KR20150058892A - System for controlling lane keeping of vehicle using exponential smoothing, and apparatus for generating virtual lane using exponential smoothing - Google Patents

Abstract

The present invention proposes a lane keeping control and virtual lane generation system which virtually generates an unrecognized/misrecognized lane and controls a vehicle to stay in a lane based on a recognized lane or a virtual lane using an exponential smoothing method. The lane keeping control and virtual lane generation system according to the present invention comprises a front image recording unit to obtain an image in front of a vehicle in motion; a lane detection unit to detect a first lane and a second lane on left and right sides in the image; a virtual lane generation unit to generate the second lane based on the first land and width thereof when only the first lane is detected between the second lane and the first lane, and virtually generate the first lane and the second lane using the exponential smoothing method when neither of two lanes is detected; and a lane keeping control unit to control the vehicle to stay in the lane based on one of three groups. Three groups are a first group which includes the first and second lanes detected; a second group which includes the first land and the virtually generated second lane; and a third group which includes the virtually generated first and second lanes.

Description

Technical Field [0001] The present invention relates to a lane keeping control system and a virtual lane generating system using an exponential smoothing method,

The present invention relates to a virtual lane generating device for virtually generating a false / unrecognized lane. The present invention also relates to a lane-keeping control system for controlling lane-keeping based on a recognized lane or a virtually generated lane.

Lane Keeping Assistance System (LKAS) is a system that detects a lane through a sensor and changes position information of the detected lane to a torque value to prevent lane departure of the vehicle.

Such lane maintenance assist system is greatly influenced by the accuracy of the camera in helping the vehicle not depart from the lane by using the road information measured through the camera.

When the lane is not recognized or mistaken by the camera due to environmental factors such as guard rail, brake mark, double lane, snow, rain, etc., frequent occurrence occurs in the road, and the LKAS control malfunctions It can not only make the driver feel a sense of heterogeneity but also can cause dangerous situations.

Korean Laid-Open Patent Application No. 2009-0053412 describes a lane departure prevention system having a lane recognition function. However, in the case of the lane estimation, the device compensates the lane using the lane width information and the noise level. However, the lane width is not an actual road information but an arbitrary value.

SUMMARY OF THE INVENTION The present invention has been conceived to solve the problems described above, and it is an object of the present invention to provide a method and apparatus for generating virtual / unrecognized lanes by using exponential smoothing and detecting a lane keeping based on a recognized lane or a virtual lane And a virtual lane generating device for controlling the lane keeping control system.

However, the objects of the present invention are not limited to those mentioned above, and other objects not mentioned can be clearly understood by those skilled in the art from the following description.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and it is an object of the present invention to provide a forward image obtaining unit that obtains a forward image in a vehicle under running. A lane detecting unit detecting first lanes and second lanes located on both sides of the forward image; If only the first lane of the first lane and the second lane is detected, the second lane is virtually created on the basis of the first lane and the lane width, and if neither the first lane nor the second lane is detected A virtual lane generator for virtually generating the first lane and the second lane using an exponential smoothing method; A first group including the detected first lane and the detected second lane, a second group including the detected first lane and the virtually generated second lane, and a second group including the detected first lane and the second lane detected, And a third group including the second generated lane, the lane keeping control unit controlling the lane keeping of the vehicle based on the lane keeping control unit. I suggest.

Preferably, the virtual lane generator virtually creates the first lane and the second lane on the basis of the previously predicted both lanes and the previously obtained both lanes.

Preferably, the virtual lane generator multiplies the difference value between the previously obtained first lane and the previously predicted first lane by a weight, adds the previously predicted first lane to generate the first lane virtually , The difference between the previously obtained second lane and the previously predicted second lane is multiplied by a weight and the previously predicted second lane is added to virtually create the second lane.

Preferably, the virtual lane generator uses a smoothing constant as the weight.

Preferably, the lane detecting unit determines whether each lane is recognized based on whether each lane is sensed, and detects the recognized lane.

Preferably, the lane-keeping control unit determines whether or not the vehicle leaves the lane on the basis of the traveling state information obtained from any one of the first group, the second group, and the third group and the vehicle sensors And controls the lane keeping of the vehicle if it is determined that the vehicle has departed from the lane.

Preferably, the lane-keeping control system is activated when the speed of the vehicle is equal to or greater than 60 KPH (kilometer per hour).

In addition, the present invention is characterized in that when only the first lane is detected from the first lane and the second lane located on both sides of the front image, the first virtual lane generating unit generates the second lane based on the first lane and the lane width, ; And a second imaginary generating unit for virtually generating the first lane and the second lane using an exponential smoothing method if neither the first lane nor the second lane is detected, We propose a virtual lane generator.

Preferably, the second virtual generation unit multiplies the difference value between the first lane previously obtained and the previously predicted first lane by a weight, adds the previously predicted first lane, and virtually creates the first lane And multiplying the difference value between the previously obtained second lane and the previously predicted second lane by a weight and adding the previously predicted second lane to virtually create the second lane.

The present invention can obtain the following effect by virtually generating the misidentified / unrecognized lane using exponential smoothing and controlling the lane keeping based on the recognized lane or the virtual generated lane.

First, it is possible to estimate a virtual lane with high accuracy and high accuracy, and it is possible to implement a stable lane keeping assist function.

Second, the risk of lane departure due to driver's carelessness can be reduced, and unnecessary steering intervention can eliminate the driver's sense of disturbance and anxiety.

1 is a conceptual diagram of a lane maintenance assist system according to an embodiment of the present invention.
FIG. 2 is a flowchart sequentially illustrating an operation method of the lane-keeping assist system shown in FIG. 1. FIG.
3 and 4 are graphs comparing virtual lanes generated according to the conventional method and virtual lanes generated according to the present invention.
5 is a block diagram schematically illustrating a lane-keeping control system according to a preferred embodiment of the present invention.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the drawings, the same reference numerals are used to designate the same or similar components throughout the drawings. In the following description of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear. In addition, the preferred embodiments of the present invention will be described below, but it is needless to say that the technical idea of the present invention is not limited thereto and can be variously modified by those skilled in the art.

The present invention proposes a virtual lane generating apparatus and method for maintaining a lane keeping assistant system function when a camera installed in a Lane Keeping Assist System (LKAS) is a lane unrecognized and a false recognition. More particularly, it is possible to predict and compensate future lanes based on past lane information when it is not easy to detect (both) lanes in a car on which the vehicle is traveling, thereby achieving a stable lane keeping assist function.

A lane keeping assist system (LKAS) is a system that monitors the lane of a car sensor mounted on a car and alerts the driver when sudden steering, drowsy driving, cell phone calls, It is a system to prevent lane departure.

Such a lane maintenance assist system is greatly influenced by the performance and accuracy of the camera because it prevents the lane departure of the vehicle using the road information measured by the camera.

The cameras used for lane recognition frequently occur in the case of unrecognized and misunderstood lane due to environmental factors such as weather, illumination, road surface condition, double lane, guard rail at the time of driving the vehicle. The unrecognition and misidentification of these lanes can lead to LKAS degradation and dangerous situations.

Accordingly, the apparatus and method for generating virtual lanes of the present invention can accurately recognize lanes without being affected by noise by predicting the past lane trends.

1 is a conceptual diagram of a lane maintenance assist system according to an embodiment of the present invention.

The lane keeping assist system 100 according to the present invention includes a front camera sensor 110, a lane recognition unit 120, a virtual lane setting unit 130, and a lane keeping support unit 140.

The front camera sensor 110 performs a function of acquiring a lane departure information and extracting lane information.

The lane recognition unit 120 processes the image input through the front camera sensor 110 to detect lane information.

The virtual lane setting unit 130 calculates a lane estimation value so as not to deviate from the lane according to whether the lane is recognized or not.

The lane-keeping support unit 140 performs necessary braking force or steering control through the virtual lane estimation value input from the virtual lane setting unit 130 to assist the vehicle from leaving the lane.

FIG. 2 is a flowchart sequentially illustrating an operation method of the lane-keeping assist system shown in FIG. 1. FIG. The following description refers to Fig. 1 and Fig.

If the lane is not recognized through the image generated by the front camera sensor 110 or the lane condition is not satisfied, the virtual lane setting unit 130 generates a virtual lane to determine whether or not the lane departure occurs.

The lane recognition is activated when the lane keeping assist system (LKAS) is turned on (S205) and the vehicle is traveling at 60 KPH (Kilometer Per Hour) or more (S210).

The lane recognition unit 120 performs a lane recognition function on the camera image (S215), whether the left lane and the right lane are all recognized (S220), whether only one of the left lane and the right lane is recognized (S230) It is determined whether both lanes are mistaken and unrecognized (S240).

When the lane recognition unit 120 recognizes both lanes normally (S220), the lane-keeping control is normally executed without generating a virtual lane (S225).

If one of the left lane and the right lane is unrecognized / mistakenly recognized, a lane is arbitrarily generated in the opposite side of the left lane and the right lane based on one of the left lane and the right lane (S235) (S225).

Finally, if both lanes are unrecognized / misidentified, lane information and exponential smoothing are used (S245) so that lane keeping control can be continuously performed (S225).

Exponential smoothing is a method of giving the largest weight to recent data and geometrically reducing the weight over time. When the exponential smoothing method is applied to the present invention, there is an advantage that the problem of the conventional average shifting method, that is, an incorrect signal reflecting phenomenon due to a time delay can be prevented.

In the case where the left lane and the right lane while driving are unrecognized / false recognition, the virtual lane setting unit 130 of the present invention calculates the lane estimation value through the following equation.

F _t = F _t-1 + 留 (A _t-1 + F _t-1 )

In the above _equation , F _t denotes predicted value of the current period, that is, lane information predicted at the present time. F _t-1 means the predicted value of electricity, that is, the lane information predicted at the previous time point, and A _t-1 means the actual value (observed value) of the electricity, that is, the lane information measured at the previous point. α is the smoothing constant.

The present invention creates a virtual lane using objective data with the assumption that the tendency of the lane continues in the future.

The lane-based compensation method using an average moving filter is a method of estimating a lane by averaging a plurality of lane values inputted during a predetermined time in the lane information. This method calculates the lane estimation value through the following equation.

F _t = S _t-1 = (A _t-1 + A _t-2 + ... + A _tn ) / n

In the above, S _t denotes the moving average of t period, and t denotes the current data sequence number. A _t denotes the actual value (observation value) of t period, and n denotes the number of data to be averaged.

Moving average filters are useful for removing noise in certain data. However, if the average number of data (n) is large, the noise removal performance is improved, but there is a phenomenon in which a change in the measurement signal is not reflected in time and a time delay occurs. Conversely, if the number of data (n) is small, the change of the measurement signal follows well, but the noise is not removed well. Therefore, the characteristics of the lane signal (or road) to be measured should be well understood and the number of data of the moving average should be selected. And the moving average filter does not have any advantage in changing the algorithm to recursive.

The present invention proposes a virtual lane generation method that calculates a lane estimation value by using an exponential smoothing method which is easy to use, has a high accuracy of short-term guesses, and can control a weight with elasticity.

(Smoothing constant, α) is given to the most recent data and the weight (smoothing constant, α) decreases with time. The data are weighted according to the exponential function of time. When the value of the smoothing constant alpha is set to be large, the sensitivity of the prediction becomes high, and when the value of the smoothing constant alpha is made small, the stability of prediction is high. This makes it possible to detect the change of the signal better than the moving average filter.

If the left lane and the right lane in the lane recognition unit 120 are unrecognized / false, the virtual lane setting unit 130 generates a virtual lane by using an exponential smoothing filter to perform lane keeping control.

3 and 4 are graphs comparing virtual lanes generated according to the conventional method and virtual lanes generated according to the present invention.

3 and 4 are graphs comparing virtual lane generation methods. MovingAvg represents the average moving method, WeightedMovingAvg represents the recursive average moving method, and ExpSmooth represents the exponential smoothing method.

The lane values estimated by different methods differ from the actual lane values. The lane estimation value of the exponential smoothing method is more accurate than the lane estimation value of the average moving method. In addition, the recursive average moving method has a better lane value than the average moving method, but there is still a time delay. Therefore, it is desirable to use a more stable exponential smoothing filter for lane estimation for more efficient lane keeping control.

In the conventional lane keeping assist system (LKAS), only the lane-based vehicle behavior state is considered when the steering control intervention is determined. Therefore, when the driver is traveling in a lateral direction in a lane, such as when driving on a lane edge or on a curved road, steering intervention occurs even when the driver normally steers the steering wheel to cause the driver to feel uneasy .

In addition, since the conventional driver condition monitoring system (DSM) only warns the driver in case of inattentive driving, when the driver leaves the lane due to sleepiness of the driver or the like, the driver must directly correct the steering, have.

In the present invention, steering control is performed only when the steering intervention is required by reflecting not only the lane-changing vehicle behavior but also the driver's condition at the time of the control intervention, thereby reducing the risk of lane departure due to driver's carelessness, The steering intervention can alleviate the driver's sense of discomfort and anxiety.

The preferred embodiments of the present invention can be configured as follows from the embodiments described with reference to Figs. 1 to 4 above. 5 is a block diagram schematically illustrating a lane-keeping control system according to a preferred embodiment of the present invention.

5, the lane-keeping control system 500 includes a front image obtaining unit 510, a lane detecting unit 520, a virtual lane generating unit 530, a lane-keeping control unit 540, a power source unit 550, 560).

The forward image obtaining unit 510 performs a function of obtaining a forward image in a vehicle under driving.

The lane detecting unit 520 detects the first lane and the second lane located on both sides of the front image obtained by the front image obtaining unit 510. The lane detecting unit 520 may determine whether each side lane is recognized according to whether each side lane is sensed, and may detect the recognized lane.

The virtual lane generator 530 generates a virtual second lane based on the first lane and the lane width when only the first lane of the first lane and the second lane is detected. The virtual lane generator 530 also functions to virtually generate the first lane and the second lane using the exponential smoothing method if neither the first lane nor the second lane is detected. The former can be performed by the first virtual creator of the virtual lane generator 530 and the latter can be performed by the second virtual creator of the virtual lane generator 530. [

The virtual lane generator 530 can virtually generate the first lane and the second lane based on the previously predicted both lanes and the previously obtained both lanes. In this case, the virtual lane generator 530 multiplies the difference value between the previously obtained first lane and the previously predicted first lane by a weight, adds the previously predicted first lane, and virtually creates the first lane , The difference value between the previously obtained second lane and the previously predicted second lane may be multiplied by a weight and the previously predicted second lane may be added to virtually create the second lane. The virtual lane generator 530 may use a smoothing constant as a weight.

The lane-keeping control unit 540 includes a first group including a detected first lane and a detected second lane, a second group including a detected first lane and a virtual-generated second lane, And controls the lane keeping of the vehicle based on any one of the third group including the lane and the second generated lane.

The lane-keeping control unit 540 determines whether or not the vehicle leaves the lane on the basis of the traveling state information obtained from any one of the first group, the second group, and the third group and the vehicle sensors, It is possible to control the lane keeping of the vehicle.

The lane-keeping control system 500 described above can be activated when the vehicle speed is 60 KPH (kilometer per hour) or more.

It is to be understood that the present invention is not limited to these embodiments, and all elements constituting the embodiment of the present invention described above are described as being combined or operated in one operation. That is, within the scope of the present invention, all of the components may be selectively coupled to one or more of them. In addition, although all of the components may be implemented as one independent hardware, some or all of the components may be selectively combined to perform a part or all of the functions in one or a plurality of hardware. As shown in FIG. In addition, such a computer program may be stored in a computer readable medium such as a USB memory, a CD disk, a flash memory, etc., and read and executed by a computer to implement an embodiment of the present invention. As the recording medium of the computer program, a magnetic recording medium, an optical recording medium, a carrier wave medium, and the like can be included.

Furthermore, all terms including technical or scientific terms have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs, unless otherwise defined in the Detailed Description. Commonly used terms, such as predefined terms, should be interpreted to be consistent with the contextual meanings of the related art, and are not to be construed as ideal or overly formal, unless expressly defined to the contrary.

It will be apparent to those skilled in the art that various modifications, substitutions and substitutions are possible, without departing from the scope and spirit of the invention as disclosed in the accompanying claims. will be. Therefore, the embodiments disclosed in the present invention and the accompanying drawings are intended to illustrate and not to limit the technical spirit of the present invention, and the scope of the technical idea of the present invention is not limited by these embodiments and the accompanying drawings . The scope of protection of the present invention should be construed according to the following claims, and all technical ideas within the scope of equivalents should be construed as falling within the scope of the present invention.

Claims (9)

A forward image acquiring unit for acquiring a forward image in a vehicle under running;
A lane detecting unit detecting first lanes and second lanes located on both sides of the forward image;
If only the first lane of the first lane and the second lane is detected, the second lane is virtually created on the basis of the first lane and the lane width, and if neither the first lane nor the second lane is detected A virtual lane generator for virtually generating the first lane and the second lane using an exponential smoothing method; And
A first group including the detected first lane and the detected second lane, a second group including the detected first lane and the virtually generated second lane, and a second group including the detected first lane and the second lane, A lane keeping control unit for controlling lane keeping of the vehicle based on any one of a third group including the second generated lane,
A lane keeping control system using an exponential smoothing method. The method according to claim 1,
Wherein the virtual lane generator virtually generates the first lane and the second lane on the basis of previously predicted both lanes and previously obtained both lanes. 3. The method of claim 2,
The virtual lane generator multiplies the difference value between the previously obtained first lane and the previously predicted first lane by a weight and adds the previously predicted first lane to virtually create the first lane, Wherein the second lane is virtually generated by multiplying the difference between the second lane and the previously predicted second lane by a weight and adding a second predicted second lane to the lane keeping control using the exponential smoothing method system. The method of claim 3,
Wherein the virtual lane generator uses a smoothing constant as the weight value. The method according to claim 1,
Wherein the lane detecting unit determines whether or not each side lane is recognized based on whether each side lane is sensed, and detects the recognized lane. The method according to claim 1,
The lane-keeping control unit determines whether or not the vehicle leaves the lane on the basis of the traveling state information obtained from any one of the first group, the second group, and the third group and the vehicle sensors, And controlling the lane keeping of the vehicle when it is determined that the lane departure has occurred. The method according to claim 1,
Wherein the lane-keeping control system is activated when the speed of the vehicle is equal to or greater than 60 KPH (kilometer per hour). A first virtual generator for generating the second lane based on the first lane and the lane width when only the first lane is detected from the first lane and the second lane located on both sides of the forward image; And
And generating a first lane and a second lane by virtue of an exponential smoothing method if neither the first lane nor the second lane is detected,
Wherein the virtual lane generator generates the virtual lane using the exponential smoothing method. 9. The method of claim 8,
The second virtual generator multiplies the difference value between the first lane previously obtained and the previously predicted first lane by a weight and adds the previously predicted first lane to virtually generate the first lane, Wherein the second lane is virtually generated by multiplying the difference between the obtained second lane and the previously predicted second lane by a weight and adding a second predicted second lane to the imaginary lane using the exponential smoothing method Generating device.

Images