete, or replace part of the configuration of each embodiment with other configurations.

[0083] This application claims priority to Japanese Patent Application No. 2023-045993, filed March 22, 2023. The entire disclosure of Japanese Patent Application No. 2023-045993, filed March 22, 2023, including the specification, claims, drawings, and abstract, is incorporated herein by reference in its entirety. [Explanation of symbols]

[0084] 1: Vehicle behavior control device, 2: Vehicle body, 3: Wheel, 5: Suspension device, 7: Adjustable damping force shock absorber (variable damper), 8: Variable damping force actuator, 11: Controller, 15, 21, 31: Vehicle behavior estimation unit, 16, 22: First AI estimation unit, 17, 23: Second AI estimation unit, 18, 24: Non-AI estimation unit, 19, 25, 32: AI estimation selection unit

Claims

1. A vehicle behavior estimation method for estimating state quantities of sprung and unsprung masses of each wheel using artificial intelligence, comprising: The state quantities of the front sprung and unsprung masses and the sprung mass of one rear wheel are estimated using artificial intelligence. The remaining rear sprung and unsprung state quantities are calculated from the front sprung and unsprung state quantities and the sprung state quantity for one rear wheel, A vehicle behavior estimation method that uses artificial intelligence to estimate the state quantities of the rear sprung and unsprung masses as needed.

2. 2. The vehicle behavior estimation method according to claim 1, Under normal conditions, the state quantities of the front sprung and unsprung masses and the sprung mass of one rear wheel are estimated using artificial intelligence. A vehicle behavior estimation method that uses artificial intelligence to estimate the state quantities of the rear sprung and unsprung masses only when the trajectories of the front and rear wheels do not match.

3. 2. The vehicle behavior estimation method according to claim 1, A vehicle behavior estimation method in which, of eight state quantities of the front sprung and unsprung masses and the rear sprung and unsprung masses, only the state quantities of the front sprung and unsprung masses and the state quantities of the left or right rear sprung masses are estimated using artificial intelligence, and the remaining quantities are calculated from the state quantities of the front sprung and unsprung masses and the state quantities of the sprung mass for one rear wheel.

4. 2. The vehicle behavior estimation method according to claim 1, A vehicle behavior estimation method that, when it is determined that the trajectories of the front wheels and the rear wheels are different from each other, estimates the state quantities of the sprung and unsprung masses for both the front and rear wheels using artificial intelligence.

5. The vehicle behavior estimation method according to claim 2 or 4, A vehicle behavior estimation method that calculates the trajectory of the front wheels and the trajectory of the rear wheels based on the steering angle and determines the deviation of the trajectories.

6. The vehicle behavior estimation method according to claim 2 or 4, A vehicle behavior estimation method that calculates the trajectory of the front wheels and the trajectory of the rear wheels based on lateral acceleration and determines the deviation of the trajectories.

7. A vehicle behavior estimation device that estimates state quantities of sprung and unsprung masses of each wheel using artificial intelligence, The state quantities of the front sprung and unsprung masses and the sprung mass of one rear wheel are estimated using artificial intelligence. The remaining rear sprung and unsprung state quantities are calculated from the front sprung and unsprung state quantities and the sprung state quantity for one rear wheel, A vehicle behavior estimation device that uses artificial intelligence to estimate the state quantities of the rear sprung and unsprung masses as needed.

8. The vehicle behavior estimation device according to claim 7, Under normal conditions, the state quantities of the front sprung and unsprung masses and the sprung mass of one rear wheel are estimated using artificial intelligence. A vehicle behavior estimation device that uses artificial intelligence to estimate the state quantities of the rear sprung and unsprung masses only when the trajectories of the front and rear wheels do not match.

9. The vehicle behavior estimation device according to claim 7, A vehicle behavior estimation device that uses artificial intelligence to estimate only the state quantities of the front sprung and unsprung masses and the state quantities of the left or right rear sprung masses out of eight state quantities of the front sprung and unsprung masses and the rear sprung and unsprung masses, and calculates the rest from the state quantities of the front sprung and unsprung masses and the state quantity of the sprung mass for one rear wheel.

10. The vehicle behavior estimation device according to claim 7, A vehicle behavior estimation device that uses artificial intelligence to estimate the state quantities of sprung and unsprung masses for both the front and rear wheels when it determines that the trajectories of the front and rear wheels are different from each other.

11. The vehicle behavior estimation device according to claim 8 or 10, A vehicle behavior estimation device that calculates the trajectory of the front and rear wheels based on the steering angle and determines the deviation of the trajectories.

12. The vehicle behavior estimation device according to claim 8 or 10, A vehicle behavior estimation device that calculates the trajectory of the front and rear wheels based on lateral acceleration and determines the deviation of the trajectories.

Images