Vehicle control device and steering angle calculation method

Abstract

A geometric calculation unit (1) calculates a traveling direction angle of a vehicle from a vehicle speed V and a yaw angular velocity of a gravity direction axis calculated from a yaw angular velocity ωy measured by an IMU (300). In addition, a steering angle δH is calculated by using the traveling direction angle and a roll angle φ. In addition, the geometric calculation unit (1) preferably calculates the traveling direction angle on the basis of the height of the IMU (300) from the road surface. According to this configuration, the steering angle δh can be calculated without using a steering angle sensor, and the configuration of the control system can be simplified.

Description

Vehicle control device and steering angle calculation method

[0001] The present invention relates to a vehicle control device and a steering angle calculation method.

[0002] In recent years, technological developments relating to attitude control of vehicles such as motorcycles have progressed, and related inventions have been published. For example, Patent Document 1 discloses a steering control device that determines the steering angle of a four-wheeled vehicle traveling along a target route from yaw acceleration, slip angle, and vehicle speed.

[0003] Japanese Patent Application Laid-Open No. 2022-112301

[0004] However, in the case of a two-wheeled vehicle, the steering angle is affected by the roll angle when turning, so there is still room for improvement in the calculation of the steering angle. In view of this situation, an object of the present invention is to improve the performance of the calculation of the steering angle of a vehicle.

[0005] The present invention, which solves the above-mentioned problems, is a vehicle control device characterized by including a geometric calculation unit that calculates a vehicle's traveling direction angle from the yaw angular velocity of the gravity direction axis and the wheel speed, and calculates a steering angle using the traveling direction angle and roll angle.

[0006] According to the present invention, the performance of calculating the steering angle of a vehicle can be improved.

[0007] It is a functional configuration diagram of a vehicle control device in this embodiment. It is an explanatory diagram regarding calculation of a static roll angle. It is an explanatory diagram regarding change in static roll angle due to riding posture. It is an explanatory diagram regarding calculation of a dynamic roll angle. It is a graph showing change in vertical acceleration when performing a wheelie. It is an explanatory diagram of a riding posture map.

[0008] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. Each figure is merely a schematic illustration to allow a sufficient understanding of the present invention. Therefore, the present invention is not limited to the illustrated examples. In each figure, common or similar components are denoted by the same reference numerals, and redundant explanations thereof will be omitted. In this embodiment, a motorcycle will be used as an example of a vehicle. For convenience of explanation, the traveling direction of a motorcycle is defined as "forward," and "rearward," "left and right" (lateral), "upward," and "downward" (direction of gravity) will be defined.

[0009] [Configuration] FIG. 1 is a functional configuration diagram of a vehicle control device according to this embodiment. The vehicle control device 100 is mounted on a motorcycle (a bar-handle vehicle). The vehicle control device 100 is a computer that performs calculations related to the attitude control of the motorcycle. In particular, the vehicle control device 100 functions as a roll angle calculation device that calculates the roll angle of a motorcycle traveling in a turn. The vehicle control device 100 includes hardware such as an input unit, an output unit, a control unit, and a storage unit. For example, if the control unit is configured with a CPU (Central Processing Unit), information processing by a computer including the control unit is realized by program execution processing by the CPU. Furthermore, a storage unit included in the computer stores various programs for implementing the functions of the computer in response to commands from the CPU. This allows collaboration between software and hardware.

[0010] The vehicle control device 100 is communicably connected to a vehicle speed sensor 200 and an IMU 300 (Inertial Measurement Unit). The vehicle speed sensor 200 detects the vehicle speed V of the motorcycle. For example, the vehicle speed sensor 200 is attached to a wheel of the motorcycle and can detect the vehicle speed V based on the wheel speed obtained by counting the number of rotations per unit time of the wheel. The vehicle speed sensor is generally provided on either the front or rear wheel of the wheels, or on either the front or rear wheel, but in this embodiment, the description will be given assuming that the vehicle speed sensor is provided on at least one wheel. The vehicle speed V is input to the vehicle control device 100 from the vehicle speed sensor 200. The IMU 300 is, for example, a six-axis IMU, and detects the longitudinal acceleration A of the motorcycle.ｘ , lateral acceleration A ｙ , vertical acceleration A ｚ , roll angular velocity ω ｒ , pitch angular velocity ω ｐ , and yaw angular velocity ω ｙ The vehicle control device 100 detects (measures) the longitudinal acceleration A ｘ , lateral acceleration A ｙ , vertical acceleration A ｚ , roll angular velocity ω ｒ , pitch angular velocity ω ｐ , and yaw angular velocity ω ｙ The longitudinal acceleration A input from the IMU 300 to the vehicle control device 100 is ｘ , lateral acceleration A ｙ , vertical acceleration A ｚ , roll angular velocity ω ｒ , pitch angular velocity ω ｐ , and yaw angular velocity ω ｙ is a value obtained by converting the output of the sensor axis output from the IMU 300 into the vehicle axis. For example, when the vehicle is tilted, the vertical acceleration A ｚ The direction of is not the direction of gravity but a direction inclined relative to the direction of gravity.

[0011] The vehicle control device 100 includes a geometric calculation unit 1, a static roll angle calculation unit 2 (first roll angle calculation unit), a dynamic roll angle calculation unit 3 (second roll angle calculation unit), a riding attitude correction value calculation unit 4 (attitude angle calculation unit), an error estimation unit 5 (roll angle error estimation unit), and an output processing unit 6. The geometric calculation unit 1 outputs values ​​used in calculations for vehicle control based on geometric information representing the shape and attitude of the motorcycle. The static roll angle calculation unit 2 calculates the static roll angle φ of the motorcycle. ｓ Calculate the static roll angle φ ｓ is a roll angle that does not depend on time. The dynamic roll angle calculation unit 3 calculates the dynamic roll angle φ of the motorcycle. ｄ Calculate the dynamic roll angle φ ｄ is the time-dependent roll angle. ｄ is the roll angular velocity ω ｒ On the other hand, the static roll angle φ ｓ is the roll angular velocity ω ｒcan be obtained by calculation without integrating with time. The riding posture correction value calculation unit 4 calculates a correction value due to the riding posture of the motorcycle as a riding posture correction value. The riding posture correction value is calculated based on the static roll angle φ ｓ The riding posture correction value calculation unit 4 can also calculate the riding posture of the motorcycle. The error estimating unit 5 can use the dynamic roll angle φ calculated by the dynamic roll angle calculation unit 3 to correct the ｄ The output processing unit 6 estimates the dynamic roll angle φ from which the influence of the error has been removed. ｄ is output as the roll angle φ of the motorcycle.

[0012] [Estimation of Dynamic Roll Angle Error] As shown in Fig. 1, the error estimation unit 5 includes a filter reset determination unit 51 (special driving determination unit, driving determination unit) and a complementary filter calculation unit 52 (roll angle error estimation unit). The filter reset determination unit 51 determines whether or not to perform calculations by the complementary filter calculation unit 52 depending on the driving state of the motorcycle. The complementary filter calculation unit 52 calculates the static roll angle φ ｓ and dynamic roll angle φ ｄ The calculation error included in is estimated by a Kalman filter. The Kalman filter is a calculation method for estimating an internal state from observed values ​​(including errors and noise) acquired by sensors (in this embodiment, the vehicle speed sensor 200 and the IMU 300). For example, the filter reset determination unit 51 can function as a calculation unit that calculates values ​​to be input to the complementary filter calculation unit 52, but is not limited to this. In this embodiment, the description will continue assuming that the filter reset determination unit 51 is such a calculation unit.

[0013] The static roll angle φ calculated by the static roll angle calculation unit 2 ｓ and the dynamic roll angle φ calculated by the dynamic roll angle calculation unit 3 ｄ is input to the filter reset determination unit 51. The filter reset determination unit 51 determines the static roll angle φ ｓ and dynamic roll angle φ ｄThe difference between the static roll angle φ and the static roll angle φ is calculated and output as a roll angle difference. The roll angle difference is input to a complementary filter calculation unit 52. The complementary filter calculation unit 52 can estimate the calculation error included in the roll angle difference using a Kalman filter. The calculation error included in the roll angle difference is calculated by subtracting the static roll angle φ from the static roll angle φ. ｓ and the calculation error and dynamic roll angle φ ｄ The complementary filter calculation unit 52 calculates the static roll angle φ from the calculation errors included in the roll angle difference. ｓ and the calculation error and dynamic roll angle φ ｄ In this embodiment, the complementary filter calculation unit 52 estimates the calculation errors included in the static roll angle φ ｓ and dynamic roll angle φ ｄ Based on the roll angle difference, the dynamic roll angle φ ｄ and outputs the calculated error contained in the calculated value to the dynamic roll angle calculation unit 3 as the dynamic roll angle error.

[0014] The dynamic roll angle calculation unit 3 calculates the dynamic roll angle φ using the dynamic roll angle error output from the complementary filter calculation unit 52 of the error estimation unit 5. ｄ The dynamic roll angle φ can be calculated by subtracting the dynamic roll angle error. ｄ The obtained dynamic roll angle φ ｄ is input to the output processing unit 6. The output processing unit 6 outputs the dynamic roll angle φ ｄ can be output as the roll angle φ of the motorcycle. Also, the obtained dynamic roll angle φ ｄ is input to the error estimation unit 5 and can be used for estimation by the complementary filter calculation unit 52. As a result, the dynamic roll angle φ ｄ For example, the complementary filter calculation unit 52 can output to the dynamic roll angle calculation unit 3 a dynamic roll angle error whose estimation accuracy has been improved to a certain level or more.

[0015] When estimating the sensor error of the IMU using a Kalman filter to calculate the roll angle of a motorcycle, conventionally, it was necessary to perform arithmetic operations using a 5×5 matrix with five types of observation values ​​(roll acceleration, roll angular velocity, yaw acceleration, yaw angular velocity, and longitudinal acceleration) output by the IMU as input, which posed a problem of a very large calculation load. According to this embodiment, the input values ​​used for estimation by the Kalman filter are the static roll angle φ ｓ and dynamic roll angle φ ｄ Therefore, the calculations performed in the sensor error estimation of the IMU 300 are arithmetic operations using a 2 × 2 matrix, and the calculation load due to matrix calculations can be minimized. As a result, the calculation load for calculating the roll angle of the motorcycle can be reduced.

[0016] It is preferable to appropriately adjust the calculations and controls performed by the vehicle control device 100 in order to maintain a certain level of accuracy in the finally calculated roll angle φ of the motorcycle while reducing the amount of calculations involved in error estimation using the Kalman filter. Methods of such adjustments are described below.

[0017] [Static roll angle φ ｓ Calculation details of static roll angle φ ｓ can be calculated using the following formula (1).

[0018] ...Formula (1)

[0019] where R is the turning radius of the motorcycle during a turn, and g is the gravitational acceleration. φ is a geometric bank angle conversion coefficient, which is a constant that can be uniquely determined from the shape and posture of the motorcycle. ２ / R) is the centripetal acceleration, directed towards the turning center. g is the direction of gravity. In other words, according to equation (1), the static roll angle φ ｓ is the direction of the turning center (V ２ The vehicle speed V can be obtained by using the arctangent of the vehicle speed (V / R) and the direction of gravity g. φ is known, so the static roll angle φ ｓTo calculate the turning radius R, it is necessary to calculate the static roll angle. Fig. 2 is an explanatory diagram regarding the calculation of the static roll angle. In order to calculate the turning radius R (radius of a virtual steady circle), the yaw angular velocity ω of the vehicle axis detected by the IMU 300 (not shown in Fig. 2) mounted on the turning motorcycle M is calculated. ｙ The yaw angular velocity of the gravity axis should be used instead of the yaw angular velocity of the horizontal axis (the yaw angular velocity of the horizontal axis does not contribute to turning). ｙ The yaw angular velocity of the gravity direction axis can be calculated by performing coordinate transformation so that the bank angle of the motorcycle M is rotated by the bank angle of the motorcycle M. The bank angle is the roll angle φ, or more accurately, the previous value of the static roll angle. The calculated yaw angular velocity of the gravity direction axis is input to the static roll angle calculation unit 2. Note that the previous value of the roll angle φ (the roll angle φ last output by the output processing unit 6) can also be used.

[0020] The static roll angle calculation unit 2 calculates the static roll angle φ using equation (1). ｓ At this time, the pitch angular velocity ω measured by the IMU 300 can be calculated. ｐ It is advisable to use the yaw angular velocity of the gravity direction axis calculated using

[0021] Also, the yaw angular velocity ω ｙ changes depending on the riding position. Riding position refers to, for example, whether the rider is leaning in or leaning out. Riding position refers to the position of the moving vehicle, taking into account not only the rider riding on the motorcycle, but also the load, including luggage carried on the motorcycle. FIG. 3 is an explanatory diagram showing the change in static roll angle depending on riding position. If the motorcycle is lean-with, the static roll angle of the motorcycle M indicates the true roll angle φ2 calculated based on the center of gravity of the vehicle. If the motorcycle is lean-in as shown in FIG. 3, the static roll angle of the motorcycle M increases and indicates the static roll angle φ1 calculated based on the center of gravity generated by the riding position. In other words, even if the inclination (bank angle) of the motorcycle M is the same, the turning radius when leaning in is smaller than when leaning with. Therefore, when the motorcycle is leaning in, the yaw angular velocity ω ｙ As a result, the static roll angle increases when leaning in.

[0022] In view of the above, if the influence of riding posture can be estimated, the yaw angular velocity ω ｙ Since the change in the static roll angle φ can be calculated, ｓ The influence of riding posture can be corrected by the lateral acceleration A measured by the IMU 300. ｙ Change in and vertical acceleration A ｚ As shown in FIG. 1 , the riding posture correction value calculation unit 4 includes a first posture angle estimation unit 41, a second posture angle acquisition unit 42, a posture angle difference calculation unit 43, and a riding posture map 44.

[0023] The first posture angle estimator 41 estimates a first posture angle φ1 based on the center of gravity determined by the riding posture. The first posture angle φ1 is calculated based on the lateral acceleration A acting in the lateral direction of the motorcycle. ｙ , and the vertical acceleration A acting in the direction of gravity of the motorcycle ｚ The second attitude angle acquisition unit 42 acquires the second attitude angle φ2, which is determined based on the vehicle center of gravity. The attitude angle difference calculation unit 43 calculates the difference α (see FIG. 3) between the first attitude angle φ1 and the second attitude angle φ2. The riding attitude map 44 is a map used to acquire the second attitude angle φ2. The riding attitude map 44 calculates the roll angle φ acting on the center of gravity of the two-wheeled vehicle M, the lateral acceleration A acting in the lateral direction of the two-wheeled vehicle M, and the like. ｙ , and the vertical acceleration A acting in the direction of gravity on the motorcycle M ｚ This map was created based on the above (details below).

[0024] The riding attitude correction value calculation unit 4 outputs the difference α as a riding attitude correction value to the static roll angle calculation unit 2. The static roll angle calculation unit 2 calculates the static roll angle φ ｓ The static roll angle φ (provisional first roll angle) can be corrected by the riding posture correction value. ｓ (the determined first roll angle) is output to the error estimation unit 5. The riding posture correction value (difference α) is a value that indicates the influence of the riding posture itself, and the yaw angular velocity ω ｙThe static roll angle calculation unit 2 can correct the yaw angular velocity of the gravity direction axis (see FIG. 2) input from the geometric calculation unit 1 with the riding attitude correction value. As a result, the static roll angle calculation unit 2 calculates the static roll angle φ using equation (1). ｓ The static roll angle calculation unit 2 and the riding attitude correction value calculation unit 4 may be combined into a functional unit called the static roll angle calculation unit 2 (first roll angle calculation unit).

[0025] [Dynamic roll angle φ ｄ Calculation details] <Travel direction correction> Dynamic roll angle φ ｄ is the roll angular velocity ω detected by the IMU 300 ｒ The integral interval is, for example, the period from the time when the previous value of the roll angle φ is output to the time when the current value of the roll angle φ is output. When a motorcycle turns, the banking direction changes. In other words, the roll angular velocity ω ｒ The axis of the dynamic roll angle φ ｄ The roll angular velocity ω is integrated in the calculation of ｒ is the roll angular velocity of the traveling axis. Therefore, by calculating the roll angular velocity of the traveling axis, the dynamic roll angle φ ｄ This can improve the calculation accuracy.

[0026] FIG. 4 is an explanatory diagram relating to the calculation of the dynamic roll angle. FIG. 4 is a plan view showing geometric information relating to a motorcycle (only the front wheel Tf and the rear wheel Tr are shown) that has been steered without banking. As shown in FIG. 4, a heading angle D occurs with respect to the motorcycle when turning. The heading angle D is the angle formed between the heading axis of the motorcycle M when turning and a line passing through the center of the front wheel Tf and the center of the rear wheel Tr, with the position of the IMU 300 (not shown in FIG. 4) as the center of rotation. In FIG. 4, R is the turning radius, C is the distance between the position of the IMU 300 (not shown in FIG. 4) and the center of the rear wheel Tr, WB is the distance between the center of the front wheel Tf and the center of the rear wheel Tr, and δ is the turning radius. ｗ is the steering angle.

[0027] According to the geometric information shown in FIG. 4, the travel direction angle D is expressed as D=sin -1(C / R). In other words, the heading angle D can be calculated from the turning radius R. Note that R = V / ω ｙ ω ｙ is the yaw angular velocity of the vehicle axis, but it is preferable to use the yaw angular velocity of the gravity axis (see Figure 2). The roll angular velocity of the traveling axis is the roll angular velocity ω of the vehicle axis ｒ can be calculated by performing coordinate transformation so that the roll angular velocity of the traveling direction axis is rotated by the traveling direction angle D. This calculation is performed by the geometric calculation unit 1. The geometric calculation unit 1 outputs the calculated roll angular velocity of the traveling direction axis to the dynamic roll angle calculation unit 3. The dynamic roll angle calculation unit 3 integrates the roll angular velocity of the traveling direction axis with respect to time to calculate the dynamic roll angle φ ｄ Furthermore, the dynamic roll angle calculation unit 3 calculates the dynamic roll angle φ using the roll angular velocity of the traveling direction axis and the dynamic roll angle error output by the error estimation unit 5. ｄ can be calculated (calculated with feedback).

[0028] At this time, the dynamic roll angle calculation unit 3 calculates the pitch angular velocity ω measured by the IMU 300. ｐ Further, the dynamic roll angle φ ｄ It is preferable to calculate the pitch angular velocity ω ｐ This is because the horizontal component of the rotation axis (extending diagonally upward or downward) of the roll angular velocity of the traveling axis can be calculated by using the dynamic roll angle φ ｄ Since it does not contribute to the pitch angular velocity ω measured by IMU 300, ｐ By using the dynamic roll angle φ ｄ The calculation accuracy can be further improved.

[0029] <Steering correction for estimation error by Kalman filter> When turning (when the steering is turned), that is, when a steering angle is generated, the static roll angle φ ｓTherefore, the calculation accuracy of the error estimated by the Kalman filter in the complementary filter calculation unit 52 also decreases. Therefore, in order to maintain the error estimation accuracy at a certain level or above while suppressing the calculation load, it is preferable to adjust the error when a steering angle is generated. Specifically, it is preferable that the complementary filter calculation unit 52 adjusts the term related to the error estimated by the Kalman filter with the steering angle. By such adjustment, the error estimator 5 can correct the dynamic roll angle error using the steering angle of the motorcycle. Furthermore, the dynamic roll angle calculation unit 3 uses the corrected dynamic roll angle error to calculate the dynamic roll angle φ ｄ can be calculated.

[0030] <Disabling the Kalman filter during special driving> Static roll angle φ ｓ Equation (1) used to calculate the error is not designed for special maneuvers such as jumps or wheelies. Therefore, the accuracy of the error estimation by the Kalman filter is poor during special maneuvers. Also, it is best not to activate controls such as the ABS (Antilock Braking System) during jumps. In light of these circumstances, rather than improving the accuracy of error estimation during special maneuvers, it is better to disable the Kalman filter during special maneuvers. Disabling the Kalman filter in this way makes it possible to achieve both the calculation accuracy and the calculation load required for cornering ABS control.

[0031] Specifically, first, the filter reset determination unit 51 shown in FIG. 1 determines whether the motorcycle is performing special driving. The determination of whether the motorcycle is performing special driving will be described later. If the filter reset determination unit 51 determines that the motorcycle is not performing special driving, the filter reset determination unit 51 outputs a reset determination result indicating this to the complementary filter calculation unit 52. The complementary filter calculation unit 52 enables the Kalman filter in accordance with the reset determination result and performs error estimation. On the other hand, if the filter reset determination unit 51 determines that the motorcycle is performing special driving, the filter reset determination unit 51 outputs a reset determination result indicating this to the complementary filter calculation unit 52. The complementary filter calculation unit 52 disables the Kalman filter in accordance with the reset determination result and does not perform error estimation. In this case, the dynamic roll angle error is not output to the dynamic roll angle calculation unit 3.

[0032] [Determining whether or not special driving is in progress] Whether or not special driving is in progress is determined by the vertical acceleration A ｚ This can be achieved by tracking the change in vertical acceleration (A). Figure 5 is a graph showing the change in vertical acceleration when performing a wheelie. The horizontal axis of the graph in Figure 5 is time, and the vertical axis is vertical acceleration (A ｚ ) The vertical axis of the graph is positive (+) downward and negative (-) upward. As shown in the graph in Figure 5, when a wheelie is performed, the vertical acceleration A ｚ Therefore, the filter reset determination unit 51 determines whether the vertical acceleration A ｚ Four thresholds Th1 to Th4 are prepared for the thresholds Th1 to Th4, which satisfy the relation Th1>Th4>0>Th3>Th2.

[0033] Vertical acceleration A measured by IMU300 ｚ is input to the filter reset determination unit 51. ｚ ≤ Th1 or A ｚ If the condition Th2 (first condition) is satisfied, the filter reset determination unit 51 determines that the motorcycle is performing a wheelie. At this time, it is not necessary to perform an error estimation calculation using the Kalman filter (this calculation is not performed in this embodiment). ｚ ≧Th2 indicates when the wheel is lifted. ｚ ≦Th1 indicates that the lifted wheel has landed. Also, while the filter reset determination unit 51 determines that the vehicle is performing a wheelie, if the vertical acceleration A ｚ Th4≧A ｚ If the condition Th3 (second condition) is met for a certain period of time, the filter reset determination unit 51 determines that the motorcycle has returned from a wheelie to a steady running state. At this time, an error estimation calculation using a Kalman filter should be performed (in this embodiment, it is performed).

[0034] More generally, the above judgment can be applied to special driving conditions other than wheelies (e.g., jumps, rear lifts). ｚIf the first condition is satisfied within a predetermined period, the filter reset determination unit 51 determines that the vehicle is in a special driving state in which at least one wheel is lifted up. ｚ If the second condition is continuously satisfied for a certain period of time, the filter reset determination unit 51 determines that the vehicle is in a steady running state in which all wheels are in contact with the ground. Note that the threshold values ​​Th1 to Th4 can be changed as appropriate depending on the type of special running state.

[0035] Note that Th3 and Th4 function as branching points for determining whether the vehicle is in a steady driving state or an unusual driving state, and can serve as the steady-state determination thresholds shown in Figure 5. Furthermore, Th1 and Th2 function as branching points for determining whether the vehicle is in a unusual driving state or an abnormal state (a state deviating from unusual driving), and can serve as the abnormality determination thresholds shown in Figure 5. The unusual driving determination section shown in Figure 5 is the period from the time when the first condition is met for the first time to the end of the certain period during which the second condition is continuously met. Furthermore, the normal driving determination section shown in Figure 5 is the period other than the unusual driving determination section.

[0036] It is preferable that the filter reset determination unit 51 determine that the special driving state is occurring when the vehicle is traveling straight ahead. When turning, the accuracy of determining whether or not the vehicle is traveling in a special driving state decreases, and if it is determined that the vehicle is not traveling in a special driving state, this may lead to a decrease in the accuracy of the error estimation by the Kalman filter. By limiting the determination of whether or not the vehicle is traveling in a special driving state to when the vehicle is traveling straight ahead, the determination accuracy can be maintained at a certain level or above, and the accuracy of the error estimation by the Kalman filter can be maintained at a certain level or above.

[0037] [Calculation of Steering Angle] The vehicle control device 100 can calculate the steering angle, which affects the accuracy of error estimation by the Kalman filter in the complementary filter calculation unit 52, using only the vehicle speed sensor 200 and the IMU 300. In other words, the vehicle control device 100 can calculate the steering angle without using a steering angle sensor. By configuring a system in which a steering angle sensor is not mounted on the motorcycle, it is possible to reduce the calculation load caused by weight changes due to the mounting of sensors, and to facilitate the implementation of the vehicle control device 100 with as few types of sensors as possible. Note that the present invention does not exclude the use of a steering angle sensor or the mounting of a steering angle sensor on the motorcycle. According to the Automotive Technology Handbook, the steering angle δ Ｈ can be calculated using the following formula (2).

[0038] ...Formula (2)

[0039] where δ ｗ is the steering angle (see Figure 4). φ is the roll angle. ε is the caster angle on the front wheel side. "atan" is "arctan" or "tan -1 " means. The previous value of the static roll angle can be used as the roll angle φ. Alternatively, the two-wheeled vehicle may be equipped with a roll angle sensor, and the roll angle φ may be acquired from the roll angle sensor and used. The roll angle sensor may be, for example, a 9-axis IMU, but is not limited to this.

[0040] According to FIG. 4, the steering angle δ ｗ can be calculated using the following formula (3).

[0041] ...Formula (3)

[0042] From equations (2) and (3), the steering angle δ Ｈ The geometric calculation unit 1 calculates the steering angle δ using equations (2) and (3). ＨIn other words, the geometric calculation unit 1 calculates the traveling direction angle D of the two-wheeled vehicle from the yaw angular velocity of the gravity axis and the wheel speed (vehicle speed V), and calculates the steering angle δ using the traveling direction angle D and the roll angle φ. Ｈ The calculated steering angle δ can be calculated. Ｈ is input to the complementary filter calculation unit 52, and the term relating to the error estimated by the Kalman filter is converted into the steering angle δ Ｈ By such adjustment, the error estimating unit 5 can estimate the steering angle δ of the motorcycle. Ｈ The dynamic roll angle calculation unit 3 can correct the dynamic roll angle error by using the corrected dynamic roll angle error. ｄ can be calculated.

[0043] Here, since the geometric calculation unit 1 calculates the traveling direction angle using the yaw angular velocity measured by the IMU 300, it is preferable to calculate the traveling direction angle based on the height from the road surface of the IMU 300. In other words, when the height from the road surface of the IMU 300 is h, the traveling direction angle D=sin -1 The turning radius R used in the calculation of (C / R) is replaced by R - h·sinφ to calculate the heading angle D, and the steering angle δ Ｈ In this way, the heading angle D is calculated after eliminating the turning radius error in the height direction. Ｈ This can improve the calculation accuracy.

[0044] [Estimation of riding posture] The vehicle control device 100 estimates the static roll angle φ ｓ The riding posture that affects the lean-in angle can be estimated using only the IMU 300. FIG. 6 is an explanatory diagram of a riding posture map. As shown in the riding posture map 44 of FIG. 6, for five types of riding postures, namely lean-with, shallow lean-in riding posture A, deep lean-in riding posture B, shallow lean-out riding posture C, and deep lean-out riding posture D, the horizontal axis is the bank angle (same as the roll angle φ) and the vertical axis is a tan(A ｙ / A ｚ ) the following characteristics 1 and 2 can be seen.

[0045] Feature 1: Even with the same roll angle φ, atan(Aｙ / A ｚ ) values ​​are different. Feature 2: If the riding posture is constant, atan(A ｙ / A ｚ ) is constant. In other words, for any roll angle φ, the change in the roll angle and atan(A ｙ / A ｚ ) (the slope of the linear line when the plots for each riding posture are fitted with a linear line) will be roughly the same as long as the same riding posture is maintained.

[0046] Based on feature 1 and feature 2, atan(A ｙ / A ｚ ) and atan(A ｙ / A ｚ ) is calculated in advance and recorded in the riding posture map 44. The riding posture correction value calculation unit 4 also refers to the riding posture map 44 and calculates the lean-with atan(A ｙ / A ｚ Furthermore, the riding posture correction value calculation unit 4 calculates the value of the current lateral acceleration A ｙ and vertical acceleration A ｚ is obtained from IMU300, and the current atan(A ｙ / A ｚ ) value. This will return the current calculated atan(A ｙ / A ｚ ) value is atan(A ｙ / A ｚ ) (deviation) from the value of the riding posture correction value calculation unit 4, the riding posture correction value calculation unit 4 can specify the riding posture.

[0047] More specifically, atan(A ｙ / A ｚ The value of atan(A ) corresponds to the second attitude angle φ2 estimated by the second attitude angle acquisition unit 42 (see FIG. 3). ｙ / A ｚ The value of atan(A) corresponds to the first attitude angle φ1 estimated by the first attitude angle estimator 41 (see FIG. 3).ｙ / A ｚ ) and the current calculated atan(A ｙ / A ｚ The degree of deviation from the value of (a) corresponds to the difference α calculated by the attitude angle difference calculation unit 43 (see FIG. 3).

[0048] When a two-wheeled vehicle is a moving vehicle including a vehicle and its load (including a passenger and luggage), the first attitude angle estimation unit 41 calculates the lateral acceleration A acting in the lateral direction of the moving vehicle. ｙ , and the vertical acceleration A acting in the direction of gravity on the moving vehicle ｚ Based on this, a first attitude angle φ1 of the moving vehicle relative to the vertical direction of the center of gravity can be estimated. Furthermore, a second attitude angle acquisition unit 42 acquires a second attitude angle φ2 of the vehicle (two-wheeled vehicle) relative to the vertical direction of the center of gravity. An attitude angle difference calculation unit 43 calculates the difference α between the first attitude angle φ1 and the second attitude angle φ2. A riding attitude correction value calculation unit 4 outputs the difference α to the static roll angle calculation unit 2 as a riding attitude correction value. Therefore, the lateral acceleration A ｙ and vertical acceleration A ｚ Based on the static roll angle φ ｓ The difference α, which is the riding posture correction value for correcting the static roll angle φ, can be easily calculated. ｓ can be easily corrected.

[0049] Here, the vertical acceleration A ｚ When becomes extremely small, atan(A ｙ / A ｚ ) diverges, which may cause the first posture angle φ1 to become unable to be output. ｚ When the vertical acceleration A becomes equal to or smaller than the first predetermined value, it is preferable to estimate the first posture angle φ1 using the value of the first predetermined value. ｚ By setting the lower limit value of the first posture angle φ1, it is possible to reliably output the first posture angle φ1 without being affected when the motorcycle jumps.

[0050] In addition, special driving such as jumping increases the vertical acceleration A ｚ When becomes extremely large, atan(A ｙ / A ｚ) becomes extremely small, which may result in the output of a difference α that is inappropriate for the current riding posture. To avoid such a situation, it is preferable that the posture angle difference calculation unit 43 sets the difference α to the second predetermined value when the difference α between the first posture angle φ1 and the second posture angle φ2 becomes equal to or greater than the second predetermined value. By setting an upper limit value for the difference α in this way, it is possible to reliably output the first posture angle φ1 without being affected by jumps of the motorcycle.

[0051] The second attitude angle acquisition unit 42 also acquires the roll angle φ acting on the center of gravity of the vehicle (two-wheeled vehicle), the lateral acceleration A acting in the lateral direction of the vehicle, ｙ , and the vertical acceleration A acting on the vehicle in the direction of gravity ｚ It is preferable to obtain the second posture angle φ2 from the riding posture map 44 created based on the above. By mapping the second posture angle φ2 in this manner, the difference α can be obtained efficiently.

[0052] [Summary] According to this embodiment, the static roll angle φ ｓ and dynamic roll angle φ ｄ After calculating the static roll angle φ, the dynamic roll angle error is estimated, thereby reducing the calculation load for calculating the roll angle φ. ｓ In addition, by calculating the roll angular velocity of the traveling axis, the dynamic roll angle φ can be calculated. ｄ can be corrected in the direction of travel, so the dynamic roll angle φ ｄ In addition, the calculation accuracy of the steering angle δ of the vehicle can be improved. Ｈ By correcting the dynamic roll angle error using ｄ In addition, when it is determined that the vehicle is not traveling in a special mode, the calculation accuracy of the dynamic roll angle φ ｄ If the vehicle is judged to be in a special driving state, the dynamic roll angle φ ｄ Since no error estimation is performed, the impact of special driving can be reduced.

[0053] Furthermore, according to this embodiment, the vertical acceleration A ｚSince the special driving state and the steady driving state can be determined based on the IMU 300 sensor value, a simple configuration can be realized that determines the special driving state using only the sensor value of the IMU 300. Furthermore, by determining the special driving state during straight-ahead driving, the accuracy of the determination can be improved. Furthermore, by setting multiple types of thresholds to determine the special driving state and the steady driving state, the accuracy of the determination as to whether the special driving state has started can be improved, and the accuracy of the determination as to whether the special driving state has ended can also be improved.

[0054] According to this embodiment, the steering angle δ is calculated using the traveling direction angle D and the roll angle φ. Ｈ In order to calculate the steering angle δ without a steering angle sensor, the steering angle δ is calculated using only the sensor value of the IMU 300. Ｈ Furthermore, since the heading angle D is calculated based on the height of the IMU 300 from the road surface, errors in the turning radius R caused by the height direction can be eliminated, and the calculation accuracy of the heading angle D can be improved.

[0055] Furthermore, according to this embodiment, the lateral acceleration A ｙ and vertical acceleration A ｚ Based on the static roll angle φ ｓ The difference α, which is the riding posture correction value for correcting the static roll angle φ, can be easily calculated. ｓ can be easily corrected. ｚ By setting a lower limit value (first predetermined value) of the difference α, the first posture angle φ1 can be reliably output without being affected by the motorcycle jumping. Also, by setting an upper limit value (second predetermined value) of the difference α, the first posture angle φ1 can be reliably output without being affected by the motorcycle jumping. Also, by mapping the second posture angle φ2, the difference α can be efficiently determined.

[0056] [Method Invention] A first method invention is a roll angle calculation method for calculating a roll angle of a vehicle, the roll angle calculation method comprising: calculating a provisional first roll angle using a yaw angular velocity and a pitch angular velocity measured by an inertial measurement unit, and a yaw angular velocity about a gravity direction axis and a wheel speed calculated from previous values ​​of physical quantities related to the roll angle; calculating a riding attitude correction value based on a difference between a first attitude angle calculated based on a center of gravity generated by a riding attitude and a second attitude angle calculated based on a vehicle center of gravity, the first attitude angle being an angle obtained from a lateral acceleration acting in a lateral direction of the vehicle and a vertical acceleration acting in a gravity direction of the vehicle; calculating a definitive first roll angle by correcting the provisional first roll angle with the riding attitude correction value; and estimating a second roll angle error based on the difference between the definitive first roll angle and a second roll angle different from the definitive first roll angle. Further, there is provided a roll angle calculation method according to a first aspect of the present invention, which is characterized in that a traveling direction angle of the vehicle is calculated from a yaw angular velocity about the gravity direction axis and a wheel speed, a roll angular velocity about the traveling direction axis is calculated based on a roll angular velocity measured by an inertial measurement unit and the traveling direction angle, and the second roll angle is calculated using the roll angular velocity about the traveling direction axis and the second roll angle error.

[0057] The second invention of the method is a driving condition determination method for determining the driving condition of a vehicle, which is characterized in that if the vertical acceleration of the vehicle satisfies a first condition, it is determined that the vehicle is in a special driving condition in which at least one wheel is off the ground, and if the vertical acceleration continues to satisfy a second condition for a certain period of time while it is determined that the vehicle is in the special driving condition, it is determined that the vehicle is in a steady driving condition in which all wheels are on the ground.

[0058] A third aspect of the present invention is a steering angle calculation method, which is characterized by calculating a vehicle's traveling direction angle from the yaw angular velocity of the gravity direction axis and the wheel speed, and calculating the steering angle using the traveling direction angle and roll angle.

[0059] A fourth invention of the method is a method for estimating a riding attitude correction value for a vehicle used to estimate the roll angle of the vehicle, the method comprising the steps of: estimating a first attitude angle of the center of gravity of the moving vehicle relative to the vertical direction based on a lateral acceleration acting in the lateral direction of the moving vehicle including the vehicle and its cargo, and a vertical acceleration acting in the direction of gravity of the moving vehicle; obtaining a second attitude angle of the center of gravity of the vehicle relative to the vertical direction; determining a difference between the first attitude angle and the second attitude angle; and setting the difference as the riding attitude correction value.

[0060] [Modification] (a): In this embodiment, the error estimator 5 (complementary filter calculator 52) calculates the dynamic roll angle error, and the dynamic roll angle calculator 3 calculates the dynamic roll angle φ ｄ However, the error estimation unit 5 (complementary filter calculation unit 52) ​​calculates the static roll angle φ ｓ Since the error of the static roll angle can also be estimated, the static roll angle error is calculated, and the static roll angle calculation unit 2 calculates the static roll angle φ ｓ (b): In this embodiment, the error estimating unit 5 may provide a feedback correction to the steering angle δ Ｈ was input to adjust the error estimation by the Kalman filter. However, for example, when there is a large steering, that is, when the steering angle δ Ｈ If the static roll angle φ is equal to or greater than a predetermined value, the accuracy of the error estimation by the Kalman filter is expected to be significantly reduced, so control may be performed to prevent the error estimation from being performed in the first place. ｓ and dynamic roll angle φ ｄ is input, the roll angle difference is calculated, and the result is output to the complementary filter calculation unit 52. However, the static roll angle φ ｓ and dynamic roll angle φ ｄ Alternatively, the complementary filter calculation unit 52 may calculate the roll angle difference and perform error estimation using a Kalman filter, without calculating the roll angle difference. ｓ and dynamic roll angle φ ｄ An error estimate may be made for each of

[0061] (d): It is also possible to realize a technology that appropriately combines the various technologies described in this embodiment. (e): The software described in this embodiment can be realized as hardware, and the hardware can be realized as software. (f): In addition, the hardware, software, flowcharts, etc. can be appropriately modified within the scope that does not deviate from the spirit of the present invention.

[0062] 100 Vehicle control device 200 Vehicle speed sensor 300 IMU (Inertial Measurement Unit) 1 Geometry calculation unit 2 Static roll angle calculation unit (first roll angle calculation unit) 3 Dynamic roll angle calculation unit (second roll angle calculation unit) 4 Riding posture correction value calculation unit (attitude angle calculation unit) 41 First attitude angle estimation unit 42 Second attitude angle acquisition unit 43 Attitude angle difference calculation unit 44 Riding posture map 5 Error estimation unit (roll angle error estimation unit) 51 Filter reset determination unit (special driving determination unit, driving determination unit) 52 Complementary filter calculation unit (roll angle error estimation unit) 6 Output processing unit

Claims

1. A vehicle control device comprising: a geometric calculation unit that calculates a vehicle's heading angle from the yaw angular velocity of the gravity axis and the wheel speed, and calculates a steering angle using the heading angle and roll angle.

2. A vehicle control device as claimed in claim 1, further comprising an inertial measurement unit that measures the yaw angular velocity, and wherein the geometric calculation unit calculates the travel direction angle based on the height of the inertial measurement unit from the road surface.

3. A steering angle calculation method comprising: calculating a vehicle's heading angle from a yaw angular velocity of a gravity axis and a wheel speed; and calculating a steering angle using the heading angle and roll angle.

Images