Suspension control device and anomaly detection device

Abstract

This can suppress the increase in estimation errors. [Solution] The vehicle state quantity calculation unit 26 is an abnormality detection device that detects abnormalities in the relative speed between the vehicle body 2 and the wheels 3 in the vehicle 1. The vehicle state quantity calculation unit 26 includes a first relative speed estimation unit 27, a second relative speed estimation unit 28, a third relative speed estimation unit 29, and an AI estimated value comparison processing unit 30 (estimated value processing unit). The AI ​​estimated value comparison processing unit 30 receives the relative speeds (first relative speed, second relative speed, third relative speed) from the first relative speed estimation unit 27, the second relative speed estimation unit 28, and the third relative speed estimation unit 29, and outputs one relative speed from the relative speeds of the first relative speed estimation unit 27, the second relative speed estimation unit 28, and the third relative speed estimation unit 29.

Description

Technical Field

[0001] The present disclosure relates to a suspension control device and an abnormality detection device that control a suspension device provided between a vehicle body and wheels of a vehicle.

Background Art

[0002] Patent Document 1 discloses a control device including: an acquisition unit that acquires an acceleration of a vehicle body in the vertical direction of the vehicle; and an estimation unit that estimates a relative speed of a wheel with respect to the vehicle body in the vertical direction of the vehicle in response to an input of a numerical value obtained by multiplying the acceleration by the mass of the vehicle, using a neural network trained to perform such estimation. This control device controls a suspension device based on the estimated relative speed.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] By the way, in the control device according to Patent Document 1, a neural network is used to estimate a relative speed. However, it is unclear whether the accuracy of the estimated relative speed is ensured. Therefore, for example, due to correspondence to unlearned data, overlearning, etc., the estimation error may increase, and the riding comfort may deteriorate.

[0005] An object of an embodiment of the present invention is to provide a suspension control device and an abnormality detection device capable of suppressing an increase in an estimation error.

Means for Solving the Problems

[0006] A suspension control device according to one embodiment of the present invention includes: a suspension device provided between the vehicle body and the wheels in a vehicle to suppress vibrations between the vehicle body and the wheels; a first relative speed estimation unit that receives signals from a physical sensor mounted on the vehicle body and estimates the relative speed between the vehicle body and the wheels based on learning results obtained by machine learning in advance; a second relative speed estimation unit that receives signals from the physical sensor and estimates the relative speed between the vehicle body and the wheels in a manner different from that of the first relative speed estimation unit; a third relative speed estimation unit that receives signals from the physical sensor and estimates the relative speed between the vehicle body and the wheels in a manner different from that of the first and second relative speed estimation units; an estimated value processing unit that receives the relative speeds from the first relative speed estimation unit, the second relative speed estimation unit and the third relative speed estimation unit and outputs a single relative speed from the relative speeds from the first relative speed estimation unit, the second relative speed estimation unit and the third relative speed estimation unit; and a control unit that receives the relative speed from the estimated value processing unit and outputs a control signal for the suspension device.

[0007] One embodiment of the present invention is an abnormality detection device for detecting abnormalities in the relative speed between a vehicle body and its wheels, comprising: a first relative speed estimation unit that receives signals from a physical sensor mounted on the vehicle body and estimates the relative speed between the vehicle body and its wheels; a second relative speed estimation unit that receives signals from the physical sensor and estimates the relative speed between the vehicle body and its wheels in a manner different from that of the first relative speed estimation unit; a third relative speed estimation unit that receives signals from the physical sensor and estimates the relative speed between the vehicle body and its wheels in a manner different from that of the first and second relative speed estimation units; and an estimated value processing unit that receives the respective relative speeds from the first, second, and third relative speed estimation units, compares the respective relative speeds from the first, second, and third relative speed estimation units, and outputs a relative speed that is determined to be normal.

[0008] One embodiment of the present invention includes: a suspension device provided between the vehicle body and the wheels of a vehicle to suppress vibrations between the vehicle body and the wheels; a first relative speed estimation unit that receives signals from physical sensors mounted on the vehicle body and estimates the relative speed between the vehicle body and the wheels based on learning results obtained by machine learning in advance; a second relative speed estimation unit that receives signals from the physical sensors and estimates the relative speed between the vehicle body and the wheels in a manner different from that of the first relative speed estimation unit; an estimation value processing unit that receives the relative speeds of the first relative speed estimation unit and the second relative speed estimation unit, diagnoses whether the relative speed of the first relative speed estimation unit is normal or not, outputs the relative speed of the first relative speed estimation unit if the relative speed of the first relative speed estimation unit is normal, and outputs an abnormality detection signal indicating that the relative speed is abnormal if the relative speed of the first relative speed estimation unit is not normal; and a control unit that, upon receiving the relative speed from the estimation value processing unit, outputs a control signal for the suspension device, and upon receiving the abnormality detection signal from the estimation value processing unit, outputs a control stop signal to stop the control of the suspension device. [Effects of the Invention]

[0009] According to one embodiment of the present invention, the increase in estimation error can be suppressed. [Brief explanation of the drawing]

[0010] [Figure 1] This is an explanatory diagram showing a vehicle to which a suspension control device according to the first embodiment is applied. [Figure 2] This is a block diagram showing a damper control unit according to the first embodiment. [Figure 3] This is an explanatory diagram showing an example of a neural network in the first relative velocity estimation unit. [Figure 4] This is a block diagram of the second relative velocity estimation unit. [Figure 5] This flowchart shows the AI ​​estimate comparison process performed by the AI ​​estimate comparison processing unit. [Figure 6]This is an explanatory diagram showing a vehicle to which a suspension control device according to the second embodiment is applied. [Figure 7] This is a block diagram showing a damper control unit according to the second embodiment. [Figure 8] This is an explanatory diagram showing a vehicle to which a suspension control device according to the third embodiment is applied. [Figure 9] This is a block diagram showing a damper control unit according to the third embodiment. [Figure 10] This flowchart shows the AI ​​estimated value anomaly detection process performed by the AI ​​estimated value anomaly detection unit. [Modes for carrying out the invention]

[0011] The following will describe in detail, with reference to the attached drawings, an example of how the suspension control device and abnormality detection device according to embodiments of the present invention can be applied to a four-wheeled vehicle.

[0012] Figure 1 shows a vehicle 1 to which the suspension control device of the present invention is applied. Vehicle 1 includes, for example, a suspension device 5, an IMU 9, and a controller 21 (ECU). In Figure 1, the lower side of the vehicle body 2 that constitutes the body of vehicle 1 is provided with, for example, left and right front wheels and left and right rear wheels (hereinafter collectively referred to as wheels 3). These wheels 3 include tires 4, and the tires 4 act as springs that absorb fine irregularities in the road surface.

[0013] The suspension device 5 is a shock absorber provided by the vehicle 1. The suspension device 5 is installed between the vehicle body 2 and the wheels 3 and suppresses vibrations between the vehicle body 2 and the wheels 3. This suspension device 5 consists of a suspension spring 6 (hereinafter referred to as the spring 6) and a damping force adjustable shock absorber (hereinafter referred to as the variable damper 7) installed in parallel with the spring 6 between the vehicle body 2 and the wheels 3.

[0014] In addition, in FIG. 1, a case where a set of suspension devices 5 are provided between the vehicle body 2 and the wheels 3 is shown. However, the suspension devices 5 can be provided in a total of four sets, individually and independently, between, for example, the four wheels 3 and the vehicle body 2, and only one set of them is schematically shown in FIG. 1.

[0015] Here, the variable damper 7 of the suspension device 5 is configured using a damping force adjustment type hydraulic shock absorber interposed between the vehicle body 2 and the wheels 3. The variable damper 7 is a force generation mechanism of the suspension device 5. The variable damper 7 is provided between the vehicle body 2 and the wheels 3, and constitutes a relative displacement suppression device that changes the force for suppressing the relative displacement between the vehicle body 2 and the wheels 3.

[0016] The variable damper 7 is attached with a damping force variable actuator 8 composed 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 a hard characteristic (hard characteristic) to a soft characteristic (soft characteristic). 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, for example, one that can adjust the damping force in a plurality of steps of two or more steps. Further, the variable damper 7 may be of a pressure control type or a flow rate control type. The variable damper 7 may be of a type that controls viscosity such as magnetorheological fluid or electrorheological fluid.

[0017] The inertial measurement unit 9 (hereinafter referred to as IMU9) is a vehicle motion detection means for detecting the six-axis momentum of the vehicle 1. The IMU9 is provided on the vehicle body 2 which is the so-called upper side of the spring. The IMU9 includes, for example, a three-axis angular velocity sensor and a three-axis acceleration sensor. The IMU9 is a multi-purpose sensor of the vehicle 1. The detection signal of the IMU9 is used, for example, for traction control, brake control, skid prevention control, automatic driving control, etc. of the vehicle 1.

[0018] The IMU9 is attached to an arbitrary sensor position on the vehicle body 2. The IMU9 detects the vertical spring acceleration at the sensor position, the roll rate which is the angular velocity in the roll direction, and the pitch rate which is the angular velocity in the pitch direction, and outputs the detection signals to the controller 21 described later. By grasping in advance the mounting position (sensor position) of the IMU9, the controller 21 can grasp the vehicle motion. That is, considering the vehicle body 2 on the spring as a rigid body, the vertical absolute velocity on the spring (spring - on velocity) of each wheel can be geometrically calculated based on the vertical spring acceleration, roll rate, and pitch rate detected at any one location of the vehicle body 2. Note that the sensor position of the IMU9 may be, for example, the center - of - gravity position of the vehicle body 2 or a position other than the center of gravity.

[0019] The CAN10 (Controller Area Network) is connected to the controller 21 and is also connected to a wheel speed sensor 11 that detects the rotational speed of the wheel 3. At this time, the wheel speed sensor 11 and the IMU9 constitute a physical sensor. Also, the CAN10 is connected to various multi - purpose sensors of the vehicle, such as a vehicle speed sensor, a steering angle sensor, etc. The CAN10 transmits various vehicle information including the rotational speed of the wheel 3 (wheel speed) and the vehicle speed which is the speed of the vehicle. Thereby, the controller 21 can obtain information such as wheel speed and vehicle speed through the CAN10.

[0020] In this embodiment, the CAN is taken as an example for the in - vehicle network for explanation, but other in - vehicle networks may also be used. The in - vehicle network may be, for example, CAN FD (CAN with Flexible Data rate), FlexRay, in - vehicle Ethernet, etc.

[0021] The controller 21 is configured as a control device for controlling the damping characteristics of the variable damper 7, and is, for example, a microcomputer. The controller 21 has a storage unit 22 consisting of ROM, RAM, non-volatile memory, etc. The storage unit 22 of the controller 21 stores various programs, information (sensor position, vehicle specifications), data, etc. for controlling the variable damper 7. The input side of the controller 21 is connected to the IMU 9 and also to the CAN 10, which is, for example, a communication network necessary for data communication. The output side of the controller 21 is connected to the variable damping force actuator 8 of the variable damper 7.

[0022] The controller 21 calculates (estimates) the relative speed (stroke speed) between the vehicle body 2 (sprung mass) and the wheel 3 (unsprung mass) from the values ​​detected by the IMU 9. At this time, the controller 21 calculates two relative speeds (first relative speed and second relative speed) from the values ​​detected by the IMU 9 using two different methods. Specifically, the controller 21 estimates the relative speed (first relative speed) from the values ​​detected by the IMU 9 using artificial intelligence (AI). The controller 21 estimates the relative speed (second relative speed) from the values ​​detected by the IMU 9 using a Kalman filter. In addition, the controller 21 calculates (estimates) the relative speed (third relative speed) between the vehicle body 2 and the wheel 3 from, for example, the values ​​detected by the wheel speed sensor 11.

[0023] The controller 21 compares the first relative velocity, the second relative velocity, and the third relative velocity to determine whether the first relative velocity, which is an AI-estimated value, is normal or not. Unless the first relative velocity is abnormal, the controller 21 controls the variable damper 7 based on the first relative velocity. That is, when the first relative velocity is normal, the controller 21 controls the variable damper 7 based on the first relative velocity. When the first relative velocity is abnormal, the controller 21 controls the variable damper 7 based on the second relative velocity estimated from the IMU9's detection value.

[0024] As shown in Figures 1 and 2, the controller 21 includes a damper control unit 23, a weight parameter storage unit 24, and a data reading unit 25. The damper control unit 23 includes a vehicle state quantity calculation unit 26 and a damping force control unit 31. The weight parameter storage unit 24 may be part of the storage unit 22, or it may be a separate unit from the storage unit 22.

[0025] The vehicle state quantity calculation unit 26 is an abnormality detection device that detects abnormalities in the relative speed between the vehicle body 2 and the wheels 3 in the vehicle 1. The vehicle state quantity calculation unit 26 acquires data on the sprung mass acceleration, roll rate, and pitch rate at the sensor position from the IMU 9. The vehicle state quantity calculation unit 26 acquires data related to the vehicle's behavior (hereinafter referred to as behavior information) via the CAN 10. At this time, the behavior information includes, for example, longitudinal acceleration (longitudinal G), lateral acceleration (lateral G), steering angle, yaw rate, and the wheel speed of each vehicle. Note that the controller 21 does not need to acquire data from the IMU 9 directly, but may acquire it via the CAN 10.

[0026] The vehicle state quantity calculation unit 26 calculates (estimates) the sprung speed and relative speed as vehicle state quantities based on data acquired from, for example, the IMU 9 and CAN 10. The vehicle state quantity calculation unit 26 includes a first relative speed estimation unit 27, a second relative speed estimation unit 28, a third relative speed estimation unit 29, and an AI estimated value comparison processing unit 30. Based on data acquired from the IMU 9 and CAN 10, the vehicle state quantity calculation unit 26 estimates the data related to suspension control required by the damping force control unit 31, specifically, instantaneous values ​​related to suspension control (sprung speed, relative speed).

[0027] The first relative velocity estimation unit 27 is equipped with artificial intelligence and estimates the relative velocity (first relative velocity) between the vehicle body 2 and the wheels 3 based on the values ​​detected by the IMU 9. The first relative velocity estimation unit 27 receives IMU signals from the IMU 9. The first relative velocity estimation unit 27 calculates the sprung mass acceleration, roll rate, and pitch rate of the sensor position from the data acquired from the IMU 9.

[0028] The first relative speed estimation unit 27 receives signals from the IMU 9 (physical sensor) mounted on the vehicle body 2 and estimates the relative speed between the vehicle body 2 and the wheels 3 based on the learning results obtained by machine learning in advance. At this time, the first relative speed estimation unit 27 calculates the stroke speed (relative speed) of the suspension device 5 of each wheel based on the sprung mass acceleration, roll rate, and pitch rate obtained from the IMU signal. Specifically, the first relative speed estimation unit 27 estimates the stroke speed of each wheel using a neural network that has been trained to estimate the stroke speed of each wheel by inputting sprung mass acceleration, roll rate, pitch rate, etc. The first relative speed estimation unit 27 reads the weight parameters stored in the weight parameter storage unit 24 via the data reading unit 25. The first relative speed estimation unit 27 estimates the stroke speed based on the sprung mass acceleration, roll rate, pitch rate and the weight parameters.

[0029] As an example, the specific configuration of the first relative velocity estimation unit 27 will be described with reference to Figure 3. As shown in Figure 3, the first relative velocity estimation unit 27 is configured similarly to the neural network disclosed in, for example, Japanese Patent Application Publication No. 2022-191913. The neural network is composed of 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~I, j=1~J), and each element of the hidden layer 102 and each element of the output layer 103 are connected by weights W2jk (j=1~J, k=1~K). This weight information (hereinafter referred to as weight parameters) is represented by the matrix of weights W1ij and weights W2jk. The weight parameters are determined in advance by machine learning and are stored in the weight parameter storage unit 24. In this example, we have shown a fully connected neural network with a single hidden layer 102, but this is not the only option. For example, a neural network with two or more hidden layers 102 is also possible.

[0030] The input layer 101 of the neural network receives time-series data, such as sprung acceleration calculated based on the IMU signal. The output layer 103 outputs instantaneous values ​​of the stroke velocity of the suspension device 5 for each wheel of the vehicle 1. The number of elements in the hidden layer 102 is generally determined from the number of elements in the input layer 101 and the output layer 103, but it is set to the number that maximizes the accuracy of state estimation by the neural network. The number of elements in the output layer 103 is determined by the output specifications for stroke velocity (relative velocity) estimation.

[0031] The neural network's machine learning is performed based on data of vehicle state variables (e.g., sprung speed, unsprung speed, stroke speed, etc.) acquired in advance by a data acquisition vehicle, and data such as sprung acceleration included in the IMU signal. The data acquisition vehicle has the same specifications as the vehicle on which the first relative speed estimation unit 27 is installed, and is equipped with various sensors (accelerometers, etc.) for acquiring vehicle state variables. In the neural network's machine learning, the weight parameters are adjusted to learn the correlation between the vehicle state variable data (sprung speed, unsprung speed, stroke speed, etc.) acquired by the data acquisition vehicle and data such as sprung acceleration. The weight parameters obtained as a result of the learning are stored in the weight parameter storage unit 24.

[0032] In this embodiment, the first relative speed estimation unit 27 performs machine learning to correlate data of vehicle state quantities acquired by a data acquisition vehicle with data of IMU signals (such as sprung mass acceleration), but the present invention is not limited to this. For example, a vehicle model corresponding to the vehicle on which the first relative speed estimation unit 27 is installed may be constructed, and the first relative speed estimation unit 27 may perform machine learning based on data acquired by simulation using this vehicle model.

[0033] The first relative velocity estimation unit 27 estimates the stroke velocity of the suspension device 5 for each wheel based on the sprung mass acceleration, roll rate, and pitch rate obtained from the IMU signal, but the present invention is not limited to this. The first relative velocity estimation unit 27 may, for example, calculate the sprung mass acceleration of each wheel based on data obtained from the IMU 9, and estimate the stroke velocity of the suspension device 5 for each wheel based on this sprung mass acceleration.

[0034] In this case, the first relative velocity estimation unit 27 calculates the vertical sprung mass acceleration of each wheel based on data acquired from the IMU 9, the sensor position of the IMU 9, and vehicle specifications information such as the shape, size, weight, wheelbase, and position of each wheel 3 of the vehicle body 2. Specifically, the first relative velocity estimation unit 27 calculates the roll angular acceleration and pitch angular acceleration of the sensor position from data acquired from the IMU 9, and also calculates the vertical sprung mass acceleration of each wheel based on the sprung mass acceleration, roll angular acceleration, and pitch angular acceleration of the sensor position and the relationship between the sensor position and the position (tire position) of each wheel 3. The first relative velocity estimation unit 27 receives the sprung mass acceleration of each wheel as input and estimates the stroke speed of each wheel using a neural network that has been trained to estimate the stroke speed of each wheel.

[0035] Furthermore, the first relative velocity estimation unit 27 may use a neural network that takes as input values ​​a value indicating the force (damping force) generated by the variable damper 7, in addition to the sprung mass acceleration of each wheel, to determine the relative velocity (stroke velocity) of the suspension device 5 for each wheel.

[0036] In this case, the first relative velocity estimation unit 27 acquires the damping force by taking the control amount from the damping force control unit 31 as input, as a value indicating the force generated by the variable damper 7 (force generating device). That is, the first relative velocity estimation unit 27 calculates the damping force generated by the variable damper 7 based on the stroke velocity output from the first relative velocity estimation unit 27 and the command current value as a control signal output from the damping force control unit 31.

[0037] The first relative velocity estimation unit 27 calculates the sprung velocity of each wheel in addition to the stroke velocity (relative velocity) of each wheel. In this case, the first relative velocity estimation unit 27 calculates the sprung velocity by integrating the sprung acceleration of each wheel. At this time, the first relative velocity estimation unit 27 is equipped with a filter (not shown) to compensate for the phase of the output signal. Therefore, the first relative velocity estimation unit 27 integrates the sprung acceleration of each wheel, performs filtering, and calculates the sprung velocity of each wheel. The first relative velocity estimation unit 27 outputs the sprung velocity of each wheel and the relative velocity as state variables of the vehicle 1.

[0038] The first relative velocity estimation unit 27 is not limited to calculating the sprung velocity by integrating the sprung acceleration of each wheel. The first relative velocity estimation unit 27 may also estimate the sprung velocity from the sprung acceleration, etc., using a neural network, similar to the stroke velocity estimation unit.

[0039] The second relative speed estimation unit 28 estimates the relative speed (second relative speed) between the vehicle body 2 and the wheels 3 based on the values ​​detected by the IMU 9. The second relative speed estimation unit 28 includes a Kalman filter 28A and a damping force variable component calculation unit 28B disclosed in Japanese Patent No. 5158333 (see Figure 4). The Kalman filter 28A calculates the relative speed between the vehicle body 2 and the wheels 3 based, for example, the sprung mass acceleration and the damping force variable component of the variable damper 7 calculated by the damping force variable component calculation unit 28B.

[0040] At this time, the second relative speed estimation unit 28 receives an IMU signal from the IMU 9. The second relative speed estimation unit 28 calculates the vertical sprung mass acceleration of each wheel based on the data acquired from the IMU 9, the sensor position of the IMU 9, and vehicle specifications information such as the shape, size, weight, wheelbase, and position of each wheel 3 of the vehicle body 2. Specifically, the second relative speed estimation unit 28 calculates the roll angular acceleration and pitch angular acceleration of the sensor position from the data acquired from the IMU 9, and also calculates the vertical sprung mass acceleration of each wheel based on the sprung mass acceleration, roll angular acceleration, and pitch angular acceleration of the sensor position, and the relationship between the sensor position and the position (tire position) of each wheel 3.

[0041] In addition, the second relative speed estimation unit 28 receives control command values ​​output from the damping force control unit 31. The damping force variable component calculation unit 28B calculates the damping force change (damping force variable component) using the estimated relative speed output from the Kalman filter 28A and the control command values ​​calculated by the damping force control unit 31, and feeds this back to the Kalman filter 28A. The Kalman filter 28A calculates the stroke speed (relative speed) of the suspension device 5 for each wheel based on the sprung mass acceleration of each wheel and the calculated damping force variable component. As a result, the second relative speed estimation unit 28 receives signals from the IMU 9 (physical sensor) and estimates the relative speed (second relative speed) between the vehicle body 2 and the wheels 3 in a different way than the first relative speed estimation unit 27.

[0042] The third relative speed estimation unit 29 receives a signal from the wheel speed sensor 11 (physical sensor) and estimates the relative speed (third relative speed) between the vehicle body 2 and the wheels 3 in a different manner than the first relative speed estimation unit 27 and the second relative speed estimation unit 28. As shown in Figure 2, the third relative speed estimation unit 29 estimates the third relative speed based on the wheel speed. The third relative speed estimation unit 29 receives the wheel speed signal for each wheel. The third relative speed estimation unit 29 calculates (estimates) the relative speed (third relative speed) between the vehicle body 2 and the wheels 3 from the detected value of the wheel speed sensor 11, which acts as a wheel speed detection means for detecting the rotational speed of the wheels 3. Specifically, the third relative speed estimation unit 29 receives the wheel speed for each wheel and estimates the stroke speed of each wheel using a neural network that has been trained to estimate the stroke speed of each wheel. The neural network of the third relative speed estimation unit 29 is configured similarly to the neural network of the first relative speed estimation unit 27. The third relative speed estimation unit 29 reads the weight parameters stored in the weight parameter storage unit 24 via the data reading unit 25. Based on the wheel speed and the weight parameters, the third relative speed estimation unit 29 estimates the stroke speed of each wheel.

[0043] In this embodiment, the third relative speed estimation unit 29 estimates the relative speed of each wheel from behavioral information including the wheel speed of each wheel using AI, but the present invention is not limited to this. For example, the third relative speed estimation unit 29 may calculate the relative speed of each wheel from behavioral information including the wheel speed of each wheel using a mathematical formula based on a Kalman filter or the like.

[0044] The AI ​​estimated value comparison processing unit 30 is an estimated value processing unit. The AI ​​estimated value comparison processing unit 30 receives the relative velocities (first relative velocity, second relative velocity, and third relative velocity) from the first relative velocity estimation unit 27, the second relative velocity estimation unit 28, and the third relative velocity estimation unit 29, and outputs one relative velocity from the relative velocities of the first relative velocity estimation unit 27, the second relative velocity estimation unit 28, and the third relative velocity estimation unit 29.

[0045] The AI ​​estimated value comparison processing unit 30 compares the first relative velocity estimated by the first relative velocity estimation unit 27, the second relative velocity estimated by the second relative velocity estimation unit 28, and the third relative velocity estimated by the third relative velocity estimation unit 29. Based on the differences between the first relative velocity, the second relative velocity, and the third relative velocity, the AI ​​estimated value comparison processing unit 30 determines whether the first relative velocity, the second relative velocity, and the third relative velocity are normal or not.

[0046] Specifically, the AI ​​estimated value comparison processing unit 30 calculates the difference a between the first relative velocity and the second relative velocity, the difference b between the first relative velocity and the third relative velocity, and the difference c between the second relative velocity and the third relative velocity. It then diagnoses whether the first relative velocity, second relative velocity, and third relative velocity are normal or not based on the thresholds a0, b0, and c0 set for differences a, b, and c, respectively. At this time, the thresholds a0, b0, and c0 are all judgment thresholds, and are positive values ​​(a0, b0, c0 > 0) within the range in which each relative velocity can be judged as normal. These thresholds are set appropriately considering the actual detected values ​​and calculated values.

[0047] When the first, second, and third relative velocities are all normal, their estimation accuracy generally follows the order of first relative velocity, second relative velocity, and third relative velocity. That is, the estimation accuracy of the first relative velocity is the highest, the estimation accuracy of the second relative velocity is the second highest, and the estimation accuracy of the third relative velocity is the lowest.

[0048] Taking this into consideration, the AI ​​estimated value comparison processing unit 30 selects the first relative speed and outputs it to the damping force control unit 31 when it determines that all relative speeds are normal. When the AI ​​estimated value comparison processing unit 30 determines that only the first relative speed is abnormal, it selects the second relative speed and outputs it to the damping force control unit 31. When the AI ​​estimated value comparison processing unit 30 determines that only the second relative speed is abnormal, it selects the first relative speed and outputs it to the damping force control unit 31. When the AI ​​estimated value comparison processing unit 30 determines that only the third relative speed is abnormal, it selects the first relative speed and outputs it to the damping force control unit 31. Furthermore, if the AI ​​estimated value comparison processing unit 30 determines that two or more relative speeds are abnormal, it outputs a fail judgment as an unexpected abnormality. In this case, the AI ​​estimated value comparison processing unit 30 identifies the abnormal relative speed or sensor and outputs the identified information as failure information, while the damping force control unit 31 executes processing based on the fail judgment.

[0049] The damping force control unit 31 is a control unit that receives the relative speed from the AI ​​estimated value comparison processing unit 30 and outputs a control signal for the suspension device 5. The damping force control unit 31 determines the control amount for controlling the variable damper 7 from, for example, the sprung speed and the relative speed. Specifically, the damping force control unit 31 calculates a control command value (command current value) for controlling the damping force of the variable damper 7 of the suspension device 5 based on the sprung speed calculated (estimated) by the first relative speed estimation unit 27 and the relative speed selected by the AI ​​estimated value comparison processing unit 30. At this time, the control amount for controlling the variable damper 7 is the control command value for controlling the damping force of the variable damper 7. Specifically, the damping force control unit 31 calculates a command current value as a control command value for improving the ride comfort of the vehicle, based on, for example, bilinear optimal control, skyhook control, H∞ control, etc. Based on the command current value, the damping force control unit 31 outputs a command current as a control signal to the variable damping force actuator 8 of the variable damper 7. Furthermore, the damping force control unit 31 may calculate control command values ​​not only for improving the ride comfort of the vehicle, but also for improving handling stability.

[0050] Next, the AI ​​estimated value comparison process performed by the AI ​​estimated value comparison processing unit 30 will be explained with reference to Figure 5. Each step in the flowchart shown in Figure 5 will be denoted as "S" (for example, step 1 will be "S1"). The same applies to the flowchart shown in Figure 10.

[0051] In S1, the AI ​​estimated value comparison processing unit 30 calculates the differences a, b, and c between the first relative velocity, the second relative velocity, and the third relative velocity. Specifically, the AI ​​estimated value comparison processing unit 30 calculates the difference a between the first relative velocity and the second relative velocity. The AI ​​estimated value comparison processing unit 30 calculates the difference b between the first relative velocity and the third relative velocity. The AI ​​estimated value comparison processing unit 30 calculates the difference c between the second relative velocity and the third relative velocity.

[0052] In S2, the AI estimated value comparison processing unit 30 determines whether the first relative speed is abnormal. Specifically, the AI estimated value comparison processing unit 30 uses, as determination condition J11, that the absolute value of difference a is greater than or equal to threshold value a0 (|a|≥a0), as determination condition J12, that the absolute value of difference b is greater than or equal to threshold value b0 (|b|≥b0), and as determination condition J13, determines whether the absolute value of difference c is less than threshold value c0 (|c|<c0). When differences a, b, and c satisfy the three determination conditions J11, J12, and J13, since the first relative speed is abnormal, it is determined as "YES" in S2, and the process proceeds to S3. On the other hand, when differences a, b, and c do not satisfy at least one of the three determination conditions J11, J12, and J13, it is determined as "NO" in S2, and the process proceeds to S5.

[0053] In S3, the AI estimated value comparison processing unit 30 executes abnormal processing of the first relative speed, which is the AI estimated value. The abnormal processing of the first relative speed is, for example, various error processes such as outputting a signal indicating that the first relative speed is abnormal and prompting the user to confirm the operation of the first relative speed estimation unit 27. After executing the abnormal processing of the first relative speed in S3, the process proceeds to S4. In S4, the AI estimated value comparison processing unit 30 selects the second relative speed, which is the calculated value of the Kalman filter based on the detected value of the IMU 9, and outputs it to the damping force control unit 31. At this time, the damping force control unit 31 controls the damping force of the variable damper 7 based on the second relative speed.

[0054] In S5, the AI estimated value comparison processing unit 30 determines whether the second relative speed is abnormal. Specifically, the AI estimated value comparison processing unit 30 uses, as determination condition J21, that the absolute value of difference a is greater than or equal to threshold value a0 (|a|≥a0), as determination condition J22, that the absolute value of difference b is less than threshold value b0 (|b|<b0), and as determination condition J23, determines whether the absolute value of difference c is greater than or equal to threshold value c0 (|c|≥c0). When differences a, b, and c satisfy the three determination conditions J21, J22, and J23, since the second relative speed is abnormal, it is determined as "YES" in S5, and the process proceeds to S6. On the other hand, when differences a, b, and c do not satisfy at least one of the three determination conditions J21, J22, and J23, it is determined as "NO" in S5, and the process proceeds to S8.

[0055] In S6, the AI estimated value comparison processing unit 30 performs abnormality processing on the second relative speed which is the IMU estimated value. The abnormality processing of the second relative speed is, for example, various error processes such as outputting a signal indicating that the second relative speed is abnormal and prompting the user to confirm the operation of the second relative speed estimation unit 28. After performing the abnormality processing of the second relative speed in S6, the process proceeds to S7. In S7, the AI estimated value comparison processing unit 30 selects the first relative speed which is the AI estimated value and outputs it to the damping force control unit 31. At this time, the damping force control unit 31 controls the damping force of the variable damper 7 based on the first relative speed.

[0056] In S8, the AI estimated value comparison processing unit 30 determines whether the third relative speed is abnormal. Specifically, the AI estimated value comparison processing unit 30 determines whether the absolute value of the difference a is smaller than the threshold value a0 (|a| < a0) as the determination condition J31, the absolute value of the difference b is greater than or equal to the threshold value b0 (|b| ≥ b0) as the determination condition J32, and the absolute value of the difference c is greater than or equal to the threshold value c0 (|c| ≥ c0) as the determination condition J33. When the differences a, b, and c satisfy the three determination conditions J31, J32, and J33, since the third relative speed is abnormal, it is determined as "YES" in S8 and the process proceeds to S9. On the other hand, when the differences a, b, and c do not satisfy at least one of the three determination conditions J31, J32, and J33, it is determined as "NO" in S8 and the process proceeds to S11.

[0057] In S9, the AI estimated value comparison processing unit 30 performs abnormality processing on the third relative speed which is the wheel speed estimated value. The abnormality processing of the third relative speed is, for example, various error processes such as outputting a signal indicating that the second relative speed is abnormal and prompting the user to confirm the operation of the third relative speed estimation unit 29 and the maintenance of the wheel speed sensor 11. After performing the abnormality processing of the third relative speed in S9, the process proceeds to S10. In S10, the AI estimated value comparison processing unit 30 selects the first relative speed which is the AI estimated value and outputs it to the damping force control unit 31. At this time, the damping force control unit 31 controls the damping force of the variable damper 7 based on the first relative speed.

[0058] In S11, the AI estimated value comparison processing unit 30 determines whether all relative speeds are normal. Specifically, the AI estimated value comparison processing unit 30 determines whether the absolute value of the difference a is smaller than the threshold value a0 (|a| < a0) as the determination condition J41, whether the absolute value of the difference b is smaller than the threshold value b0 (|b| < b0) as the determination condition J42, and whether the absolute value of the difference c is smaller than the threshold value c0 (|c| < c0) as the determination condition J43. When the differences a, b, and c satisfy the three determination conditions J41, J42, and J43, since the first relative speed, the second relative speed, and the third relative speed are all normal, it is determined as "YES" in S11 and the process proceeds to S12. On the other hand, when the differences a, b, and c do not satisfy at least one of the three determination conditions J41, J42, and J43, it is determined as "NO" in S11 and the process proceeds to S14.

[0059] In S12, the AI estimated value comparison processing unit 30 executes normal processing for all relative speeds. The normal processing for all relative speeds outputs, for example, a signal indicating that all relative speeds are normal. After executing the normal processing for all relative speeds in S12, the process proceeds to S13. In S13, the AI estimated value comparison processing unit 30 selects the first relative speed, which is the AI estimated value, and outputs it to the damping force control unit 31. At this time, the damping force control unit 31 controls the damping force of the variable damper 7 based on the first relative speed.

[0060] When the system determines "NO" in S11, it means that an anomaly has occurred in two or more relative speeds. Therefore, in S14, the AI ​​estimated value comparison processing unit 30 performs unexpected anomaly processing. Unexpected anomaly processing includes various error processing, such as outputting a signal indicating that multiple relative speeds are abnormal and prompting the user to perform maintenance on the IMU 9, wheel speed sensor 11, etc. After performing unexpected anomaly processing in S14, the system proceeds to S15. In S15, the AI ​​estimated value comparison processing unit 30 outputs a signal based on the fail determination to the damping force control unit 31. At this time, the damping force control unit 31 controls the damping force of the variable damper 7, for example, so that the damping force becomes hard. Note that the damping force control unit 31 is not limited to setting the damping force to hard, for example, but may also output a current command equivalent to a conventional passive damper (hereinafter referred to as a conventional equivalent current command).

[0061] Thus, the suspension control device according to this embodiment includes a suspension device 5 installed between the vehicle body 2 and the wheels 3 in the vehicle 1 to suppress vibrations between the vehicle body 2 and the wheels 3; a first relative speed estimation unit 27 that receives signals from a physical sensor (IMU 9) mounted on the vehicle body 2 and estimates the relative speed (first relative speed) between the vehicle body 2 and the wheels 3 based on learning results obtained by machine learning in advance; a second relative speed estimation unit 28 that receives signals from the IMU 9 included in the physical sensor and estimates the relative speed (second relative speed) between the vehicle body 2 and the wheels 3 in a different way from the first relative speed estimation unit 27; and a first phase The system includes a third relative speed estimation unit 29 that estimates the relative speed (third relative speed) between the vehicle body 2 and the wheels 3 in a different manner from the relative speed estimation unit 27 and the second relative speed estimation unit 28; an AI estimated value comparison processing unit 30 (estimated value processing unit) that receives the relative speeds (first relative speed, second relative speed, third relative speed) from the first relative speed estimation unit 27, the second relative speed estimation unit 28, and the third relative speed estimation unit 29, and outputs one relative speed from the relative speeds of the first relative speed estimation unit 27, the second relative speed estimation unit 28, and the third relative speed estimation unit 29; and a damping force control unit 31 (control unit) that receives the relative speed from the AI ​​estimated value comparison processing unit 30 and outputs a control signal for the suspension device 5.

[0062] At this time, the vehicle state quantity calculation unit 26 is an abnormality detection device that detects abnormalities in the relative speed between the vehicle body 2 and the wheels 3 in the vehicle 1. It includes a first relative speed estimation unit 27 that receives signals from the IMU 9 and wheel speed sensor 11, which are physical sensors mounted on the vehicle body 2, and estimates the relative speed (first relative speed) between the vehicle body 2 and the wheels 3, and a second relative speed estimation unit 28 that receives signals from the IMU 9 included in the physical sensor and estimates the relative speed (second relative speed) between the vehicle body 2 and the wheels 3 in a different way from the first relative speed estimation unit 27, and a signal from the wheel speed sensor 11 included in the physical sensor. The system includes a third relative speed estimation unit 29 that receives data and estimates the relative speed between the vehicle body 2 and the wheels 3 (third relative speed) in a different manner from the first relative speed estimation unit 27 and the second relative speed estimation unit 28, and an AI estimated value comparison processing unit 30 (estimated value processing unit) that receives the relative speeds (first relative speed, second relative speed, third relative speed) from the first relative speed estimation unit 27, the second relative speed estimation unit 28, and the third relative speed estimation unit 29, compares the relative speeds from the first relative speed estimation unit 27, the second relative speed estimation unit 28, and the third relative speed estimation unit 29, and outputs a relative speed that is determined to be normal.

[0063] As a result, the AI ​​estimated value comparison processing unit 30 can determine whether the accuracy of three relative speeds (first relative speed, second relative speed, and third relative speed) calculated using different methods is ensured by comparing them. Therefore, if, for example, the estimation error of the first relative speed, which is an AI estimated value, increases, the AI ​​estimated value comparison processing unit 30 can output other relative speeds for which estimation accuracy is ensured. This ensures stable estimated values ​​of relative speeds and suppresses the increase in estimation errors of relative speeds. As a result, the damping force control unit 31 can control the damping force of the suspension device 5 using relative speeds for which estimation accuracy is ensured, thereby ensuring the ride comfort of the vehicle 1.

[0064] In the first embodiment, the physical sensors are an IMU 9 and a wheel speed sensor 11 that detects the wheel speed of the wheel 3. The first relative speed estimation unit 27 and the second relative speed estimation unit 28 receive signals from the IMU 9, and the third relative speed estimation unit 29 receives wheel speed signals from the wheel speed sensor 11.

[0065] Therefore, the first relative velocity estimation unit 27 and the second relative velocity estimation unit 28 can estimate relative velocities (first relative velocity, second relative velocity) with high accuracy based on data acquired from the IMU signal (e.g., sprung mass acceleration). In addition, the third relative velocity estimation unit 29 estimates the relative velocity (third relative velocity) based on the wheel speed acquired from the wheel speed signal. At this time, the third relative velocity estimation unit 29 estimates the relative velocity based on a different physical quantity (wheel speed) than the first relative velocity estimation unit 27 and the second relative velocity estimation unit 28. Therefore, the third relative velocity estimation unit 29 can estimate the relative velocity (third relative velocity) in a different way than the first relative velocity estimation unit 27 and the second relative velocity estimation unit 28.

[0066] The AI ​​estimated value comparison processing unit 30 receives the relative velocities of the first relative velocity estimation unit 27, the second relative velocity estimation unit 28, and the third relative velocity estimation unit 29, and diagnoses whether the relative velocity of the first relative velocity estimation unit 27 is normal or not.

[0067] As a result, the AI ​​estimated value comparison processing unit 30 can diagnose whether the first relative velocity, which is the AI ​​estimated value, is normal or not by comparing three relative velocities (first relative velocity, second relative velocity, and third relative velocity) that were calculated using different methods.

[0068] The AI ​​estimated value comparison processing unit 30 calculates the difference a between the first relative velocity of the first relative velocity estimation unit 27 and the second relative velocity of the second relative velocity estimation unit 28, calculates the difference b between the first relative velocity of the first relative velocity estimation unit 27 and the third relative velocity of the third relative velocity estimation unit 29, calculates the difference c between the second relative velocity of the second relative velocity estimation unit 28 and the third relative velocity of the third relative velocity estimation unit 29, and diagnoses whether the relative velocities from the first relative velocity estimation unit 27, the second relative velocity estimation unit 28, and the third relative velocity estimation unit 29 are normal or not based on the thresholds a0, b0, and c0 set for differences a, b, and c, respectively.

[0069] In this case, when only the estimation error of the first relative velocity increases, the differences a and b become larger, and the difference c becomes smaller. When only the estimation error of the second relative velocity increases, the differences a and c become larger, and the difference b becomes smaller. When only the estimation error of the third relative velocity increases, the differences b and c become larger, and the difference a becomes smaller. Therefore, the AI ​​estimated value comparison processing unit 30 can diagnose whether the first relative velocity, second relative velocity, and third relative velocity are normal or not by comparing the differences a, b, c with the thresholds a0, b0, c0.

[0070] The AI ​​estimated value comparison processing unit 30 outputs the first relative velocity of the first relative velocity estimation unit 27 if it diagnoses the first relative velocity of the first relative velocity estimation unit 27 as normal, and outputs the second relative velocity of the second relative velocity estimation unit 28 if it diagnoses the first relative velocity of the first relative velocity estimation unit 27 as abnormal.

[0071] Therefore, as long as the estimation accuracy of the first relative speed is ensured, the AI ​​estimated value comparison processing unit 30 diagnoses the first relative speed as normal and outputs the first relative speed. At this time, the damping force control unit 31 can use the first relative speed with high estimation accuracy to control the damping force of the suspension device 5 and improve the ride comfort of the vehicle 1.

[0072] On the other hand, if the estimation error of the first relative speed increases, the AI ​​estimated value comparison processing unit 30 diagnoses the first relative speed as abnormal and outputs a relative speed other than the first relative speed, for example, the second relative speed. This suppresses the increase in the estimation error of the relative speed and ensures a stable estimated value of the relative speed. At this time, the damping force control unit 31 controls the damping force of the suspension device 5 using the second relative speed calculated by another method, rather than using the low-precision first relative speed. This prevents the deterioration of ride comfort that occurs when using the low-precision first relative speed and ensures the ride comfort of the vehicle 1.

[0073] In the first embodiment, the AI ​​estimated value comparison processing unit 30 outputs the relative velocity of the second relative velocity estimation unit 28 (second relative velocity) when it diagnoses the first relative velocity of the first relative velocity estimation unit 27 as abnormal, but the present invention is not limited to this. The AI ​​estimated value comparison processing unit 30 may output the relative velocity of the third relative velocity estimation unit 29 (third relative velocity) when it diagnoses the first relative velocity of the first relative velocity estimation unit 27 as abnormal.

[0074] Next, Figures 6 and 7 show a second embodiment. The characteristic of the second embodiment is that the physical sensor includes an imaging device (camera) that images the area in front of the vehicle, and the third relative speed estimation unit receives an image signal (camera signal) from the imaging device. In the second embodiment, the same reference numerals are used for the same components as in the first embodiment described above, and their descriptions are omitted.

[0075] Camera 40 is an imaging device installed at the front of the vehicle body 2 (see Figure 6). Camera 40 and IMU 9 constitute a physical sensor. Camera 40 measures the road surface conditions in front of the vehicle and outputs a camera signal including image data. Camera 40 is composed of a stereo camera, for example, which is becoming common in automated driving support systems. The image data of the area in front of the vehicle captured by camera 40 is output to the controller 21. Note that the imaging device is not limited to camera 40, which consists of a stereo camera, but may also be a combination of, for example, a millimeter-wave radar and a monaural camera.

[0076] As shown in Figures 6 and 7, the vehicle state quantity calculation unit 41 according to the second embodiment is configured similarly to the vehicle state quantity calculation unit 26 according to the first embodiment. The vehicle state quantity calculation unit 41 is an abnormality detection device that detects abnormalities in the relative speed between the vehicle body 2 and the wheels 3 in the vehicle 1. The vehicle state quantity calculation unit 41 calculates (estimates) the sprung speed and relative speed as vehicle state quantities based on data acquired from, for example, the IMU 9, CAN 10, and camera 40. The vehicle state quantity calculation unit 41 includes a first relative speed estimation unit 27, a second relative speed estimation unit 28, a third relative speed estimation unit 42, and an AI estimated value comparison processing unit 30.

[0077] The third relative speed estimation unit 42 receives camera signals from the camera 40 (physical sensor) and estimates the relative speed (third relative speed) between the vehicle body 2 and the wheels 3 in a different manner from the first relative speed estimation unit 27 and the second relative speed estimation unit 28. The third relative speed estimation unit 42 is configured similarly to the vehicle behavior estimation unit disclosed in, for example, Japanese Patent Application Publication No. 2022-191913. The third relative speed estimation unit 42 processes image data from the camera 40 and estimates road surface information based on the image information from the camera 40. Based on the estimated road surface information and information within the vehicle network acquired from the CAN 10 (vehicle network information), the third relative speed estimation unit 42 estimates the vehicle behavior, including the relative speed between the vehicle body 2 and the wheels 3. In image processing, for example, vertical vibration of the vehicle 1 can be detected by comparing images frame by frame.

[0078] The third relative speed estimation unit 42 receives vehicle network information, road surface information, etc., and estimates the stroke speed of each wheel using a neural network that has been trained to estimate the stroke speed of each wheel. The neural network of the third relative speed estimation unit 42 is configured in the same way as the neural network of the first relative speed estimation unit 27. The third relative speed estimation unit 42 reads the weight parameters stored in the weight parameter storage unit 24 via the data reading unit 25. The third relative speed estimation unit 42 estimates the stroke speed of each wheel based on the road surface displacement, etc., and the weight parameters.

[0079] In this embodiment, the third relative speed estimation unit 42 estimates the relative speed of each wheel from road surface information, etc., using a neural network (AI), but the present invention is not limited to this. For example, the third relative speed estimation unit 42 may calculate the relative speed of each wheel from road surface information, etc., using a mathematical formula based on a vehicle model, a Kalman filter, etc.

[0080] Thus, the second embodiment can achieve substantially the same effects as the first embodiment. In the second embodiment, the physical sensors are an IMU 9 and a camera 40 (imaging device) that images the area in front of the vehicle 1. The first relative speed estimation unit 27 and the second relative speed estimation unit 28 receive signals from the IMU 9, and the third relative speed estimation unit 42 receives image signals from the camera 40.

[0081] Therefore, the first relative velocity estimation unit 27 and the second relative velocity estimation unit 28 can estimate relative velocities (first relative velocity, second relative velocity) with high accuracy based on data acquired from the IMU signal (e.g., sprung mass acceleration). In addition, the third relative velocity estimation unit 42 estimates relative velocity (third relative velocity) based on road surface displacement acquired from the image signal (camera signal). At this time, the third relative velocity estimation unit 42 estimates relative velocity based on a different signal (camera signal) than the first relative velocity estimation unit 27 and the second relative velocity estimation unit 28. Therefore, the third relative velocity estimation unit 42 can estimate relative velocity (third relative velocity) in a different way than the first relative velocity estimation unit 27 and the second relative velocity estimation unit 28.

[0082] Next, Figures 8 to 10 show a third embodiment. The characteristic of the third embodiment is that the suspension control device has an estimated value processing unit that receives the relative speeds of a first relative speed estimation unit and a second relative speed estimation unit, diagnoses whether the relative speed of the first relative speed estimation unit is normal or not, outputs the relative speed of the first relative speed estimation unit if the relative speed of the first relative speed estimation unit is normal, and outputs an abnormality detection signal indicating that the relative speed is abnormal if the relative speed of the first relative speed estimation unit is not normal, and a control unit that outputs a control signal for the suspension device when it receives a relative speed from the estimated value processing unit, and outputs a control stop signal to stop the control of the suspension device when it receives an abnormality detection signal from the estimated value processing unit. In the third embodiment, the same reference numerals are used for the same components as in the first embodiment described above, and their descriptions are omitted.

[0083] As shown in Figure 8, the vehicle state quantity calculation unit 51 according to the third embodiment is configured similarly to the vehicle state quantity calculation unit 26 according to the first embodiment. The vehicle state quantity calculation unit 51 is an abnormality detection device that detects abnormalities in the relative speed between the vehicle body 2 and the wheels 3 in the vehicle 1. The vehicle state quantity calculation unit 51 calculates (estimates) the sprung speed and relative speed as vehicle state quantities based on data acquired from, for example, the IMU 9 and CAN 10. The vehicle state quantity calculation unit 51 includes a first relative speed estimation unit 27, a second relative speed estimation unit 52, and an AI estimated value abnormality determination unit 53.

[0084] The second relative speed estimation unit 52 according to the third embodiment is configured similarly to the third relative speed estimation unit 29 according to the first embodiment. The second relative speed estimation unit 52 receives a signal from the wheel speed sensor 11 (physical sensor) and estimates the relative speed (second relative speed) between the vehicle body 2 and the wheels 3 in a different manner than the first relative speed estimation unit 27. As shown in Figure 9, the second relative speed estimation unit 52 estimates the second relative speed based on the wheel speed. The wheel speed signal of each wheel is input to the second relative speed estimation unit 52. The second relative speed estimation unit 52 calculates (estimates) the relative speed (second relative speed) between the vehicle body 2 and the wheels 3 from the detected value of the wheel speed sensor 11, which acts as a wheel speed detection means for detecting the rotational speed of the wheels 3.

[0085] The AI ​​Estimated Value Anomaly Determination Unit 53 is an estimated value processing unit. The AI ​​Estimated Value Anomaly Determination Unit 53 receives the relative velocities (first relative velocity, second relative velocity) of the first relative velocity estimation unit 27 and the second relative velocity estimation unit 52, and diagnoses whether the first relative velocity of the first relative velocity estimation unit 27 is normal or not. If the first relative velocity of the first relative velocity estimation unit 27 is normal, the AI ​​Estimated Value Anomaly Determination Unit 53 outputs the first relative velocity of the first relative velocity estimation unit 27. If the relative velocity of the first relative velocity estimation unit is not normal, the AI ​​Estimated Value Anomaly Determination Unit 53 outputs an anomaly detection signal indicating that the first relative velocity is abnormal.

[0086] The AI ​​estimated value anomaly determination unit 53 calculates the difference Δ between the first relative velocity and the second relative velocity, and diagnoses whether the first relative velocity is normal or not based on the difference Δ and the threshold Δ0. At this time, the threshold Δ0 is a determination threshold, which is a positive value (Δ0>0) within the range in which the first relative velocity can be judged as normal, and is set appropriately considering the actual detected value, calculated value, etc.

[0087] The AI ​​estimated value anomaly detection unit 53 determines that the first relative velocity is normal when the absolute value of the difference Δ is less than the threshold Δ0 (|Δ|<Δ0). The AI ​​estimated value anomaly detection unit 53 determines that the first relative velocity is abnormal when the absolute value of the difference Δ is greater than or equal to the threshold Δ0 (|Δ|≧Δ0). If the first relative velocity is normal, the AI ​​estimated value anomaly detection unit 53 outputs the first relative velocity. If the first relative velocity is abnormal, the AI ​​estimated value anomaly detection unit 53 outputs an anomaly detection signal indicating that the first relative velocity is abnormal.

[0088] The damping force control unit 54 is a control unit that outputs a control signal for the suspension device 5 when it receives a first relative velocity from the AI ​​estimated value anomaly determination unit 53, and outputs a control stop signal to stop the control of the suspension device 5 when it receives an anomaly detection signal from the AI ​​estimated value anomaly determination unit 53. Similar to the damping force control unit 31 in the first embodiment, the damping force control unit 54 determines the control amount for controlling the variable damper 7.

[0089] When the first relative speed is normal, the damping force control unit 54 determines a control amount to control the variable damper 7, for example, from the sprung speed and the relative speed. Specifically, the damping force control unit 54 calculates a control command value (command current value) for controlling the damping force of the variable damper 7 of the suspension device 5, based on the sprung speed calculated (estimated) by the first relative speed estimation unit 27 and the first relative speed output from the AI ​​estimated value abnormality determination unit 53. At this time, the control amount for controlling the variable damper 7 is the control command value for controlling the damping force of the variable damper 7. Specifically, the damping force control unit 54 calculates a command current value as a control command value to improve the ride comfort of the vehicle, for example, based on bilinear optimal control, skyhook control, H∞ control, etc. Based on the command current value, the damping force control unit 54 outputs a command current as a control signal to the variable damping force actuator 8 of the variable damper 7. Note that the damping force control unit 54 may also calculate control command values ​​to improve handling stability, not just the ride comfort of the vehicle.

[0090] In response to this, if the first relative speed is abnormal, the damping force control unit 54 receives an abnormality detection signal from the AI ​​estimated value abnormality determination unit 53. In this case, the damping force control unit 54 outputs a control stop signal to stop the control of the suspension device 5. When the damping force control unit 54 outputs a control stop signal, the damper control unit 23 outputs a current command (hereinafter referred to as a conventional damper equivalent current command) to the damping force variable actuator 8 that corresponds to a conventional passive damper. However, when the damping force control unit 54 outputs a control stop signal, the damper control unit 23 may also output a current command to the damping force variable actuator 8 that generates a constant damping force on the hardware side, for example.

[0091] Next, with reference to Figure 10, the AI ​​estimated value anomaly determination process by the AI ​​estimated value anomaly determination unit 53 will be explained.

[0092] In S21, the AI ​​estimated value anomaly determination unit 53 calculates the difference Δ between the first relative velocity and the second relative velocity. In S22, the AI ​​estimated value anomaly determination unit 53 determines whether the first relative velocity is abnormal or not. Specifically, the AI ​​estimated value anomaly determination unit 53 determines whether the absolute value of the difference Δ is greater than or equal to the threshold Δ0 (|Δ|≧Δ0). If the absolute value of the difference Δ is greater than or equal to the threshold Δ0, the first relative velocity is abnormal, so in S22 it determines "YES" and proceeds to S23. On the other hand, if the absolute value of the difference Δ is less than the threshold Δ0, the first relative velocity is normal, so in S22 it determines "YES" and proceeds to S24.

[0093] In S23, the AI ​​estimated value anomaly determination unit 53 determines that the first relative speed is abnormal and outputs an anomaly detection signal indicating that the first relative speed is abnormal. At this time, the damping force control unit 54 outputs a control stop signal to stop the control of the suspension device 5, and the damper control unit 23 outputs a conveyor-equivalent current command to the damping force variable actuator 8, causing the variable damper 7 to operate as a conventional passive damper.

[0094] In S24, the AI ​​estimated value anomaly determination unit 53 determines that the first relative velocity is normal and outputs the first relative velocity, which is the AI ​​estimated value, to the damping force control unit 54. At this time, the damping force control unit 54 controls the damping force of the variable damper 7 based on the first relative velocity.

[0095] Thus, the third embodiment can also obtain substantially the same effects as the first embodiment. Furthermore, the suspension control device according to the third embodiment includes a suspension device 5 installed between the vehicle body 2 and the wheels 3 in the vehicle 1 to suppress vibrations between the vehicle body 2 and the wheels 3, a first relative speed estimation unit 27 that receives signals from a physical sensor (IMU 9) mounted on the vehicle body 2 and estimates the relative speed (first relative speed) between the vehicle body 2 and the wheels 3 based on learning results obtained by machine learning in advance, a second relative speed estimation unit 52 that receives signals from a wheel speed sensor 11 included in the physical sensor and estimates the relative speed (second relative speed) between the vehicle body 2 and the wheels 3 in a different manner from the first relative speed estimation unit 27, and the relative speeds of the first relative speed estimation unit 27 and the second relative speed estimation unit 52 respectively The system includes an AI estimated value anomaly determination unit 53 (estimated value processing unit) that receives speed (first relative speed, second relative speed), diagnoses whether the first relative speed of the first relative speed estimation unit 27 is normal or not, outputs the first relative speed of the first relative speed estimation unit 27 if the first relative speed of the first relative speed estimation unit 27 is normal, and outputs an anomaly detection signal indicating that the first relative speed is abnormal if the first relative speed of the first relative speed estimation unit 27 is not normal, and a damping force control unit 54 (control unit) that outputs a control signal for the suspension device 5 when it receives the first relative speed from the AI ​​estimated value anomaly determination unit 53, and outputs a control stop signal to stop the control of the suspension device 5 when it receives an anomaly detection signal from the AI ​​estimated value anomaly determination unit 53.

[0096] Therefore, in the third embodiment, the third relative velocity estimation unit can be omitted compared to the first embodiment, and the configuration of the controller 21 can be simplified. In addition, in the third embodiment, the soundness of the first relative velocity, which is an AI estimated value, can be determined using the first relative velocity and the second relative velocity, and the damping force of the variable damper 7 can be controlled using the first relative velocity for which estimation accuracy has been ensured.

[0097] In the third embodiment, the second relative speed estimation unit 52 estimates the second relative speed based on the wheel speed, but the present invention is not limited thereto. The second relative speed estimation unit 52 may, for example, estimate the relative speed (second relative speed) between the vehicle body 2 (sprung mass) and the wheel 3 (unsprung mass) via a Kalman filter from the sprung mass acceleration signal included in the IMU signal. The second relative speed estimation unit 52 may also estimate the relative speed (second relative speed) between the vehicle body 2 and the wheel 3 by receiving a camera signal from the camera 40 (physical sensor), similar to the third relative speed estimation unit 42 in the second embodiment.

[0098] In the embodiments described above, the neural network of the first relative velocity estimation unit 27 calculates the stroke velocity (relative velocity) of the suspension device 5 of each wheel by taking a signal from the IMU 9 (IMU signal) as input, but the present invention is not limited thereto. The neural network of the first relative velocity estimation unit 27 may, for example, calculate the stroke velocity of the suspension device 5 of each wheel based on the sprung mass acceleration and the variable damping force of each wheel, similar to the second relative velocity estimation unit 28.

[0099] In the embodiments described above, the controller 21 acquires data (wheel speed data) from the wheel speed sensor 11 via CAN 10, but the present invention is not limited thereto. For example, the controller may acquire wheel speed data directly from the wheel speed sensor 11. That is, the controller 21 may, for example, acquire detection values ​​from various sensors directly from those sensors. Furthermore, the controller 21 may acquire behavioral information, including wheel speed, from other controllers or the like.

[0100] In the embodiments described above, the IMU 9 is installed outside the controller 21 and is received by the controller 21 for control purposes. The present invention is not limited to this, and the IMU may be mounted inside the controller.

[0101] In the embodiments described above, the case in which a variable damper 7 consisting of a semi-active damper is used as the force generating mechanism was explained as an example. 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 embodiments, the case in which a force generating mechanism that generates an adjustable force between the vehicle body 2 and the wheel 3 is configured with a variable damper 7 consisting of a damping force adjustable hydraulic shock absorber was explained as an example. The present invention is not limited to this, and for example, the force generating mechanism may be configured with an air suspension, stabilizer (Kinesus), electromagnetic suspension, etc., in addition to a hydraulic shock absorber.

[0102] In the embodiments described above, a control device for a suspension system used in a four-wheeled automobile was used as an example. However, the present invention is not limited to this, and may be applied to two-wheeled vehicles, three-wheeled vehicles, and also to work vehicles, transport vehicles such as trucks and buses. [Explanation of symbols]

[0103] 1: Vehicle, 2: Body, 3: Wheels, 5: Suspension system, 7: Variable damper, 8: Variable damping actuator, 9: Inertial Measurement Unit (IMU), 11: Wheel speed sensor, 21: Controller, 23: Damper control unit, 26, 41, 51: Vehicle state quantity calculation unit (anomaly detection device), 27: First relative speed estimation unit, 28, 52: Second relative speed estimation unit, 28A: Kalman filter, 28B: Variable damping component calculation unit, 29, 42: Third relative speed estimation unit, 30: AI estimated value comparison processing unit (estimated value processing unit), 31, 54: Damping force control unit (control unit), 40: Camera (imaging device), 53: AI estimated value anomaly determination unit (estimated value processing unit)

Claims

1. A suspension device provided between the vehicle body and the wheels to suppress vibrations between the vehicle body and the wheels, A first relative speed estimation unit receives signals from physical sensors mounted on the vehicle body and estimates the relative speed between the vehicle body and the wheels based on learning results obtained by machine learning in advance. A second relative speed estimation unit receives signals from the physical sensor and estimates the relative speed between the vehicle body and the wheels in a manner different from that of the first relative speed estimation unit, A third relative speed estimation unit receives signals from the physical sensors and estimates the relative speed between the vehicle body and the wheels in a manner different from that of the first relative speed estimation unit and the second relative speed estimation unit, An estimation value processing unit receives the relative speeds from the first relative speed estimation unit, the second relative speed estimation unit, and the third relative speed estimation unit, and outputs a single relative speed from the relative speeds from the first relative speed estimation unit, the second relative speed estimation unit, and the third relative speed estimation unit, A control unit that receives the relative speed from the estimated value processing unit and outputs a control signal for the suspension device, A suspension control device having

2. The suspension control device according to claim 1, The physical sensors include an inertial measurement unit and a wheel speed sensor that detects the wheel speed of the wheel. The first relative velocity estimation unit and the second relative velocity estimation unit receive signals from the inertial measurement unit, The third relative speed estimation unit receives the wheel speed signal from the wheel speed sensor. Suspension control device.

3. The suspension control device according to claim 1, The physical sensor comprises an inertial measurement unit and an imaging device that captures images of the area in front of the vehicle. The first relative velocity estimation unit and the second relative velocity estimation unit receive signals from the inertial measurement unit, The third relative velocity estimation unit receives an image signal from the imaging device. Suspension control device.

4. The suspension control device according to claim 1, The estimation processing unit receives the relative speeds from the first relative speed estimation unit, the second relative speed estimation unit, and the third relative speed estimation unit, and diagnoses whether the relative speed of the first relative speed estimation unit is normal or not. Suspension control device.

5. The suspension control device according to claim 4, The aforementioned estimated value processing unit is: The difference a between the relative velocity of the first relative velocity estimation unit and the relative velocity of the second relative velocity estimation unit is calculated. The difference b between the relative velocity of the first relative velocity estimation unit and the relative velocity of the third relative velocity estimation unit is calculated. The difference c between the relative velocity of the second relative velocity estimation unit and the relative velocity of the third relative velocity estimation unit is calculated. The threshold values ​​set for each of the differences a, b, and c are used to diagnose whether the relative velocity from the first relative velocity estimation unit, the second relative velocity estimation unit, and the third relative velocity estimation unit is normal or not. Suspension control device.

6. The suspension control device according to claim 4, The aforementioned estimated value processing unit is: If the relative velocity of the first relative velocity estimation unit is diagnosed as normal, the relative velocity of the first relative velocity estimation unit is output. If the first relative velocity estimation unit diagnoses the relative velocity as abnormal, the second relative velocity estimation unit or the third relative velocity estimation unit outputs the relative velocity. Suspension control device.

7. This is an anomaly detection device that detects abnormalities in the relative speed between the vehicle body and the wheels. A first relative speed estimation unit receives signals from a physical sensor mounted on the vehicle body and estimates the relative speed between the vehicle body and the wheels, A second relative speed estimation unit receives signals from the physical sensor and estimates the relative speed between the vehicle body and the wheels in a manner different from that of the first relative speed estimation unit, A third relative speed estimation unit receives signals from the physical sensors and estimates the relative speed between the vehicle body and the wheels in a manner different from that of the first relative speed estimation unit and the second relative speed estimation unit, An estimation value processing unit receives the relative speeds from the first relative speed estimation unit, the second relative speed estimation unit, and the third relative speed estimation unit, compares the relative speeds from the first relative speed estimation unit, the second relative speed estimation unit, and the third relative speed estimation unit, and outputs a relative speed that is determined to be normal. An anomaly detection device having the following features.

8. The abnormality detection device according to claim 7, The physical sensors include an inertial measurement unit and a wheel speed sensor that detects the wheel speed of the wheel. The first relative velocity estimation unit and the second relative velocity estimation unit receive signals from the inertial measurement unit, The third relative speed estimation unit receives the wheel speed signal from the wheel speed sensor. Anomaly detection device.

9. The abnormality detection device according to claim 7, The physical sensor comprises an inertial measurement unit and an imaging device that captures images of the area in front of the vehicle. The first relative velocity estimation unit and the second relative velocity estimation unit receive signals from the inertial measurement unit, The third relative velocity estimation unit receives an image signal from the imaging device. Anomaly detection device.

10. The abnormality detection device according to claim 7, The aforementioned estimated value processing unit is: The difference a between the relative velocity of the first relative velocity estimation unit and the relative velocity of the second relative velocity estimation unit is calculated. The difference b between the relative velocity of the first relative velocity estimation unit and the relative velocity of the third relative velocity estimation unit is calculated. The difference c between the relative velocity of the second relative velocity estimation unit and the relative velocity of the third relative velocity estimation unit is calculated. The threshold values ​​set for each of the differences a, b, and c are used to diagnose whether the relative velocity from the first relative velocity estimation unit, the second relative velocity estimation unit, and the third relative velocity estimation unit is normal or not. Anomaly detection device.

11. The abnormality detection device according to claim 7, The aforementioned estimated value processing unit is: If the relative velocity of the first relative velocity estimation unit is diagnosed as normal, the relative velocity of the first relative velocity estimation unit is output. If the first relative velocity estimation unit diagnoses the relative velocity as abnormal, the second relative velocity estimation unit or the third relative velocity estimation unit outputs the relative velocity. Anomaly detection device.

12. A suspension device provided between the vehicle body and the wheels to suppress vibrations between the vehicle body and the wheels, A first relative speed estimation unit receives signals from physical sensors mounted on the vehicle body and estimates the relative speed between the vehicle body and the wheels based on learning results obtained by machine learning in advance. A second relative speed estimation unit receives signals from the physical sensor and estimates the relative speed between the vehicle body and the wheels in a manner different from that of the first relative speed estimation unit, An estimation value processing unit receives the relative speeds of the first relative speed estimation unit and the second relative speed estimation unit, diagnoses whether the relative speed of the first relative speed estimation unit is normal or not, outputs the relative speed of the first relative speed estimation unit if it is normal, and outputs an abnormality detection signal indicating that the relative speed is abnormal if it is not normal. A control unit that, upon receiving the relative speed from the estimated value processing unit, outputs a control signal for the suspension device, and upon receiving the abnormality detection signal from the estimated value processing unit, outputs a control stop signal to stop the control of the suspension device, A suspension control device having

Images