Vehicle control device

Abstract

The present invention pertains to a vehicle control device and a vehicle control method that generate an assist control input on the basis of a zeroing control barrier function in response to an operation input for changing a traveling state of a vehicle, and control the traveling state on the basis of the operation input and the assist control input. The vehicle control device and the vehicle control method find a subset of a safe set, which represents a predetermined safe area considered safe, on the basis of predetermined constraints related to the operation input and the control input, and determine whether an initial state of the vehicle at the start of control belongs to the found subset or not, and consequently, when it is determined that the initial state belongs to the subset, control the traveling state on the basis of the operation input and the assist control input.

Description

Vehicle control device and method

[0001] The present invention relates to a vehicle control device and a vehicle control method for controlling the running of a vehicle.

[0002] Vehicles are typically driven by a driver's driving operations. However, in recent years, technologies to assist the driver in driving the vehicle, such as following a vehicle in a traffic jam or avoiding a collision with an object, have been researched and developed. One control theory for such systems in which the driver's driving operations intervene is human-assist control using a so-called control barrier function (CBF). CBFs are broadly classified into reciprocal CBFs, which have the property that the function value of the CBF diverges to infinity at the boundary of a safety region set as a safe region, and zero-type CBFs (zeroing CBFs), which have the property that the function value of the CBF becomes zero at the boundary of the safety region. Human-assist control using this zeroing CBF is disclosed, for example, in Non-Patent Document 1.

[0003] In this non-patent document 1, the driving operation of the vehicle by the driver is ｈ (t) is an input affine nonlinear control system; dx / dt = f(x) + g(x)(u + u ｈ In (t), a human assist control u is designed by using a Zero-type CBF.

[0004] However, in actual vehicles, there are constraints on acceleration (including not only positive acceleration for acceleration but also negative acceleration for deceleration), steering, etc. due to the vehicle's performance, etc. Alternatively, there may be constraints on acceleration, steering, etc. due to the vehicle's ride comfort, etc. Non-Patent Document 1 does not take such constraints into consideration.

[0005] Kazunari Tezuka, Fumikazu Nakamura, "Human-assisted control using zero-type control barrier functions," 64th Joint Conference on Automatic Control, November 13th and 14th, 2021, online, pp. 105-112

[0006] The present invention has been made in consideration of the above circumstances, and its object is to provide a vehicle control device and a vehicle control method that can control a vehicle with assist control using zeroed CBF, taking into account constraints.

[0007] The vehicle control device and vehicle control method according to the present invention are a device and method that generate an assist control input based on a zeroing control barrier function in response to an operation input that changes the running state of a vehicle, and control the running state based on the operation input and the assist control input, and determine a subset of a safety set that represents a predetermined safety region that is considered safe based on predetermined constraints related to the operation input and the control input, determine whether the initial state of the vehicle at the start of control belongs to this determined subset, and if it is determined as a result that the initial state belongs to the subset, control the running state based on the operation input and the assist control input.

[0008] These and other objects, features and advantages of the present invention will become apparent from the following detailed description and accompanying drawings.

[0009] 1 is a block diagram showing the configuration of a vehicle control device in an embodiment. FIG. 2 is a diagram for explaining a target system in a first embodiment. FIG. 3 is a flowchart showing the operation of a vehicle control device in an embodiment. FIG. 4 is a diagram for explaining a target system in a second embodiment. FIG. 5 is a graph showing the time changes in acceleration and speed when the vehicle behavior is a first pattern. FIG. 6 is a graph showing the time changes in acceleration and speed when the vehicle behavior is a second pattern. FIG. 7 is a graph showing an example of a simulation result in the target system in the second embodiment. FIG. 8 is a diagram for explaining a target system in a third embodiment. FIG. 9 is a graph showing the time changes in steering angle velocity and steering angle when the vehicle behavior is a third pattern. FIG. 10 is a graph showing the time changes in steering angle velocity and steering angle when the vehicle behavior is a fourth pattern. FIG. 11 is a graph showing an example of a simulation result in the target system in the third embodiment.

[0010] An embodiment of the present invention will be described below with reference to the drawings. In each drawing, components with the same reference numerals are designated as identical components, and their description will be omitted where appropriate. In this specification, generic reference numerals are used without subscripts, and individual components are designated with subscripts.

[0011] (Preparation) First, human assist control using a zero-type (zeroing) control barrier function disclosed in Non-Patent Document 1 will be described.

[0012] Regarding notation, for simplicity of notation, we use the continuous function f:R ｎ ×R→R ｎ , g:R ｎ ×R→R ｎ×ｍ C against １ class differentiable function h:R ｎ → Lie derivative of R L ｆ h(x, t), L ｇ h(x, t) is defined by the following equations 1 and 2, respectively.

[0013]

[0014]

[0015] A continuous function γ １ :R ≧０ →R ≧０ is in class κ if γ １ (・) is a strictly monotonically increasing function and γ １ (0) = 0. R represents a real number, and R ≧０ represents a non-negative real number, and R ＞０ represents a positive real number.

[0016] A continuous function γ ２ : R → R is included in the extension class κ if γ ２ (・) is a strictly monotonically increasing function and γ ２ (0) = 0.

[0017] The target system of human assist control is the vehicle operation input by the driver. ｈ (t) is an input affine nonlinear control system expressed by the following equation 3.

[0018]

[0019] where x∈R ｎ represents a state, t∈R represents time, and u∈R ｍ represents the control input (control input amount, input amount of human assist control), and u ｈ ∈R→R ｍ represents the feedforward input (the input amount of the vehicle driving operation by the driver). ｎ ×R→R ｎ and g: R ｎ ×R→R ｎ×ｍ Let be a locally Lipschitz continuous map, and u ｈ is a continuous map.

[0020] The domain of the state x in the input affine nonlinear control system expressed in the above formula 3 is the safe set X⊂R. ｎ and the set of residuals X ｕ ⊂R ｎ \X ⊂ R ｎ The safe set X is divided into C shown in the following formula 4-1 and formula 4-2. １ class differentiable function h:R ｎ → Let ∂X be an open and time-invariant set defined using R. Here, ∂X is the boundary of the safe set X. When x∈X, the target system is safe.

[0021]

[0022] In the above-mentioned formulas 3, 4-1 and 4-2, the human assist control input u is a control input u∈R ｍ The following conditions A1 and A2 are satisfied.

[0023] Condition A1 is the arbitrary operation input u ｈ ∈R→R ｍ , the control input u∈R ｍ The purpose is to guarantee the positive invariance of the safe set X.

[0024] Condition A2 is that the control input u∈R ｍ is minimized under predefined evaluation criteria.

[0025] Human assist control using a zero-type (zeroing) control barrier function is to obtain a control input u=k(x, t) that satisfies the above conditions A1 and A2 using a zeroing CBF.

[0026] That is, in the above formulas 3, 4-1 and 4-2, the zeroed CBF is C １ class differentiable function h:R ｎ →R satisfies the following conditions B1 and B2.

[0027] Condition B1 is proper for x∈X, that is, for any L∈R ≧０ For x, the superset {x|h(x)≧L} is a compact set.

[0028] Condition B2 states that for any continuous mapping u ｈ ∈R→R ｍ , there exists an extended class κ of locally Lipschitz continuous functions α that satisfy the following equation 5:

[0029]

[0030] The derivation and the like are disclosed in Non-Patent Document 1, and the control input u = k(x, t) is given by the following equation 6, and an arbitrary continuous mapping u ｈ : R → R ｍ For any time t≧0, x(t)εX.

[0031]

[0032] Here, the function I:R ｎ ×R×R ｍ →R, J: R ｎ →R are functions defined by the following equations 7-1 and 7-2, respectively.

[0033]

[0034] Note that the control input u = k(x, t) is any (x, t) ∈ X × R ≧０ is the smallest among the control inputs u that satisfy the following equation 8. Therefore, the design problem of the control input u that minimizes ||u|| and satisfies the following equation 8 becomes the problem of minimizing ||u|| under the constraint condition of the following equation 8.

[0035]

[0036] (Principle of the embodiment) Next, human assist control using a zeroing control barrier function in consideration of constraints will be described.

[0037] The target system represented by the above formula 3 is an operation input u ｈ and a given constraint (input constraint) on the control input u, the initial state x is in the safe set X. ０ Even if there is a certain safe set X, safety is not necessarily guaranteed. This is because the condition B2 of the CBF cannot be satisfied within the input constraint v∈U in a subset of the safe set X. This point is solved here by using the Viability Kernel.

[0038] In the target system and safe set X expressed by the above formula 3, Viability Kernel V:=Viab(X) is defined by the following formula 9.

[0039]

[0040] That is, the initial state x of the target system expressed by the above formula 3 ０ Ga x ０ ∈V, the control input u and the operation input u make the safe set X an invariant set. ｈ exists within the input constraint ν, where for all x ０ When x(t)∈X, ∀t≧0 holds for ∈X, the safe set X is said to be an invariant set.

[0041] In this embodiment, the object is to find a human assist control u=k(x, t) that satisfies x∈X while satisfying the input constraint v∈U. When using the above-mentioned Viability Kernel, this can be rephrased as finding a human assist control u=k(x, t) that makes the Viability Kernel V⊂X an invariant set.

[0042] The input constraint ν∈U is a continuously differentiable convex function g ｉIn this case, in the target system and safe set X expressed by the above formula 3, the viability control barrier function (V-CBF) that combines the concepts of the Viability Kernel and CBF satisfies the following conditions C1, C2, and C3: １ class differentiable function h:R ｎ →R can be defined as

[0043] Condition C1 is that for a Viability Kernel V for a safe set X, h(x) satisfies the following equations 11-1 to 11-3. Here, ∂V is the boundary of V, and V ｃ is the complement of V.

[0044] Condition C2 is proper for x∈V, i.e., for any L∈R ≧０ For x, the upper level set {x|h(x)≧L} is compact.

[0045] Condition C3 states that for any continuous mapping u ｈ : R → U ｈ For x, there exists an extended class κ of locally Lipschitz continuous functions α:R→R that satisfy the following inequality 12. Here, the time derivative of h(x) is defined by the following equation 13.

[0046]

[0047]

[0048]

[0049] For the V-CBF defined in this way, the set of inputs S νｃｂｆ makes the Viability Kernel V an immutable set, thereby guaranteeing the logical safety of the target system.

[0050]

[0051] The input constraint ν∈U is a continuously differentiable convex function g ｉ Therefore, the human assist control u=k(x, t) can be obtained by solving the convex optimization problem given by the following equation 15. Here, the function I:R ｎ ×R ｍ→R is defined by the following equation 16.

[0052]

[0053]

[0054] When the target system expressed by the above formula 3 is a one-input system, the input constraint is given by the norm constraint expressed by the following formula 17, and when the convex optimization problem is solved, an arbitrary continuous mapping u ｈ : R → [-a ｍａｘ , a ｍａｘ ], the human assist control u=k(x, t) is given by the following equation 18. Here, the function α:R→R is an extended class κ locally Lipschitz continuous function, and the function I:R ｎ ×R→R is defined by the following equation 19.

[0055]

[0056]

[0057]

[0058] In the above description, from the above equation 9, the initial state x ０ ∈V, the initial state x ０ is in the Viability Kernel, and the target system is safe. On the other hand, in the initial state x ０ is outside the Viability Kernel, i.e., the initial state x ０ When ∈X\V⊂X, safety is lacking under the input constraints expressed by the above-mentioned equation 10, but the damage can be mitigated by using a control input that maximizes the derivative of V-CBF.

[0059] First Embodiment In the first embodiment, V-CBF is applied to a vehicle system represented by a double integral system having input constraints as a target system.

[0060] FIG. 1 is a block diagram showing the configuration of a vehicle control device according to an embodiment. The configurations of the vehicle control devices Sa to Sc in the first to third embodiments are basically the same, and therefore FIG. 1 illustrates not only the vehicle control device Sa in the first embodiment, but also the configurations of the vehicle control devices Sb and Sc in the second and third embodiments. Therefore, in FIG. 1, the configuration of the vehicle control device Sa in the first embodiment is denoted by a reference symbol with a suffix a, the configuration of the vehicle control device Sb in the second embodiment is denoted by a reference symbol with a suffix b, the configuration of the vehicle control device Sc in the third embodiment is denoted by a reference symbol with a suffix c, and the same configurations of the vehicle control devices Sa to Sc in the first to third embodiments are denoted by only the reference symbols. This is also true for FIG. 3, which will be described later.

[0061] 1, the vehicle control device Sa in the first embodiment is mounted on a vehicle VC, and includes, for example, a target detection unit 1, a running state detection unit 2, an operation input unit 3, a control processing unit 4a, a memory unit 5a, a first power unit (drive wheel drive unit) 6, a second power unit (steering wheel drive unit) 7, a braking unit 8, and an interface unit (IF unit) 9. The vehicle control method in the first embodiment is implemented in this vehicle control device Sa.

[0062] The target detection unit 1 is connected to the control processing unit 4a, and is a device that detects targets present around a vehicle (host vehicle) VC under the control of the control processing unit 4a, and detects the position of the detected target (relative target position) relative to the vehicle VC. The target detection unit 1 outputs the detection result to the control processing unit 4a. The target detection unit 1 detects the target within a predetermined detection range. The relative target position is represented by the distance (relative distance) from the host vehicle VC to the target and the direction (azimuth) of the target as seen from the host vehicle VC. The vehicle VC equipped with the vehicle control device S (Sa to Sc) will be referred to as the "host vehicle" as appropriate. Such a target detection unit 1 includes, for example, a radar disposed at the front end of the host vehicle VC (for example, at the center position of the bumper) and detecting targets within a predetermined detection range, and a camera disposed within the host vehicle VC on the ceiling surface (inner surface of the roof) near the windshield so as to cover the detection range and capture images of the area ahead of the host vehicle VC, and detects targets from images generated by the camera using a machine-learned learning model, and determines and outputs the relative target position (relative distance and direction) of the detected target from the detection result of the radar. The target detection unit 1 may further include a sonar disposed, for example, in a square on the vehicle VC in order to detect targets relatively close to the host vehicle VC.

[0063] In this embodiment, the target detection unit 1 further detects at least one of the velocity (relative target velocity) of the detected target relative to the vehicle (host vehicle) VC and the acceleration (relative target acceleration) of the detected target relative to the vehicle (host vehicle) VC, and outputs the detection result to the control processing unit 4a. The target detection unit 1 further detects at least one of the relative target velocity and the relative target acceleration depending on the target system. When detecting the relative target velocity, for example, the radar determines the relative target velocity of the target relative to the host vehicle VC based on the Doppler shift of a reflected wave relative to a transmitted wave. When detecting the relative target acceleration, for example, the relative target acceleration can be determined by time-differentiating the relative target velocity.

[0064] In this embodiment, the target detection unit 1 is not limited to being configured with a camera and a radar, but may be configured with a stereo camera instead of a camera and a radar as long as it can detect the relative target position. In this case, one of the cameras of the stereo camera functions as the camera, and the stereo camera functions as the radar.

[0065] The driving condition detection unit 2 is connected to the control processing unit 4a and is a device that detects the driving condition of the vehicle (host vehicle) VC under the control of the control processing unit 4a. The driving condition detection unit 2 outputs the detection result to the control processing unit 4a. The driving condition detection unit 2 includes, for example, at least one of a speed detection unit (vehicle speed sensor) that detects the speed (vehicle speed) of the host vehicle VC, an acceleration detection unit (acceleration sensor) that detects the acceleration of the host vehicle VC, an attitude angle detection unit (yaw rate sensor) that detects the attitude angle (yaw) of the host vehicle VC, and a steering angle detection unit (first steering angle sensor) that detects the steering angle (steering angle of the steering wheels) of the host vehicle VC. The driving condition detection unit 2 is configured to include at least one of the speed detection unit, the acceleration detection unit, the attitude angle detection unit, and the steering angle detection unit depending on the target system. Of course, the speed detection unit, the acceleration detection unit, the attitude angle detection unit, and the steering angle detection unit are merely examples, and the driving condition detection unit 2 may include other detection units depending on the target system.

[0066] The operation input unit 3 is connected to the control processing unit 4a and is a device that outputs operation inputs (operation input amount, feedforward input, input amount of feedforward input) that change the driving state of the vehicle (host vehicle) VC to the control processing unit 4a. For example, when an operation input is input by the driver and this operation input by the driver is output to the control processing unit 4a, the operation input unit 3 is configured to include, for example, an accelerator sensor that measures the depression amount (stroke amount) of an accelerator pedal, a brake sensor that measures the depression amount (stroke amount) of a brake pedal, and a steering angle sensor (second steering angle sensor) that measures the rotation angle (steering angle) of a steering wheel, and outputs the measurement results measured by each sensor to the control processing unit 4a. Alternatively, for example, the operation input unit 3 is configured to include a known driving system, such as a constant-speed driving system that performs constant-speed driving or an automatic driving system that performs automatic driving, and outputs operation inputs generated by the driving system to the control processing unit 4a.

[0067] The first power unit (drive wheel drive unit) 6 is connected to the control processing unit 4a and is a device that drives the drive wheels of the vehicle (host vehicle) VC in accordance with the control of the control processing unit 4a. The first power unit 6 is configured, for example, to include a power source (first power source) that generates power and a power transmission mechanism (transmission, etc.) that transmits the power generated by the first power source to the drive wheels, and rotates the drive wheels using the power. The first power source is configured, for example, to include a prime mover such as an engine, a motor, or a hybrid device thereof, and its accessories.

[0068] The second power unit (steering wheel drive unit) 7 is connected to the control processing unit 4a and is a device that drives the steering wheels of the vehicle (host vehicle) VC in accordance with the control of the control processing unit 4a. The second power unit 7 is configured to include, for example, a power source (second power source) that generates power, and a drive mechanism, such as a rack-and-pinion type, that uses the power generated by the second power source to rotate the steering shaft of the steering wheel, thereby rotating the steering shaft using the power. The second power source is configured to include, for example, a motor, a pump that is operated by the engine as the first power source and generates hydraulic pressure as power, etc.

[0069] The braking unit 8 is connected to the control processing unit 4a and is a device that applies a braking force to the vehicle (host vehicle) VC under the control of the control processing unit 4a to decelerate the host vehicle VC. The braking unit 8 can stop the host vehicle VC as a result of decelerating the host vehicle VC, and can also maintain the host vehicle VC at a standstill by continuing to apply a braking force while the host vehicle VC is stopped. The braking unit 8 is configured to include, for example, a braking device such as a disc brake and a regenerative brake, and its associated devices.

[0070] The acceleration of the host vehicle VC is adjusted, and the speed (vehicle speed) of the host vehicle VC is adjusted by controlling at least one of the first power unit 6 and the braking unit 8. The traveling direction of the host vehicle VC is adjusted by controlling the second power unit 7.

[0071] The IF unit 9 is connected to the control processing unit 4a and is a device for inputting and outputting data between the control processing unit 4a and an external device connected to the IF unit 9, and is, for example, an interface circuit of RS-232C, which is a serial communication method, an interface circuit using the USB standard, etc. The external device is, for example, an input device for inputting data, such as a numeric keypad or keyboard, or a display device for displaying data, such as a liquid crystal display (LCD).

[0072] The storage unit 5a is connected to the control processing unit 4a and is a circuit that stores various predetermined programs and various predetermined data under the control of the control processing unit 4a. The various predetermined programs include, for example, a control processing program, which includes, for example, a control program that controls each of the units 1 to 3, 5a, 6 to 9 of the vehicle control device Sa according to the function of each unit; a subset calculation program that calculates a subset of a safety set representing a predetermined safety area that is considered safe based on predetermined constraints related to operation inputs and control inputs; an inclusion / inclusion determination program that determines whether the initial state of the vehicle at the start of control belongs to the subset calculated by the subset calculation program; and a driving control program that controls the driving state based on the operation input and the assist control input when the inclusion / inclusion determination program determines that the initial state belongs to the subset. The various predetermined data include, for example, data necessary for executing each of these programs, such as the constraints, detection results from the target detection unit 1, detection results from the driving state detection unit 2, and operation input from the operation input unit 3. The storage unit 5a includes, for example, a ROM (Read Only Memory) which is a nonvolatile storage element, an EEPROM (Electrically Erasable Programmable Read Only Memory) which is a rewritable nonvolatile storage element, a flash memory, etc. The storage unit 5a also includes a RAM (Random Access Memory) which serves as a working memory for the control processing unit 4a and stores data generated during execution of the predetermined program, etc.

[0073] The storage unit 5a functionally includes a constraint information storage unit 51a that stores constraint information representing the constraints. As described above, the constraints are predetermined constraints related to operation inputs and control inputs, in other words, constraints related to the traveling state of the vehicle (host vehicle) VC. The constraints include at least one of a first constraint on acceleration (including positive acceleration and negative acceleration), a second constraint on jerk, which is the time derivative of acceleration, a third constraint on steering angle, and a fourth constraint on steering angular velocity, which is the time derivative of steering angle. The constraints include at least one of the first to fourth constraints depending on the target system. Of course, the first to fourth constraints are merely examples, and other constraints may be present depending on the target system. The constraints may also be related values ​​indirectly representing the acceleration, jerk, steering angle, and steering angular velocity, respectively. The constraints are determined appropriately according to the specifications of the vehicle VC, for example, and are input from an input device or a display device connected to the IF unit 9 at the timing of, for example, the manufacturing stage, delivery stage, or inspection stage of the vehicle VC while referring to the display device, and are stored in the constraint information storage unit 51 a. Alternatively, for example, the constraints may be calculated according to an algorithm that is appropriately set in advance according to the condition of the vehicle itself or the condition of the environment around the vehicle.

[0074] The control processing unit 4a is a circuit that controls each of the units 1 to 3, 5a, 6 to 9 of the vehicle control device Sa according to the function of each unit, generates an assist control input based on a nulling control barrier function in response to an operation input that changes the running state of the vehicle VC, and controls the running state based on the operation input and the assist control input. The control processing unit 4a is configured to include, for example, a CPU (Central Processing Unit) and its peripheral circuits. By executing a control processing program, the control processing unit 4a functionally includes a control unit 41, a subset calculation unit 42a, a membership determination unit 43a, and a running control unit 44a.

[0075] The control unit 41 controls each of the units 1 to 3, 5a, and 6 to 9 of the vehicle control device Sa in accordance with the function of each unit, and is in charge of overall control of the vehicle control device Sa.

[0076] The subset calculation unit 42a calculates a subset of a safety set representing a predetermined safety region that is considered safe based on predetermined constraints related to operation inputs and control inputs. In this embodiment, the subset calculation unit 42a further calculates the safety set by setting a safety region between the target detected by the target detection unit 1 and the vehicle (host vehicle) VC.

[0077] The belonging / non-belonging determination unit 43 a determines whether or not the initial state of the vehicle at the start of control belongs to the subset obtained by the subset calculation unit 42 a. The initial state of the vehicle at the start of control is the first state when the assist control input is generated based on the zeroing control barrier function to control the running state, and is the state of the vehicle at the current timing when assist control based on the zeroing control barrier function is repeatedly started at predetermined time intervals.

[0078] The driving control unit 44a controls the driving state based on the operation input and the assist control input when the belonging / non-belonging determination unit 43a determines that the initial state belongs to the subset. When the belonging / non-belonging determination unit 43a determines that the initial state does not belong to the subset, the driving control unit 44a generates an assist control input that maximizes the derivative of the zeroing control barrier function and controls the driving state based on the operation input and the assist control input. When accelerating or decelerating the vehicle (host vehicle) VC to change the driving state of the host vehicle VC, the driving control unit 44a controls the driving state by controlling at least one of the first power unit 6 and the braking unit 8. When changing the driving state of the host vehicle VC by changing the steering angle of the vehicle (host vehicle) VC, the driving control unit 44a controls the driving state by controlling the second power unit 7.

[0079] The subset calculation unit 42a, the belonging / non-belonging determination unit 43a, and the driving control unit 44a will be described in more detail.

[0080] The vehicle system that is the target system of the vehicle control device Sa in the first embodiment is expressed by the following equation 20-1, and its input constraint is expressed by the following equation 20-2.

[0081]

[0082] Here, x = [x １ , x２ ] Ｔ ∈R ２ represents the state of the vehicle (x １ represents the vehicle position, and x ２ represents the vehicle speed), t∈R represents time, u∈R represents the control input (control input amount, input amount of assist control), and u ｈ :R ≧０ → (-a ｍａｘ , ∞) is a continuous mapping and represents the feedforward input (operation input, operation input amount). ｍａｘ >0 is the maximum acceleration in the deceleration direction, and u ｈ ≧-a ｍａｘ It should be noted that equation 20-1 is given by f(x)=[x ２ , 0] Ｔ , g(x)=[0,1] Ｔ This is the case when

[0083] In order to ensure safety, it is preferable that the initial value of the input basically satisfies Expression 20-2. The same applies to the second and third embodiments described later. However, even if the initial value of the input constraint does not satisfy Expression 20-2, damage can be mitigated.

[0084] 2 is a diagram for explaining the target system in the first embodiment. In this vehicle system, as shown in FIG. 2, a vehicle VC is traveling in the direction of an obstacle Ob. The obstacle Ob is, for example, a target detected by the target detection unit 1. This traveling direction is x １ The direction of the axis is assumed to be the x-axis, and the coordinate origin is set to the boundary of the obstacle Ob. １ The direction of the axis is x １ The direction of the axis may be arbitrary, and in this case, the same explanation can be given by considering the component in the direction of travel. The same applies to the second embodiment described below.

[0085] In the case shown in FIG. 2, the safe set X is X=(−∞, 0)×R⊂R ２ The remaining set X ｕ is X ｕ =R ２ \X ⊂ R ２ That is, in this vehicle system, the position x １ Ga x１ If ∈(−∞, 0), the vehicle system is safe. This safe set X represents a safety area set by the subset calculation unit 42a between the target (here, the obstacle Ob) detected by the target detection unit 1 and the vehicle VC.

[0086] Therefore, in the first embodiment, the initial state x of the vehicle system expressed by Equation 20-1 ０ Ga x ０ ∈X, then any u ｈ For any t≧0, a control input (input for assist control) u is obtained that satisfies x(t)εX and the input constraint expressed by Equation 20-2. １０ Let the initial velocity be x ２０ Then, the initial state x ０ is x ０ = [x １０ , x ２０ ] Ｔ is.

[0087] First, the Viability Kernel is calculated. In the first embodiment, the subset calculation unit 42a calculates the Viability Kernel as a subset Viab(X) of the safety set X. More specifically, the maximum acceleration -a ｍａｘ If the time it takes for the vehicle VC to stop is τ, then x ２ <0, the vehicle VC is moving away from the boundary of the obstacle Ob, so x ２ ≧0, the following equation 21 holds, and the distance d(x) traveled by the vehicle VC up to this time τ is given by the following equation 22.

[0088]

[0089]

[0090] From the above, we can obtain the Viability Kernel V for the safe set X. ｘ is given by the following equation 23.

[0091]

[0092] That is, x ０ ∈V ｘ⊆X, then under the input constraints expressed by Equation 20-2, for any t≧0, x １ In the first embodiment, there is at least one control input u that ensures that (t)∈(−∞, 0). ０ (= [x １０ , x ２０ ] Ｔ ) is a subset V ｘ It is determined whether the object belongs to the group.

[0093] In addition, x ２ Considering <0, Viability Kernel V ｘ is given by the following equation 24.

[0094]

[0095] And this Viability Kernel V ｘ More specifically, the control input u is calculated from the Viability Kernel V expressed by Equation 23. ｘ In this case, the function h(x): ２ →R is expressed as the following formula 25, and L ｆ h(x) and L ｇ From equation 25, h(x) becomes equations 26-1 and 26-2, respectively, and therefore the control input (input of assist control) u becomes equation 27 below.

[0096]

[0097]

[0098]

[0099] Here, γ:R ２ →R, I:R ２ ×R → R and J:R ２ ×R→R is defined as in the following equations 28-1, 28-2, and 28-3, respectively, for any non-negative constants K and C≧0.

[0100]

[0101] Note that as the value of constant K increases, the intervention time of the control input becomes slower, and therefore a larger control input is applied. As the value of constant C increases, the intervention time of the control input becomes slower, and therefore a larger control input is applied. Compared to constant K, constant C has a greater effect on the time when the control input begins to work.

[0102] In the first embodiment, the traveling control unit 44a determines the subset V ｘ In the initial state x ０ If it is determined that the vehicle speed belongs to the vehicle speed range, first, the magnitude relationship between I and J is compared from Equation 27 to determine whether or not to perform assist control. If I is equal to or greater than J (I≧J), the assist control is not performed and the operation input u by the driver or the driving system is used. ｈ The driving state is controlled by controlling at least one of the first power unit 6, the second power unit 7, and the braking unit 8 based on the above (non-assist control driving control), and when I is less than J (I<J), in order to perform assist control, an assist control input u is calculated using equations 27, 28-1 to 28-3, 25, 26-1, and 26-2, and an operation input u by the driver or the driving system is calculated. ｈ The traveling control unit 44a controls the traveling state by controlling at least one of the first power unit 6, the second power unit 7, and the braking unit 8 based on the assist control input u (first mode of traveling control). ｘ In the initial state x ０ does not belong to the zeroing control barrier function, an assist control input u is generated that maximizes the derivative of the zeroing control barrier function, and the operation input u ｈ The driving state is controlled by controlling at least one of the first power unit 6, the second power unit 7, and the braking unit 8 based on the assist control input u (a second mode of driving control).

[0103] In addition, when the vehicle system shown in FIG. 2 is the target system, the driving state detection unit 2 includes at least the speed detection unit (vehicle speed sensor), and in the driving control of the first and second modes, the driving control unit 44a detects the operation input u ｈThe driving state is controlled by controlling at least one of the first power unit 6 and the braking unit 8 based on the assist control input u.

[0104] The control processing unit 4a and the storage unit 5a can be configured by a computer mounted on the vehicle VC, which is called an ECU (Electronic Control Unit).

[0105] Next, the operation of the first embodiment will be described with reference to a flowchart shown in FIG.

[0106] When the vehicle VC starts operating, the vehicle control device Sa initializes the necessary parts and starts its operation. By executing the control processing program, the control processing unit 4a is functionally configured with a control unit 41, a subset calculation unit 42a, a membership determination unit 43a, and a driving control unit 44a.

[0107] The vehicle control device Sa repeatedly executes each of the processes S1 to S8a shown in FIG. 3 at each timing of a predetermined time interval until the vehicle VC is terminated, for example, until the vehicle VC is terminated.

[0108] 3, at this timing, first, the vehicle control device Sa, by using the control unit 41 of the control processing unit 4a, acquires the detection result from the target detection unit 1 and stores the acquired detection result in the memory unit 5a, and acquires the detection result from the running state detection unit 2 and stores the acquired detection result in the memory unit 5a (S1). The position of the vehicle in the initial state at the start of control (the position of the host vehicle relative to the target) (here, the initial position) is calculated as the inverse of the relative target position detected by the target detection unit 1, and the speed of the vehicle in the initial state at the start of control (here, the initial speed) is the detection result of the speed detection unit in the running state detection unit 2.

[0109] Next, the vehicle control device Sa determines whether or not a target has been detected by referring to the detection result acquired from the target detection unit 1 using the control unit 41 (S2). If the result of this determination is that a target has been detected (Yes), the vehicle control device Sa then executes process S3a. On the other hand, if the result of the above determination is that a target has not been detected (No), the vehicle control device Sa then executes process S6.

[0110] In this process S3a, the vehicle control device Sa determines a safety set X by using the subset calculation unit 42a of the control processing unit 4a, by setting a safety region between the target detected by the target detection unit 1 and the vehicle (host vehicle) VC, and determines a subset Viab(X) by calculating a Viability Kernel. In the example shown in FIG. 2, the subset calculation unit 42a determines a safety set X=(-∞, 0) by setting a range from an obstacle Ob detected as a target by the target detection unit 1 to the vehicle (host vehicle) VC as a safety region, and determines a subset V given by Equation 23. ｘ Ask for.

[0111] Next, the vehicle control device Sa determines whether or not the subset V obtained by the subset calculation unit 42a is in the inclusion / inclusion determination unit 43a of the control processing unit 4a. ｘ , the initial state x of the vehicle at the start of control acquired in process S1 ０ As a result of this determination, it is determined whether the subset V ｘ , the initial state x ０ If it is determined that the subset V belongs to the vehicle control device Sa (Yes), the vehicle control device Sa then executes the process S5a. ｘ , the initial state x ０ If it is determined that the vehicle control device Sa does not belong to the group (No), the vehicle control device Sa then executes the process S8a.

[0112] In this process S5a, the magnitude relationship between I and J is compared to determine whether or not to perform assist control. In the example shown in FIG. 2, I is expressed by Equation 28-2, and J is expressed by Equation 28-3. If the result of this determination is that I is equal to or greater than J (I≧J, No), the vehicle control device Sa causes the cruise control unit 44a of the control processing unit 4a to execute the non-assist cruise control (S6), and ends this process for the current timing. On the other hand, if I is less than J (I<J, Yes), the vehicle control device Sa causes the cruise control unit 44a of the control processing unit 4a to execute the first mode of cruise control for assist control (S7a), and ends this process for the current timing.

[0113] In the process S8a, the vehicle control device Sa executes the second mode of driving control in the assist control by the driving control unit 44a of the control processing unit 4a, and ends this process at the current timing.

[0114] The non-assist control, the first mode control, and the second mode control are each continued until the next timing. If the first and second modes of driving control have ended by the next timing, the assist control is ended. The same applies to the second and third embodiments described below.

[0115] As described above, the vehicle control device Sa and the vehicle control method implemented therein in the first embodiment generate an assist control input based on the zeroing control barrier function by dividing the subset Viab(X) of the safety set X by the operation input u ｈ Since the vehicle VC is determined based on predetermined constraints related to the control input u, the vehicle VC can be controlled by assist control using the zeroed CBF, taking the constraints into consideration.

[0116] The vehicle control device Sa and the vehicle control method described above determine a safe set X by setting a predetermined safe area between the target (obstacle Ob in the example shown in FIG. 2) detected by the target detection unit 1 and the vehicle (host vehicle) VC, and therefore can determine the safe set X according to the situation around the vehicle (host vehicle) VC.

[0117] The vehicle control device Sa and the vehicle control method are configured such that the inclusion / inclusion determination unit 43a determines the initial state x of the vehicle at the start of control in the subset Viab(X). ０ does not belong to the range, an assist control input u that maximizes the derivative of the zeroing control barrier function h(x) is generated. This lacks safety, but can mitigate the damage.

[0118] Second Embodiment Next, another embodiment will be described. In the second embodiment, V-CBF is applied to a vehicle system represented by a double integral system having input constraints of acceleration a and jerk j. The jerk j is defined as the time derivative of acceleration a.

[0119] 1, the vehicle control device Sb in the second embodiment is mounted on a vehicle VC and includes, for example, a target detection unit 1, a running state detection unit 2, an operation input unit 3, a control processing unit 4b, a memory unit 5b, a first power unit (drive wheel drive unit) 6, a second power unit (steering wheel drive unit) 7, a braking unit 8, and an interface unit (IF unit) 9. The vehicle control method in the second embodiment is implemented in this vehicle control device Sb.

[0120] The target detection unit 1, running state detection unit 2, operation input unit 3, first power unit (driving wheel drive unit) 6, second power unit (steering wheel drive unit) 7, braking unit 8 and IF unit 9 in the vehicle control device Sb of the second embodiment are similar to the target detection unit 1, running state detection unit 2, operation input unit 3, first power unit (driving wheel drive unit) 6, second power unit (steering wheel drive unit) 7, braking unit 8 and IF unit 9 in the vehicle control device Sa of the first embodiment, respectively, and therefore their description will be omitted.

[0121] The control processing unit 4b and its functional configuration 41, 42b, 43b, 44b, and the memory unit 5b and its functional configuration 51b in the vehicle control device Sb of the second embodiment are basically the same as the control processing unit 4a and its functional configuration 41, 42a, 43a, 44a, and the memory unit 5a and its functional configuration 51a in the vehicle control device Sa of the first embodiment, except that the target systems are different and therefore the constraints and formulas are different.Therefore, only the above-mentioned differences will be described and the remaining explanations will be omitted.

[0122] The vehicle system that is the target system of the vehicle control device Sb in the second embodiment is expressed by the following formula 29-1, and F ｒ is given by the following equation 29-2, the input constraint of the acceleration a is expressed by the following equation 29-3, and the input constraint of the jerk j is expressed by the following equation 29-4. The jerk j is a parameter that affects the ride comfort and the safety of the occupants, and in this embodiment, the constraints are set as described above.

[0123]

[0124] Here, x １ represents the position of the vehicle VC [m], and x ２ represents the speed [m / s] of the vehicle VC, and the state x of the vehicle is expressed as x = [x １ , x ２ ] Ｔ t represents time [s], and u represents the control input, here the required acceleration [m / s ２ ], and u ｈ is the feedforward input (operation input) [m / s ２ ]. The feedforward input (operation input) u ｈ satisfies the jerk constraint or the control period is sufficiently short with respect to the acceleration change. ｒ represents the running resistance [N], M represents the mass [kg] of the vehicle VC, and R ０ , R １ , R ２ represents the rolling resistance coefficient. ｍｉｎ represents the lower limit of acceleration, and a ｍａｘ represents the upper limit of acceleration. ｍｉｎ represents the lower limit of the jerk, and j ｍａｘ represents the upper limit of the jerk.

[0125] 4 is a diagram illustrating a target system in the second embodiment. In this vehicle system, as shown in FIG. 4, a vehicle VC moves in the direction of an obstacle Ob at a speed x ２ The running resistance F ｒ This direction of travel is x １ The initial position of the vehicle is the coordinate origin, and the position C ｗAssume that the obstacle Ob is placed in a stationary state so that the boundary of the obstacle Ob is located at .

[0126] The obstacle Ob is, for example, a target detected by the target detection unit 1. In the second embodiment, the traveling state detection unit 2 includes at least the speed detection unit (vehicle speed sensor) and the acceleration detection unit (acceleration sensor). The storage unit 5b stores the mass M of the vehicle (host vehicle) VC from Equation 29-1 as the various predetermined data, and the running resistance coefficient R ０ , R １ , R ２ The mass M of the vehicle (host vehicle) VC may be measured by a weight sensor provided on the vehicle VC. ０ , R １ , R ２ is determined in advance as appropriate from a plurality of samples, for example. The constraints of Equation 29-3 and Equation 29-4 are stored in the constraint information storage unit 51b. Note that if there is a gradient on the road surface, a force corresponding to the gradient acts on the vehicle (host vehicle) VC, and therefore the traveling state detection unit 2 may be provided with an inclination detection unit (gradient detection unit, inclination sensor) that detects the inclination of the vehicle (host vehicle) VC with respect to the horizontal.

[0127] In the vehicle system shown in Figure 4, the behavior of the vehicle VC is divided into two patterns, a first and a second, until the vehicle VC stops. In the first pattern, the acceleration a is lower than its lower limit a ｍｉｎ In the second pattern, the deceleration slows down and the vehicle stops before the acceleration a reaches its lower limit a ｍｉｎ Whether the behavior of the vehicle VC falls into the first or second pattern can be determined by a discriminant (first discriminant) D1 expressed by the following equation 30. When the value of the first discriminant D1 is equal to or less than 0 (non-positive, D1≦0), the behavior of the vehicle VC falls into the first pattern, and when the value of the first discriminant D1 is greater than 0 (positive, D1>0), the behavior of the vehicle VC falls into the second pattern.

[0128]

[0129] When the behavior of the vehicle VC is the first pattern, the acceleration at a certain point in time is a ０ , speed x ２０ When the vehicle VC's behavior is the first pattern, the travel distance d [m] until the vehicle VC stops within the input constraints is given by the following equations 31-1 to 31-3, and when the vehicle VC's behavior is the second pattern, the travel distance d [m] until the vehicle VC stops within the input constraints is given by the following equations 32-1 to 32-5.

[0130]

[0131]

[0132] As an example, Fig. 5 shows a graph of acceleration change over time and a graph of speed change over time when the vehicle behavior is in the first pattern. Fig. 6 shows a graph of acceleration change over time and a graph of speed change over time when the vehicle behavior is in the second pattern. The horizontal axis of each of Figs. 5 and 6 represents elapsed time, the upper side of each vertical axis represents acceleration a, and the lower side of each vertical axis represents speed x. ２ is.

[0133] The boundary position of the obstacle Ob is C ｗ Therefore, the Viability Kernel is given by the following equation 33, and the function h(x) is given by the following equation 34. As described above, the travel distance d(x) in equations 33 and 34 is expressed by one of the following equations 31-1 to 31-3 and 32-1 to 32-5 depending on the behavior pattern of the vehicle VC.

[0134]

[0135]

[0136] Lie differential L ｆ h and L ｇ When the behavior of the vehicle VC is in the first pattern, h is given by the following equations 35-1 and 35-2, and when the behavior of the vehicle VC is in the second pattern, h is given by the following equations 36-1 and 36-2.

[0137]

[0138]

[0139] Therefore, the control input u is expressed by the following equations 37-1 to 37-3.

[0140]

[0141] In the second embodiment, the subset calculation unit 42b calculates the Viability Kernel of Equation 33 as a subset Viab(X) of the safety set X. The inclusion / non-inclusion determination unit 43b adds the initial state x of the vehicle at the start of control to this subset Viab(X). ０ Then, the traveling control unit 44b determines whether the initial state x belongs to the subset Viab(X). ０ If the assist control is not performed, the magnitude relationship between I given by the equation 37-2 and J given by the equation 37-3 is compared. If the assist control is not performed, the non-assist control driving control is performed. If the assist control is performed, the operation input u ｈ , and execute the first mode of driving control based on the assist control input u given by Equation 37-1, and add the initial state x ０ does not belong to the category, the second mode of travel control based on the zeroing control barrier function h(x) of Equation 34 is executed.

[0142] The operation of the second embodiment will be described. In Fig. 3, at this timing, the vehicle control device Sb first executes processing S1 similar to that of the first embodiment by the control unit 41 of the control processing unit 4b, and then executes processing S2. If the result of the determination in processing S2 is that a target object has been detected (Yes), the vehicle control device Sb then executes processing S3b. On the other hand, if the result of the determination is that a target object has not been detected (No), the vehicle control device Sb then executes processing S6 similar to that of the first embodiment.

[0143] In this process S3b, the vehicle control device Sb sets a safety area between the target detected by the target detection unit 1 and the vehicle (host vehicle) VC by the subset calculation unit 42b of the control processing unit 4b, and calculates a safety set X=(−∞, C ｗ ) and obtain the subset Viab(X) by obtaining the Viability Kernel of Equation 33.

[0144] Next, the vehicle control device Sb uses the inclusion / inclusion determination unit 43a of the control processing unit 4b to add the initial state x of the vehicle at the start of control acquired in the process S1 to the subset Viab(X) calculated by the subset calculation unit 42b. ０ As a result of this determination, it is determined whether the initial state x belongs to the subset Viab(X) (S4b). ０ If it is determined that the initial state x belongs to the subset Viab(X) (Yes), the vehicle control device Sb then executes the process S5b. ０ If it is determined that the vehicle control device Sb does not belong to the group (No), the vehicle control device Sb then executes processing S8b.

[0145] In this process S5b, whether or not to perform assist control is determined by comparing the magnitude relationship between I given by equation 37-2 and J given by equation 37-3. If the result of this determination is that I is equal to or greater than J (I≧J, No), the vehicle control device Sb causes the driving control unit 44b of the control processing unit 4b to execute the non-assist control driving control (S6), and ends this process at this timing. On the other hand, if I is less than J (I<J, Yes), the vehicle control device Sb causes the driving control unit 44b to execute the operation input u ｈ , and the assist control input u given by equation 37-1, the first mode of travel control is executed (S7a), and the process for this timing is terminated.

[0146] In the process S8b, the vehicle control device Sb executes the second mode of driving control based on the zeroing control barrier function h(x) of Equation 34 by the driving control unit 44b, and ends this process at the current timing.

[0147] The non-assist control, the first mode control, and the second mode control are each continued until the next timing.

[0148] An example of a simulation result will be described. FIG. 7 is a graph showing an example of a simulation result in the target system in the second embodiment. FIG. 7A is a graph showing a change in position over time, with the horizontal axis representing elapsed time [s] and the vertical axis representing position [m]. FIG. 7B is a graph showing a change in velocity over time, with the horizontal axis representing elapsed time [s] and the vertical axis representing velocity [m / s]. FIG. 7C is a graph showing a change in acceleration over time, with the horizontal axis representing elapsed time [s] and the vertical axis representing acceleration [m / s]. ２ FIG. 7D is a graph showing the change in jerk over time, with the horizontal axis representing elapsed time [s] and the vertical axis representing jerk [m / s ３ In this simulation, each value is C ｗ =10 [m], initial position x １０ =0 [m], initial speed x ２０ =5 [m / s], initial acceleration a ０ = 3 [m / s ２ ], the upper limit of acceleration a ｍａｘ = 3 [m / s ２ ], the lower limit of acceleration a ｍｉｎ = -3 [m / s ２ ], upper limit of jerk j ｍａｘ = 3.92 [m / s ３ ], the lower limit of jerk j ｍｉｎ =-3.92 [m / s ３ 7, it can be seen that the vehicle VC is safely stopped while complying with the acceleration constraints and the jerk constraints.

[0149] The vehicle control device Sb and the vehicle control method implemented therein in the second embodiment have the same effects as the vehicle control device Sa and the vehicle control method implemented therein in the first embodiment.

[0150] According to the second embodiment, a vehicle control device Sb and a vehicle control method can be provided that use assist control using a zeroed CBF, taking into account the first constraint on acceleration a (constraint on acceleration a) and the second constraint on jerk j (constraint on jerk j) as constraints.

[0151] Third Embodiment Next, another embodiment will be described. In the third embodiment, the V-CBF is applied to a vehicle system represented by a double integral system having input constraints of the steering angle φ and the steering velocity Ψ as a target system. The steering velocity Ψ is defined as the time derivative of the steering angle φ.

[0152] 1, the vehicle control device Sc in the third embodiment is mounted on a vehicle VC and includes, for example, a target detection unit 1, a running state detection unit 2, an operation input unit 3, a control processing unit 4c, a memory unit 5c, a first power unit (drive wheel drive unit) 6, a second power unit (steering wheel drive unit) 7, a braking unit 8, and an interface unit (IF unit) 9. The vehicle control method in the third embodiment is implemented in this vehicle control device Sc.

[0153] The target detection unit 1, running state detection unit 2, operation input unit 3, first power unit (driving wheel drive unit) 6, second power unit (steering wheel drive unit) 7, braking unit 8 and IF unit 9 in the vehicle control device Sc of the third embodiment are similar to the target detection unit 1, running state detection unit 2, operation input unit 3, first power unit (driving wheel drive unit) 6, second power unit (steering wheel drive unit) 7, braking unit 8 and IF unit 9 in the vehicle control device Sa of the first embodiment, respectively, and therefore their description will be omitted.

[0154] The control processing unit 4c and its functional configuration 41, 42c, 43c, 44c, and the memory unit 5c and its functional configuration 51c in the vehicle control device Sc of the third embodiment are basically the same as the control processing unit 4a and its functional configuration 41, 42a, 43a, 44a, and the memory unit 5a and its functional configuration 51a in the vehicle control device Sa of the first embodiment, except that the target systems are different and therefore the constraints and formulas are different.Therefore, only the differences will be described and the remaining explanations will be omitted.

[0155] The vehicle system that is the target system of the vehicle control device Sc in the third embodiment is expressed by the following equation 38-1, the input constraint of the steering angle Φ is expressed by the following equation 38-2, and the input constraint of the steering angular velocity Ψ is expressed by the following equation 38-3. The steering angular velocity Ψ is a parameter that affects the ride comfort, and in this embodiment, the input constraint is set as described above.

[0156]

[0157] Here, x １ is the x of the vehicle VC １ represents the coordinate position [m], and y １ is the y of the vehicle VC １ represents the coordinate position [m], and x ２ represents the speed [m / s] of the vehicle VC. Here, since it is two-dimensional, x ２ is a vector (x ２ = (dx １ / dt, dy １ / dt) where θ represents the attitude angle [rad] of the vehicle VC, φ represents the steering angle [rad], and Ψ represents the steering angle speed [rad / s]. t represents time [s], u represents the control input, here the required steering angle [rad], and u ｈ represents the feedforward input (operation input) [rad]. ｍｉｎ represents the lower limit of the steering angle, and φ ｍａｘ represents the upper limit of the steering angle. ｍｉｎ represents the lower limit of the steering angular velocity, and ψ ｍａｘ represents the upper limit of the steering angle velocity. The vehicle state is determined by the vehicle position (x １ , y １ ), speed x ２ and the attitude angle θ.

[0158] 8 is a diagram illustrating a target system in the third embodiment. In this vehicle system, as shown in FIG. 8, a vehicle VC moves in the direction of an obstacle Ob at a speed x ２ The vehicle VC has a width W and a length L ｆ +L+L ｒ It is composed of the first car length L ｆ is the length from the front end of the vehicle VC to the center of the front wheels, the second vehicle length L is the length from the center of the front wheels to the center of the rear wheels, and the third vehicle length L ｒ is the length from the center of the rear wheel to the rear end of the vehicle VC. In this target system, a Cartesian coordinate system is set with the initial position of the vehicle as the coordinate origin, and the position y １ =C ｗ The boundary of the obstacle Ob is located at x １Assume that the obstacle Ob is stationary and positioned along the x axis. The attitude angle θ is １ Axis and vector velocity x ２ The steering angle φ is the angle formed by the vector velocity x ２ (vehicle length direction) and the steering wheel (front wheel in this case). For simplicity, the extension direction of the obstacle Ob is x １ The direction of the axis is x １ The direction of the axis may be arbitrary, and in this case, the same explanation can be given by considering the component in the extension direction and the component in the direction perpendicular thereto.

[0159] The obstacle Ob is, for example, a target detected by the target detection unit 1. In the third embodiment, the traveling condition detection unit 2 includes at least the speed detection unit (vehicle speed sensor), the attitude angle detection unit (yaw rate sensor), and the steering angle detection unit (first steering angle sensor). The storage unit 5c stores the second vehicle length L of the vehicle (host vehicle) VC from Equation 38-1 as the various predetermined data. The constraint information storage unit 51b stores the constraints of Equations 38-2 and 38-3.

[0160] In the vehicle system shown in Figure 8, if the traveling vehicle VC becomes parallel to the obstacle Ob, it can avoid a collision with the obstacle Ob by steering, so the behavior of the vehicle VC is divided into two patterns, a third and a fourth. In the third pattern, steering begins before the steering angle φ reaches its constraint, and the vehicle VC becomes parallel to the obstacle Ob. In the fourth pattern, steering begins after the steering angle φ reaches its constraint, and the vehicle VC becomes parallel to the obstacle Ob. Whether the behavior of the vehicle VC falls into either the third or fourth pattern can be determined by a discriminant (second discriminant) D2 expressed by the following equation 39, and the value of the second discriminant D2 is determined based on the lower limit value φ of the steering angle: ｍｉｎ If D2≧φ ｍｉｎ ), the behavior of the vehicle VC is in the fourth pattern, and the value of the second discriminant D2 is equal to or smaller than the lower limit value φ of the steering angle. ｍｉｎ If D2<φ ｍｉｎ ), the behavior of the vehicle VC is in the fourth pattern.

[0161]

[0162] When the behavior of the vehicle VC is the third pattern, the attitude angle at a certain point in time is θ ０ , the steering angle is φ ０ When the vehicle VC's behavior is the third pattern, the travel distance d [m] until the attitude angle θ becomes parallel to the obstacle Ob within the input constraints is given by the following equations 40-1 to 40-3, and when the vehicle VC's behavior is the fourth pattern, the travel distance d [m] until the attitude angle θ becomes parallel to the obstacle Ob is given by the following equations 41-1 to 41-4.

[0163]

[0164]

[0165] As an example, Fig. 9 shows a graph of the change in steering angle velocity over time and a graph of the change in steering angle over time when the vehicle behavior is Pattern 3. Fig. 10 shows a graph of the change in steering angle velocity over time and a graph of the change in steering angle over time when the vehicle behavior is Pattern 4. The horizontal axis of each of Figs. 9 and 10 represents elapsed time, the upper side of each vertical axis represents the steering angle velocity ψ, and the lower side of each vertical axis represents the steering angle φ.

[0166] The boundary position of the obstacle Ob is y = C ｗ Therefore, the Viability Kernel is given by the following equation 42, and the function h(x) is given by the following equation 43. As described above, the travel distance d(x) in equations 42 and 43 is expressed by one of the following equations 40-1 to 40-3 and 41-1 to 41-4 depending on the behavior pattern of the vehicle VC.

[0167]

[0168]

[0169] Lie differential L ｆ h and L ｇ h are given by the following equations 44-1 and 44-2, respectively, and must be calculated according to the behavior pattern of the vehicle VC.

[0170]

[0171] Therefore, the control input u is expressed by the following equations 45-1 to 45-3.

[0172]

[0173] In the third embodiment, the subset operation unit 42c calculates the safe set (y=C ｗ The Viability Kernel of Equation 42 is found as a subset Viab of the region (the set of regions on the -y side from the y-axis). The belonging / non-belonging determination unit 43c determines whether the initial state of the vehicle at the start of control belongs to this subset Viab. If the initial state belongs to the subset Viab, the cruise control unit 44c compares the magnitude relationship between I given by Equation 45-2 and J given by Equation 45-3, and if assist control is not to be performed, executes non-assisted cruise control, and if assist control is to be performed, executes non-assisted cruise control. ｈ , and executes a first mode of driving control based on the assist control input u given by equation 45-1, and if the initial state does not belong to the subset Viab, executes a second mode of driving control based on the zeroing control barrier function h(x) of equation 43.

[0174] The operation of the third embodiment will be described. In Fig. 3, at this timing, the vehicle control device Sc first executes processing S1 similar to that of the first embodiment by using the control unit 41 of the control processing unit 4c, and then executes processing S2. If the result of the determination in processing S2 is that a target object has been detected (Yes), the vehicle control device Sc then executes processing S3c. On the other hand, if the result of the determination is that a target object has not been detected (No), the vehicle control device Sc then executes processing S6 similar to that of the first embodiment.

[0175] In this process S3c, the vehicle control device Sc sets a safety area between the target detected by the target detection unit 1 and the vehicle (host vehicle) VC by the subset calculation unit 42c of the control processing unit 4c, thereby calculating a safety set (y=C ｗ The subset Viab is obtained by finding the Viability Kernel of Equation 42.

[0176] Next, the vehicle control device Sc determines whether the initial state of the vehicle at the start of control acquired in process S1 belongs to the subset Viab calculated by the subset calculation unit 42c using the belonging / non-belonging determination unit 43c of the control processing unit 4c (S4c). If it is determined that the initial state belongs to the subset Viab (Yes), the vehicle control device Sc then executes process S5c. On the other hand, if it is determined that the initial state does not belong to the subset Viab (No), the vehicle control device Sb then executes process S8c.

[0177] In this process S5c, a comparison is made between I given by equation 45-2 and J given by equation 45-3 to determine whether or not to perform assist control. If the result of this determination is that I is equal to or greater than J (I≧J, No), the vehicle control device Sc executes the non-assist control driving control (S6) using the driving control unit 44c of the control processing unit 4c, and ends this process at this timing. On the other hand, if I is less than J (I<J, Yes), the vehicle control device Sc executes the non-assist driving control using the driving control unit 44c based on the operation input u ｈ , and the assist control input u given by equation 45-1, the first mode of travel control is executed (S7c), and the process for this timing is terminated.

[0178] In the process S8c, the vehicle control device Sc executes the second mode of driving control based on the zeroing control barrier function h(x) of Equation 43 by the driving control unit 44c, and ends this process at the current timing.

[0179] The non-assist control, the first mode control, and the second mode control are each continued until the next timing.

[0180] An example of a simulation result will be described. Fig. 11 is a graph showing an example of a simulation result in the target system according to the third embodiment. １ The graph shows the change in position over time on the x-axis, with the horizontal axis representing elapsed time [s] and the vertical axis representing １ The position [m] on the y axis. １Graph showing the change in position over time on the axis, the horizontal axis being elapsed time [s] and the vertical axis being y １ 11C is a graph showing the time change of the attitude angle, the horizontal axis of which is elapsed time [s] and the vertical axis of which is the attitude angle [rad]. FIG. 11D is a graph showing the time change of the steering angle, the horizontal axis of which is elapsed time [s] and the vertical axis of which is the steering angle [rad]. FIG. 11E is a graph showing the time change of the steering angular velocity, the horizontal axis of which is elapsed time [s] and the vertical axis of which is the steering angular velocity [rad / s]. In this simulation, each value is C ｗ =4 [m], initial position x １０ = 0, y １０ =0 [m], initial speed x ２０ = 4 [m / s], y ２０ =0 [m / s], initial attitude angle θ ０ = π / 6 [rad], upper limit of steering angle φ ｍａｘ = π / 4 [rad], lower limit of steering angle φ ｍｉｎ =-π / 4 [rad], upper limit value of steering angular velocity ψ ｍａｘ =π / 24 [rad / s], lower limit value of steering angular velocity ψ ｍｉｎ 11, it can be seen that the vehicle VC is safely parallel to the obstacle Ob while the steering angle constraints and steering angle speed constraints are observed.

[0181] The vehicle control device Sc and the vehicle control method implemented therein in the third embodiment have the same effects as the vehicle control device Sa and the vehicle control method implemented therein in the first embodiment.

[0182] According to the third embodiment, it is possible to provide a vehicle control device Sc and a vehicle control method that use assist control using a zeroing CBF, taking into account the third constraint on the steering angle φ (constraint on the steering angle φ) and the fourth constraint on the steering angular velocity ψ (constraint on the steering angular velocity ψ) as constraints.

[0183] This specification discloses various aspects of the technology as described above, but the main technologies among them are summarized below.

[0184] A vehicle control device according to one aspect generates an assist control input based on a zeroing control barrier function in response to an operation input that changes a driving state of a vehicle, and controls the driving state based on the operation input and the assist control input. The vehicle control device includes a subset calculation unit that calculates a subset of a safety set that represents a predetermined safety region that is considered safe based on the operation input and predetermined constraints related to the control input, a membership determination unit that determines whether an initial state of the vehicle at the start of control belongs to the subset calculated by the subset calculation unit, and a driving control unit that controls the driving state based on the operation input and the assist control input when the membership determination unit determines that the initial state belongs to the subset.

[0185] Note that the operational input includes not only the operational input by the driver but also the operational input by a driving system such as an automated driving system, as described above, and therefore is referred to as "assist control" rather than "human assist control." Therefore, the principle of "assist control using zeroing CBF (assist control using zero-type CBF)" is the same as the principle of "human assist control using zeroing CBF (human assist control using zero-type CBF)."

[0186] When generating an assist control input based on a zeroing control barrier function, such a vehicle control device determines a subset of the safe set based on predetermined constraints related to the operation input and the control input, so that the vehicle can be controlled using assist control using a zeroing CBF that takes the constraints into consideration.

[0187] Preferably, the above-mentioned vehicle control device further includes a target detection unit that detects targets present around the vehicle and detects positions of the detected targets relative to the vehicle (relative target positions), and the subset calculation unit further determines the safety set by setting the safety region between the targets detected by the target detection unit and the vehicle. Preferably, the target detection unit further detects at least one of a speed of the detected target relative to the vehicle (relative target speed) and an acceleration of the detected target relative to the vehicle (relative target acceleration). Preferably, in the above-mentioned vehicle control device, the constraints are constraints related to the running state of the vehicle. Preferably, in the above-mentioned vehicle control device, the constraints include at least one of a first constraint on acceleration (including positive acceleration and negative acceleration), a second constraint on jerk which is the time derivative of acceleration, a third constraint on steering angle, and a fourth constraint on steering angular velocity which is the time derivative of steering angle. Preferably, in the above-mentioned vehicle control device, the subset calculation unit calculates the subset by calculating a Viability Kernel. Preferably, the above-mentioned vehicle control device further includes a driving condition detection unit that detects a driving condition of the vehicle. Preferably, the driving condition detection unit includes at least one of a speed detection unit that detects the speed (vehicle speed) of the vehicle, an acceleration detection unit that detects the acceleration of the vehicle, an attitude angle detection unit that detects the attitude angle of the vehicle, and a steering angle detection unit that detects the steering angle of the vehicle. Preferably, the subset calculation unit sets a coordinate system having a coordinate origin that is a boundary of the target detected by the target detection unit, and sets a range from the boundary of the target to the vehicle as a safety region. Preferably, the subset calculation unit sets a coordinate system having a coordinate origin that is a current vehicle position, and sets a coordinate origin that is a range from the boundary of the target to the current vehicle position as a safety region. Preferably, the above-mentioned vehicle control device further includes a target detection unit that detects targets present around the vehicle and detects the position of the detected targets relative to the vehicle, and a speed detection unit that detects the speed of the vehicle, and the initial state of the vehicle is represented by the initial position and initial speed of the vehicle.Preferably, the above-mentioned vehicle control device further includes a target detection unit that detects targets present around the vehicle and detects the position of the detected targets relative to the vehicle, a speed detection unit that detects the speed of the vehicle, and an attitude angle detection unit that detects the attitude angle of the vehicle, and an initial state of the vehicle is represented by an initial position, initial speed, initial acceleration, and initial attitude angle of the vehicle. Preferably, the above-mentioned vehicle control device further includes a first power unit that drives drive wheels of the vehicle and a braking unit that brakes the vehicle, and when accelerating or decelerating the vehicle to change the running state of the vehicle, the traveling control unit controls the running state by controlling at least one of the first power unit and the braking unit. Preferably, the above-mentioned vehicle control device further includes a second power unit that drives steering wheels of the vehicle, and when changing the steering angle of the vehicle to change the running state of the vehicle, the traveling control unit controls the running state by controlling the second power unit.

[0188] In another aspect, the above-mentioned vehicle control device further includes a target detection unit that detects targets present around the vehicle and detects the position of the vehicle relative to the detected targets, and the subset calculation unit further determines the safety set by setting the safety area between the targets detected by the target detection unit and the vehicle.

[0189] Such a vehicle control device is equipped with a target detection unit, and determines the safety set by setting the safety area between the target detected by the target detection unit and the vehicle, so that the safety set can be determined according to the situation around the vehicle.

[0190] In another aspect, in the above-described vehicle control device, when the belongingness determination unit determines that the initial state does not belong to the subset, the traveling control unit generates an assist control input that maximizes the derivative of the zeroing control barrier function, and controls the traveling state based on the operation input and the assist control input.

[0191] Such a vehicle control device generates an assist control input that maximizes the derivative of the zeroing control barrier function when the belonging / non-belonging determination unit determines that the initial state does not belong to the subset, so although it lacks safety, it can mitigate the damage.

[0192] In another aspect, in the above-described vehicle control device, the constraints are a first constraint on acceleration and a second constraint on jerk, which is a time derivative of acceleration.

[0193] This makes it possible to provide a vehicle control device that performs assist control using zeroed CBF, taking into consideration the first constraint on acceleration (constraint on acceleration) and the second constraint on jerk (constraint on jerk) as constraints.

[0194] In another aspect, in the above-described vehicle control device, the constraints are a third constraint on the steering angle and a fourth constraint on the steering angular velocity, which is a time derivative of the steering angle.

[0195] This makes it possible to provide a vehicle control device that performs assist control using a zeroed CBF, taking into consideration the third constraint on the steering angle (constraint on the steering angle) and the fourth constraint on the steering velocity (constraint on the steering velocity) as constraints.

[0196] According to another aspect, a vehicle control method generates an assist control input based on a zeroing control barrier function for an operation input that changes a vehicle running state, and controls the vehicle running state based on the operation input and the assist control input. This vehicle control method includes a subset calculation step of obtaining a subset of a safety set that represents a predetermined safety region that is considered safe based on the operation input and predetermined constraints related to the control input, a membership determination step of determining whether an initial state of the vehicle at the start of control belongs to the subset obtained in the subset calculation step, and a running control step of controlling the vehicle running state based on the operation input and the assist control input when it is determined in the membership determination step that the initial state belongs to the subset.

[0197] In this vehicle control method, when generating an assist control input based on a zeroing control barrier function, a subset of the safe set is determined based on predetermined constraints related to the operation input and the control input, so that the vehicle can be controlled by assist control using a zeroing CBF that takes the constraints into consideration.

[0198] In order to express the present invention, the present invention has been properly and sufficiently described above through the embodiments with reference to the drawings, but it should be recognized that those skilled in the art can easily change and / or improve the above-mentioned embodiments. Therefore, unless the changes or improvements made by those skilled in the art are at a level that causes departure from the scope of the claims described in the claims, such changes or improvements are interpreted as being included in the scope of the claims.

[0199] According to the present invention, a vehicle control device and a vehicle control method for controlling the running of a vehicle can be provided.

Claims

1. A vehicle control device that generates an assist control input based on a nulling control barrier function for an operation input that changes the driving state of a vehicle, and controls the driving state based on the operation input and the assist control input, a subset operation unit that obtains a subset of a safety set representing a predetermined safe region based on predetermined constraints related to the operation input and the assist control input, a membership determination unit that determines whether an initial state of the vehicle at the start of control belongs to the subset obtained by the subset operation unit, and a travel control unit that controls the travel state based on the operation input and the assist control input when the membership determination unit determines that the initial state belongs to the subset, wherein when the membership determination unit determines that the initial state does not belong to the subset, the travel control unit generates an assist control input that maximizes the derivative of the nulling control barrier function, and controls the travel state based on the operation input and the assist control input. Vehicle control device.

2. The vehicle control device further includes an object detection unit that detects an object existing around the vehicle and detects the position of the detected object with respect to the vehicle, wherein the subset operation unit further obtains the safety set by setting the safe region between the vehicle and the object detected by the object detection unit. The vehicle control device according to claim 1.

3. The constraints are a first constraint on acceleration and a second constraint on jerk, which is the time derivative of acceleration. The vehicle control device according to claim 1 or claim 2.

4. The constraints are a third constraint on steering angle and a fourth constraint on steering angular velocity, which is the time derivative of steering angle. The vehicle control device according to claim 1 or claim 2.