KR20170035142A - Apparatus and method for estimating position of vehicle using nonlinear tire model - Google Patents

Abstract

The present invention relates to an apparatus and method for estimating a position of a vehicle using a nonlinear tire model, and more particularly, to an apparatus for estimating a position of a vehicle using a nonlinear tire model according to an embodiment of the present invention. A position measuring unit for measuring a GPS position of the vehicle; And a vehicle position estimator for estimating a position of the vehicle by applying the obtained state information of the vehicle and the GPS position of the measured vehicle to a nonlinear tire model reflecting the slip angle of each tire.

Description

TECHNICAL FIELD [0001] The present invention relates to an apparatus and method for estimating a position of a vehicle using a nonlinear tire model,

The present invention relates to an apparatus and method for estimating a position of a vehicle, and more particularly to an apparatus and method for estimating a position of a vehicle using a nonlinear tire model.

In recent years, research has been actively conducted on an ADAS (Advanced Driver Assistance System) for improving the convenience and stability of a driver. In order to operate such a system stably, it is necessary to provide a technique for providing accurate vehicle position information.

Generally, the position information of a vehicle depends on a GPS (Global Positioning System). However, since GPS uses satellite signals, the reliability of location information due to the surrounding environment is low. In order to overcome these limitations, information convergence technology that fuses with sensor information in the vehicle has been studied extensively.

Dead Reckoning (DR) is a method of integrating the Inertial Measurement Unit (IMU) and GPS position information into the vehicle. The DR method has an advantage of obtaining accurate position information continuously, but there is a limit that accumulation error of IMU accumulates with time. In order to overcome these limitations, a model-based sensor fusion algorithm is used. Particularly, an extended Kalman filter (EKF) based on a Bayesian filter and capable of directly applying a nonlinear model is widely used.

On the other hand, in order to obtain high position accuracy, the fusion algorithm is important, but the model selection of the vehicle is also important. Kinematic and dynamic models are widely used as vehicle models for position estimation. The kinematic model is a simple model that does not consider the vehicle's tire slip and is suitable for use in low-speed driving conditions. On the other hand, the dynamics model is a complicated model considering the tire slip of a vehicle and is suitable for use in a high-speed driving condition.

A bicycle model has been used as a dynamic model for estimating the position of a vehicle. The bicycle model is a model approximated by a two-wheeled two-wheeled vehicle, in which the wheel angle is approximated to the average value of the left and right front wheel angles of the vehicle, the front wheel force is approximated to the average value of the left and right front wheel forces, The force is approximated by the average of the left and right rear wheel forces. This model is expressed more simply than the four wheel vehicle model (FWVM), and has a limitation in not reflecting the slip angle of each tire although it is advantageous in operation speed. It can be seen that the position error occurs.

Also, a tire model approximated to a small slip angle state is mainly used. The vehicle dynamic model using this model shows a large position error in a state of a large tire slip angle. Therefore, it is necessary to select a tire model that can be used even in a large tire slip angle state. Tires with high nonlinearity are modeled through empirical methods, and various tire models are proposed. This nonlinear tire model is used only for the characterization of the vehicle speed, acceleration, and slip angle.

As described above, the conventional vehicle position estimation system uses a dead reckoning (DR) algorithm using an expensive inertial sensor (IMU) and an expensive DGPS (Differential Global Position System) Inertial sensors and expensive DGPS are essential. For example, accuracy only increases when high precision DGPS with a position error of less than 3 m is used.

The conventional vehicle position estimation system uses an information fusion algorithm based on a vehicle model, but accuracy is degraded by using only a model approximated to a bicycle model.

In addition, since the conventional vehicle position estimation system uses a linear model in which the nonlinearity of the tire and the slip of the road are simply approximated, the position error becomes large when the vehicle is traveling at a high speed and a severe cornering. For example, the position error in this conventional vehicle position estimation system is more than 3 m on average.

The embodiments of the present invention estimate the position of the vehicle by applying the vehicle state information and the GPS position of the vehicle to the four wheel vehicle model (FWVM) to which the nonlinear tire model is applied, so that even in a high speed running or cornering state in which a large slip angle occurs An apparatus and method for estimating a position of a vehicle using a nonlinear tire model capable of estimating a position with a high accuracy.

Specifically, in the embodiments of the present invention, a four-wheel vehicle model that considers the forces of each wheel, rather than a bicycle model, And a method of estimating a position of a vehicle using the method.

Embodiments of the present disclosure use a nonlinear model (e.g., a Dugoff model) in which the slip of tires and roads is modeled to accurately estimate the slip of tires and roads, The present invention provides an apparatus and method for estimating a position of a vehicle using a nonlinear tire model that can accurately calculate a direction force of a vehicle.

Embodiments of the present invention are based on the assumption that the speed of each wheel, the vehicle speed and the speed of each wheel from various modules (for example, anti-lock brake system (ABS), electronic stability control (ESC) The position of the vehicle can be accurately estimated without acquiring additional devices or components by acquiring the state information of the vehicle such as acceleration, steering angle, and the like through the basic communication network (e.g., CAN (Controller Area Network) bus) An apparatus and method for estimating a position of a vehicle using a nonlinear tire model.

Embodiments of the present invention use an extended Kalman filter (EKF) that stochastically fuses the state information of various vehicles and the four-wheel vehicle model (FWVM) to which the non-linear tire model is applied, An apparatus and method for estimating a position of a vehicle using a nonlinear tire model capable of accurately estimating a position of a vehicle even if a GPS having a large error (for example, a position error of 10 m) is used.

According to a first aspect of the present invention, there is provided an information processing apparatus comprising: an information obtaining unit obtaining state information of a vehicle; A position measuring unit for measuring a GPS position of the vehicle; And a vehicle position estimator for estimating a position of the vehicle by applying the obtained state information of the vehicle and the GPS position of the measured vehicle to a nonlinear tire model reflecting a slip angle of each tire, A device may be provided.

The information obtaining unit may obtain the acceleration in the longitudinal and lateral directions, the yaw rate, the wheel speed of each wheel, and the steering angle from the sensors installed in the vehicle as the state information of the vehicle.

The vehicle position estimating unit may include a slip angle calculating unit for calculating a slip angle in a longitudinal direction and a lateral direction for each tire using the obtained state information of the vehicle; A vertical drag calculation unit for calculating a vertical drag per tire relative to a contact point between the tire and the road surface; A force calculator calculating the longitudinal and lateral forces of each tire by applying the computed slip angle in the longitudinal and lateral directions of each tire and the vertical drag per tire to the nonlinear tire model reflecting the slip angle of each tire; And a position estimating unit for estimating a position of the vehicle by applying the calculated longitudinal and lateral forces of each tire to a four wheel vehicle model (FWVM) in which the force of each tire is reflected.

The slip angle calculator calculates the slip angle in the longitudinal direction using the longitudinal and lateral speeds of the angular velocity at the center of the wheel, the tires, and the road surface from among the obtained state information of the vehicle, The slip angle in the lateral direction can be calculated using the longitudinal and transverse speeds of the steering angle, the tires and the contact points of the road surface, respectively.

The vertical drag calculation unit can calculate the vertical drag per tire with respect to the contact points of the tire and the road surface in consideration of the torque balance, the length from the center of gravity of the vehicle to the front and rear wheels, the width between the left and right wheels, .

The vehicle position estimating unit may estimate the position of the vehicle on the basis of the extended Kalman filter that stochastically fuses the obtained GPS position of the vehicle with the position of the vehicle estimated from the four-wheel vehicle model.

The vehicle position estimator may include a time update module for calculating a (k + 1) th system state and a (k + 1) th error covariance; And a measurement update module updating the kth state-estimate and the kth error covariance using the obtained state information of the vehicle.

The time update module may apply a first order Euler approximation to discretize the vehicle dynamics model.

The time update module may calculate the matrix A in the Jacoby of the nonlinear system model and calculate the (k + 1) th error covariance using the calculated matrix A. [

The measurement update module calculates a matrix H as a Jacobian based on a measurement model including a vehicle's GPS position, longitudinal and lateral acceleration, and yaw rate, and calculates a Kalman gain from the calculated matrix H, Lt; th > state-estimation can be calculated.

Meanwhile, according to a second aspect of the present invention, there is provided a method for controlling a vehicle, Measuring a GPS position of the vehicle; And estimating a position of the vehicle by applying the obtained state information of the vehicle and the GPS position of the measured vehicle to a nonlinear tire model reflecting the slip angle of each tire, Can be provided.

The step of acquiring the state information of the vehicle may acquire the longitudinal and lateral acceleration, the yaw rate, the wheel speed of each wheel and the steering angle from the sensors installed in the vehicle as the state information of the vehicle.

The step of estimating the position of the vehicle may include calculating a slip angle in a longitudinal direction and a lateral direction for each tire using the obtained state information of the vehicle; Calculating a vertical drag per tire relative to a contact point of the tire and the road surface; Calculating longitudinal and lateral forces of each tire by applying the computed slip angle in the longitudinal and lateral directions of each tire and the vertical drag per tire to a nonlinear tire model reflecting a slip angle for each tire; And estimating a position of the vehicle by applying the calculated longitudinal and lateral forces of each tire to a four wheel vehicle model (FWVM) in which the force of each tire is reflected.

Wherein the calculating of the slip angle includes calculating longitudinal slip angles using the longitudinal and lateral speeds of the angular velocity at the center of the wheel, the tires and the road surface contact points among the obtained state information of the vehicle, The slip angle in the lateral direction can be calculated using the longitudinal and lateral speeds of the steering angle, the tires, and the road surface contact point, respectively.

Wherein the calculating of the vertical drag calculates the vertical drag per tire relative to the contact points of the tire and the road surface in consideration of the torque balance, the length from the center of gravity of the vehicle to the front and rear wheels, the width between the left and right wheels, Can be calculated.

The step of estimating the position of the vehicle may estimate the position of the vehicle on the basis of the extended Kalman filter that stochastically fuses the GPS position of the obtained vehicle and the position of the vehicle estimated from the four-wheel vehicle model.

Estimating the position of the vehicle includes: calculating a (k + 1) -th system state and a (k + 1) -th error covariance; And updating the measurement by updating the kth state-estimate and the kth error covariance using the obtained state information of the vehicle.

Updating the time may apply a first order Euler approximation to discretize the vehicle dynamics model.

The updating of the time may calculate the matrix A in the Jacoby of the nonlinear system model and calculate the (k + 1) th error covariance using the calculated matrix A.

The step of updating the measurement may include calculating a matrix H to a Jacobian based on a measurement model including a vehicle's GPS position, longitudinal and lateral acceleration, and yaw rate, and calculating a Kalman gain from the calculated matrix H To calculate the kth state-estimate.

The embodiments of the present invention estimate the position of the vehicle by applying the vehicle state information and the GPS position of the vehicle to the four wheel vehicle model (FWVM) to which the nonlinear tire model is applied, so that even in a high speed running or cornering state in which a large slip angle occurs A position with high accuracy can be estimated.

Embodiments of the present invention can improve the positional accuracy of position estimation by using a four wheel vehicle model that considers the forces of each wheel rather than a bicycle model.

Embodiments of the present disclosure use a nonlinear model (e.g., a Dugoff model) in which the slip of tires and roads is modeled to accurately estimate the slip of tires and roads, Direction force can be accurately calculated.

Embodiments of the present invention are based on the assumption that the speed of each wheel, the vehicle speed and the speed of each wheel from various modules (for example, anti-lock brake system (ABS), electronic stability control (ESC) The position of the vehicle can be accurately estimated without acquiring additional devices or components by acquiring the state information of the vehicle such as acceleration, steering angle, and the like through the basic communication network (e.g., CAN (Controller Area Network) bus) .

Embodiments of the present invention use an extended Kalman filter (EKF) that stochastically fuses the state information of various vehicles and the four-wheel vehicle model (FWVM) to which the non-linear tire model is applied, It is possible to accurately estimate the position of the vehicle even if a GPS having a large error (e.g., a position error of 10 m) is used.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram of a vehicle position estimating apparatus using a nonlinear tire model according to an embodiment of the present invention. FIG.
2 is an explanatory diagram of a four-wheel vehicle model applied to a vehicle position estimating unit according to an embodiment of the present invention.
3 is an explanatory diagram of an extended Kalman filter applied to a vehicle position estimating unit according to an embodiment of the present invention.
4 is an explanatory diagram of a road environment for simulation in which vehicle position estimation is applied according to the embodiment of the present invention.
FIG. 5 is a simulation result of the trajectory in the longitudinal and lateral directions according to the embodiment of the present invention. FIG.
FIGS. 6 to 8 are simulation results of the position estimation of the vehicle according to the embodiment of the present invention.
FIGS. 9 and 10 are simulation results of estimated slip angles of the tire in the longitudinal and transverse directions according to the embodiment of the present invention. FIG.
FIGS. 11 and 12 are simulation results of estimated slip errors for each tire in the longitudinal and lateral directions according to the embodiment of the present invention. FIG.
13 is a flowchart illustrating a method of estimating a position of a vehicle using a nonlinear tire model according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

In describing the embodiments, descriptions of techniques which are well known in the technical field to which this specification belongs and which are not directly related to this specification are not described. This is for the sake of clarity without omitting the unnecessary explanation and without giving the gist of the present invention.

For the same reason, some of the components in the accompanying drawings are exaggerated, omitted, or schematically illustrated. Also, the size of each component does not entirely reflect the actual size. In the drawings, the same or corresponding components are denoted by the same reference numerals.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram of a vehicle position estimating apparatus using a nonlinear tire model according to an embodiment of the present invention. FIG.

1 is a block diagram of a vehicle position estimating apparatus 100 using a nonlinear tire model according to an embodiment of the present invention. The vehicle position estimating apparatus 100 includes an information obtaining unit 110, a position measuring unit 120, and a vehicle position estimating unit 130.

The specific configuration and operation of each component of the vehicle position estimation apparatus 100 using the nonlinear tire model of FIG. 1 will be described below.

The information obtaining unit 110 obtains the state information of the vehicle. The information obtaining unit 110 obtains the longitudinal and lateral accelerations, the yaw rate, the wheel speed of each wheel, and the steering angle as state information of the vehicle. Here, the information obtaining unit 110 may include an ABS (Anti-lock Brake System) module, an ESC (Electronic Stability Control) module, and an EPS (Electric Power Steering) module installed in the vehicle. therefore. The information obtaining unit 110 obtains the wheel speed, the vehicle speed, and the acceleration from the wheel speed sensor, the vehicle speed sensor, the acceleration sensor, the yaw rate sensor, and the steering angle sensors included in each module installed in the vehicle, acceleration, yaw rate, and steering angle. The information acquiring unit 110 can acquire, through the controller area network (CAN) bus, information about each sensor installed in the vehicle as the parts of the vehicle become electronic.

The position measuring unit 120 measures the GPS position including the X-coordinate and the Y-coordinate of the vehicle using a low-cost GPS module. Here, a low-cost GPS module is a navigation system based on GPS which is used because of its low price and convenience, although its error range is wider than that of high-precision GPS.

The vehicle position estimating unit 130 applies the state information of the vehicle obtained by the information obtaining unit 110 and the GPS position of the vehicle measured by the position measuring unit 120 to a nonlinear tire model reflecting the slip angle of each tire, Estimate the location.

More specifically, the vehicle position estimating unit 130 estimates the vehicle position using a four wheel vehicle model (FWVM) and a dugoff nonlinear tire model based on an extended Kalman filter (EKF) do. Since the nonlinear tire model of the four-wheel vehicle model (FWVM) and the dugoff (dugoff) is implemented in units of modules, the present invention can be applied to a position estimation method using various vehicle state information. In addition, there is an advantage in that it can be designed for each function in a low-cost micro-controller unit (MCU).

The vehicle position estimating unit 130 estimates the position of the vehicle using a four wheel vehicle model (FWVM) reflecting the force per wheel. The vehicle position estimating unit 130 estimates the position of the vehicle using a tire model that reflects a slip angle of each tire having a strong nonlinearity. The position estimation method according to the four-wheel vehicle model (FWVM) can exhibit high performance.

1, the vehicle position estimating unit 130 includes a slip angle calculating unit 131, a vertical drag calculating unit 132, a force calculating unit 133, and a position estimating unit 134 can do.

The specific configuration and operation of each component of the vehicle position estimation unit 130 shown in FIG. 1 will be described below.

The slip angle calculator 131 calculates the slip angle in the longitudinal direction and the lateral direction for each tire using the state information of the vehicle obtained by the information obtaining unit 110. [

Specifically, the slip angle calculator 131 can calculate slip angles in the longitudinal and lateral directions of each tire as shown in Equation (1) below. The slip angle calculator 131 can calculate the longitudinal slip angle using the longitudinal and transverse speeds of the angular velocity at the center of the wheel, the contact points of the tire and the road surface. Further, the slip angle calculating section 131 may calculate the slip angles in the lateral direction using the longitudinal and transverse speeds of the contact points of the steering angle, the tire, and the road surface.

Here, the slip angle in the longitudinal direction and the longitudinal slip and sideslip angle in the lateral direction can be calculated as shown in the following equation (1).

는 각 타이어의 종 및 횡 방향의 종 방향의 슬립각 및 횡 방향의 사이드 슬립각,

은 타이어의 반지름,

는 각각 타이어 및 도로의 접촉점(tire/road contact point)의 종 및 횡 방향(longitudinal, lateral)의 속도,

는 각 휠 중심에서의 각속도,

는 조향각,

는 차량 CG에서의 종 및 횡방향 속도를 이용하여 슬립각을 계산하는 것으로 atan2는 tan의 역함수를 나타낸다.here,

The slip angle in the vertical direction of the longitudinal and transverse directions of each tire and the slip angle between the slip angle in the longitudinal direction,

The radius of the tire,

Are the longitudinal and lateral speeds of the tires and road contact points, respectively,

The angular velocity at each wheel center,

The steering angle,

Calculates the slip angle using the longitudinal and transverse velocities in the vehicle CG, and atan2 represents the inverse of tan.

The vertical drag calculation unit 132 calculates the vertical drag per tire with respect to the contact point between the tire and the road surface. The vertical drag calculation unit 132 calculates the vertical drag per tire relative to the contact point between the tire and the road surface in consideration of the torque balance, the length from the center of gravity of the vehicle to the front and rear wheels, the width between the left and right wheels, Is calculated as shown in the following equation (2).

는 각 타이어 및 도로의 접촉점의 수직 항력(Normal forces),

는 종 및 횡 방향의 가속도,

는 CG로부터 전방 휠까지의 길이,

는 CG로부터 후방 휠까지의 길이,

는 좌측 및 우측 휠 간의 너비,

는 중력 가속도(9.8m/s²),

는 지면에서 차량의 CG까지의 높이를 나타낸다.here,

Normal forces of the contact points of each tire and road,

The acceleration in the longitudinal and transverse directions,

Is the length from the CG to the front wheel,

The length from the CG to the rear wheel,

The width between the left and right wheels,

(9.8 m / s < ² >),

Represents the height from the ground to the CG of the vehicle.

)은 CG에서의 종 및 횡 방향의 가속도(

)에 의해 결정되고, 수직 항력(

)은 각 타이어의 수직 항력(normal force)의 합과 같게 된다. 따라서 각 타이어의 수직 항력은 토크 밸런스(Torque balance)에 의해 상기의 [수학식 2]와 같이 계산될 수 있다.If there is no effect of the vehicle's roll, pitch and suspension dynamics, the vertical drag at the CG of the vehicle

) Is the acceleration in the longitudinal and transverse directions in CG

), And the vertical drag (

) Is equal to the sum of the normal forces of each tire. Therefore, the vertical drag force of each tire can be calculated as shown in Equation (2) by the torque balance.

The force calculator 133 calculates the force and the lateral direction of each tire using the slip angle in the longitudinal and lateral directions of each tire calculated in the slip angle calculator 131 and the vertical drag per tire calculated in the vertical drag calculator 132 Lt; / RTI >

The force calculation unit 133 uses the tire model to calculate the longitudinal and lateral forces of each tire. Tires are very nonlinear. Therefore, the position estimating unit 134 according to the embodiment of the present invention can estimate a tire having a similar performance to that of the experimental results with only a small number of parameters among quasi-static tire models, Model (dugoff tire model) is used. The force calculation unit 133 can calculate the force in the longitudinal direction and the lateral direction of each tire through the tire model of the dug-off according to the following equation (3).

는 타이어의 종 방향의 강성(longitudinal stiffness),

는 타이어의 코너링의 강성(cornering stiffness),

는 각 타이어의 종 및 횡 방향의 종 방향의 슬립각 및 횡 방향의 사이드 슬립각을 나타낸다. 또한,

는 미끄러짐에 따라 그 영역을 구분하는 인자(Sliding Boundary Point)로

에 따라

가 결정되며, 각각 하기의 [수학식 4] 및 [수학식 5]와 같이 나타낼 수 있다.here,

The longitudinal stiffness of the tire,

The cornering stiffness of the tire,

Represents the longitudinal slip angle in the longitudinal direction and the lateral slip angle in the lateral direction of each tire. Also,

(Sliding Boundary Point) that divides the area according to the slip

Depending on the

And can be expressed by the following equations (4) and (5), respectively.

는 i번째 각 타이어 및 도로의 접촉점의 수직 항력(Normal forces), 여기서,

는 타이어의 종 방향의 강성(longitudinal stiffness),

는 타이어의 코너링의 강성(cornering stiffness),

는 각 타이어의 종 및 횡 방향의 종 방향의 슬립각 및 횡 방향의 사이드 슬립각,

는 타이어와 도로간 접촉면의 마찰 계수(tire/road friction coefficient)를 나타낸다. here,

Is the normal forces of the contact points of each ith tire and road,

The longitudinal stiffness of the tire,

The cornering stiffness of the tire,

The slip angle in the vertical direction of the longitudinal and transverse directions of each tire and the slip angle between the slip angle in the longitudinal direction,

(Tire / road friction coefficient) between the tire and the road.

On the other hand, the position estimating unit 134 applies the four wheel vehicle model (FWVM), which is calculated by the force calculating unit 133, to the four-wheel vehicle model in which the force of each tire in the longitudinal direction and the lateral direction is reflected, .

2 is an explanatory diagram of a four-wheel vehicle model applied to the vehicle position estimating unit 130 according to an embodiment of the present invention.

The vehicle position estimating unit 130 according to an embodiment of the present invention estimates the position of the vehicle using the dynamic model shown in FIG. 2, in order to consider the slip and the force of each tire. From the dynamic model shown in FIG. 2, the force of each tire can be expressed by the following equation (6).

는 각 휠의 종 및 횡 방향의 힘, m은 차량의 무게 중심점(CG, Center of Gravity)에서의 질량,

는 무게 중심점에서의 요 레이트,

는 각각 차량의 CG에서의 종 및 횡 방향 속도,

는 차량의 CG에서의 종 및 횡 방향 가속도,

은 CG에서 Z축 방향의 모멘트(moment),

는 CG에서의 관성모멘트 상수를,

는 CG에서의 요 가속도를 나타낸다.here,

Is the force in the longitudinal and lateral directions of each wheel, m is the mass at the center of gravity (CG) of the vehicle,

Is the yaw rate at the center of gravity,

Are the longitudinal and transverse speeds in the CG of the vehicle, respectively,

The longitudinal and transverse acceleration at the CG of the vehicle,

Is a moment in the Z-axis direction of the CG,

Is the moment of inertia in CG,

Represents the yaw acceleration in the CG.

Here, the sum of the forces of the respective tires can be calculated as shown in the following equation (7).

는 각 휠의 종 및 횡 방향의 힘, 조향각으로 좌측 및 우측 모두 동일하다고 가정한다.

는 CG로부터 전방 휠까지의 길이,

는 CG로부터 후방 휠까지의 길이,

는 좌측 및 우측 휠 간의 너비,

는 조향각을 나타낸다.here,

Are assumed to be the same in both the left and right sides due to the longitudinal and lateral forces and steering angles of each wheel.

Is the length from the CG to the front wheel,

The length from the CG to the rear wheel,

The width between the left and right wheels,

Represents the steering angle.

3 is an explanatory diagram of an extended Kalman filter applied to the vehicle position estimating unit 130 according to an embodiment of the present invention.

The Extended Kalman Filter (EKF) applied to the vehicle position estimation unit 130 may fuse various vehicle state information based on a Bayesian filter. Here, the extended Kalman filter can be applied directly in a nonlinear system without any special linearization operation.

As shown in FIG. 3, the extended Kalman filter is divided into a time update module and a measurement update module.

)와 k+1번째의 오류 공분산(error covariance ahead)(

)을 계산한다. 여기서, k+1번째의 시스템 상태(

)와 k+1번째의 오류 공분산(

)는 시간 스텝(time step) k마다 계산되는 이산 시스템을 나타낸다. 따라서, 상기의 [수학식 6 및 7]과 같이 나타난 동역학 모델(dynamic model)이 실시간으로 동작하기 위해서는 이산-시간 시스템(discrete-time system) 모델로 변환되어야 한다. 시간 업데이트 모듈은 하기의 [수학식 8]과 같이, 1차 오일러 근사화(first-order Euler approximation)를 적용하여 차량의 동역학 모델에 대한 수학식(Vehicle dynamic model equation)들을 이산화할 수 있다.Looking at the time update module, the time update module updates the (k + 1) th system state ahead (

) And (k + 1) -th error covariance ahead (

). Here, the (k + 1) th system state

) And (k + 1) th error covariance (

) Represents the discrete system that is calculated for each time step k. Therefore, the dynamic model shown in Equations (6) and (7) above must be converted to a discrete-time system model in order to operate in real time. The time update module may discretize the vehicle dynamic model equations for a vehicle by applying a first-order Euler approximation as in Equation (8) below.

는 k번째 상태공간백터(State Space Vector),

는 k-1번째 상태공간백터,

는 상태공간백터의 미분치,

는 시스템의 샘플링 구간(sampling period)을 나타낸다.here,

Is a kth state space vector,

K-1 < th > state space vector,

Is the derivative of the state space vector,

Represents the sampling period of the system.

Therefore, a continuous vehicle model can be approximated to a discrete vehicle model using an Euler approximation method, as shown in Equation (9) below.

는 각 휠의 종 및 횡 방향의 힘, m은 차량의 무게 중심점(CG, Center of Gravity)에서의 질량,

는 무게 중심점에서의 요 레이트,

는 각각 차량의 CG에서의 종 및 횡 방향 속도,

은 CG에서 Z축 방향의 모멘트(moment),

는 CG에서 Z축 방향의 관성모멘트(moment of inertia),

는 CG에서의 요 가속도,

는 x축 방향의 차량위치의 미분치,

는 x축 방향의 차량속도의 미분치,

는 y축 방향의 차량위치의 미분치,

는 y축 방향의 차량속도의 미분치,

는 x축 방향의 추정된 차량위치,

는 x축 방향의 추정된 차량속도,

는 y축 방향의 추정된 차량위치,

는 y축 방향의 추정된 차량속도,

는 추정된 요 각도,

는 추정된 요레이트,

는 시스템의 샘플링 구간(sampling period)을 나타낸다.here,

Is the force in the longitudinal and lateral directions of each wheel, m is the mass at the center of gravity (CG) of the vehicle,

Is the yaw rate at the center of gravity,

Are the longitudinal and transverse speeds in the CG of the vehicle, respectively,

Is a moment in the Z-axis direction of the CG,

Is the moment of inertia in the Z-axis direction of the CG,

The yaw acceleration in CG,

The derivative of the vehicle position in the x-axis direction,

The derivative of the vehicle speed in the x-axis direction,

The derivative of the vehicle position in the y-axis direction,

The derivative of the vehicle speed in the y-axis direction,

The estimated vehicle position in the x-axis direction,

The estimated vehicle speed in the x-axis direction,

The estimated vehicle position in the y-axis direction,

The estimated vehicle speed in the y-axis direction,

Is the estimated yaw angle,

The estimated yaw rate,

Represents the sampling period of the system.

)은 하기의 [수학식 10]과 같이 나타낼 수 있다.And the system model expressed by discrete state vector (

) Can be expressed by the following equation (10).

는 각각 차량 CG의 종 및 횡 방향 속도,

는 차량 CG의 요 각도,

는 차량 CG의 요 레이트를 나타낸다.Here, x and y are the longitudinal and transverse positions of the vehicle CG, respectively,

Respectively denote the longitudinal and transverse speeds of the vehicle CG,

The yaw angle of the vehicle CG,

Represents the yaw rate of the vehicle CG.

)를 구하기 위해서는 매트릭스 A(matrix A)가 필요하다. 매트릭스 A는 비선형 시스템 모델의 자코비안(Jacobian)으로 하기의 [수학식 11]과 같이 계산될 수 있다.k + 1 < th > error covariance ahead (

) A matrix A (matrix A) is required to obtain the. Matrix A can be calculated as Jacobian of the nonlinear system model as shown in Equation (11) below.

는 비선형 시스템 모델의 자코비안(Jacobian),

는 각각 차량의 CG에서의 종 및 횡 방향 속도,

는 차량 CG의 요 각도(진행방향),

는 차량 CG의 요 레이트를 나타낸다.here,

Jacobian of the nonlinear system model,

Are the longitudinal and transverse speeds in the CG of the vehicle, respectively,

(Advancing direction) of the vehicle CG,

Represents the yaw rate of the vehicle CG.

Meanwhile, a measurement update module will be described.

)과 오류 공분산(error covariance)(

)을 업데이트한다. 측정 벡터(measurement vector)(

)는 하기의 [수학식 12]와 같이 나타낼 수 있다.The measurement update module of the vehicle position estimating unit 130 uses the sensor data to estimate a state-estimate

) And error covariance (

). Measurement vector (

) Can be expressed by the following equation (12).

,

는 ESC로부터 획득된 종 및 횡 방향의 가속도,

는 ESC로부터 요 레이트를 나타낸다. 측정 업데이트 모듈에 적용되는 측정 모델(measurement model)(

)은 하기의 [수학식 13]과 같이 설정될 수 있다.Where X and Y are the vehicle's longitudinal and lateral GPS positions from the GPS,

,

The acceleration in the horizontal direction and the acceleration in the horizontal direction obtained from the ESC,

Represents the yaw rate from ESC. The measurement model applied to the measurement update module (

Can be set as shown in the following equation (13).

는 측정 모델(measurement model), X, Y는 GPS로부터 차량의 GPS 위치,

는 각 휠의 종 및 횡 방향의 힘,

는 ESC로부터 획득된 요 레이트를 나타낸다.here,

X , Y are the GPS position of the vehicle from the GPS,

The longitudinal and lateral forces of each wheel,

Represents the yaw rate obtained from the ESC.

)을 계산하고, 그 계산된 칼만 이득을 통해 상태-추정(State-estimate)(

)를 계산할 수 있다.The measurement update module uses Kalman gain (

), And calculates a state-estimate ("

) Can be calculated.

The matrix H (matrix H) is required to calculate the Kalman gain. H is the Jacobian of the measurement model, and is expressed by Equation (14) below.

를 시스템 모델

로 편미분한 매트릭스를 나타낸다.Here,

System model

.

4 is an explanatory diagram of a road environment for simulation in which vehicle position estimation is applied according to the embodiment of the present invention.

The method of estimating the position of the vehicle according to the embodiment of the present invention is analyzed through Monte Carlo simulation. Monte Carlo simulation uses the commercial program "ASM (Automotive Simulation Models)". "ASM" is a kinetic model of a vehicle consisting of a 24-DOF multi-body system. The parameters of the simulation vehicle are set as shown in [Table 1] below. The parameters and noise of the sensors attached to the vehicle are set as shown in [Table 2] below. The noise of each sensor is assumed to be Gaussian white noise. In addition, the sample rate at which the position estimation algorithm is performed is set to 50 Hz.

The simulation scenario is carried out in the handling track of 3000m of the road model of "ASM".

When the vehicle is running on the track, set the vehicle's longitudinal and lateral acceleration to a maximum of 5 m / s2 to generate large slip angles and ratios. In addition, the maximum speed of the vehicle is limited to 80 km / h so that the vehicle does not depart from the course.

As shown in Table 1 above, the vehicle parameters include the vehicle mass, the yaw moment of inertia, the distance from the center of gravity to the front wheel, The distance from the center of gravity to the rear wheel (distance from CG to the rear wheel), the track width, the longitudinal stiffness of each tire, and the cornering stiffness of each tire each tire).

As shown in Table 2, the wheel speed sensor, the acceleration sensor, the gyro sensor, and the GPS sensor are 0.1 rad / s and 1.0 m / s, respectively, s2, 0.1 rad / s, and 10 m of noise.

FIG. 5 is a simulation result of longitudinal and lateral trajectories according to an embodiment of the present invention. FIG.

As shown in Fig. 5, the vehicle position estimation apparatus 100 according to the embodiment of the present invention implements a simulation on a handling track so that a large slip angle and a ratio are generated. You can see that the GPS signal is distributed around the reference signal. It can be seen that the position of the vehicle estimated by the position estimating apparatus 100 according to the embodiment of the present invention is similar to the reference signal.

FIGS. 6 to 8 are simulation results of the position estimation of the vehicle according to the embodiment of the present invention.

Simulated results of the estimated vehicle state and the estimated vehicle state error are as shown in FIGS. 6 and 7. FIG.

As shown in FIGS. 6 (a) and 6 (b), it can be seen that the longitudinal and lateral positions accurately estimate the reference signal.

As shown in FIGS. 7A and 7B, it can be seen that the longitudinal and lateral position errors occur at about 1 m or so.

As shown in FIGS. 6 (c) and 6 (d), it can be seen that the longitudinal and lateral velocities also well estimate the reference signal.

As shown in FIGS. 7 (c) and 7 (d), longitudinal and lateral velocity errors in the longitudinal and lateral directions are about 1 m / s.

As shown in FIGS. 6 (e) and 6 (f), it can be seen that the yaw and yaw rates similarly estimate the reference signal.

As shown in FIGS. 7 (e) and 7 (f), it can be seen that the yaw and yaw rate have a noise of about 20 deg / s.

)를 1차 오일러 근사화로 적분하는 모델이기 때문이다. 1차 오일러 근사화(first-order Euler approximation)으로 적분하는 모델이기 때문이다.However, as shown in FIGS. 8 (a) and 8 (b), it is seen that the estimated yaw has an error of -30 degrees maximum up to 22s. However, from 2 seconds onwards, it converges to within about 5 deg. The reason why the yaw error occurs in the early stage is simply the yaw rate

) With a first-order Euler approximation. Because it is a model that integrates into a first-order Euler approximation.

FIGS. 9 and 10 are simulation results of estimated slip angles of the tire in the longitudinal and lateral directions according to the embodiment of the present invention. FIG.

As shown in Figs. 9A and 9B and Figs. 10A and 10B, the longitudinal slip angle in the longitudinal direction of the front tire and the longitudinal slip angle in the transverse direction (lateral slip angle) occurs in the range of 5deg.

On the other hand, as shown in Figs. 9 (c) and 9 (d), a slip in the longitudinal direction of the rear tire occurs at a maximum of 50 deg.

10 (c) and 10 (d), it can be seen that a lateral slip occurs in a range of 10 deg. These results show that the vehicle is running rapidly.

FIGS. 11 and 12 are simulation results of estimated slip errors for each tire in the longitudinal and lateral directions according to the embodiment of the present invention. FIG.

As shown in Figs. 11 (a), 11 (b), 11 (c), and 11 According to the simulation results, the slip errors of each tire have slip angle errors of about 4 deg in both the longitudinal and transverse directions.

However, when the vehicle is stopped, it can be seen that the slip angle of each tire increases. The reason is due to the limitations of the model for estimating the tire slip angle. That is, a model for estimating the slip angle of a tire can be expressed by the following equation (1).

)를 더해주었다. 여기서, 엡실론(

)은 시스템에 영향을 주지 않는 범위 내에서, 시뮬레이션을 통해 구한 것이다.[Equation 1] is calculated by dividing the speed of the contact point between the tire and the road. Therefore, when the speed of the contact point between the tire and the road becomes 0, the vehicle diverges. In this specification, in order to solve the problem of numerical instability, the speed of the contact point of the tire and the road is set to zero so that the speed of the epsilon

). Here, epsilon (

) Is obtained through simulation within a range that does not affect the system.

13 is a flowchart illustrating a method of estimating a position of a vehicle using a nonlinear tire model according to an embodiment of the present invention.

The information obtaining unit 110 in the vehicle position estimating apparatus 100 obtains the state information of the vehicle (S1302). Further, the position measurement unit in the vehicle position estimation apparatus 100 measures the GPS position of the vehicle.

The vehicle position estimating unit 130 in the position estimating apparatus 100 calculates the slip angles in the longitudinal and lateral directions of each wheel by using the vehicle position information and the GPS position obtained by the information obtaining unit 110 S1304).

The vehicle position estimating unit 130 in the position estimating apparatus 100 calculates the vertical drag per wheel by using the state information of the vehicle and the GPS position acquired by the information obtaining unit 110 (S1306).

Subsequently, the vehicle position estimating unit 130 calculates longitudinal / lateral forces of each tire using the nonlinear tire model from the computed slip angles of the longitudinal / lateral slip and the vertical slip of each wheel (1308).

The vehicle position estimation unit 130 estimates the position of the vehicle using the extended Kalman filter (S1310).

1 to 3, the vehicle position estimation apparatus 100 according to the embodiment of the present invention estimates the position of the vehicle in accordance with a vehicle position estimation method necessary for the ADAS to be stably driven. The vehicle position estimation apparatus 100 estimates the position using a low-cost GPS sensor and sensor information attached to the vehicle, instead of using an expensive sensor to estimate a high-precision position without interruption. In addition, we applied FWVM, a vehicle dynamics model, and an extended Kalman filter (EKF) algorithm based on a Dugoff tire model, which modeled a tire with high nonlinear characteristics with a small number of parameters.

As described in FIGS. 4 to 12, the performance of the proposed algorithm was evaluated using a commercial simulation ASM. As a result, it was confirmed that the species and lateral position were estimated to be within 1m. The longitudinal and lateral velocities were also estimated to be within the range of 1 m / s.

The simulation results show that Yaw estimates the error of about 5deg over time, although the error of the initial position of the yaw occurred up to -30deg. It was confirmed that the yaw rate was estimated at an error of 20 deg / s. Also, it can be confirmed that the tire slip angle in the longitudinal and transverse directions is estimated by an error of 6 deg and 4 deg, respectively.

The vehicle position estimation apparatus 100 according to the embodiment of the present invention can increase the position accuracy of the vehicle by applying the nonlinear vehicle model.

It will be understood by those skilled in the art that the present specification may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. It is therefore to be understood that the above-described embodiments are illustrative in all aspects and not restrictive. The scope of the present specification is defined by the appended claims rather than the foregoing detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalents are included in the scope of the present specification Should be interpreted.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, It is not intended to limit the scope of the specification. It will be apparent to those skilled in the art that other modifications based on the technical idea of the present invention are possible in addition to the embodiments disclosed herein.

100:
110: Information obtaining unit
120:
130: vehicle position estimating unit
131: Slip angle calculating unit
132: vertical force calculation unit
133: Force calculation unit
134:

Claims (20)

An information obtaining unit for obtaining status information of the vehicle;
A position measuring unit for measuring a GPS position of the vehicle; And
Estimating a position of the vehicle by applying the obtained state information of the vehicle and the GPS position of the measured vehicle to a nonlinear tire model reflecting a slip angle of each tire,
And estimating a position of the vehicle using the nonlinear tire model. The method according to claim 1,
The information obtaining unit
An apparatus for estimating a position of a vehicle using a nonlinear tire model that acquires longitudinal and lateral accelerations, yaw rates, wheel speeds of respective wheels, and steering angles from sensors installed in a vehicle as vehicle state information. The method according to claim 1,
The vehicle position estimating unit
A slip angle calculation unit for calculating a slip angle in a longitudinal direction and a lateral direction for each tire using the obtained state information of the vehicle;
A vertical drag calculation unit for calculating a vertical drag per tire relative to a contact point between the tire and the road surface;
A force calculator calculating the longitudinal and lateral forces of each tire by applying the computed slip angle in the longitudinal and lateral directions of each tire and the vertical drag per tire to the nonlinear tire model reflecting the slip angle of each tire; And
And calculates the position of the vehicle by applying the calculated longitudinal and lateral forces of each tire to a four wheel vehicle model (FWVM)
And estimating a position of the vehicle using the nonlinear tire model. The method of claim 3,
The slip angle calculating unit
Calculating longitudinal slip angles using longitudinal and lateral speeds of the angular velocity at the center of the wheel, the contact points of the tire and the road surface among the obtained state information of the vehicle, and calculating a steering angle, A device for estimating a position of a vehicle using a nonlinear tire model that calculates slip angles in the lateral direction using the longitudinal and transverse velocities of the contact points on the road surface. The method of claim 3,
The vertical drag calculation unit
A vehicle using a nonlinear tire model that calculates the vertical drag per tire relative to the contact point between the tire and the road surface in consideration of the torque balance, the length from the center of gravity of the vehicle to the front and rear wheels, the width between the left and right wheels, . The method according to claim 1,
The vehicle position estimating unit
And estimating the position of the vehicle on the basis of the extended Kalman filter for stochastically fusing the obtained GPS position of the vehicle and the position of the vehicle estimated from the four-wheel vehicle model. The method according to claim 6,
The vehicle position estimating unit
a time update module for calculating a (k + 1) -th system state and a (k + 1) -th error covariance; And
And a measurement update module for updating the kth state-estimate and the kth error covariance using the obtained state information of the vehicle,
And estimating a position of the vehicle using the nonlinear tire model. 8. The method of claim 7,
The time update module
An apparatus for estimating the position of a vehicle using a nonlinear tire model that discretizes a vehicle dynamics model by applying a first order Euler approximation. 8. The method of claim 7,
The time update module
Calculating a matrix A in a Jacobian of a nonlinear system model, and calculating a (k + 1) th error covariance using the calculated matrix A; 8. The method of claim 7,
The measurement update module
Estimating the kth state-estimate from the calculated matrix H by calculating the matrix H based on the measured model including the vehicle's GPS position, longitudinal and lateral acceleration, and yaw rate, An apparatus for estimating a position of a vehicle using a nonlinear tire model to be calculated. Obtaining status information of the vehicle;
Measuring a GPS position of the vehicle; And
Estimating the position of the vehicle by applying the obtained state information of the vehicle and the GPS position of the measured vehicle to a nonlinear tire model reflecting a slip angle for each tire
And estimating a position of the vehicle using the nonlinear tire model. 12. The method of claim 11,
The step of acquiring the state information of the vehicle
A method for estimating a position of a vehicle using a nonlinear tire model that acquires longitudinal and lateral acceleration, yaw rate, wheel speed of each wheel, and steering angle from the sensors installed in the vehicle as vehicle state information. 12. The method of claim 11,
The step of estimating the position of the vehicle
Calculating a slip angle in a longitudinal direction and a lateral direction for each tire by using the obtained state information of the vehicle;
Calculating a vertical drag per tire relative to a contact point of the tire and the road surface;
Calculating longitudinal and lateral forces of each tire by applying the computed slip angle in the longitudinal and lateral directions of each tire and the vertical drag per tire to a nonlinear tire model reflecting a slip angle for each tire; And
Estimating the position of the vehicle by applying the calculated longitudinal and lateral forces of the respective tires to a four wheel vehicle model (FWVM)
And estimating a position of the vehicle using the nonlinear tire model. 14. The method of claim 13,
The step of calculating the slip angle
Calculating longitudinal slip angles using longitudinal and lateral speeds of the angular velocity at the center of the wheel, the contact points of the tire and the road surface among the obtained state information of the vehicle, and calculating a steering angle, A method for estimating a position of a vehicle using a nonlinear tire model that calculates a lateral slip angle using the longitudinal and transverse speeds of the contact points of the road surface, respectively. 14. The method of claim 13,
The step of calculating the vertical drag
A vehicle using a nonlinear tire model that calculates the vertical drag per tire relative to the contact point between the tire and the road surface in consideration of the torque balance, the length from the center of gravity of the vehicle to the front and rear wheels, the width between the left and right wheels, / RTI > 12. The method of claim 11,
The step of estimating the position of the vehicle
And estimating a position of the vehicle on the basis of an extended Kalman filter that stochastically fuses the obtained GPS position of the vehicle with the position of the vehicle estimated from the four-wheel vehicle model. 17. The method of claim 16,
The step of estimating the position of the vehicle
updating a time by calculating a (k + 1) th system state and a (k + 1) -th error covariance; And
Updating the measurement by updating the kth state-estimate and the kth error covariance using the obtained state information of the vehicle
And estimating the position of the vehicle using the nonlinear tire model. 18. The method of claim 17,
The step of updating the time
A Method for Estimating Vehicle Location Using Nonlinear Tire Model Discretizing Vehicle Dynamics Model Using First Order Euler Approximation. 18. The method of claim 17,
The step of updating the time
Calculating a matrix A in a Jacobian of a nonlinear system model, and calculating a (k + 1) th error covariance using the calculated matrix A. A method for estimating a position of a vehicle using a nonlinear tire model. 18. The method of claim 17,
The step of updating the measurement
Estimating the kth state-estimate from the calculated matrix H by calculating the matrix H based on the measured model including the vehicle's GPS position, longitudinal and lateral acceleration, and yaw rate, A method for estimating the position of a vehicle using a nonlinear tire model.

Images