JP2015217707A - Vehicle control device and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To control a vehicle without making an occupant of the vehicle feel wobble.SOLUTION: A target locus calculation part 26 calculates a target locus of a vehicle on the basis of a detected lane boundary line. A state estimation part 28 estimates curvature of the target locus, lateral position deviation of the vehicle to the target locus, and yaw angle deviation of the vehicle to the target locus. A curvature noise calculation part 30 calculates curvature noise included in the curvature of the target locus. A feedback gain setting part 34 sets feedback gain which is predetermined so that a field angular rate of a fixation point of the vehicle's driver becomes smaller than a perception threshold value. A locus tracking controller 36 calculates a target steering angle of the vehicle on the basis of the set feedback gain according to feed forward control of the curvature of the target locus, and feedback control of the lateral position deviation and yaw angle deviation of the vehicle. A steering angle controller 38 controls the vehicle so as to achieve the calculated target steering angle.

Description

  The present invention relates to a vehicle control device and a program, and more particularly, to a vehicle control device and a program for calculating a driving operation amount of a vehicle so as to follow a target locus.

  Conventionally, a control technique related to prevention of vehicle wobbling is known. For example, in the technique described in Patent Document 1, the filtered curvature is used for control in order to prevent the vehicle from wobbling due to the noise of the curvature detected from the image. However, if the filtered curvature is always used for the control, the turning start timing is delayed due to the phase delay, so the detected curvature is adopted when the deviation between the detected curvature and the filtering curvature tends to increase.

  In the technique described in Patent Document 2, it is considered that the estimation accuracy of the corresponding state quantity is lower as the diagonal value of the estimation error covariance matrix of the Kalman filter is larger. In this technique, the feedback gain is determined using the LQ control theory, and by setting the weight of the state quantity with low estimation accuracy in the evaluation function to be small, the feedback amount of the state quantity with low estimation accuracy is reduced. The effect of estimation error is suppressed.

JP 2006-347461 A JP 2007-302204 A

  However, in the technique described in Patent Document 1, when the detection curvature and the filtering curvature are switched, the value of the curvature used for control may change abruptly and affect the vehicle behavior.

  Moreover, in the technique described in Patent Document 2, the weight of the evaluation function in the LQ control theory is merely limited to the relative size of each state quantity and the manipulated variable, and is perceived by the occupant. It does not limit the size of physical quantities (gazing angle viewing angle, lateral jerk). For this reason, the sense of wandering may not be suppressed.

  The present invention has been made in view of the above circumstances, and provides a vehicle control device and a program capable of controlling the vehicle so that the vehicle follows the target trajectory without the vehicle occupant feeling a sense of wander. For the purpose.

  To achieve the above object, a vehicle control device according to the present invention detects a road lane boundary line based on a sensor that detects road information ahead of the vehicle and the road information detected by the sensor. On the basis of the target trajectory calculated by the detection means, the target trajectory calculation means for calculating the target trajectory of the vehicle based on the lane boundary line detected by the detection means, and the target trajectory calculated by the target trajectory calculation means, Based on the detection result of the lane boundary line by the detection means, the estimation means for estimating the curvature of the target locus, the lateral position deviation of the vehicle with respect to the target locus, and the yaw angle deviation of the vehicle with respect to the target locus, Calculated by curvature noise calculation means for calculating curvature noise included in the curvature of the target locus estimated by the estimation means, and calculated by the curvature noise calculation means On the basis of the curvature noise, the curvature noise is determined in advance so that at least one of the viewing angular velocity of the gazing point of the driver of the vehicle and the lateral jerk obtained by differentiating the lateral acceleration of the vehicle is smaller than a threshold value. Further, feedback gain setting means for setting a feedback gain of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle, feedforward control of the curvature of the target locus, and the lateral position deviation of the vehicle and the yaw angle of the vehicle According to deviation feedback control, the curvature of the target locus estimated by the estimation means, the lateral position deviation of the vehicle, and the yaw angle deviation of the vehicle, and the lateral position deviation of the vehicle set by the feedback gain setting means And a feedback gain of yaw angle deviation of the vehicle, The driving operation amount calculation means for calculating the driving operation amount of the vehicle so as to follow the target locus, and the vehicle is controlled so as to realize the driving operation amount calculated by the driving operation amount calculation means. And a control means.

  The program according to the present invention includes a detection unit that detects a lane boundary line of a road based on the road information detected by a sensor that detects road information ahead of the vehicle, and the lane boundary line detected by the detection unit. A target trajectory calculating means for calculating a target trajectory of the vehicle, a curvature of the target trajectory based on the target trajectory calculated by the target trajectory calculating means, a lateral position deviation of the vehicle with respect to the target trajectory, And an estimation means for estimating a yaw angle deviation of the vehicle with respect to the target trajectory, and curvature noise included in the curvature of the target trajectory estimated by the estimation means based on a detection result of the lane boundary line by the detection means. Curvature noise calculating means for calculating, based on the curvature noise calculated by the curvature noise calculating means, the vehicle driver The lateral position deviation of the vehicle and the yaw angle of the vehicle, which are predetermined for the curvature noise so that at least one of the viewing angular velocity of the gazing point and the lateral jerk obtained by differentiating the lateral acceleration of the vehicle is smaller than a threshold value. The target estimated by the estimating means according to feedback gain setting means for setting a feedback gain of deviation, feedforward control of curvature of the target locus, and feedback control of lateral position deviation of the vehicle and yaw angle deviation of the vehicle Based on the curvature of the trajectory, the lateral position deviation of the vehicle, and the yaw angle deviation of the vehicle, and the lateral position deviation of the vehicle and the feedback gain of the yaw angle deviation of the vehicle set by the feedback gain setting means, The driving operation amount of the vehicle is set so that the vehicle follows the target locus. Driving maneuver amount calculator for output, and to realize the driving operation amount calculated by the driving operation amount calculating means, a program for functioning as a control means for controlling the vehicle.

  According to the present invention, the road information ahead of the vehicle is detected by the sensor. And a detection means detects the lane boundary line of a road based on the road information detected by the sensor.

  Then, the target trajectory calculating means calculates the target trajectory of the vehicle based on the lane boundary line detected by the detecting means.

  Then, the estimation means estimates the curvature of the target locus, the lateral position deviation of the vehicle with respect to the target locus, and the yaw angle deviation of the vehicle with respect to the target locus based on the target locus calculated by the target locus calculation means.

  Then, the curvature noise included in the curvature of the target locus estimated by the estimation means is calculated by the curvature noise calculation means based on the detection result of the lane boundary line by the detection means.

  Then, based on the curvature noise calculated by the curvature noise calculating means by the feedback gain setting means, at least one of the viewing angular velocity of the gaze point of the vehicle driver and the lateral jerk obtained by differentiating the lateral acceleration of the vehicle is smaller than the threshold value. Thus, the feedback gains of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle, which are predetermined with respect to the curvature noise, are set.

  Then, the curvature of the target trajectory estimated by the estimating means according to the feedforward control of the curvature of the target trajectory and the feedback control of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle by the driving operation amount calculating means, and the lateral position of the vehicle Based on the deviation, the yaw angle deviation of the vehicle, and the feedback gain of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle set by the feedback gain setting means, the vehicle operation is performed so that the vehicle follows the target locus. Calculate the operation amount.

  Then, the vehicle is controlled by the control means so as to realize the driving operation amount calculated by the driving operation amount calculating means.

  Thus, based on the detection result of the lane boundary line, the curvature noise included in the curvature of the estimated target trajectory is calculated, and based on the calculated curvature noise, the viewing angular velocity of the gaze point of the vehicle driver and the vehicle The target trajectory curvature is set by setting a feedback gain for the lateral deviation of the vehicle and the yaw angle deviation of the vehicle so that at least one of the lateral jerk obtained by differentiating the lateral acceleration of the vehicle becomes smaller than the threshold value. According to the feedforward control of the vehicle and the feedback control of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle, the curvature of the estimated target trajectory, the lateral position deviation of the vehicle, and the yaw angle deviation of the vehicle, Based on the position deviation and the feedback gain of the yaw angle deviation of the vehicle, the driving operation amount of the vehicle is calculated so that the vehicle follows the target locus, By controlling the vehicle so as to realize the rotation operation amount, without feeling a sense of fluctuation occupant of the vehicle, it is possible to control the vehicle so that the vehicle follows the target locus.

  The feedback gain setting means of the present invention is such that the larger the curvature noise calculated by the curvature noise calculation means, the smaller the feedback gain of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle. The lateral position deviation and the yaw angle deviation feedback gain of the vehicle can be set.

  Further, the detection means of the present invention further calculates a detection reliability indicating a degree of reliability of the detected lane boundary line, and the curvature noise calculation means has the detection reliability calculated by the detection means. The curvature noise can be calculated so that the curvature noise becomes larger as the value is lower.

  The driving operation amount of the present invention may be at least one of a steering angle of the vehicle and a yaw moment of the vehicle.

  The program of the present invention can also be provided by being stored in a storage medium.

  As described above, according to the vehicle control device and program of the present invention, the curvature noise included in the curvature of the estimated target trajectory is calculated based on the detection result of the lane boundary, and the calculated curvature noise is calculated. Based on the vehicle lateral position deviation and the vehicle's lateral deviation determined in advance so that at least one of the lateral jerk obtained by differentiating the viewing angular velocity of the gazing point of the vehicle driver and the lateral acceleration of the vehicle is smaller than a threshold value. Set the feedback gain of the yaw angle deviation, and according to the feedforward control of the curvature of the target locus, and the feedback control of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle, the estimated curvature of the target locus, the lateral position deviation of the vehicle, Based on the yaw angle deviation of the vehicle and the feedback gain of the set lateral position deviation of the vehicle and the yaw angle deviation of the vehicle. By calculating the driving operation amount of the vehicle so as to follow the target trajectory and controlling the vehicle so as to realize the driving operation amount, the vehicle follows the target trajectory without feeling a sense of wandering of the vehicle occupant. Thus, the effect that the vehicle can be controlled is obtained.

It is a figure which shows schematic structure of the vehicle control apparatus which concerns on the 1st Embodiment of this invention. It is a block diagram which shows the functional structure of the vehicle control apparatus which concerns on the 1st Embodiment of this invention. It is explanatory drawing for demonstrating the positional relationship of the vehicle and turning lane which are turning. It is explanatory drawing for demonstrating the extraction process of an edge point. It is explanatory drawing for demonstrating the extraction process of a lane boundary line. It is explanatory drawing for demonstrating the calculation process of S / N ratio. It is explanatory drawing for demonstrating the curvature noise calculation map. It is a block diagram which shows the detail of the locus | trajectory tracking controller 36 of the vehicle control apparatus which concerns on the 1st Embodiment of this invention. It is a block diagram which shows the functional structure of the feedback gain design apparatus which concerns on the 1st Embodiment of this invention. It is a block diagram which shows the model for feedback gain design in the feedback gain design apparatus which concerns on the 1st Embodiment of this invention. It is a flowchart which shows the content of the vehicle control processing routine in the vehicle control apparatus which concerns on the 1st Embodiment of this invention. It is a flowchart which shows the content of the feedback gain design processing routine in the feedback gain design apparatus which concerns on the 1st Embodiment of this invention. It is a block diagram which shows the model for feedback gain design in the feedback gain design apparatus which concerns on the 2nd Embodiment of this invention. It is a block diagram which shows the model for feedback gain design in the feedback gain design apparatus which concerns on the 3rd Embodiment of this invention.

<Overview>
In the embodiment of the present invention, the vehicle is controlled so as to suppress the wobbling feeling felt by the passengers of the vehicle. FIG. 1 shows the positional relationship between a turning vehicle and a traveling lane. 1 represents the lateral position deviation of the vehicle with respect to the target locus. Further, θ represents the yaw angle deviation of the vehicle with respect to the target locus, C represents the curvature of the target locus, and L represents the distance from the vehicle to the driver's gazing point (hereinafter referred to as the gaze distance). Here, r _g (viewing angular speed of the gazing point) yaw direction rotational speed of the road surface as viewed from the interior of the driver is expressed by the following equation.

In the above formula (1), θ ^· represents the yaw rate deviation of the vehicle, and e ^· represents the lateral velocity deviation of the vehicle. More viewing angular velocity r _g in the formula (1) is large, fluctuation feeling increases the vehicle driver feels. On the other hand, the field of view angular velocity r _g is if the following perception threshold r _s, the driver does not feel a sense of fluctuation.

  In addition, the vehicle occupant feels not only by vision but also by somatic sensation. In somatic sensation, it is known that a vehicle occupant perceives a sense of wobbling from a lateral jerk obtained by differentiating the lateral acceleration of the vehicle.

  Therefore, in the embodiment of the present invention, an example will be described in which the present invention is applied to a vehicle control device that controls a vehicle so as to suppress the feeling of wobbling felt by a vehicle occupant.

  Hereinafter, specific description will be given with reference to the drawings.

[First Embodiment]
<Configuration of vehicle control device>
In the first embodiment of the present invention, a case will be described as an example in which a feedback gain is set so that the viewing angular velocity of the gaze point of the driver of the vehicle is smaller than the perception threshold, and the vehicle is controlled using the feedback gain. . As shown in FIG. 2, the vehicle control apparatus 10 according to the first embodiment detects a camera 12 that captures an image in front of the vehicle, a vehicle speed sensor 14 that detects the vehicle speed of the vehicle, and a yaw rate of the vehicle. A yaw rate sensor 16, a steering angle sensor 18 for detecting a steering angle of a steering wheel as an operation state when the driver operates the host vehicle, a computer (ECU) 20, and an electric power steering for controlling a steering angle of a wheel 40.

  The camera 12 sequentially captures an image ahead of the vehicle as road information ahead of the vehicle.

  The computer 20 includes a CPU, a RAM, and a ROM that stores a program for executing a vehicle control processing routine to be described later, and is functionally configured as follows. As shown in FIG. 3, the computer 20 includes a front image captured by the camera 12, a vehicle speed detected by the vehicle speed sensor 14, a yaw rate detected by the yaw rate sensor 16, and a steering detected by the steering angle sensor 18. An information acquisition unit 22 that acquires a corner, a lane boundary detection unit 24 that detects a lane boundary based on the acquired front image, a target trajectory calculation unit 26 that calculates a target trajectory of the vehicle, and a target of the vehicle A state estimation unit 28 that estimates the state of the vehicle based on the trajectory, a curvature noise calculation unit 30 that calculates curvature noise based on the detection reliability representing the degree of reliability of the lane boundary, and a feedback designed in advance Based on the feedback gain database 32 in which the gain is stored and the feedback gain designed in advance, feedback is performed. A feedback gain setting unit 34 for setting a gain, a trajectory tracking controller 36 for calculating a target steering angle of the vehicle based on the set feedback gain, and a target steering angle based on the target steering angle of the vehicle. And a steering angle controller 38 for controlling the vehicle. The camera 12 is an example of a sensor, the lane boundary detection unit 24 is an example of a detection unit, the state estimation unit 28 is an example of an estimation unit, and the trajectory tracking controller 36 is an example of a driving operation amount calculation unit. The steering angle controller 38 is an example of a control means.

  The information acquisition unit 22 sequentially acquires the front image captured by the camera 12, the vehicle speed detected by the vehicle speed sensor 14, the yaw rate detected by the yaw rate sensor 16, and the steering angle detected by the steering angle sensor 18. To do.

  The lane boundary detection unit 24 detects the lane boundary of the road based on the front image acquired by the information acquisition unit 22. Specifically, the lane boundary detection unit 24 extracts luminance change points (edge points) of the road image as shown in FIG. Then, the lane boundary detection unit 24 extracts edge groups (L1, R1, etc.) that can be regarded as straight lines using the Hough transform as shown in FIG. 5, and the innermost parallel edge group L2 from within the areas LA, RA. , R2 is selected as the lane boundary line.

  Further, the lane boundary detection unit 24 calculates a detection reliability indicating the reliability of the detected lane boundary. Specifically, as shown in FIG. 6, the lane boundary detection unit 24 regards the number of edge points through which the lane boundary passes as a signal amount S and the number of edge points that do not pass through as a noise amount N. Let N ratio be the detection reliability. Since the method of setting the S / N ratio as the detection reliability is the same as the method described in the reference document (Japanese Patent Laid-Open No. 2013-105179), detailed description thereof is omitted.

  The target locus calculation unit 26 calculates a target locus of the vehicle based on the lane boundary line detected by the lane boundary line detection unit 24. For example, the target locus calculation unit 26 sets the center of the detected left and right lane boundary lines as the target locus. Further, the target trajectory calculation unit 26 may use a line that is closer to the left or right as the target trajectory according to the traveling condition.

  The state estimating unit 28 estimates the curvature of the target locus, the lateral position deviation of the vehicle with respect to the target locus, and the yaw angle deviation of the vehicle with respect to the target locus based on the target locus calculated by the target locus calculating unit 26. In the present embodiment, a case will be described as an example in which the curvature of the target locus, the lateral position deviation of the vehicle with respect to the target locus, and the yaw angle deviation of the vehicle with respect to the target locus are estimated using the extended Kalman filter.

The state estimation unit 28 assumes that the shape of the target locus y is a clothoid curve, as shown in the following equation (2). Then, the lateral position deviation e, the yaw angle deviation θ, the vehicle position curvature C ₀ , and the curvature change rate C ₁ are estimated using an extended Kalman filter. And the state estimation part 28 calculates the curvature C ahead of T second according to the following formula | equation (3). Here, T is set to a value corresponding to the delay time of the control device / vehicle response.

  Here, x in the above formula (2) indicates the front-rear position of the vehicle as shown in FIG. V represents the vehicle speed.

In the present embodiment, the lateral position deviation e _t at time t, the yaw angle deviation θ _t at time t, the curvature C _0t of the vehicle position at time t, and the curvature change rate C _1t at time t are expressed as state vector X _t = [E _t , θ _t , C _0t , C _1t ] Also, the edge point at time t on the target locus calculated by the target locus calculation unit 26, the observed value Y _t.

First, the state estimation unit 28 is based on the state vector X _t−1 = [e _t−1 , θ _t−1 , C _0t−1 , C _1t−1 ] at time t−1 in the prediction step of the extended Kalman filter. Thus, the state vector X _{t | t−1} = [e _t , θ _t , C _0t , C _1t ] at time _t is calculated. X _{t | t−1} represents a prior estimated value of time t predicted based on data available up to time t−1. Further, the value to which the subscript t | t (or t-1 | t-1) is given represents the post-estimation value at time t estimated based on the data available up to time t.

Next, in the filtering step of the extended Kalman filter, the state estimation unit 28 is calculated according to the above equation (2) from the state vector [e _t , θ _t , C _0t , C _1t ] calculated at the prediction step. Filtering is performed based on the target trajectory y and the observed value Y _t that is the edge point of the target trajectory calculated by the target trajectory calculation unit 26, and the state vector X _{t | t−} output at the prediction step is output. ₁ is corrected and X _{t | t} = [^ e _t , ^ θ _t , ^ C _0t , ^ C _1t ] is estimated.

Then, the state estimation unit 28 _calculates the above expression (3) based on [^ C _0t , ^ C _1t ], which are elements of the state vector estimated in the filtering step, and the vehicle speed V acquired by the information acquisition unit 22. ) To calculate the curvature C ahead of T seconds.

The curvature noise calculation unit 30 calculates the curvature noise included in the curvature of the target trajectory estimated by the state estimation unit 28 based on the detection result of the lane boundary line by the lane boundary detection unit 24.
Specifically, the curvature noise calculation unit 30 obtains the relationship between the detection reliability and the curvature noise created in advance so that the curvature noise increases as the detection reliability calculated by the lane boundary detection unit 24 decreases. The curvature noise is calculated using the represented map. As shown in FIG. 7, the map is created in advance so that the curvature noise is smaller as the detection reliability is higher.

  Here, the reason why the curvature noise is taken into account is that vehicle wobble may occur due to the curvature noise included in the estimated curvature of the target locus. When driving support is performed and curvature noise is included in the curvature of the target locus, the amount of operation of the vehicle (target yaw rate, etc.) changes regardless of the actual lane curvature. Therefore, since curvature noise is included in the curvature of the target trajectory, the lane cannot be smoothly followed (that is, fluctuation occurs).

  In the feedback gain database 32, feedback gains of steering angle deviation, lateral position deviation, lateral speed deviation, yaw angle deviation, and yaw rate deviation, which are designed in advance for each of the curvature noises by the feedback gain design device 100 described later, are stored. A combination is stored. The combination of feedback gains stored in the feedback gain database 32 is designed in advance for each of the curvature noises so that the viewing angular velocity of the gaze point of the vehicle driver is smaller than the perceptual threshold. Further, the feedback gain of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle is set in advance so that the feedback gain of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle becomes smaller as the curvature noise increases.

  Based on the curvature noise calculated by the curvature noise calculation unit 30, the feedback gain setting unit 34 calculates a steering angle deviation, a lateral position deviation, and a lateral velocity deviation with respect to the curvature noise from the feedback gain stored in the feedback gain database 32. The combination of the feedback gain of the yaw angle deviation and the yaw rate deviation is acquired and set.

  The trajectory tracking controller 36 performs T seconds estimated by the state estimation unit 28 according to feedforward control of the curvature of the target trajectory and feedback control of steering angle deviation, lateral position deviation, lateral speed deviation, yaw angle deviation, and yaw rate deviation. The curvature of the target trajectory ahead, the lateral deviation of the vehicle, and the yaw angle deviation of the vehicle, the vehicle speed, yaw rate, and steering angle acquired by the information acquisition unit 22, and the feedback gain set by the feedback gain setting unit 34 Based on the above, the target steering angle u is calculated so that the vehicle follows the target trajectory. The target steering angle u is an example of a driving operation amount of the vehicle.

A detailed configuration of the trajectory tracking controller 36 is shown in FIG. Here, K _i (i = 1 to 5) shown in FIG. 8 represents a feedback gain, V represents a vehicle speed, and K _v represents a vehicle yaw gain. The trajectory tracking controller 36 includes feed forward of the curvature C of the target trajectory T seconds ahead and feedback such as lateral position deviation. As shown in FIG. 8, the feedback gain K _i (i = 1 to 5) corresponds to each of the steering angle deviation, the lateral position deviation, the lateral speed deviation, the yaw angle deviation, and the yaw rate deviation.

  Specifically, the trajectory tracking controller 36 calculates the target steering angle u according to the following equation (4).

  The steering angle controller 38 controls the electric power steering 40 so as to realize the target steering angle u calculated by the trajectory tracking controller 36. For example, the steering angle controller 38 may control the electric power steering 40 using a PD control law.

  The electric power steering 40 controls the steering angle of the wheel according to the control of the steering angle controller 38.

<Configuration of feedback gain design device>
Next, the configuration of the feedback gain design apparatus 100 will be described. The feedback gain design device 100 calculates a combination of feedback gains used in the vehicle control device 10 according to the curvature noise.

  The feedback gain design apparatus 100 according to the first embodiment is configured by a computer including a CPU, a RAM, and a ROM that stores a program for executing a feedback gain design processing routine described later. The feedback gain design apparatus 100 is functionally configured as follows. As shown in FIG. 9, the feedback gain designing apparatus 100 includes a signal receiving unit 102 that receives a signal, a feedback gain calculating unit 104, and a feedback gain database 106.

  The signal receiving unit 102 receives an amplitude value of curvature noise.

  The feedback gain calculation unit 104 calculates a feedback gain using the H∞ control theory based on the amplitude value of the curvature noise received by the signal reception unit 102.

  FIG. 10 is a block diagram of a feedback gain design model. The feedback gain calculation unit 104 calculates the feedback gain using the feedback gain design model shown in FIG.

Here, the steering system in FIG. 10 is a first-order lag model, and the vehicle model is a linear two-wheel model. The steering system is a model that calculates the steering angle δ based on the target steering angle u calculated by the track following controller. The vehicle model is a model for calculating the lateral position deviation e, the lateral speed deviation e ^· , the yaw angle deviation θ, the yaw rate deviation θ ^·, and the yaw rate r based on the steering angle δ calculated by the steering system.

Also, C _n in FIG. 10 represents the curvature noise, r _g represents the viewing angular velocity, L represents a gaze distance, W _C represents the constant weight of curvature noise, W _r denotes the constant weight of the viewing angular velocity, W _e represents a constant weight of the lateral position deviation, and W _u (s) represents a frequency weight of the target steering angle.

In the feedback gain design model, consider a case where the target locus is a straight line. Therefore, since the true value of the curvature is zero, all the curvature to be inputted is the noise C _n. Further, since the yaw rate deviation theta ^· and the vehicle yaw rate r is equal viewing angular velocity r _g is expressed by the following equation.

  Specifically, the feedback gain calculation unit 104 satisfies the following constraints (1) to (3) based on the amplitude value of the curvature noise received by the signal receiving unit 102 using the H∞ control theory. Thus, a combination of feedback gains is calculated.

(1) field of view angular velocity _{r g} is equal to or less than the perception threshold _{r s.}
(2) The lateral position deviation e is not more than an upper limit value e _max (for example, 0.1 m).
(3) The high frequency component of the target steering angle is small (conditions for preventing the steering from becoming busy).

To meet the above (1) to (3) sets the amplitude value of the curvature noise constant weight W _C, set the following (1 ') to (3').

(1 ') to set a constant weight _{W r} of the field of view angular velocity to 1 / _{r s.}
(2 ') to set the constant weight _{W e} of the transverse position to _{1 / e max.}
(3 ′) The frequency weight W _u (s) of the target steering angle is set to a filter having a large high frequency gain.

Then, the feedback gain calculation unit 104 only needs to obtain a combination of feedback gains in which the gains from the disturbance w to the evaluation amounts z _r , z _e , and z _u are 1 or less using the H∞ control theory. The stability of the feedback system is guaranteed by the H∞ control theory.

Specifically, first, the feedback gain calculation unit 104, and a constant weight W _c of curvature noise, and constant weights W _r of the field angular velocity and a constant weight W _e of the transverse position, the frequency weight of the target steering angle W _u And set.

Next, the feedback gain calculation unit 104 uses the H∞ control theory so that the norm of the transfer function from the disturbance w to the evaluation amount z _u , the evaluation amount z _e , and the evaluation amount z _r becomes 1 or less. The combination of feedback gains (K _i (i = 1 to 5)) is calculated.

Note that the feedback gain calculation unit 104 increases the combination of feedback gains so that the feedback gain of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle decreases as the curvature noise C _n characterized by the constant weight W _c increases. (K _i (i = 1 to 5)) is calculated.

The reason why the feedback gain of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle becomes smaller as the curvature noise C _n becomes larger will be described below. As the curvature noise C _n increases, the follow-up deviation (lateral position deviation, yaw angle deviation) increases, and the amount of operation for reducing the follow-up deviation increases. When the operation amount is large, the yaw rate generated in the vehicle increases. As a result, the viewing angular velocity increases and the feeling of wobbling becomes stronger. Therefore, in order to suppress the feeling of wobbling, it is necessary to determine the operation amount so that the viewing angular velocity / yaw rate does not become excessive. In order to prevent the operation amount from becoming excessive even if the noise is large, the feedback gain of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle may be reduced.

The feedback gain calculation unit 104 stores the combination of the calculated feedback gain (K _i (i = 1 to 5)) and the amplitude value of the curvature noise in the feedback gain database 106.

  As described above, by using the feedback gain satisfying the above (1) to (3), the vehicle can be made to follow the target locus without a sense of wander even when the front image (road image) is noisy.

  In the feedback gain database 106, combinations of feedback gains calculated by the feedback gain calculation unit 104 are stored for each amplitude value of curvature noise.

<Operation of Feedback Gain Design Device 100>
Next, the operation of the feedback gain design apparatus 100 according to the first embodiment will be described. First, when an amplitude value of curvature noise is input to the feedback gain design apparatus 100, the feedback gain design process routine shown in FIG.

  In step S <b> 100, the signal receiving unit 102 receives an input of an amplitude value of curvature noise.

In step S102, the feedback gain calculation unit 104, sets a constant weight _{W c} of curvature noise.

In step S104, the feedback gain calculation unit 104 sets a constant weight _{W r} of the field angular velocity and a constant weight _{W e} of the transverse position, and a frequency weight _{W u} of the target steering angle.

In step S106, a feedback gain combination (K _i (i = 1 to 5)) in which the norm of the transfer function from disturbance w to evaluation amount z _u , evaluation amount z _e , and evaluation amount z _r is 1 or less is used. calculate.

In step S <b> 108, the combination of the feedback gain combination (K _i (i = 1 to 5)) calculated in step S <b> 106 and the amplitude value of the curvature noise is stored in the feedback gain database 106.

  In step S110, it is determined whether or not the processing of steps S102 to S108 has been executed for all constant weights of curvature noise that can be taken. When the processing of step S102 to step S108 is executed for all constant weights of curvature noise that can be taken, the process proceeds to step S112. On the other hand, if the processing of step S102 to step S108 is not executed for all possible constant weights of curvature noise, the process returns to step S102, and the processing of step S102 to step S108 is repeated.

In step S112, a feedback gain combination process (K _i (i = 1 to 5)) corresponding to each of the constant weights W _c stored in the feedback gain database 106 in step S108 is interpolated to perform feedback gain design processing. End the routine.

Since the constant weight W _c of curvature noise set in the step S102 is discrete, by performing the interpolation in step S112, it may correspond to a continuous curvature noise.

<Operation of Vehicle Control Device 10>
Next, the operation of the vehicle control device 10 according to the present embodiment will be described. First, when a combination of feedback gain and curvature noise stored in the feedback gain database 106 of the feedback gain design apparatus 100 is input to the vehicle control apparatus 10, it is stored in the feedback gain database 32. Then, when the vehicle travels and front images of the vehicle are sequentially captured by the camera 12, a vehicle control processing routine shown in FIG.

  First, in step S200, the information acquisition unit 22 is detected by the front image captured by the camera 12, the vehicle speed detected by the vehicle speed sensor 14, the yaw rate detected by the yaw rate sensor 16, and the steering angle sensor 18. Get the steering angle.

  In step S202, the lane boundary detection unit 24 detects the lane boundary of the road based on the front image acquired in step S200.

  In step S204, the lane boundary detection unit 24 calculates a detection reliability indicating the reliability of the lane boundary detected in step S202.

  In step S206, the target locus calculation unit 26 calculates the target locus of the vehicle based on the lane boundary detected in step S202.

  In step S208, the state estimation unit 28 estimates the curvature of the target locus, the lateral position deviation of the vehicle with respect to the target locus, and the yaw angle deviation of the vehicle with respect to the target locus based on the target locus calculated in step S206.

  In step S210, the curvature noise is calculated by the curvature noise calculation unit 30 based on the detection reliability calculated in step S204, using a map representing the relationship between the detection reliability and the curvature noise created in advance.

  In step S212, based on the curvature noise calculated in step S210 by the feedback gain setting unit 34, the steering angle, lateral position deviation, lateral direction corresponding to the curvature noise is calculated from the feedback gain stored in the feedback gain database 32. Set the feedback gain combination of speed deviation, yaw angle deviation, and yaw rate deviation.

  In step S214, the trajectory tracking controller 36 obtains the curvature of the target trajectory according to the feedforward control and the feedback control of the steering angle, lateral position deviation, lateral speed deviation, yaw angle deviation, and yaw rate deviation in step S200. Based on the vehicle speed, the yaw rate, the steering angle, the curvature of the target locus estimated in step S208, the lateral position deviation of the vehicle, the yaw angle deviation of the vehicle, and the feedback gain set in step S212, the vehicle The target steering angle u of the vehicle is calculated so as to follow the target locus.

  In step S216, the steering angle controller 38 controls the vehicle so as to realize the target steering angle u of the vehicle calculated by the trajectory tracking controller 36, and ends the vehicle control processing routine.

  As described above, according to the vehicle control device according to the first embodiment, the curvature noise included in the curvature of the estimated target trajectory is calculated based on the detection result of the lane boundary line, and is calculated. Based on the curvature noise, set a feedback gain for the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle, which is predetermined for the curvature noise so that the viewing angular velocity of the gaze point of the driver of the vehicle is smaller than the threshold value, According to the feedforward control of the curvature of the target trajectory and the feedback control of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle, the estimated curvature of the target trajectory, the lateral position deviation of the vehicle, and the yaw angle deviation of the vehicle are set. The target steering angle amount of the vehicle is calculated so that the vehicle follows the target locus based on the lateral position deviation of the vehicle and the feedback gain of the yaw angle deviation of the vehicle, By controlling the vehicle to achieve the target steering angle, without the driver feeling a sense of fluctuation can control the vehicle so that the vehicle follows the target locus.

  Further, even when the road image (front image) is noisy, the vehicle can be made to follow the target locus without a sense of wobbling.

  Further, when the detection reliability of the lane boundary line is low, the curvature noise of the curvature estimation value becomes larger than usual. Therefore, by determining the feedback gain based on the noise corresponding to the detection reliability, even when the road image is noisy, the vehicle can be made to follow the target locus without a sense of wobbling.

  Further, by using a feedback gain that minimizes the viewing angular velocity of the gazing point with respect to the curvature noise that is the noise of the curvature estimation value, it is possible to suppress the sense of wandering.

[Second Embodiment]
Next, a second embodiment of the present invention will be described. In addition, about the part which becomes the structure similar to 1st Embodiment, the same code | symbol is attached | subjected and description is abbreviate | omitted.

  The second embodiment is different from the first embodiment in that the feedback gain is designed in consideration of the lateral jerk obtained by differentiating the lateral acceleration of the vehicle instead of the viewing angular velocity. Vehicle occupants feel wandering not only visually but also by somatic sensations. In somatic sensation, it has been found that the vehicle occupant perceives wandering from the side jerk. In the second embodiment, it is possible to suppress the feeling of wobbling by using a feedback gain in which the lateral jerk is smaller than the threshold with respect to the noise of the curvature estimation value.

  As shown in FIG. 9, the feedback gain design device 2100 according to the second embodiment includes a feedback gain design device 100, a signal reception unit 102 that receives signals, a feedback gain calculation unit 2104, and a feedback gain database 106. And.

FIG. 13 is a block diagram of a feedback gain design model used in the second embodiment. The feedback gain calculation unit 2104 calculates the feedback gain using the feedback gain design model shown in FIG. Here, s in FIG. 13 represents a differential operator, and W _j represents a constant weight of the lateral jerk.

  The feedback gain calculation unit 2104 calculates a combination of feedback gains so as to satisfy the following constraint (4).

(4) The value of the lateral jerk is not more than the perception threshold r _j .

  In order to satisfy the above (4), the following (4 ') is set.

(4 ′) The horizontal jerk constant weight W _j is set to 1 / r _j .

Then, the feedback gain calculation unit 2104 may use the H∞ control theory to obtain a combination of feedback gains where the gain from the disturbance w to the evaluation quantities z _u , z _j , and z _e is 1 or less. The stability of the feedback system is guaranteed by the H∞ control theory.

Specifically, the feedback gain calculation unit 2104 calculates a constant weight W _{c for} curvature noise, a constant weight W _{j for} lateral jerk, a constant weight W _{e for} lateral position deviation, and a frequency weight W _u for a target steering angle. Set.

Next, the feedback gain calculation unit 2104 uses the H∞ control theory so that the norm of the transfer function from the disturbance w to the evaluation amount z _u , the evaluation amount z _j , and the evaluation amount z _e is 1 or less. The combination of feedback gains (K _i (i = 1 to 5)) is calculated.
Note that the feedback gain calculation unit 2104 combines the feedback gains so that the feedback gain of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle becomes smaller as the curvature noise C _n characterized by the constant weight W _c is larger. (K _i (i = 1 to 5)) is calculated.

The feedback gain calculation unit 2104 then stores the combination of the calculated feedback gain combination (K _i (i = 1 to 5)) and the amplitude value of the curvature noise in the feedback gain database 106.

  The vehicle control apparatus 10 according to the second embodiment controls the vehicle using a combination of feedback gains designed by the feedback gain design apparatus 2100.

  Note that other configurations and operations of the vehicle control device 10 and the feedback gain design device 2100 according to the second embodiment are the same as those of the first embodiment, and thus description thereof is omitted.

  As described above, according to the vehicle control apparatus according to the second embodiment, the curvature noise included in the curvature of the estimated target trajectory is calculated based on the detection result of the lane boundary line, and is calculated. Based on the curvature noise, set a feedback gain for the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle, which is predetermined with respect to the curvature noise so that the lateral jerk obtained by differentiating the lateral acceleration of the vehicle becomes smaller than a threshold value, According to the feedforward control of the curvature of the target trajectory and the feedback control of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle, the estimated curvature of the target trajectory, the lateral position deviation of the vehicle, and the yaw angle deviation of the vehicle are set. The target steering angle of the vehicle is calculated so that the vehicle follows the target trajectory based on the feedback gain of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle. By controlling the vehicle so as to realize the steering angle, without feeling a sense of fluctuation occupant of the vehicle, it is possible to control the vehicle so that the vehicle follows the target locus.

  Further, by using the feedback gain that minimizes the lateral jerk against the curvature noise that is the noise of the curvature estimation value, it is possible to suppress the feeling of wobbling.

[Third Embodiment]
Next, a third embodiment will be described. In addition, about the part which becomes the same structure as 1st and 2nd embodiment, the same code | symbol is attached | subjected and description is abbreviate | omitted.

  The third embodiment is different from the first and second embodiments in that the feedback gain is designed in consideration of the viewing angular velocity and the lateral jerk.

  As shown in FIG. 9, the feedback gain design apparatus 3100 according to the third embodiment includes a signal reception unit 102 that receives a signal, a feedback gain calculation unit 3104, and a feedback gain database 106.

  FIG. 14 is a block diagram of a feedback gain design model used in the third embodiment. The feedback gain calculation unit 3104 calculates the feedback gain using the feedback gain design model shown in FIG.

Specifically, the feedback gain calculation unit 3104, and a constant weight W _c of curvature noise, and constant weights W _r of the field angular velocity and a constant weight W _j of the lateral jerk, and a constant weight W _e of the transverse position, the target A frequency weight W _{u of the} steering angle is set.

Next, the feedback gain calculation unit 3104 uses the H∞ control theory, and the norm of the transfer function from the disturbance w to the evaluation amount z _u , the evaluation amount z _j , the evaluation amount z _e , and the evaluation amount z _r is 1. A combination of feedback gains (K _i (i = 1 to 5)) is calculated so as to be as follows.
Note that the feedback gain calculation unit 3104 combines the feedback gains so that the feedback gain of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle decreases as the curvature noise C _n characterized by the constant weight W _c increases. (K _i (i = 1 to 5)) is calculated.

The feedback gain calculation unit 3104 then stores the combination of the calculated feedback gain combination (K _i (i = 1 to 5)) and the amplitude value of the curvature noise in the feedback gain database 106.

  The vehicle control apparatus 10 according to the third embodiment controls the vehicle using a combination of feedback gains designed by the feedback gain design apparatus 3100.

  In addition, since the other structure and effect | action of the vehicle control apparatus 10 and the feedback gain design apparatus 3100 which concern on 3rd Embodiment are the same as that of 1st Embodiment, description is abbreviate | omitted.

  As described above, according to the vehicle control device according to the third embodiment, the curvature noise included in the curvature of the estimated target trajectory is calculated based on the detection result of the lane boundary line, and is calculated. The lateral position deviation of the vehicle and the vehicle, which are determined in advance with respect to the curvature noise so that the lateral jerk obtained by differentiating the viewing angular velocity of the gazing point of the driver of the vehicle and the lateral acceleration of the vehicle becomes smaller than the perception threshold based on the curvature noise. The target trajectory curvature and the vehicle lateral position deviation are set according to the feedforward control of the curvature of the target trajectory and the feedback control of the vehicle lateral position deviation and the vehicle yaw angle deviation. Based on the yaw angle deviation of the vehicle and the set lateral gain of the vehicle and the feedback gain of the yaw angle deviation of the vehicle. Thus, by calculating the target steering angle of the vehicle and controlling the vehicle so as to realize the target steering angle, the vehicle can be adjusted so that the vehicle follows the target trajectory without feeling a sense of wobbling. Can be controlled.

  In the above embodiment, the case where the camera 12 is used as the sensor for detecting road information ahead of the vehicle has been described as an example. However, the present invention is not limited to this, and a laser radar device or the like is used. Road information ahead of the vehicle may be detected.

  In the above embodiment, the case where the target steering angle is calculated as the driving operation amount of the vehicle has been described as an example. However, the present invention is not limited to this, and the yaw moment of the vehicle is used as the driving operation amount of the vehicle. It may be calculated.

Further, the feedback gain setting unit 34 in the above embodiment has been described as an example in which a combination of feedback gains of steering angle deviation, lateral position deviation, lateral speed deviation, yaw angle deviation, and yaw rate deviation is set. It is not limited to this. A combination (K ₂ , K ₄ ) of at least a lateral position deviation and a yaw angle deviation feedback gain may be set. For example, only a combination of a lateral position deviation and a yaw angle deviation feedback gain (K ₂ , K ₄ ) is set. You may make it do.
Further, the feedback gain calculation units 104, 2104, and 3104 in the above-described embodiment may calculate at least a combination (K ₂ , K ₄ ) of the lateral position deviation and the yaw angle deviation feedback gain. , And only the combination of the yaw angle deviation feedback gain (K ₂ , K ₄ ) may be calculated.

  In the above embodiment, the extended Kalman filter is used as an example to estimate the curvature of the target locus, the lateral position deviation of the vehicle with respect to the target locus, and the yaw angle deviation of the vehicle with respect to the target locus. However, the curvature of the target locus, the lateral position deviation of the vehicle with respect to the target locus, and the yaw angle deviation of the vehicle with respect to the target locus may be estimated by other estimation methods.

  The program of the present invention can be provided by being stored in a recording medium.

DESCRIPTION OF SYMBOLS 10 Vehicle control apparatus 12 Camera 14 Vehicle speed sensor 16 Yaw rate sensor 18 Steering angle sensor 20 Computer 22 Information acquisition part 24 Lane boundary detection part 26 Target locus calculation part 28 State estimation part 30 Curvature noise calculation part 32 Feedback gain database 34 Feedback gain setting Unit 36 Trajectory tracking controller 38 Steering angle controller 40 Electric power steering 100, 2100, 3100 Feedback gain design device 102 Signal receiving units 104, 2104, 3104 Feedback gain calculation unit 106 Feedback gain database

Claims (5)

A sensor for detecting road information ahead of the vehicle;
Detection means for detecting a lane boundary of a road based on the road information detected by the sensor;
Target trajectory calculating means for calculating a target trajectory of the vehicle based on the lane boundary detected by the detecting means;
Estimating means for estimating a curvature of the target locus, a lateral position deviation of the vehicle with respect to the target locus, and a yaw angle deviation of the vehicle with respect to the target locus based on the target locus calculated by the target locus calculating means; ,
Curvature noise calculation means for calculating curvature noise included in the curvature of the target trajectory estimated by the estimation means based on the detection result of the lane boundary line by the detection means;
Based on the curvature noise calculated by the curvature noise calculating means, the curvature so that at least one of the viewing angular velocity of the gaze point of the driver of the vehicle and the lateral jerk obtained by differentiating the lateral acceleration of the vehicle is smaller than a threshold value. Feedback gain setting means for setting a feedback gain of a lateral position deviation of the vehicle and a yaw angle deviation of the vehicle, which is predetermined with respect to noise;
According to the feedforward control of the curvature of the target locus and the feedback control of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle, the curvature of the target locus estimated by the estimating means, the lateral position deviation of the vehicle, and Based on the yaw angle deviation of the vehicle, the lateral position deviation of the vehicle set by the feedback gain setting means, and the feedback gain of the yaw angle deviation of the vehicle, so that the vehicle follows the target locus, Driving operation amount calculating means for calculating the driving operation amount of the vehicle;
Control means for controlling the vehicle so as to realize the driving operation amount calculated by the driving operation amount calculating means;
A vehicle control apparatus. The feedback gain setting means has a lateral position of the vehicle such that the larger the curvature noise calculated by the curvature noise calculation means, the smaller the feedback gain of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle. The vehicle control device according to claim 1, wherein a feedback gain of the deviation and the yaw angle deviation of the vehicle is set. The detection means further calculates a detection reliability representing a degree of reliability of the detected lane boundary line,
The vehicle control device according to claim 1, wherein the curvature noise calculation unit calculates the curvature noise such that the curvature noise increases as the detection reliability calculated by the detection unit decreases.   The vehicle control device according to any one of claims 1 to 3, wherein the driving operation amount is at least one of a steering angle of the vehicle and a yaw moment of the vehicle. Detecting means for detecting a lane boundary of a road based on the road information detected by a sensor for detecting road information ahead of the vehicle;
Target trajectory calculating means for calculating a target trajectory of the vehicle based on the lane boundary line detected by the detecting means;
Estimating means for estimating a curvature of the target locus, a lateral position deviation of the vehicle with respect to the target locus, and a yaw angle deviation of the vehicle with respect to the target locus based on the target locus calculated by the target locus calculating means;
Curvature noise calculation means for calculating curvature noise included in the curvature of the target locus estimated by the estimation means based on the detection result of the lane boundary line by the detection means;
Based on the curvature noise calculated by the curvature noise calculating means, the curvature so that at least one of the viewing angular velocity of the gaze point of the driver of the vehicle and the lateral jerk obtained by differentiating the lateral acceleration of the vehicle is smaller than a threshold value. A feedback gain setting means for setting a feedback gain of a lateral position deviation of the vehicle and a yaw angle deviation of the vehicle, which is predetermined with respect to noise;
According to the feedforward control of the curvature of the target locus and the feedback control of the lateral position deviation of the vehicle and the yaw angle deviation of the vehicle, the curvature of the target locus estimated by the estimating means, the lateral position deviation of the vehicle, and Based on the yaw angle deviation of the vehicle, the lateral position deviation of the vehicle set by the feedback gain setting means, and the feedback gain of the yaw angle deviation of the vehicle, so that the vehicle follows the target locus, A program for functioning as a control means for controlling the vehicle so as to realize the driving operation amount calculated by the driving operation amount calculating means, and a driving operation amount calculating means for calculating the driving operation amount of the vehicle.

Images