KR20140133994A - Apparatus and method for alarming impact - Google Patents

Abstract

The present invention relates to a method and an apparatus for warning of a collision, which includes a relative position input step of receiving relative position information of a peripheral object; a relative position information processing step of calculating at least one of relative velocity vector information and relative acceleration vector information using the relative position information; a position estimation step of using at least one among the calculated relative position information, relative velocity vector information, and relative acceleration vector information to estimate the relative position information of the peripheral object after a predetermined amount of time; and a collision prediction step of using the estimated relative position information of the peripheral object after the predetermined amount of time to predict the possibility of collision between the peripheral object and a user′s vehicle. According to the present invention, Kalman filter is applied to the position, velocity, and acceleration of the peripheral object to estimate the relative position of the peripheral object after the predetermined amount of time and to predict the possibility of collision with the peripheral object, thereby reducing measurement noise and properly dealing with an unexpected situation.

Description

[0001] APPARATUS AND METHOD FOR ALARMING IMPACT [0002]

The present invention relates to a crash warning method and apparatus.

In general, a cruise control system (ACC) or a smart parking assist system (SPAS) mounted on a vehicle measures a distance in a traveling direction by using a vehicle radar or an ultrasonic sensor, Control the driving speed by directly controlling the driving device such as the engine and the brake.

This conventional technique simply measures distances to other vehicles or obstacles and utilizes them in collision prediction. However, the prior art is very vulnerable to measurement noise due to temperature, humidity, and vibration that may occur in a distance measuring sensor. In addition, there is a problem that it is difficult to cope with an unexpected situation such as a vehicle entering a lane change from another lane or a vehicle entering a direction of 90 degrees at an intersection or a sudden acceleration of another vehicle.

In the present invention, a Kalman filter is applied to the position, velocity, and acceleration of a surrounding object to estimate a relative position of a surrounding object after a certain period of time and to estimate a possibility of collision with a surrounding object, thereby reducing measurement noise, And an object of the present invention is to provide a collision warning method and apparatus.

The objects of the present invention are not limited to the above-mentioned objects, and other objects and advantages of the present invention which are not mentioned can be understood by the following description and more clearly understood by the embodiments of the present invention. It will also be readily apparent that the objects and advantages of the invention may be realized and attained by means of the instrumentalities and combinations particularly pointed out in the appended claims.

According to another aspect of the present invention, there is provided a method of estimating a relative position of a peripheral object of a running vehicle and warning a collision, the method comprising: a relative position input step of inputting relative position information of the peripheral object; A relative position information processing step of calculating at least one of relative speed vector information and relative acceleration vector information by using at least one of the relative position information, relative speed vector information, and relative acceleration vector information, And a collision prediction step of predicting collision probability between the neighboring object and the vehicle using the relative position information of the neighboring object after the estimated time period .

The present invention also provides an apparatus for estimating a relative position of a peripheral object of a running vehicle and warning a collision, the apparatus comprising: a relative position input unit for receiving relative position information of the peripheral object; A relative position information processing unit for calculating at least one of relative position information, relative velocity vector information, and relative acceleration vector information, and relative position information of the surrounding object after a predetermined time using at least one of the calculated relative position information, relative velocity vector information, And a collision predicting unit for predicting collision probability between the surrounding object and the vehicle using the relative position information of the surrounding object after the estimated time.

According to the present invention as described above, the Kalman filter is applied to the position, velocity, and acceleration of the peripheral object to estimate the relative position of the surrounding objects after a predetermined time, and the possibility of collision with the surrounding object is predicted, It is possible to cope with the problem.

Fig. 1 shows a vehicle running on the road and other vehicles running in the vicinity. Fig.
2 is a configuration diagram of a collision warning apparatus according to an embodiment of the present invention;
3 is a graph showing a relative distance to an estimated peripheral object over time.
4 is a flowchart illustrating a collision warning process according to an exemplary embodiment of the present invention.
5 illustrates an initial operation process according to an embodiment of the present invention.
FIG. 6 is a flowchart for explaining a collision warning method according to an embodiment of the present invention; FIG.

The above and other objects, features, and advantages of the present invention will become more apparent by describing in detail exemplary embodiments thereof with reference to the attached drawings, which are not intended to limit the scope of the present invention. In the following description, well-known functions or constructions are not described in detail since they would obscure the invention in unnecessary detail. 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 denote the same or similar elements.

A collision warning device is employed in order to prevent an accident from occurring due to collision of a running vehicle with an object such as another vehicle or obstacle in the vicinity.

Generally, such a collision warning apparatus measures a distance to a dangerous object and warns a collision risk when an object is detected within the distance.

On the other hand, there is a measurement noise in the sensor used for measuring the distance to the object, but it is very vulnerable to such noise when the conventional method of measuring only the distance is used. In addition, there is a problem that it is difficult to adapt to unexpected situations such as a risk of collision with an incoming vehicle or a vehicle entering the intersection at 90 degrees by changing lanes in another lane.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and it is an object of the present invention to provide a method and apparatus for estimating a relative position of a surrounding object after a predetermined time by applying a Kalman filter to a position, velocity, And to provide a collision warning method and apparatus capable of appropriately responding to an unexpected situation.

Fig. 1 shows a vehicle running on the road and other vehicles running in the vicinity thereof. Referring to FIG. 1, a vehicle is traveling on a four-lane road (102), and each of the traveling vehicles is recognized as one object (104). These objects can be represented by coordinates. For example, the coordinates of the vehicle 108 as the reference vehicle may be set to the origin coordinates (0, 0), and other vehicles may be set to any coordinates (x, y) according to the distance from the origin. Each vehicle is recognized as a peripheral object, and each of the objects has relative position information, relative velocity vector information, and relative acceleration vector information, respectively.

2 is a block diagram of a collision warning apparatus according to an embodiment of the present invention. Referring to FIG. 2, the collision warning apparatus includes an input unit 202, a relative position information processing unit 212, a position estimating unit 214, and a collision predicting unit 220. The position estimating unit 214 may include a presumption estimating unit 216 and a compensating unit 218. The colliding estimating unit 220 may include a computing unit 222 and a control unit 224. [ The upper collision warning device may further include a warning unit 226 and a driving unit 228.

The input unit 202 receives the relative position information between the peripheral object and the vehicle shown in FIG. The input unit 202 may be at least one of the vision sensor 204, the radar sensor 206, the ultrasonic sensor 208, and other communication means 210. In addition, . For example, in the case of using the ultrasonic sensor 208, when the ultrasonic wave transmitted from the transmitting unit of the ultrasonic sensor 208 attached to the vehicle is reflected by the surrounding object and multiplied by the velocity of the ultrasonic wave to return to the receiving unit, Is the distance that the peripheral object makes a round trip, and half of the round trip distance is the distance from the vehicle to the surrounding object. Also, the relative position information may be input using the other communication means 210 of the vehicle and the surrounding vehicle. For example, it is possible to acquire the positional information using the GPS device attached to the vehicle and the neighboring vehicle, and acquire the relative position information between the vehicles while exchanging the information with each other.

The relative position information processing unit 212 calculates at least one of the relative velocity vector information and the relative acceleration vector information using the relative position information. In order to calculate the relative velocity vector information and the relative acceleration vector information using the relative position information, the following equation is used.

The relative position x is input at a predetermined time interval t in the input unit 202. In Equation 1, x is a relative position before the predetermined time interval t, and x ' Relative position. The linear and rotational motion of an automobile can be said to be equivalent constant velocity motion when measured during a short time interval (t). v _x is the x-axis component of the relative velocity vector, and a _x is the x-axis component of the relative acceleration vector. When x and x 'are obtained through the input unit 202, it is assumed that the predetermined constant time interval t is known and is equivalent speed motion, so that the x axis component of the relative velocity vector and the x axis component of the relative acceleration vector are known have. The above equations apply equally to the x-axis and the y-axis. Thus, the relative position information processing unit 212 can calculate the relative velocity vector information and the relative acceleration vector information using the relative position information.

The position estimating unit 214 estimates relative position information of a surrounding object after a predetermined time using relative position information, relative velocity vector information, and relative acceleration vector information calculated by the relative position information processing unit. The position estimation unit 214 may further include a prediction value predicting unit 216 and a correction unit 218. The prediction value predicting unit 216 may include at least one of relative position information, relative velocity vector information, and relative acceleration vector information The one predictor is predicted and the compensator updates the prior estimate using at least one actual measurement value of relative position information, relative velocity vector information, and relative acceleration vector information. In order to predict the possibility of collision with surrounding objects, the relative position information after a predetermined time of the neighboring object estimated using the relative velocity vector information and the relative acceleration vector information calculated by the relative position information processing unit 212 is used. However, the calculated relative velocity vector information and relative acceleration vector information are combined with noise due to temperature, humidity, and vibration generated in the relative position measurement process through the sensors of the input unit 202. The relative position information after the certain time estimated by the relative velocity vector information or the relative acceleration vector information is not accurate because noise is combined. Therefore, it is desirable to compensate for noise in order to calculate accurate relative position information.

When estimating the relative position information after a predetermined time from the currently measured relative velocity vector information of the neighboring object and the relative acceleration vector information, the noise included in the relative velocity vector information and the relative acceleration vector information is removed through a prediction step and a correction step . This elimination of the noise included in the information is called a filter. In the apparatus and method according to the present invention, such a filter can be applied and a Kalman filter of the filter technique is applied.

The Kalman filter predicts the state of a linear system and predicts with minimization of possible errors. The relative velocity vector information and the relative acceleration vector information of the surrounding objects can be linearly simplified, so that a Kalman filter can be applied. In a linear system, it can be represented by the following two expressions, which are the measurement model functions that can be directly measured by the sensor, and the state model functions that represent the current state but can not be measured.

Equation 2 is the measurement model function and Equation 3 is the state model function. In the above equation, z _k is the measured noise, y _k is the measured value, and y is determined using the x _k value of the state model function. x is the current state of the system, such as position or speed, u _k is noise that can not be controlled by the system (for example, white noise), and w _k is the processor noise. The actual expression of the Kalman filter is:

는 현재 상대 위치 정보의 사전 추정치,

은 일정 시간 간격 이전의 사후 추정치,

는

에 대한 에러 공분산,

은 측정 잡음의 공분산, A는 상태 전이 행렬이다.The prediction step of the Kalman filter is to predict a prior estimate of the current relative position information at time t as: < EMI ID = 4.1 > Equation (4) is also referred to as a time update equation. here,

A pre-estimate of the current relative position information,

Is a posteriori estimate before a certain time interval,

The

The error covariance for < RTI ID =

Is the covariance of the measurement noise, and A is the state transition matrix.

를 갱신하기 위해 현재 상대 위치 정보의 측정값인

를 이용한다.

가 갱신된 사후 추정치인

는 다음의 수학식 5와 같이 표현된다. 수학식 5는 측정값 갱신 방정식이라고도 한다.
In the correction step,

The current value of the relative position information

.

Lt; RTI ID = 0.0 >

Is expressed by the following equation (5). Equation (5) is also referred to as a measurement value update equation.

는 현재 상대 위치 정보의 사후 추정치,

는 현재 상대 위치 정보의 사전 추정치,

는 칼만 필터의 이득(gain), 현재 상대 위치 정보의 측정값인

, H는 실제 측정치 행렬, I는 단위 행렬,

는

에 대한 에러 공분산,

는

에 대한 에러 공분산, 그리고

는 측정 잡음의 공분산이다.here,

Is a posteriori estimate of the current relative position information,

A pre-estimate of the current relative position information,

Is the gain of the Kalman filter, the measured value of the current relative position information

, H is the actual measurement value matrix, I is the unit matrix,

The

The error covariance for < RTI ID =

The

The error covariance for

Is the covariance of the measurement noise.

In the embodiment of the present invention, the Kalman filter operates while the prediction and correction steps are repeatedly performed. The prediction step is performed in the prediction value prediction unit 216 of the present invention, and the correction step can be performed in the correction unit 218. [ The position estimating unit 214 may include the prediction estimating unit 216 and the compensating unit 218. As a result, the position estimating unit 214 may calculate relative position information, relative velocity vector information, relative acceleration vector information Can be estimated.

The collision prediction unit 220 predicts collision probability between the surrounding object and the vehicle using the estimated relative position information, relative velocity vector information, and relative acceleration vector information. The collision prediction unit 220 may further include an operation unit 222 and a control unit 224. Here, the computing unit 222 calculates relative position information after a predetermined time by applying the motion equation to the relative position information, relative velocity vector information, and relative acceleration vector information of the estimated neighboring object, and the control unit 224 calculates the relative position information It is determined that the possibility of collision is high when the relative distance between the surrounding object and the vehicle is less than a predetermined distance.

3 is a graph showing a relative distance to an estimated peripheral object over time. The possibility of collision is set to be determined to be high when the relative distance between the vehicle and the surrounding object is 5 m. Referring to FIG. 3, in FIG. 302, when another vehicle travels in the direction of a point of -40 m from the point of 80 m on the X axis and the point of -20 m from the point of 50 m on the Y axis, the vehicle is at the origin point. Figure 304 shows the relative distance to the vehicle when another vehicle is in progress as shown in Figure 302. It is judged that the possibility of collision is high at 3.3 seconds when the relative distance is 5 m.

The management unit 230 stores the relative position information of the peripheral object estimated by the position estimation unit 214. If the relative position information of the surrounding object after the estimated time is new relative position information, If it is not necessary to maintain the relative position information of the neighboring object after the predetermined time, the deletion may be performed when the relative position information of the neighboring object after the estimated time is different from the previous one. By managing only relative position information that is valid while storing, adding, deleting, and updating relative position information, less memory can be used and efficient processing can be performed.

The warning unit 226 can alert the driver using at least one of display, sound, emergency light, and tail light when the collision prediction unit determines that the possibility of collision is high. More specifically, when it is determined that the possibility of collision is high and a collision warning signal is received in the warning unit, the driver of the vehicle can be warned of the risk of collision through the in-vehicle display device, or the driver can be notified of the danger. In addition, emergency lights and tail lights can be blinked to warn drivers of other vehicles of the risk of collision.

The driving unit 228 can directly control the driving system by performing at least one of an airbag control, a brake control, an engine control, an ESC control, and a steering control when it is determined that the possibility of collision is high in the collision prediction unit. More specifically, when the possibility of collision is high, the brakes can be operated to minimize the impact caused by the collision, and the driver can be protected from danger by operating the airbag in advance. In addition, by stopping the engine, it is possible to increase the stability by preventing the vehicle from accelerating or applying braking force to each wheel when the driver can not control the vehicle by the electronic stability control (ESC) control. If the possibility of collision is high in the traveling direction of the vehicle, steering can be controlled to change the traveling direction to a route with a low possibility of collision. In this way, by directly controlling the drive system, it is possible to effectively assist the driver from actively coping with the collision in an unpredictable situation.

4 is a flowchart illustrating a collision warning process according to an embodiment of the present invention. Referring to FIG. 4, relative position information, relative velocity vector information, and relative acceleration vector information of peripheral objects are input through an input unit (402). This information is then added, deleted, and updated to the tracking list (404). Such stored information is selected based on only valid information.

Next, the reliability of the current value measured using the current value 406, which is the relative position information of the peripheral object, the relative velocity vector information, and the relative acceleration vector information, and the previous value 408 stored in the tracking list is calculated 410). Q and R vectors that reflect external factors to the Kalman filter are used to calculate the reliability, and relative position information, relative velocity vector information, and relative acceleration vector information of the peripheral object reflecting external factors are estimated using the Q and R vectors (412). The estimated information is used to calculate relative position information of surrounding objects 414 after a certain time, and when there is a high possibility of collision 416, the driver is alerted or directly controls the driving system 418 to prevent collision. If the possibility of collision is not high, start from step 402 again.

5 illustrates an initial operation process according to an embodiment of the present invention. Referring to FIG. 5, an initial value is set when estimating relative position information, relative velocity vector information, and relative acceleration vector information of a neighboring object. nRun indicates the number of operations after the operation is started.

If nRun is 1 (508), since the initial operation is started, the initial time is recorded, the system error covariance is set, and the relative position is calculated. The relative position is the measured value.

If nRun is 2 (506), the time is measured to calculate the time interval with the previous time, update the trace lists, and then calculate the measurement noise. Then, the initial relative velocity vector information is estimated and the relative position information is estimated using this information.

If nRun is equal to 3 (504), the time is measured to calculate the time interval with the previous time, the tracking lists are updated, and the measurement noise is calculated. Next, the relative acceleration vector information is estimated and the relative position information and the relative speed information are estimated using this information.

If nRun is 4 (502), the time is measured to calculate the time interval with the previous time, the tracking lists are updated, and the measurement noise is calculated. Next, the relative position, relative velocity vector information, and relative acceleration vector information are estimated. When nRun is 4 or more, the operation is the same as that in the case where nRun is 4 (502).

6 is a flowchart illustrating a collision warning method according to an embodiment of the present invention. Referring to FIG. 6, in a method of estimating a relative position of a peripheral object of a running vehicle and warning a collision, relative position information of a surrounding object is received (602). The relative position information may be input using at least one of a vision sensor, a radar sensor, an ultrasonic sensor, and other communication means.

Next, relative speed vector information and relative acceleration vector information are calculated (604). The relative velocity vector information and the relative acceleration vector information may use the relative position information input in step 602. [

Next, the relative position of the surrounding objects after the predetermined time is estimated (606). Although not shown in FIG. 6, in order to estimate the relative position of the peripheral object, step 606 includes a step of predicting the relative velocity vector information, the prior estimation value of the relative acceleration vector information, the relative velocity vector information, and the measured value of the relative acceleration vector information And updating the pre-estimation value.

Next, the possibility of collision between the surrounding object and the vehicle is predicted (608). Applying a motion equation to the relative position information of surrounding objects after a predetermined time to calculate the relative position information and estimating a collision probability when the relative distance between the calculated relative position information and the vehicle is less than a predetermined distance, May be determined to be high.

Then, if it is determined that the possibility of collision is high, the driver is warned or directly controls the driving system (610). At this time, the driver can be warned by using a display, a sound, an emergency light, and a tail lamp, or can directly control the driving system by performing airbag control, brake control, engine control, ESC control and steering control. If it is determined in step 608 that the possibility of collision is not high, it can be repeated from step 602. [

Although not shown in FIG. 6, the collision warning method according to an embodiment of the present invention may further include storing at least one of relative position information, relative velocity vector information, and relative acceleration vector information of a surrounding object. Adding the relative position information of the neighboring object after the estimated time is new relative position information, deleting the relative position information of the neighboring object after the estimated time, And updating the relative position information of the neighboring object after the estimated time is different from that of the previous neighboring object.

While the present invention has been described in connection with what is presently considered to be practical exemplary embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, But the present invention is not limited thereto.

Claims (14)

A method for estimating a relative position of a peripheral object of a running vehicle and warning a collision,
A relative position input step of receiving relative position information of the peripheral object;
A relative position information processing step of calculating at least one of relative velocity vector information and relative acceleration vector information using the relative position information;
A position estimating step of estimating relative position information of the surrounding object after a predetermined time using at least one of the calculated relative position information, relative velocity vector information, and relative acceleration vector information;
A collision prediction step of predicting a collision probability between the neighboring object and the vehicle using the relative position information of the neighboring object after the estimated time,
The collision warning method.
The method according to claim 1,
The relative position information
A collision warning method using at least one of a vision sensor, a radar sensor, an ultrasonic sensor, and other communication means.
The method according to claim 1,
The position estimating step
Estimating at least one pre-estimation value among the relative position information, relative velocity vector information, and relative acceleration vector information; And
A correction step of updating the predicted value using at least one of the relative position information, the relative velocity vector information, and the relative acceleration vector information,
The collision warning method.
The method according to claim 1,
The collision prediction step
Applying an equation of motion to relative position information of the peripheral object after the estimated time;
Calculating relative position information after a predetermined time according to the motion equation; And
And determining that the possibility of collision is high if the relative distance between the vehicle and the surrounding object according to the calculated relative position information is less than a predetermined distance
The collision warning method.
The method according to claim 1,
Storing relative position information of the neighboring object after the estimated time;
If the relative position information of the neighboring object after the estimated time is new relative position information;
If it is not necessary to maintain the relative position information of the neighboring object after the estimated time, And
Updating the relative position information of the neighboring object after the estimated time is different from previous one
The collision warning method further comprising:
The method according to claim 1,
A driver warning step of warning the driver of at least one of a display, a sound, an emergency light, and a tail light when the possibility of collision is high;
The collision warning method further comprising:
The method according to claim 1,
A driving system control step of directly controlling the driving system by performing at least one of an airbag control, a brake control, an engine control, an ESC control, and a steering control when it is determined that the possibility of collision is high
The collision warning method further comprising:
An apparatus for estimating a relative position of a peripheral object of a running vehicle and warning a collision,
A relative position input unit receiving relative position information of the peripheral object;
A relative position information processing unit for calculating at least one of relative velocity vector information and relative acceleration vector information using the relative position information;
A position estimator for estimating relative position information of the peripheral object after a predetermined time using at least one of the calculated relative position information, relative velocity vector information, and relative acceleration vector information;
And a collision prediction unit for predicting a collision probability between the neighboring object and the vehicle using the relative position information of the neighboring object after the estimated time,
And a collision warning device.
9. The method of claim 8,
The position input unit
A vision sensor, a radar sensor, an ultrasonic sensor, and other communication means.
9. The method of claim 8,
The position estimating unit
A prediction value predicting unit for predicting at least one of the relative position information, the relative velocity vector information, and the relative acceleration vector information; And
And a compensator for updating the predicted value using at least one of the relative position information, the relative velocity vector information, and the relative acceleration vector information,
And a collision warning device.
9. The method of claim 8,
The collision prediction unit
An operation unit for applying the motion equation to the relative position information of the surrounding object after the estimated time and calculating relative position information after a predetermined time according to the motion equation; And
When the relative distance between the vehicle and the surrounding object according to the calculated relative position information is less than a predetermined distance,
And a collision warning device.
9. The method of claim 8,
Storing the relative position information of the neighboring object after the estimated time when the relative position information of the neighboring object after the estimated time is new relative position information, When the relative position information of the neighboring object after the estimated time is different from the previous one,
The collision warning device further comprising:
9. The method of claim 8,
A warning unit for warning the driver of at least one of a display, a sound, an emergency light, and a tail light when the possibility of collision is high,
The collision warning device further comprising:
9. The method of claim 8,
And a driving unit for directly controlling the driving system by performing at least one of an airbag control, a brake control, an engine control, an ESC control, and a steering control when it is determined that the possibility of collision is high.
The collision warning device further comprising:

Images