Rudder angle control device and rudder angle control method

Abstract

If the vehicle's position deviates from the target path while traveling along it, the driver must appropriately correct the steering. [Solution] The steering angle control device 161 controls the steering angle of a vehicle to follow a target path and includes a prediction unit 161B1 that predicts the behavior of the vehicle, including the deviation of the vehicle's position and direction from the target path, up to a predetermined time in advance, based on vehicle state quantities indicating the vehicle's motion state and information indicating the target path, and an optimization unit 161B2 that optimizes the target steering angle using the prediction results obtained by the prediction unit 161B1. The prediction unit 161B1 has a path deviation model that models the relationship between the deviation of the vehicle's position and direction from the target path and the curvature of the target path, including the steady-state relationship between the vehicle state quantities and vehicle characteristic information relating to the vehicle's motion characteristics.

Description

[Technical Field]

[0001] The present invention relates to a steering angle control device that controls the steering angle of a vehicle based on a vehicle model that models the motion state of the vehicle. [Background technology]

[0002] A technique is known for setting a target trajectory at an intersection and controlling steering, braking, and driving to travel along that target trajectory (see Patent Document 1). [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2019-167039 [Overview of the project] [Problems that the invention aims to solve]

[0004] The technology described in Patent Document 1 discloses how steering and braking / driving are controlled to travel according to a set target trajectory. However, details are not disclosed, and for example, the control of the steering when traveling along the target trajectory without error is not clarified, leaving room for improvement. Following a target trajectory improves traffic safety while minimizing the decline in traffic flow. This, in turn, can contribute to the development of a sustainable transportation system. [Means for solving the problem]

[0005] A first aspect of the present invention is a steering angle control device that controls the steering angle of a vehicle to follow a target path, comprising: a prediction unit that predicts the behavior of the vehicle, including the deviation of the vehicle's position and direction from the target path, up to a predetermined time in advance, based on vehicle state quantities indicating the vehicle's motion state and information indicating the target path; and an optimization unit that optimizes the target steering angle using the prediction results obtained by the prediction unit, wherein the prediction unit has a path deviation model that models the relationship between the deviation of the vehicle's position and direction from the target path and the curvature of the target path, including the steady-state relationship between the vehicle state quantities and vehicle characteristic information relating to the vehicle's motion characteristics. A second aspect of the present invention is a steering angle control method for controlling the steering angle of a vehicle to follow a target path, which involves predicting the behavior of the vehicle, including the deviation of the vehicle's position and direction from the target path, up to a predetermined time in advance, based on vehicle state variables indicating the vehicle's motion state and information indicating the target path, and optimizing the target steering angle using the prediction results obtained from the prediction, wherein the prediction uses a path deviation model that models the relationship between the deviation of the vehicle's position and direction from the target path and the curvature of the target path, including the steady-state relationship between the vehicle state variables and vehicle characteristic information relating to the vehicle's motion characteristics. [Effects of the Invention]

[0006] According to the present invention, when the vehicle's behavior is considered to be substantially steady state, it becomes possible to reduce the computational load compared to conventional methods without compromising the accuracy of predicting the vehicle's behavior, including deviations in the vehicle's position and direction relative to the target path. As a result, it becomes possible to implement the system in lower-spec ECUs and the like. [Brief explanation of the drawing]

[0007] [Figure 1] A diagram illustrating the configuration of a vehicle control system having a steering angle control device according to an embodiment of the present invention. [Figure 2] A block diagram showing the main components of the rudder angle control device. [Figure 3A] A block diagram illustrating the flow of rudder angle control. [Figure 3B] A block diagram illustrating the details of the rudder angle control unit in Figure 3A. [Figure 4A] A schematic diagram showing an example of a predicted path that follows the target path. [Figure 4B] A schematic diagram showing the relationship between a two-wheeled vehicle model and a target route according to the embodiment. [Figure 5] A flowchart illustrating an example of the arithmetic processing performed by the arithmetic unit based on the program. [Modes for carrying out the invention]

[0008] Embodiments of the invention will be described below with reference to the drawings. A steering angle control device according to one embodiment of the present invention controls the steering angle of the vehicle's steering device (e.g., power steering device) so that the vehicle follows a target path (which may also be called a target trajectory). The steering angle control device can be applied to, for example, a vehicle with an autonomous driving function, i.e., an autonomous vehicle. The steering angle control device according to this embodiment is applicable to both manually driven vehicles with driver assistance functions and autonomous vehicles, but for the sake of explanation, the following example will be given to the case where it is applied to an autonomous vehicle. Furthermore, in this embodiment, the vehicle equipped with the steering angle control device may be referred to as "the vehicle itself" to distinguish it from other vehicles. The vehicle itself may be an engine-powered vehicle with an internal combustion engine as its driving source, an electric vehicle with a drive motor as its driving source, or a hybrid vehicle with both an engine and a drive motor as its driving sources. The vehicle itself can be driven not only in an automated driving mode that does not require driver operation, but also in a manual driving mode with driver operation.

[0009] <Vehicle Configuration> First, the general configuration of the vehicle related to autonomous driving will be described. Figure 1 is a block diagram illustrating the configuration of the vehicle control system 100 of the vehicle having a steering angle control device according to the embodiment. As shown in Figure 1, the vehicle control system 100 mainly comprises a controller 10, a group of external sensors 1 and 2, an internal sensor group 2, an input / output device 3, a positioning unit 4, a map database 5, a navigation device 6, a communication unit 7, and an actuator AC for driving.

[0010] External sensor group 1 is a collective term for multiple sensors (external sensors) that detect external conditions, which are information about the surroundings of the vehicle. External sensor group 1 includes, for example, a lidar that measures the distance from the vehicle to surrounding obstacles by measuring scattered light from the vehicle's omnidirectional illumination; a radar that detects other vehicles and obstacles around the vehicle by emitting electromagnetic waves and detecting reflected waves; and a camera mounted on the vehicle that has an image sensor such as a CCD or CMOS sensor to capture images of the area around the vehicle (front, rear, and sides).

[0011] Internal sensor group 2 is a collective term for multiple sensors (internal sensors) that detect the vehicle's driving status. Internal sensor group 2 includes, for example, a vehicle speed sensor that detects the vehicle's speed, acceleration sensors that detect the vehicle's acceleration in the longitudinal direction (direction of travel) and the acceleration in the lateral direction (lane width direction) (lateral acceleration), a rotation speed sensor that detects the rotation speed of the driving power source, and a yaw rate sensor that detects the rotational angular velocity of the vehicle's center of gravity around the vertical axis. Sensors that detect the driver's operations in manual driving mode, such as accelerator pedal operation, brake pedal operation, and steering wheel operation, are also included in internal sensor group 2.

[0012] Input / output device 3 is a general term for devices that receive commands from the driver or output information to the driver. Input / output device 3 includes, for example, various switches that the driver uses to input commands by operating control members, a microphone that the driver uses to input commands by voice, a display that provides information to the driver via displayed images, and a speaker that provides information to the driver by voice.

[0013] The positioning unit (GNSS unit) 4 has a positioning sensor that receives positioning signals transmitted from positioning satellites. Positioning satellites are artificial satellites such as GPS satellites and quasi-zenith satellites. The positioning unit 4 uses the positioning information received by the positioning sensor to measure the current position (latitude, longitude, altitude) of the vehicle.

[0014] The map database 5 is a device that stores general map information used in the navigation device 6, and is composed of, for example, magnetic disks and semiconductor elements. The map information includes road location information, road shape information (curvature, etc.), and location information of intersections and junctions. Note that the map information stored in the map database 5 is different from the high-precision map information stored in the storage unit 12 of the controller 10.

[0015] The navigation device 6 is a device that, for example, searches for a route along the road to a destination entered by the driver and provides driving guidance along the searched route. The input of the destination and driving guidance along the searched route are performed via the input / output device 3. The route search is performed based on the current position of the vehicle measured by the positioning unit 4, the position of the entered destination, and the map information stored in the map database 5. The current position of the vehicle can also be measured using the detection values ​​of the external sensor group 1, and the route may be searched based on this current position and high-precision map information stored in the storage unit 12.

[0016] The communication unit 7 communicates with various servers (not shown) via a network including wireless communication networks such as the Internet and mobile phone networks, and obtains map information, driving history information, and traffic information from the servers periodically or at arbitrary times. In addition to obtaining driving history information, the communication unit 7 may also transmit its own vehicle's driving history information to the server. The network includes not only public wireless communication networks but also closed communication networks established for each predetermined management area, such as wireless LAN, Wi-Fi (registered trademark), Bluetooth (registered trademark), etc. The acquired map information is output to the map database 5 and the storage unit 12, and the map information is updated.

[0017] Actuator AC is a drive actuator used to control the movement of the vehicle. When the drive source is an engine, actuator AC includes a throttle actuator that adjusts the opening degree (throttle opening) of the engine's throttle valve. When the drive source is a drive motor, the drive motor is included in actuator AC. Brake actuators that operate the vehicle's braking system and steering actuators that drive the steering system are also included in actuator AC.

[0018] The controller 10 is comprised of an electronic control unit (ECU). More specifically, the controller 10 includes a computer having an arithmetic unit 11 such as a CPU (microprocessor), a storage unit 12 such as ROM and RAM, and other peripheral circuits (not shown) such as an I / O interface. Although it is possible to provide multiple ECUs with different functions, such as an ECU for engine control, an ECU for drive motor control, and an ECU for braking system, for convenience, Figure 1 shows the controller 10 as a collection of these ECUs.

[0019] The memory unit 12 stores high-precision, detailed map information for autonomous driving. The high-precision map information includes road location information, road shape information (curvature, etc.), road gradient information, intersection and branching point location information, type and location information of lane markings such as white lines, number of lanes (driving lanes), lane width and location information for each lane (information on the center position of the lane and the boundary lines of the lane positions), location information of landmarks (traffic lights, signs, buildings, etc.) as markers on the map, and road surface profile information such as road surface irregularities. The high-precision map information stored in the memory unit 12 may include high-precision map information acquired from outside the vehicle via the communication unit 7, or it may include high-precision map information created by the vehicle itself using detection values ​​from the external sensor group 1 or detection values ​​from the external sensor group 1 and the internal sensor group 2. The memory unit 12 may also store information such as various control programs and threshold values ​​used in the programs. The calculation unit 11 has a functional configuration that includes a vehicle position recognition unit 13, an external environment recognition unit 14, an action plan generation unit 15, and a driving control unit 16.

[0020] The vehicle position recognition unit 13 recognizes the vehicle's position on the map (vehicle position) based on the vehicle's position information obtained by the positioning unit 4 and the map information in the map database 5. The vehicle position may also be recognized using high-precision map information stored in the storage unit 12 and surrounding information of the vehicle detected by the external sensor group 1, thereby enabling high-precision recognition of the vehicle's position. Furthermore, the vehicle's movement information (direction of movement, distance traveled) can be calculated based on the detection values ​​of the internal sensor group 2, and the vehicle's position can be recognized accordingly. Furthermore, when the vehicle's position can be measured by sensors installed on or beside the road, the vehicle's position can also be recognized by communicating with those sensors via the communication unit 7.

[0021] The external environment recognition unit 14 recognizes the external conditions around the vehicle based on signals from the external sensor group 1, such as cameras, lidars, and radars. For example, it recognizes the position, speed, and acceleration of surrounding vehicles (vehicles in front and behind) traveling around the vehicle, the position of surrounding vehicles that are stopped or parked around the vehicle, and the position and state of other objects, and creates target information. Other objects include signs, traffic lights, roads, buildings, guardrails, utility poles, billboards, pedestrians, and bicycles. Markings on the road surface, such as lane markings (white lines, etc.) and stop lines, are also included in other objects (roads). The state of other objects includes the color of traffic lights (red, blue, yellow), the speed and direction of pedestrians and cyclists, etc. Some of the stationary objects among the other objects constitute landmarks that serve as indicators of location on the map, and the external environment recognition unit 14 also recognizes the location and type of these landmarks.

[0022] The action plan generation unit 15 generates a driving trajectory (target trajectory) for the vehicle from the present time to a predetermined time in advance, based on, for example, the route searched by the navigation device 6, the high-precision map information stored in the memory unit 12, the vehicle's position recognized by the vehicle position recognition unit 13, and the external conditions recognized by the external environment recognition unit 14. If there are multiple possible target trajectories on the route searched by the navigation device 6, the action plan generation unit 15 selects the optimal trajectory from among them that complies with laws and regulations and satisfies criteria such as efficient and safe driving, and generates the selected trajectory as the target route. The action plan generation unit 15 then generates an action plan corresponding to the generated target route. The action plan generation unit 15 generates various action plans corresponding to driving modes such as overtaking to pass a preceding vehicle, changing lanes, following a preceding vehicle, lane keeping to maintain the lane, decelerating, or accelerating. When generating a target route, the action plan generation unit 15 first determines the driving mode and then generates the target route based on the driving mode.

[0023] In autonomous driving mode, the driving control unit 16 controls each actuator AC so that the vehicle travels along the target path generated by the action plan generation unit 15. For example, in autonomous driving mode, the driving control unit 16 considers the driving resistance determined by the road gradient, etc., and calculates the required driving force to obtain the target acceleration per unit time calculated by the action plan generation unit 15. Then, it provides feedback control to the actuator AC so that the actual acceleration detected by, for example, the internal sensor group 2 becomes the target acceleration. In other words, it controls the drive actuator AC so that the vehicle travels at the target speed and target acceleration. Furthermore, in automatic driving mode, the driving control unit 16 calculates the optimal steering angle for the vehicle to follow a target route based on vehicle state quantities observed by the internal sensor group 2, etc. Then, it outputs a steering angle instruction signal corresponding to the calculated steering angle to control the steering actuator AC. In manual driving mode, the driving control unit 16 controls each actuator AC in accordance with driving commands (such as steering operations) from the driver acquired by the internal sensor group 2.

[0024] Incidentally, in lane-keeping driving, steering actuators (which can also be called steering actuators) are controlled so that the vehicle travels through a target position in the lane width direction. However, if disturbances such as changes in road surface slope (changes in gradient in the lane width direction) or strong crosswinds occur, the vehicle's position in the lane width direction may deviate from the target position, or the vehicle's orientation may deviate from the direction of travel. Therefore, in this embodiment, the steering angle control device is configured as follows to eliminate the positional deviation and orientation deviation of the vehicle that occur during driving.

[0025] <Steering angle control device> Figure 2 is a block diagram showing the main components of a steering angle control device 50 according to an embodiment. This steering angle control device 50 is configured, for example, as part of the functions of the controller 10 in Figure 1. The controller 10 is connected to a camera 1a, a steering angle sensor 2a, a steering angular velocity sensor 2b, a steering torque sensor 2c, a navigation device 6, and a steering actuator AC1.

[0026] Camera 1a is a monocular camera having an image sensor and constitutes part of the external sensor group 1 in Figure 1. Camera 1a may also be a stereo camera. Camera 1a is mounted, for example, at a predetermined position on the front of the vehicle and continuously captures images of the space in front of the vehicle to acquire images (camera images) of objects. Objects include lane markings on the road. Alternatively, or in conjunction with camera 1a, objects may be detected by radar, lidar, or the like.

[0027] The steering angle sensor 2a detects, for example, the rotation angle (steering angle) of the steering shaft connected to a steering wheel (not shown). The steering angular velocity sensor 2b detects the rotational angular velocity (steering angular velocity) of the steering shaft. Steering angular velocity may also be simply called steering angular velocity. The steering torque sensor 2c detects the steering operation by the driver, more specifically, the steering torque acting on the steering wheel. For example, the steering angle detected by the steering angle sensor 2a when the steering wheel is rotated counterclockwise from the neutral position is defined as a positive value, and the steering angle detected by the steering angle sensor 2a when the steering wheel is rotated clockwise from the neutral position is defined as a negative value. The steering angle sensor 2a, steering angular velocity sensor 2b, and steering torque sensor 2c described above constitute a part of the internal sensor group 2 shown in Figure 1. The internal sensor group 2 may also include an IMU (Inertial Measurement Unit) that detects the vehicle's three-axis translational and rotational motion.

[0028] The controller 10 has the following functional configurations, which are handled by the calculation unit 11 (Figure 1): a path error calculation unit 131, a target calculation unit 141, a track calculation unit 151, a target path calculation unit 152, and a rudder angle control unit 161. In addition, as described above, the controller 10 has a storage unit 12. The route error calculation unit 131 may constitute part of the vehicle position recognition unit 13. The target calculation unit 141 may constitute part of the external environment recognition unit 14. The road calculation unit 151 and the target route calculation unit 152 may constitute part of the action plan generation unit 15. The steering angle control unit 161 may constitute part of the driving control unit 16.

[0029] <Path error> The route error calculation unit 131 compares the position and orientation of the vehicle recognized by the vehicle position recognition unit 13 with the target route set by the target route calculation unit 152 (described later), and calculates the lateral position deviation and azimuth deviation of the vehicle relative to the target route directly beside the vehicle. The route error calculation unit 131 first recognizes the position and shape of the lane markings from the camera image of camera 1a, and recognizes the lane based on the recognition result. Next, the route error calculation unit 131 compares the recognized position, angle, and shape of the lane with the target route set by the target route calculation unit 152 (described later), and calculates the amount of deviation between the position of the vehicle and the position on the target route directly beside the vehicle (center of the lane) as the route lateral position deviation. It also calculates the amount of deviation between the direction of the vehicle and the azimuth angle of the target route directly beside the vehicle as the route azimuth angle deviation. The route error calculation unit 131 may calculate the route lateral position deviation and route azimuth angle deviation using the vehicle's position and orientation recognized based on the high-precision map information stored in the memory unit 12 and the surrounding information of the vehicle detected by the external sensor group 1, or the vehicle's position and orientation measured by the positioning unit 4.

[0030] <Target> The target calculation unit 141 calculates information indicating targets present around the vehicle. Based on signals input from the external sensor group 1, including the camera 1a, lidar, and radar, the target calculation unit 141 recognizes targets including moving objects such as other vehicles, bicycles, and pedestrians, as well as stationary objects (which may also be called terrain features) such as guardrails and signs, and outputs target information indicating the recognized targets.

[0031] <Base Track> The route calculation unit 151 calculates (searches for) a route (referred to as a base route) based on the current position of the vehicle measured by the positioning unit 4, the location of the destination entered by the driver, and the map information stored in the map database 5. The calculation of the base route is the same as the route search performed by the navigation device 6. The route calculation unit 151 may also obtain a route set by the navigation device 6 as the base route from the navigation device 6.

[0032] <Target Path> The target path calculation unit 152 sets target positions in the lane width direction that the vehicle should traverse on the base path calculated by the path calculation unit 151 and the external environment recognized by the external environment recognition unit 14. When the vehicle is, for example, maintaining its lane, the target path calculation unit 152 repeatedly sets target positions along the direction of travel. As a result, a target path (a trajectory obtained by connecting the target positions) is generated along the base path. Furthermore, if the external environment recognition unit 14 recognizes an obstacle such as a utility pole or a parked vehicle in front of the vehicle's direction of travel, the target route calculation unit 152 uses the obstacle information output from the obstacle calculation unit 141 to set a target position such that the distance between the vehicle and the obstacle in the lane width direction does not fall below a certain distance when the vehicle passes to the side of the obstacle. In addition, if the driver gives an instruction to change lanes via the turn signal (not shown), the target route calculation unit 152 sets a target position such that the vehicle's driving position gradually moves along the direction of travel towards the center of the target lane.

[0033] <Steering angle control> When, for example, the vehicle is traveling while maintaining the lane, the steering angle control unit 161 controls the steering angle via the steering actuator AC1 so that the host vehicle travels following the target path. Specifically, the steering angle control unit 161 calculates the steering angle necessary to make the position of the host vehicle follow the target path.

[0034] <Acquisition of vehicle state quantity> First, the steering angle control unit 161 acquires the target path generated by the target path calculation unit 152. Further, the steering angle control unit 161 acquires the vehicle mass m [kg] of the host vehicle, the yaw moment of inertia I f , f , [kgm 2 [m], the distance l f [m] between the center of gravity G and the front wheel axle, the distance l r [m] between the center of gravity G and the rear wheel axle, the equivalent cornering power K f [N / rad] of one front wheel, the equivalent cornering power K r [N / rad] of one rear wheel, the stability factor A [-], etc. as design information regarding the host vehicle from the storage unit 12. These may be referred to as vehicle characteristic information regarding the motion characteristics. Also, some of these correspond to the respective symbols in FIG. 4B described later. The steering angle control unit 161 further acquires, from the output values of the sensors constituting the internal sensor group 2 or by calculation using the output values of the sensors, the vehicle speed (body speed) V [m / s], the yaw angular velocity (yaw rate) γ [rad / s], the steering angle δ f [rad] of the front wheels, the gravitational acceleration g [m / s 2 [m / s f , the steering angular velocity δ [[ID=*28]]

[0035] [[ID=*29]] The above “´” indicates a time derivative. That is, the steering angular velocity δ f ´ is dδ f It seems there are some formatting issues in the original text with the "m / s" notations in lines 22 and 24. I've tried to make sense of it in the translation. If you can clarify the correct formatting, that would be great.This is synonymous with / dt. The vehicle orientation is calculated, for example, based on the direction of the target path's extension and the vehicle's length direction (sometimes called the longitudinal direction) recognized from the camera image of camera 1a, etc. The lateral path deviation e[m] is, as described above, the amount of deviation of the vehicle's position from the target path in the lane width direction. The azimuth angle deviation Δψ[rad] is the angle of deviation of the vehicle's orientation (vehicle orientation) relative to the target path. The yaw angular velocity deviation Δγ[rad / s] is the deviation between the yaw angular velocity γ[rad / s] detected by the yaw rate sensor included in the internal sensor group 2 and the target yaw angular velocity (=vehicle speed V[m / s] × path curvature κ[rad / m]). The path curvature κ[rad / m] is the curvature of the target path in front of the vehicle's direction of travel and is calculated, for example, by the steering angle control unit 161. Next, the rudder angle control unit 161 estimates state variables that cannot be observed using the internal sensor group 2 using the state estimation model (equations (1) and (2) described later), which will be explained in detail later. In this embodiment, the rudder angle disturbance δ d In addition, the vehicle body slip angle β [rad] is estimated, and the steering angle disturbance δ is also estimated. d Effective front wheel steering angle δ excluding f Calculate ^[rad]. "^" indicates an estimated value. Front wheel steering angle δ f and the effective steering angle δ of the front wheels f ^, rudder angle disturbance δ d The relationship is δ f =δ f ^+δ d It is expressed by the following equation. Here, the rudder angle disturbance δ d This refers to a steering angle that does not affect the vehicle's behavior, such as a counter-steering motion against a cant angle φ [rad]. The cant angle φ is the gradient in the lane width direction (road surface transverse gradient). Counter-steering caused by crosswinds or misalignment of the steering gear's midpoint is called a steering angle disturbance δ. d It may be included in this. The vehicle slip angle β [rad] is the angle of deviation between the direction of the vehicle speed V and the direction of the vehicle itself. Note that the above steering angle velocity δ f ' may be a sensor value from the steering angular velocity sensor 2b, or a value calculated based on the sensor value from the steering angle sensor 2a. Also, the longitudinal component of the vehicle speed V is V x [m / s], the transverse component is Vy [m / s], gravitational acceleration g[m / s 2 The transverse component of ] is g y [m / s 2 ]

[0036] <Optimal rudder angle sequence> Next, the steering angle control unit 161 uses the acquired vehicle state quantity and the above m and l values. f , the above l r , the above K f , the above K r The above A, etc., are input into a driving simulation model (hereinafter referred to as the prediction model). In this embodiment, the prediction model (equations (7) and (8) described in detail later) is used to calculate the optimal steering angle (optimal steering angle sequence) so that the future driving position of the vehicle follows the target path through model predictive control. Details of the prediction model and model predictive control will be described later.

[0037] <Steering angle instruction value> The rudder angle control unit 161 extracts the rudder angle to be indicated for a predicted time ahead from the above optimal rudder angle sequence and excludes the rudder angle disturbance δ in advance. d The target rudder angle, including the added factor, is output as a rudder angle instruction value, and the rudder angle is controlled by the steering actuator AC1. Furthermore, when the steering angle control unit 161 is performing steering angle control based on the target steering angle, if steering torque is detected by the steering torque sensor 2c, it may determine that the driver has performed a steering operation (an instruction to change the steering angle has been given) and may interrupt the steering angle control. Furthermore, the steering angle control unit 161 may continue steering angle control based on the target steering angle unless a large steering torque that clearly indicates the driver's intention to cancel path-following driving is detected by the steering torque sensor 2c while performing steering angle control based on the target steering angle; in other words, unless a steering operation that changes the steering angle by more than a predetermined value is performed.

[0038] <Flow of steering angle control> Figure 3A is a block diagram illustrating the flow of steering angle control by the steering angle control unit 161. Of the configurations exemplified in Figure 2, the target calculation unit 141, the track calculation unit 151, the target path calculation unit 152, the steering angle control unit 161, the steering actuator AC1, and the vehicle body of the vehicle 101 are shown. Figure 3B is a block diagram illustrating the details of the steering angle control unit 161 shown in Figure 3A. The steering angle control unit 161 includes a vehicle state estimation unit 161A using a Kalman filter, a model prediction control unit 161B, and a target steering angle calculation unit 161C. The vehicle state estimation unit 161A using a Kalman filter includes a state estimation model 161A1. Furthermore, the model prediction control unit 161B includes a prediction unit 161B1 and an optimization unit 161B2.

[0039] <Model-based predictive control> Refer to Figure 3B to explain model predictive control. Model predictive control is a conventional technology. Model predictive control is a control method that calculates the optimal control input using predictive estimation of the controlled object. Model predictive control uses a predictive model and an optimizer. The predictive model is a model that mimics the controlled object. In this embodiment, a prediction unit 161B1, which combines a vehicle model and a path deviation model, is used as the predictive model. Furthermore, an optimization unit 161B2, which evaluates the operation of the prediction unit 161B1 and calculates the optimal control input, is used as the optimizer.

[0040] Figure 4A illustrates the position of the vehicle 101, the target path Tr, and the predicted path Pr. Figure 4B is a schematic diagram showing the relationship between the two-wheeled model and the target path Tr according to the embodiment. Each symbol in the figures represents the vehicle speed V (vehicle longitudinal speed V) of the vehicle 101. x Vehicle lateral speed V y ), center of gravity G, yaw angular velocity γ, vehicle path azimuth deviation Δψ, vehicle path lateral position deviation e, center of gravity G - front wheel axle distance l f , center of gravity G - rear wheel axle distance l r , front wheel steering angle δ f Target path Tr, vehicle slip angle β, front wheel slip angle β f , rear wheel slip angle β r Corresponds to symbol 2F. yf The above Kf [N / rad] and the above β f This indicates the front wheel cornering force [N], which is the product of [rad]. Symbol 2F yr The above K r [N / rad] and the above β r This shows the rear wheel cornering force [N], which is the product of [rad]. In this embodiment, the motion model of the vehicle 101 is represented by an equivalent two-wheel model in which the two front wheels and two rear wheels of the vehicle 101 are moved to the central axis of the vehicle body.

[0041] The vehicle state estimation unit 161A in Figure 3B uses the state estimation model 161A1 to estimate the steering angle disturbance δ based on the vehicle state quantity. d The vehicle body slip angle β [rad] and other parameters are estimated. Equation (1) is the state equation for state estimation model 161A1. Equation (2) is the output equation (which may also be called the observation equation) for state estimation model 161A1. Equations (1) and (2) are examples of continuous-time state estimation models using a steering angle velocity input vehicle model (two-wheeled model) as the vehicle model. x' = A S x + B S u ……… (1) y = C S x ……… (2) However, the coefficient matrix in the formula is as follows:

number

[0042] The state estimation model 161A1 uses, for example, the steering angular velocity δ of the front wheels. f We model the vehicle state variables that are influenced by factors such as [rad / s], and use these factors, the observed vehicle state variables, and the previous estimate to calculate the steering angle disturbance δ d The system also calculates an estimated value of the vehicle body slip angle β. In general, the observed vehicle state variables contain stationary noise components (observational noise), as well as noise components representing the uncertainty of the model (process noise). By repeatedly performing estimation calculations using the state estimation model 161A1, the system takes into account the uncertainties of the observed values ​​and the model to obtain the estimation with the least error.

[0043] The model prediction control unit 161B takes the target path Tr and the observed or estimated vehicle state quantities of the vehicle 101 as input and performs a model prediction calculation. At predetermined calculation intervals (e.g., several milliseconds to several seconds), it predicts the speed and direction of travel (predicted path Pr) of the vehicle 101 for the next few seconds (predicted horizon) on the model, and performs a process (solution search calculation) to find the optimal steering angle sequence for the next few seconds (predicted horizon). In this embodiment, the input x to the prediction model is information indicating the state of the vehicle 101 (Δθ, e), and the input u is information indicating the input of the vehicle 101 (front wheel steering angle δ). f) and input w is information indicating the curvature κ of the target path Tr. The rudder angle δ of input u f This is the actual rudder angle δ obtained by the internal sensor group 2. f The steering angle disturbance δ estimated by the vehicle state estimation unit 161A d Excluding the above estimated value δ f Enter ^.

[0044] <Explanation of the predictive model> Equation (3) is the state equation of the prediction model of the prediction unit 161B1. Equation (4) is the output equation of the prediction model of the prediction unit 161B1. Equations (3) and (4) are examples of continuous-time prediction models using a path deviation system that includes a steering angle input vehicle model (referred to as a steady-state two-wheel model) representing a steady state as a vehicle model. x' = A C x + B C u +W C w ……… (3) y = C C x ……… (4) However, the coefficient matrix in the formula is as follows:

number

[0045] When discretizing the state equations and output equations of the above equations (3) and (4), the following equations (7) and (8) are obtained by applying the assumption of zero-order hold. Equations (7) and (8) are predictive models of the zero-order hold discrete time. By using zero-order hold, it becomes possible to maintain the accuracy of the curvature κ of the target path even when using a long sample time, for example, compared to the case of zero-order hold. The assumption of zero-order hold refers to the idea that the state changes linearly from the state at the previous sample time to the state at the next sample time. [Number] However, x(k) is a state variable (in the embodiment, Δθ, e), u(k) is an input of the vehicle (in the embodiment, the rudder angle δ f of the front wheel), w(k) is a disturbance component (in the embodiment, the path curvature κ), and y(k) is an output of the vehicle (Δθ, e). Note that C d = C C , B d0 = B d - B d1 , W d0 = W d - W d1 are as follows. The sizes are A d : n x × n x , Bd0 :n x ×n U W d0 :n x ×n W , B d1 :n x ×n U W d1 :n x ×n W , C d :n y ×n x That is the case.

[0046] According to equations (7) and (8) above, by repeatedly determining the vehicle state variables at the next step k+1 based on the vehicle state variables at step k in the model, the vehicle state variables at the next Hp step (k=Hp) from the current time (k=0) can be predicted. More specifically, the prediction unit 161B1 inputs the observed or calculated vehicle state variables of the vehicle 101 at the present time into the state equation and predicts the vehicle state variables corresponding to each step from k=1 to k=Hp.

[0047] The optimization unit 161B2 uses multiple elements for evaluation (in this embodiment, the path lateral position deviation e, the path azimuth angle deviation Δθ, and the front wheel steering angle difference Δδ). f Each of these is represented by a function, and the sum of the outputs of each function is output as the evaluation function J. Constraints that must be satisfied are also set.

[0048] Equation (9) is an example of the evaluation function J of the optimization unit 161B2. Equation (10) is an example of a constraint condition.

number

[0049] In this embodiment, the evaluation function J is defined by multiple elements. Specifically, it includes a function f1 for the lateral position deviation e of the vehicle's path, a function f2 for the azimuth angle deviation Δθ of the velocity vector's path, and the steering angle difference Δδ of the front wheels. f The function f3 is a function of the same nature as the function f1, f2, and f3. The evaluation function J is a function composed of these three elements. The optimization unit 161B2 finds the value of the evaluation function J that is determined by the sum of these three functions f1, f2, and f3, which is the smallest value.

[0050] The three functions f1, f2, and f3 will be explained in more detail. The optimization unit 161B2 calculates the result of multiplying the square of the path lateral position deviation e(k) at each step from the next step k=1 to the next Hp step (k=Hp) by the weight value We, and sets this as the function f1 with respect to the path lateral position deviation e. Furthermore, the optimization unit 161B2 calculates the result of multiplying the square of the path azimuth angle deviation Δθ(k) at each step from the next step k=1 to the next Hp step (k=Hp) by the weight value WΔθ, and sets this as a function f2 with respect to the path azimuth angle deviation Δθ. Furthermore, the optimization unit 161B2 calculates the rudder angle difference (Δδ) at each step from the next step k=1 to the next Hp step (k=Hp). f (k)-Δδ f (k-1)) squared and the weight value WΔδ f The result of multiplication with is calculated, and the rudder angle difference Δδ f Let f3 be a function related to this.

[0051] The model prediction control unit 161B, having the configuration described above, repeats the loop processing by the prediction unit 161B1 and the optimization unit 161B2 multiple times for each calculation cycle to determine the optimal rudder angle sequence that satisfies the constraints and minimizes the evaluation function J. In other words, the optimal rudder angle sequence u(1) to u(Hp) for each step from the next step k=1 to the next Hp step (k=Hp) is determined as a candidate for rudder angle instruction values.

[0052] <Change in weight values> In this embodiment, by changing the weight values, multiple elements used for evaluation (in this embodiment, the path lateral position deviation e, the path azimuth angle deviation Δθ, and the front wheel steering angle difference Δδ) can be used. f Within that context, it becomes possible to set an element that takes precedence over other elements.

[0053] For example, the weight value We can be combined with other weight values ​​WΔθ and WΔδ f By making it greater than this, steering angle control becomes possible that converges the path lateral position deviation e as quickly as possible within the range in which the vehicle 101 can respond (the range of the constraint conditions by equation (10) above). Similarly, the weight value WΔθ is the same as the other weight values ​​We and WΔδ f By making it greater than this, steering angle control becomes possible that converges the path azimuth deviation Δθ as quickly as possible within the range in which the vehicle 101 can respond (the range of the constraint conditions by equation (10) above). For the two examples above, the weight value WΔδ f By making this larger than the other weight values ​​We and WΔθ, the rudder angle difference Δδ f This enables steering angle control that minimizes steering input (in other words, minimizes steering input). Because steering input is reduced, vehicle 101 will not turn, resulting in steering angle control that does not follow the target path. Furthermore, a weight value for emergency avoidance, which is the opposite of the control that reduces steering input and maximizes steering input, can also be set. Specifically, the weight value WΔδ f By setting this to zero, vehicle 101 can prioritize avoiding danger above all else in its steering angle control, rather than the impact on the occupants (such as ride comfort).

[0054] The target steering angle calculation unit 161C calculates the steering angle disturbance δ estimated by the vehicle state estimation unit 161A. d The optimal rudder angle sequence u(1)~u(Hp) determined by the model prediction control unit 161B is input, and the rudder angle to be indicated for a predetermined predicted time tp is extracted from the optimal rudder angle sequence, and the rudder angle disturbance δ that was excluded in advance is used. dThe target rudder angle, including the specified value, is output to the steering actuator AC1 as the rudder angle instruction value for the next few milliseconds. In this embodiment, the rudder angle corresponding to the predicted time tp is obtained from the optimal rudder angle sequence u(1) to u(Hp) by linear interpolation.

[0055] <Vehicle model in predictive model> Generally, the vehicle model used in the state equation of the state estimation model 161A1 and the vehicle model used in the state equation of the prediction model of the prediction unit 161B1 are of the same type. Furthermore, the discretization method is also common. For example, if the state estimation model 161A1 uses a vehicle model with steering angular velocity input and zero-order hold discretization, the prediction model also uses a vehicle model with steering angular velocity input and zero-order hold discretization. This is because using the same type of vehicle model and discretization method results in better consistency between state estimation and model prediction in terms of vehicle behavior. However, in this prediction model, the input of the curvature κ of the target path becomes a zero-order hold discrete, which can lead to large discrepancies with the continuous-time target path over long sample times. In other words, the prediction accuracy of the behavior of the vehicle 101, including the deviation of the vehicle's position and direction relative to the target path, may deteriorate. Additionally, the large number of state variables in the prediction model increases the computational load and memory usage in the prediction unit 161B1, which is also a problem. Therefore, in this embodiment, the input of the curvature κ of the target path in the prediction model is discretized using a first-order hold so that the deviation from the target path over continuous time is kept small even over long sample times, that is, to improve the prediction accuracy of the behavior of the vehicle 101, including the deviation of the position and direction of the vehicle 101 relative to the target path. Furthermore, the vehicle model included in the prediction model is configured such that equivalent vehicle behavior can be predicted even when the discretization method is changed, by adjusting the steering angle δ of the front wheels. f The input was to be the steering angle δ of the front wheel. f A vehicle model with the input δ is a front wheel steering angular velocity f Compared to methods that use ' as input, this method allows for a reduction in the number of state variables (lower dimensionality), which can be expected to lower the computational load and memory usage. Based on the above ideas, different types of vehicle models and discretization methods were used for state estimation and model prediction, as follows. Specifically, the state estimation models represented by equations (1) and (2) above use zero-order hold discretization, while the prediction models represented by equations (3) and (4) above, adopted in the embodiment, use first-order hold discretization. Furthermore, the vehicle model in the state equation of equation (1) above consists of the three rows enclosed by the dashed line in the coefficient matrices As and Bs, with the steering angular velocity δ of the front wheels as the input u. f While the input is ', in the state equation of equation (3) above adopted in the embodiment, by representing the vehicle model in a steady state, it is not necessary to treat the vehicle body slip angle β and yaw angular velocity γ as states, and the elements corresponding to the vehicle model are incorporated into the coefficient matrix Bc of the state equation of the path deviation system, and the steering angle δ of the front wheels is used as the input u. f The difference lies in the fact that it is an input method.

[0056] Equation (11) is the state equation of the prediction model before it was improved to equation (3) as adopted in the embodiment. Equation (12) is the output equation of the prediction model before it was improved to equation (4) as adopted in the embodiment. Equations (11) and (12) are examples of a continuous-time prediction model using a steering angle velocity input vehicle model (two-wheeled model) and a path deviation system as a vehicle model. x' = oldA C x +oldB C u +oldW C w … (11) y = oldC C x …… (12) However, the coefficient matrix in the formula is as follows:

number

[0057] When discretizing the state and output equations in equations (11) and (12) above, applying the zero-order hold assumption yields equations (13) and (14). Equations (13) and (14) are zero-order hold discrete-time prediction models. As mentioned above, the zero-order hold assumption is the idea that, for example, the state at the previous sample time is maintained until the next sample time.

number

[0058] According to equation (11) above for the prediction model before improvement and equation (1) above for the state estimation model of the embodiment, the type of vehicle model is the same for both model prediction and state estimation. Furthermore, when discretizing both models, the assumption of zero-order hold is applied. Specifically, the vehicle model in the state equation of equation (11) above consists of the three rows enclosed by the dashed line from the coefficient matrices oldAc, oldBc, and oldWc, with the steering angular velocity δ of the front wheels as the input u. f In contrast to the input being ', the vehicle model of the state equation in equation (1) above consists of the three rows enclosed by the dashed line in the coefficient matrices As and Bs, with the front wheel steering angular velocity δ as the input u. f They coincide in the point where the ' is input. Therefore, the consistency between model prediction and state estimation is good. However, the size of the vehicle model in the state equation of equation (11) before improvement is equivalent to three rows enclosed by dashed lines in the coefficient matrices oldAc, oldBc, and oldWc, which occupies more than half of the matrix elements of the coefficient matrices oldAc, oldBc, and oldWc. Therefore, reducing the size is desirable to lower the computational load.

[0059] Equation (15) is the state equation of the prediction model in the process of improving the state equation from equation (11) to equation (3) adopted in the embodiment. Equation (16) is the output equation of the prediction model in the process of improving the output equation from equation (12) to equation (4) adopted in the embodiment. Equations (15) and (16) are examples of a continuous-time prediction model using a steering angle input vehicle model (two-wheeled model) and a path deviation system as vehicle models. x' = midA C x +midB C u +midW C w … (15) y = midC C x …… (16) However, the coefficient matrix in the formula is as follows:

number

[0060] When discretizing the state and output equations in equations (15) and (16) above, applying the first-order hold assumption yields equations (7) and (8). The equations for the first-order hold discrete-time prediction model remain the same before and after the refinement of the state equations. However, x(k) is a state variable (β, γ, Δψ, e), and u(k) is the vehicle input (front wheel steering angle δ). f ), where w(k) is the disturbance component (path curvature κ) and y(k) is the vehicle output (Δθ, e). In other words, in the equation of state in equation (15) above, the number of vehicle state variables treated as x(k) has decreased compared to before the improvement. Note that the input of u(k) before the improvement was the steering angular velocity of the front wheels δ. f In contrast to ', the steering angle of the front wheels δ during the improvement process f It has been changed to this.

[0061] The size of the vehicle model in the state equation (15) of the prediction model under improvement is the two rows enclosed by the dashed line among the coefficient matrices midAc, midBc, and midWc. Although this is less than the matrix elements of the coefficient matrices oldAc, oldBc, and oldWc, it occupies half of the matrix elements of midAc, midBc, and midWc. Therefore, further reduction in size is desired to lower the computational load.

[0062] To further reduce the size, the state equation (3) used in the embodiment employs a vehicle model representing the steady state, as described above. This makes it possible to incorporate elements corresponding to the vehicle model into the state equation of the path deviation system (coefficient matrix Bc), and the size of the coefficient matrix of the state equation (3) used in the embodiment can be made extremely compact compared to the state equation (15) which is still under improvement. As a result, the computational load on the prediction unit 161B1 is further reduced, making it possible to implement it in even lower-spec ECUs, etc.

[0063] <Explanation of the flowchart> Figure 5 is a flowchart showing an example of calculation processing performed by the arithmetic unit 11 of the controller 10 in Figure 2 according to a predetermined program. The processing shown in this flowchart is repeatedly performed, for example, when the vehicle 101 is driving in automatic driving mode. It is also repeatedly performed when the vehicle 101 is driving in manual driving mode and, for example, when the lane keeping function, which is one of the driver assistance functions, is enabled, i.e., when lane keeping driving is in progress.

[0064] In step S10, the controller 10 obtains the route information as the target path described above using the rudder angle control unit 161 and proceeds to step S20. In step S20, the controller 10 acquires the vehicle state quantities and other information described above using the steering angle control unit 161 and proceeds to step S30. In step S30, the controller 10 performs the aforementioned model predictive control calculation using the rudder angle control unit 161 and proceeds to step S40. In step S40, the controller 10, using the steering angle control unit 161, calculates the target steering angle as described above, outputs a steering angle instruction value, and performs steering angle control, then proceeds to step S50. As a result, the steering actuator AC1 is controlled based on the steering angle instruction value.

[0065] In step S50, the controller 10 determines whether or not to terminate the process. If, for example, the automatic driving mode is deactivated, the controller 10 determines step S50 to be positive and terminates the process shown in Figure 5. If, for example, the automatic driving mode is to be continued, the controller 10 determines step S50 to be negative and returns to step S10, and repeats the process described above.

[0066] According to the embodiments described above, the following effects and advantages are achieved. (1) The steering angle δ of the vehicle 101 is set to follow the target path Tr. f The steering angle control device 50 that controls the steering includes a prediction unit 161B1 that predicts the vehicle speed V, yaw angular velocity γ, etc., as the behavior of the vehicle 101, including the lateral position deviation e and azimuth angle deviation Δθ of the vehicle 101 as deviations in position and direction from the target path Tr, based on vehicle state quantities indicating the motion state of the vehicle 101 and information indicating the target path Tr, up to a predetermined time in advance, and an optimization unit 161B2 that optimizes the target steering angle using the prediction results obtained by the prediction unit 161B1. The prediction unit 161B1 has a path deviation model (state equation (3), output equation (4)) that models the relationship between the position and direction deviation of the vehicle 101 with respect to the target path Tr (lateral position deviation e, azimuth angle deviation Δθ) and the curvature κ of the target path Tr, including the steady state relationship between the vehicle state quantities and vehicle characteristic information (l[m], A[-], etc.). The objective of the model prediction is to accurately predict the future behavior of the vehicle 101 and to achieve this with low computational load. Using the path azimuth angle deviation Δψ, which is the angle of deviation of the vehicle orientation relative to the target path Tr, and the vehicle slip angle β, the relationship Δθ = Δψ + β holds. Therefore, conventionally (under improvement), a vehicle model that included Δψ and β as state variables was included in the state equation (15) to control Δθ to approach 0. In contrast, in this embodiment, the path deviation model is represented in a steady state, and β is not treated as a state. As a result, Δθ'=Δψ' can be treated, and the elements corresponding to the vehicle model in the state equation (15) under improvement are incorporated into the state equation (3) of the path deviation system according to this embodiment (coefficient matrix Bc). As a result, control that brings Δθ closer to 0 can be performed even without a vehicle model that includes Δψ and β as state variables. Furthermore, the accuracy of predicting the future behavior of the vehicle 101 is guaranteed even when the path deviation model is represented in a steady state, so the performance of following the target trajectory is not impaired. Since there is no vehicle model that includes Δψ and β as state variables, the size of the coefficient matrix in equation (3) above is extremely compact compared to the state equation (15) under improvement. This reduces the number of vehicle state quantities handled in the model prediction calculation and lowers the computational load on the prediction unit 161B1, making it possible to implement it in a lower-spec ECU, etc. Furthermore, by modeling the steady-state relationship, it becomes possible to ensure continuity with control in the extremely low-speed range (e.g., automatic parking control) compared to not modeling the steady-state. The above vehicle state variables include the vehicle speed V and the steering angle δ of the front wheels, which are acquired by the internal sensor group 2 of the vehicle 101. f In addition to the first vehicle state quantities such as yaw angular velocity γ, the steering angle disturbance δ, which is not acquired by the internal sensor group 2 of the vehicle 101, is estimated by the vehicle state estimation unit 161A. d This also includes a second vehicle state variable, such as the slip angle β.

[0067] (2) In the steering angle control device 50 described in (1) above, the path deviation model (state equation (3), output equation (4)) has a state equation (3) in which the above deviations are the path azimuth angle deviation Δθ as the first deviation angle in the direction in which the vehicle 101 is moving relative to the target path Tr, and the path lateral position deviation e as the amount of deviation of the position of the vehicle 101 from the target path Tr in the lane width direction. With this configuration, the model prediction control unit 161B can determine the optimal steering angle sequence to appropriately resolve the lateral position deviation e and azimuth deviation Δθ of the path, which are the deviations in the position and direction of the vehicle 101 relative to the target path Tr.

[0068] (3) In the steering angle control device 50 described in (2) above, the path deviation model has a state equation (3) in which the derivative value Δθ´ of the path azimuth angle deviation, which is the first deviation angle in the direction in which the vehicle 101 is moving relative to the target path Tr, is equal to the derivative value Δψ´ of the path azimuth angle deviation, which is the second deviation angle in the direction of the vehicle 101 relative to the target path Tr. With this configuration, it becomes possible to incorporate the elements corresponding to the vehicle model in the state equation of equation (15) above, which is a prediction model under improvement, into the state equation (3) of the path deviation system according to the embodiment (coefficient matrix Bc). As a result, the size of the coefficient matrix of equation (3) above becomes extremely compact compared to the state equation of equation (15) above, which is under improvement. Consequently, the computational load on the prediction unit 161B1 can be reduced, making it possible to implement it in lower-spec ECUs, etc.

[0069] (4) In the steering angle control device 50 described in (2) above, the path deviation model has a state equation (3) that includes a term for the first deviation angle (Δθ) and a term for its time derivative (dθ / dt). With this configuration, it becomes possible to incorporate the elements corresponding to the vehicle model in the state equation of equation (15) above, which is a prediction model under improvement, into the state equation (3) of the path deviation system according to the embodiment (coefficient matrix Bc). As a result, the size of the coefficient matrix of equation (3) above becomes extremely compact compared to the state equation of equation (15) above, which is under improvement. Consequently, the computational load on the prediction unit 161B1 can be reduced, making it possible to implement it in lower-spec ECUs, etc.

[0070] (5) In the steering angle control device 50 described in (1) above, the path deviation model (state equation (3), output equation (4)) is subject to the assumption of a first-order hold when discretized. If we assume a zero-order hold during discretization, the difference between the input from the continuous-time state equation (curvature κ of the target path) and the input from the discrete-time state equation (curvature κ of the target path) can become large, making it difficult to use long sample times. In contrast, by applying the first-order hold assumption, the difference between the input from the continuous-time state equation (curvature κ of the target path) and the input from the discrete-time state equation (curvature κ of the target path) is suppressed compared to the zero-order hold case. This makes it possible to improve the prediction accuracy of the deviation of the position and direction of the vehicle 101 relative to the target path. Note that the steering angular velocity δ f '0th-order hold discretization of the input vehicle model and steering angle δ f Since this is equivalent to the first-order hold discretization of the input vehicle model, the accuracy of predicting vehicle behavior is equivalent. In other words, it becomes possible to accurately predict the behavior of the vehicle 101 even at a distance, improving the performance of tracking the target trajectory.

[0071] The above embodiment can be modified into various forms. Modifications will be described below. (Variation 1) In the above embodiment, the vehicle state variables are the vehicle speed (vehicle body speed) V [m / s], the yaw angular velocity (yaw rate) γ [rad / s], and the steering angle δ of the front wheels. f [rad], acceleration g[m / s 2 ], steering angular velocity δ f Examples of parameters include '[rad / s], vehicle orientation, path lateral position deviation e[m], vehicle path azimuth angle deviation Δψ[rad], yaw angular velocity deviation Δγ[rad / s], and vehicle slip angle β[rad], but other state variables may also be added.

[0072] (Modification 2) In the above embodiment, the weight values ​​We, WΔθ, and WΔδ for the first and second elements are f The values ​​of can be changed as appropriate. Also, the weight values ​​We, WΔθ, and WΔδ are used for emergency avoidance. f The values ​​of these can also be changed as appropriate. In other words, the weight values ​​We, WΔθ, and WΔδ applied to the evaluation function (equation (9)) are also acceptable. f The value can be changed as needed.

[0073] The above description is merely an example, and the present invention is not limited by the embodiments and modifications described above, as long as they do not impair the features of the present invention. [Explanation of Symbols]

[0074] 1 External sensor group, 2 Internal sensor group, 6 Navigation device, 10 Controller, 11 Calculation unit, 12 Memory unit, 13 Vehicle position recognition unit, 14 External environment recognition unit, 15 Action plan generation unit, 16 Driving control unit, 50 Steering angle control device, 100 Vehicle control system, 101 Own vehicle, 131 Path error calculation unit, 141 Target calculation unit, 151 Road calculation unit, 152 Target path calculation unit, 161 Steering angle control unit, 161A Vehicle state estimation unit, 161B Model prediction control unit, 161B1 Prediction unit, 161B2 Optimization unit, 161C Target steering angle calculation unit, AC1 Steering actuator

Claims

1. A steering angle control device that controls the steering angle of a vehicle so that it follows a target path, A prediction unit predicts the behavior of the vehicle, including deviations in the vehicle's position and direction from the target path, up to a predetermined time in advance, based on vehicle state quantities indicating the vehicle's motion state and information indicating the target path. The system includes an optimization unit that optimizes the target steering angle using the prediction results obtained by the prediction unit, The prediction unit has a path deviation model that models the relationship between the deviation and the curvature of the target path, including the steady-state relationship between the vehicle state variables and vehicle characteristic information relating to the vehicle's motion characteristics. A steering angle control device characterized by the following features.

2. In the steering angle control device according to claim 1, The path deviation model has a state equation in which the deviation is defined as the first deviation angle in the direction the vehicle is traveling relative to the target path and the amount of deviation in the lane width direction of the vehicle's position from the target path. A steering angle control device characterized by the following features.

3. In the steering angle control device according to claim 2, The path deviation model has a state equation in which the derivative of the first deviation angle is equal to the derivative of the second deviation angle of the vehicle's orientation relative to the target path. A steering angle control device characterized by the following features.

4. In the steering angle control device according to claim 2, The path deviation model has a state equation that includes the first deviation angle term and its time derivative term. A steering angle control device characterized by the following features.

5. In the steering angle control device according to claim 1, The aforementioned path deviation model applies the assumption of a first-order hold during discretization. A steering angle control device characterized by the following features.

6. A steering angle control method for controlling the steering angle of a vehicle so that it follows a target path, Based on the vehicle state quantities indicating the vehicle's motion state and the information indicating the target path, predict the vehicle's behavior, including deviations in the vehicle's position and direction relative to the target path, up to a predetermined time in advance. Using the prediction results obtained from the above prediction, the target rudder angle is optimized, The prediction uses a path deviation model that models the relationship between the deviation and the curvature of the target path, including the steady-state relationship between the vehicle state variables and vehicle characteristic information relating to the vehicle's motion characteristics. A steering angle control method characterized by the following.

Images