JP3786113B2 - Approach prediction device - Google Patents

Description

  The present invention relates to an approach prediction apparatus that predicts whether a moving object such as a forward vehicle or a person traveling in front of a traveling direction of a host vehicle traveling on a road surface approaches the host vehicle.

This kind of approach prediction apparatus proposed conventionally can be roughly classified into the following four types.
(1) Detecting the distance of vehicles ahead using a radar whose detection direction is fixed within a specified range
(2) Detecting the distance and direction of the vehicle ahead using a radar whose detection direction changes according to the steering angle of the host vehicle (for example, the steering operation angle)
(3) The angle of the radar beam (irradiation wave) is narrowed, and the beam is scanned (scanned) within a predetermined angle to detect the distance and direction of the vehicle ahead.
(4) An image in front of the host vehicle captured by the imaging device is processed by the image processing device, lane information drawn on the road surface in front of the host vehicle is detected, and a radar detection range is controlled in the traveling lane of the host vehicle. What to do

  However, the conventional approach prediction devices described above have the following problems.

  (1) Since the radar detection direction is limited to one direction, if the radar beam is narrowed down narrowly, the radar detection direction may be on the road in the middle of traveling on a curved road, at the entrance / exit of the curved road, or in front. It will shift from the direction of travel, making it difficult to detect vehicles ahead. On the other hand, if the beam is widened, an unnecessary object such as an object traveling on the straight road may be detected.

  (2) If the radar detection direction is changed according to the steering angle of the host vehicle, the beam width of the radar can be made about the width of the lane. Since the direction does not change, it is effective on a curved road, but the radar detection direction may be deviated from the traveling direction of the road before or at the entrance of the curved road, and there is a possibility that the vehicle ahead cannot be detected.

  (3) If the beam is scanned, it is possible to detect a wide range of forward vehicles in front of the host vehicle, and it is possible to reliably detect the forward vehicle etc. in the middle of the curved road, at the entrance or near, but the road conditions in front of the host vehicle Since the forward vehicle or the like is detected without taking the above into consideration, there is a possibility that the degree of approach to the host vehicle cannot be properly determined. For example, a forward vehicle running in the lane next to the host vehicle along a curved road is traveling in a direction approaching the host vehicle if only the moving direction is seen, but the host vehicle is actually traveling in the same direction. So the possibility of approaching is low. However, in the conventional approach prediction apparatus, there is a possibility that it is determined that there is a high possibility of approaching such a preceding vehicle.

  (4) Although the front vehicle on the lane in which the host vehicle is traveling can be accurately detected by the image processing apparatus, the front vehicle or the like traveling in a lane other than the lane of the host vehicle or a place other than the lane (for example, a road shoulder) ) It may be difficult to detect the vehicle ahead.

  An object of the present invention is to provide an approach prediction device that can accurately determine the degree of approach of a forward vehicle or the like by grasping the road condition ahead of the host vehicle.

  The present invention will be described with reference to FIG. 1 which is a functional block diagram of the present invention. The invention of claim 1 is applied to an apparatus for predicting the degree of proximity of a moving object such as a human to a host vehicle traveling on a road surface.

The approach prediction apparatus according to claim 1 includes a pedestrian crossing detection unit 201 that detects a pedestrian crossing, a host vehicle behavior detection unit 202 that detects a behavior of the host vehicle, and a moving object that exists in front of the host vehicle in the traveling direction. and object behavior detecting means 203 for detecting the behavior, detected crosswalk, Ru and a proximity predicting means 204 based on the behavior of the behavior and the vehicle of a moving object predicting proximity of the moving object with respect to the vehicle.

  The above-described object is that the own vehicle behavior detecting unit 202 calculates the position and velocity vector of the own vehicle, and the object behavior detecting unit 203 calculates the position of the moving object and between the moving object and the pedestrian crossing. When the distance of the vehicle is less than or equal to a predetermined value, it is determined that the moving object crosses the pedestrian crossing, and when the moving object is determined to cross the pedestrian crossing, This is achieved by outputting the proximity based on the distance.

It is preferable that the pedestrian crossing detection means 201 detects a pedestrian crossing from the shape of the white line drawn on the road surface .

Alternatively, when the pedestrian crossing detection unit 201 includes a reception unit that receives road information transmitted from a sign post installed on the roadside, it is preferable to detect the pedestrian crossing from the received road information .

Alternatively, when the pedestrian crossing detection means 201 includes a receiving means for receiving a GPS signal from a satellite and a storage means for storing road map data, the pedestrian crossing is determined from the received GPS signal and road map data. It is preferable to detect .

  Moreover, it is preferable that the object behavior detection unit 203 includes an infrared detection unit that detects infrared rays emitted from the moving object.

  The pedestrian crossing detection means 201 detects a pedestrian crossing, the own vehicle behavior detection means 202 detects the behavior of the own vehicle, and the object behavior detection means 203 detects the behavior of a moving object existing ahead of the traveling direction of the own vehicle. The approach degree predicting means 204 predicts the approach degree of the moving object to the own vehicle based on the detected pedestrian crossing, the behavior of the moving object, and the behavior of the own vehicle.

  The degree of approach between the moving object and the host vehicle is predicted in consideration of the existence of a pedestrian crossing, and the degree of approach considering the road condition ahead of the host vehicle can also be predicted.

  In the section of means and action for solving the above-described problems for explaining the configuration of the present invention, the drawings of the embodiments are used for easy understanding of the present invention. However, the present invention is limited to the embodiments. It is not something.

  As described above in detail, according to the present invention, the degree of approach is calculated in consideration of travel guidance information such as lanes drawn on the road surface. Predicts the degree of proximity of moving objects accurately

-1st Example-
The first embodiment is, for example, as shown in FIG. 2, the preceding vehicle MVb traveling a traffic lane R ₊₁ in front of the vehicle MVa traveling lane R ₀ (hereinafter, referred to as front vehicle) If exists, the vehicle MVa position A, speed vector Va, forward vehicle MVb position B, speed vector Vb are calculated, and further, a direction vector Db indicating the direction of the white line WLb that is a boundary line between the lane R ₀ and the lane R ₊₁ is calculated. Then, it is determined whether or not the forward vehicle MVb changes the lane to the lane in which the vehicle MVa travels, and whether or not the vehicle MVa approaches a safe inter-vehicle distance or less after the lane change. In the following description, the X axis, the Y axis, and the Z axis are taken as shown in FIG.

  FIG. 3A is a block diagram showing a schematic configuration of an embodiment of an approach prediction apparatus according to the present invention. As shown in FIG. 3A, this approach prediction device is mounted on a vehicle MVa that travels on the road surface RD in the travel direction indicated by the arrow A.

  Reference numeral 1 denotes a wheel speed sensor that is provided in the vicinity of the left and right rear wheel tires 2 and detects the rotational speed of the left and right rear wheel tires 2. For example, the wheel speed sensor 1 rotates the tire 2 by detecting a change in the magnetic flux by outputting a magnetic flux toward the groove, and a conductor disk that is integrally formed with the tire 2 with grooves formed at a predetermined pitch at the peripheral edge. Although a magnetic sensor that outputs a pulse train proportional to the number is provided, the configuration itself is well known, and a detailed description thereof will be omitted.

  Reference numeral 3 denotes a CCD camera that is arranged in front of the vehicle MVa and images the forward field of view of the vehicle MVa. A detection signal from the CCD camera 3 is processed by the preprocessing unit 4 to detect white lines WLa to WLc. . The result is input to the control device 5. The preprocessing unit 4 can also recognize a road sign, recognize a pattern other than a white line drawn on the road surface, and detect a roadside reflector installed on the road.

  Reference numeral 6 denotes a radar device disposed at the front of the vehicle MVa, which irradiates the electromagnetic wave beam BM in front of the vehicle MVa and receives a reflected wave from the preceding vehicle existing on the beam BM. In the illustrated example, the electromagnetic wave beam BM is narrowed down relatively narrowly, and is configured to scan within an angle range (for example, ± 45 ° to 60 °) that is greater than the normal lane width. The detection signal of the radar device 6 is input to the preprocessing unit 7, where the distance from the propagation time of the reflected electromagnetic wave received by the radar device 6 to the vehicle ahead is measured, and the vehicle from the arrival direction of the reflected electromagnetic wave. The direction of the forward vehicle MVb with respect to MVa is detected. The relative coordinate position of the preceding vehicle with respect to the host vehicle is calculated from these distance and direction data. The processing result of the preprocessing unit 7 is also input to the control device 5.

  The control device 5 is mainly composed of a microcomputer, and as a function thereof, as shown in FIG. 3B, includes a data extraction unit 51, a forward vehicle behavior prediction unit 52, and an approach degree determination unit 53. . The data extraction unit 51 extracts data such as the position of the vehicle MVa, the traveling direction, the position of the preceding vehicle, and the traveling direction, as will be described later. The behavior prediction unit 52 of the preceding vehicle predicts the behavior of the preceding vehicle based on the various data extracted by the data extraction unit 51. The approach degree judgment unit 53 predicts and judges the degree of approach between the preceding vehicle and the host vehicle based on various data extracted by the data extraction unit 51 and the behavior of the preceding vehicle predicted by the preceding vehicle behavior prediction unit 52. To do.

  The approach degree determination unit 53 includes a rule memory 53a, an IF unit collation unit 53b, a rule selection unit 53c, and a rule competition unit 53d. In the illustrated example, a technique used in so-called artificial intelligence is used for predicting the degree of approach with the vehicle ahead. Briefly describing this, a large number of rules are stored in advance in the rule memory 53a, and each rule is described by IF to THEN rules used in artificial intelligence. The IF unit collation unit 53b selects a rule according to the output of the data extraction unit 51 from among a plurality of conditions stored in the rule memory 53a according to the outputs of the data extraction unit 51 and the behavior prediction unit 52 of the preceding vehicle. Then, the IF part of the selected rule is verified. Next, the rule selection unit 53c selects a specific rule that satisfies the IF unit based on the collation result of the IF unit collation unit 53b. Finally, the rule competition unit 53d causes these rules to compete when a plurality of rules are selected by the rule selection unit 53c. The competition result of the rule competition unit 53d corresponds to the degree of approach with the preceding vehicle. Details of the operation of each part constituting the proximity determination unit 53 will be described later.

  FIG. 4 is a main flowchart for explaining the operation of the approach prediction apparatus. In step S10, image data of the forward field of view of the vehicle MVa imaged by the CCD camera 3 and processed by the preprocessing unit 4 is read, and image processing is performed on the image data to detect a white line. In step S20, a detection signal from the wheel speed sensor 1 is captured. In step S30, the position and direction of the preceding vehicle calculated in advance by the preprocessing unit 7 are read based on the signal from the radar device 6. In step S40, using various signals read in this way, data necessary for predicting the behavior of the preceding vehicle, which will be described later, that is, the absolute coordinates of the own vehicle, the preceding vehicle and the white line, the own vehicle and the preceding vehicle. Each speed vector, each traveling lane of the host vehicle and the preceding vehicle, and a unit direction vector of a white line near the preceding vehicle are extracted by calculation.

  In step S50, the front vehicle behavior prediction unit 52 predicts the behavior of the vehicle ahead based on the extracted data. In step S60, after the IF part collation part 53b of the proximity determination part 53 collates the IF part of various rules, the rule selection part 53c selects a rule based on the result of the collated IF part. In step S70, the rule selected in step S60 is caused to compete in the rule competition unit 53d of the proximity determination unit 53. In step S80, the degree of approach is output based on the conflicting rule.

ここに、Ｉ，Ｊ：観測画像上の任意の点Ｐの座標値
ｉ，ｊ：Ｐ(Ｉ，Ｊ)に対応する視点変換画像上のｐ点の座標値
Ａ，Ｂ：視点変換画像の表示サイズ
ｐ，ｑ：ＣＣＤカメラの画素サイズ
θv：ＣＣＤカメラの俯角
ｆ：ＣＣＤカメラの焦点距離
ｉｎｔ( )：( )内の値の整数部分を示す関数
なお、上述の(１)，(２)式による視点変換は、ＣＣＤカメラ３により撮像された物体がすべて路面ＲＤ上に存在する平面物体であることを前提としているので、たとえば図６(ａ)に示す前方車両ＭＶｂは、同図(ｂ)ではＭＶｂ’に示すように歪んで変換される。 Next, details of each step of FIG. 4 will be described.
(1) Image Processing Routine FIG. 5 shows details of the image processing routine executed in step S10 of FIG. In step S11, image data captured by the CCD camera 3 and processed by the preprocessing unit 4 is read. In step S12, the viewpoint of the image data obtained in the preprocessing unit 4 is converted. That is, as shown in FIG. 6 (a), the CCD camera 3 images the road surface RD obliquely from the predetermined height of the vehicle MVa. However, the direction vector of the white line WL described later (Db in FIG. 2). In order to calculate (shown), it is more convenient to obtain an image in which the road surface RD and the vehicles MVa and MVb are looked down from directly above. A conversion formula for converting an observation image obtained from the CCD camera 3 shown in FIG. 6A to an image looking down from directly above (hereinafter referred to as a viewpoint conversion image) as shown in FIG. It is given by

Where I, J: coordinate values of an arbitrary point P on the observed image
i, j: Coordinate value of p point on the viewpoint conversion image corresponding to P (I, J)
A, B: Display size of viewpoint conversion image
p, q: CCD camera pixel size
θv: CCD camera depression angle
f: focal length of the CCD camera int (): function indicating the integer part of the value in () Note that the viewpoint conversion according to the above-mentioned formulas (1) and (2) is performed for all objects imaged by the CCD camera 3 on the road surface. Since it is assumed that the object is a planar object existing on the RD, for example, the forward vehicle MVb shown in FIG. 6A is distorted and converted as shown by MVb ′ in FIG. 6B.

  In step S13, a white line extraction operation for extracting a white line included in the image is performed on the image data subjected to viewpoint conversion.

  The details of the white line extraction operation are shown in the flowchart of FIG. In step S131, a region that is a white line candidate is detected from the image data whose viewpoint is converted by the white line detection filter.

The white line detection filter used in the present embodiment is, for example, as shown in FIG. 8, and the “1” portion that is widened to the left and right by the width w with respect to the target point p having the coordinate value (x, y). And “−1” portions on the left and right sides thereof. The width w may be given in advance according to the width of the white line WL to be extracted, and is set such that W <w, where W is the width of the white line to be extracted. The output F (x, y) of the white line extraction filter for the target point p (x, y) is given by the following equation.

ここに、０＜ｋ＜１：定数
Ｍ：画像の横方向画素数 If the width w is appropriately determined, the value of the output F (x, y) increases when the target point p (x, y) is positioned on the white line WL, and thus the output F calculated at each point of the entire image. By performing threshold processing on (x, y), a white line candidate region can be extracted. At this time, in order to set an appropriate threshold value for each target point p (x, y), the threshold value t is determined for each horizontal line (X-axis direction) as shown in the following equation.

Where 0 <k <1: constant
M: Number of pixels in the horizontal direction of the image

Ｌ（ｘ，ｙ）が１の値をとる領域が、白線の候補領域である。 When image data representing a white line candidate region generated using the threshold value t is L (x, y), L (x, y) is given by the following equation.

A region where L (x, y) takes a value of 1 is a white line candidate region.

  In the white line candidate region extraction operation executed in step S131 described above, the white line WL itself actually drawn on the road surface RD is extracted, and the actual white line as shown in FIGS. WL is periodically interrupted. Therefore, as shown in FIG. 9A, the white line candidate region extracted in step S131 is also (ideally) periodically interrupted. However, the concept of the white line is different from the actual white line WL, and is a series of curves that are not interrupted in the middle. If the white line is interrupted, there is a possibility of hindering the calculation of the direction vector of the white line described later. To connect the white line candidate areas.

  The procedure of step S132 will be described in detail with reference to FIG. The white line candidate area extracted in step S131 is thinned as shown in FIG. 9B, and the center line of the white line candidate area is extracted. Next, as shown in FIG. 9C, after the end points of the center line are detected, the end points having a predetermined interval or less are connected as shown in FIG. Further, if the logical sum of the end points connected as shown in FIG. 5E and the white line region candidate extracted in step S131 is taken, the white line candidate region as shown in FIG. The concatenation work is completed. In each operation of step S132, image processing such as thinning, end point detection, and logical sum may be performed by a known method.

  In step S133, a white line template is created as follows based on the white line candidate regions connected in step S132. A labeling process is performed on the connected white line candidate regions, and a quadratic curve is matched to a region having the maximum area in each labeled region to obtain a template. For example, if the white line candidate region extracted in step S131 is as shown in FIG. 10A, the white line at the approximate center is the longest among the three white line candidate regions in the connected state. Since this is a region having the largest area, a quadratic curve is matched to the substantially central white line candidate region, and a template T (x, y) as shown in FIG.

  In step S134, the white line candidate region extracted in step S131 is searched for a matching region using the template T (x, y) created in step S133. As a result, for the white line candidate region as shown in FIG. 10A, three white line regions WLRa to WLRc as shown in FIG. 10C are detected.

(2) Data Extraction FIG. 11 is a flowchart showing the data extraction routine executed in step S40 of the main flowchart shown in FIG. In step S41, based on the signal from the wheel speed sensor 1 input in step S20 of FIG. 4, the coordinates of the position of the vehicle MVa are fixed to the coordinates whose origin is the position of the vehicle MVa at a predetermined time, that is, the road surface RD. Obtained on the absolute coordinate system. As shown in FIG. 2, the absolute coordinate value of the vehicle MVa is A (x _a , y _a , 0).

  A pulse train corresponding to the vehicle speed is taken from the wheel speed sensor 1 to calculate the travel distance of the vehicle MVa. On the other hand, the moving direction of the vehicle MVa is obtained from the difference between the wheel speeds detected by the wheel speed sensors 1 of the left and right rear wheels. Therefore, the absolute coordinate value of the vehicle MVa can be obtained by sequentially adding the moving distance and moving direction of the vehicle MVa.

Next, in step S42, the absolute coordinate value of the vehicle MVa obtained in step S41 is added to the data of the white line regions WLRa to WLRc obtained in step S134 of FIG. 7 as an offset, whereby the white line regions WLRa to WLRc are added. Is converted to data on the absolute coordinate system. Similarly, the absolute coordinate value of the host vehicle MVa is added as an offset to the relative coordinate value of the preceding vehicle read in step S30, and the position of the preceding vehicle is converted to an absolute coordinate value. In the example shown in FIG. 2, the absolute coordinate value of the forward vehicle MVb is represented by B (x _b , y _b , 0).

In step S43, a difference between the absolute coordinate values of the vehicle MVa and the forward vehicle MVb is obtained from the absolute coordinate values of the vehicle MVa and the forward vehicle MVb measured last time and the absolute coordinate values of the vehicle MVa and the forward vehicle MVb measured this time. Thus, each velocity vector is calculated. As shown in FIG. 2, Va velocity vector of the vehicle MVa and the vehicle ahead MVb respectively _{(v ax, v ay, 0} ), expressed by _{Vb (v bx, v by,} 0).

In step S44, it is recognized that a plurality of lanes are divided by the white line regions WLRa to WLRc converted into the position data of the absolute coordinate system in step S42, and the vehicle MVa and the preceding vehicle MVb travel on which lane respectively. It is determined whether or not. In this embodiment, as shown in FIG. 2, it is determined that the vehicle MVa is traveling on the lane R ₀ and the front vehicle MVb is traveling on the lane immediately adjacent to the right side of the lane R ₀ on R _+1. The In this specification, the lane immediately adjacent to the left side of the lane of the host vehicle is expressed as R ₋₁ , and the lane adjacent to the right side of the vehicle is expressed as R ₊₂ .

  In step S45, a unit direction vector of the white line region WLRb in the vicinity of the forward vehicle MVb is detected. Details of the white line unit direction vector detection method are shown in the flowchart of FIG.

In step S451 of FIG. 12, as shown in FIG. 13 (a), the front vehicle MVb is on the white line area WLRb on the side close to the lane R ₀ in the white line area that divides the lane R _{+ 1} that the front vehicle MVb travels. A point C at which the distance from the position B is minimum is detected. In step S452, points C ₁ and C ₂ separated from the point C by a predetermined distance are set on both sides of the point C on the white line region WLRb. In step S453, a straight line connecting these points C ₁ and C ₂ is set as a unit direction vector of the white line. As shown in FIG. 2, the unit direction vector of the white line WLb is represented by Db.

ここに、×：ベクトルの外積を表す記号
∩：アンド条件を示す演算子
ｋ₁：定数（０＜ｋ₁＜１であり、かつ、ｋ₁は０に近い値）
ｋ₂：定数（０＜ｋ₂であり、かつ、ｋ₂は０に近い値）
⊃：左辺が真ならば右辺が真であることを表す記号
ＲＣ(+1,0)：前方車両が車線Ｒ₊₁からＲ₀に車線変更するという事実
( )_z：( )内のベクトルのｚ成分
Ａ∈Ｒ₀ ：座標Ａが車線Ｒ₀内にあるという事実
Ｂ∈Ｒ₊₁：座標Ｂが車線Ｒ₊₁内にあるという事実 (3) Forward Vehicle Behavior Prediction Details of the forward vehicle behavior prediction in step S50 of FIG. 4 are as follows. In this embodiment, it is predicted whether or not the forward vehicle MVb will change the lane to the travel lane of the vehicle MVa. This prediction formula is given by the following formula.

Where: x: Symbol for vector cross product
演算 子: Operator indicating AND condition
k ₁ : constant (0 <k ₁ <1 and k ₁ is a value close to 0)
k ₂ : Constant (0 <k ₂ and k ₂ is a value close to 0)
⊃: A symbol indicating that the right side is true if the left side is true RC (+1,0): The fact that the vehicle ahead changes lane from R ₊₁ to R ₀
() _z : z component of vector in () A∈R ₀ : fact that coordinate A is in lane R ₀ B∈R ₊₁ : fact that coordinate B is in lane R ₊₁

(Db × Vb) _z / | Vb | is the value of sin θ when the angle formed by the unit direction vector Db of the white line and the speed vector Vb of the forward vehicle MVb is θ, and if this value is positive, The vehicle MVb is approaching the white line. Also, | (Db × Vb) _z | is a speed component perpendicular to the white line (its unit direction vector Db) of the speed vector Vb of the preceding vehicle MVb. k ₁ is a constant for determining the sign of (Db × Vb) _z / | Vb |, and is set to a constant close to 0 in consideration of measurement errors. Further, k ₂ is a threshold value of | (Db × Vb) _z |, and it is not determined that there is a risk of lane change even when the vehicle ahead runs slightly in the lane R _{+ 1} . Set to value.

Therefore, in a state where the vehicle MVa is traveling on the lane R _{0 and the} preceding vehicle MVb is traveling on the lane R ₊₁ , (Db × Vb) z / | Vb | takes a positive value (> k ₁ ). If | (Db × Vb) _z | is determined to be larger than a predetermined threshold (> k ₂ ), the forward vehicle MVb is changed from the lane R ₊₁ to the lane R ₀ where the vehicle MVa is traveling. Predicting what is going to happen, the fact RC (+1,0) becomes true.

(4) IF part verification of rule The IF part verification of the rule in step S60 of FIG. 4 is performed as follows. Based on the various data calculated by the data extraction unit 51 and the behavior prediction unit 53 of the preceding vehicle, the IF unit collation unit 53b includes a rule including this data as a parameter of the IF unit in the rule memory 53a of the proximity determination unit 53. Are selected from a large number of stored rules, and various data are input for the selected rule to check the IF section, that is, whether or not the IF section of each rule is determined to be true.

ここに、Ｔ：空走時間
ルール番号２のルールのＩＦ部が真と判定されるためには、ＲＣ(+1,0)，速度ベクトルＶａ，Ｖｂおよび座標値Ａ，Ｂが入力される必要がある。 FIG. 14 shows various rules stored in the rule memory 53a. The related rule in this embodiment is the rule of rule number 2, RC (+1,0) is a variable indicating the behavior of the preceding vehicle calculated in the above step S50, and Va and Vb are both in the above step S40. The speed vectors f (Va, Vb) of the vehicle MVa and the preceding vehicle MVb calculated in (Step S43 in FIG. 11) are functions indicating the safe inter-vehicle distance, and are defined by the following equations.

Here, it is necessary to input RC (+1,0), velocity vectors Va and Vb, and coordinate values A and B in order to determine that the IF part of the rule of rule No. 2 is true. There is.

In the example shown in FIG. 2, the forward vehicle MVb is about to change the lane R ₀ of the vehicle MVa, and the speed vector Vb of the forward vehicle MVb is in the direction of the lane R _0. Is determined to be true. Further, when the distance y _b -ya between the forward vehicle MVb and the vehicle MVa along the Y-axis direction has _a positive value smaller than the safe inter-vehicle distance f (Va, Vb), the result of the lane change is that When the vehicle MVb enters the lane R ₀ , there is a risk that the distance between the vehicles MVb is narrow and the vehicle MVb may approach sufficiently. The IF part of the rule number 2 is determined to be true, the rule is established, and the approach degree = 1 is output. The here,
Approach degree = 0. . . No effect Approach = 1. . . Attention should be paid to the vehicle ahead. . . Suppose there is a risk of contacting the vehicle ahead. Incidentally, -f (Va, Vb) by obtaining the magnitude relation between the distance y _b -y _a about multiplied by the constant smaller than 1 in the section may determine the proximity = 2.

であり、式(６)の左辺が真と判定されずにＲＣ(+1,0)は偽になる。この結果、ルール番号２のルールのＩＦ部は真と判定されず、ルール番号２のルールは成立せずに接近度の出力は行われない。 On the other hand, in the case shown in FIG. 15 where the host vehicle enters the curved road from the straight road, the forward vehicle MVb has already traveled on the curved road, and in the conventional approach predicting device, the forward vehicle MVb is placed immediately before the vehicle MVa. In order to detect, although the possibility that the front vehicle MVb will actually change lanes is low, there is a possibility that an alarm indicating approach is output. In this case, according to the present embodiment, the speed vector Vb of the preceding vehicle MVb and the unit direction vector Db of the white line WLb are parallel,

RC (+1,0) becomes false without the left side of Equation (6) being determined to be true. As a result, the IF part of the rule with the rule number 2 is not determined to be true, the rule with the rule number 2 is not established, and the approach degree is not output.

(5) Rule Conflict The rule conflict in step S70 of FIG. 4 will be described. As shown in the present embodiment and the embodiments described later, there are many situations where the approach should be predicted while the vehicle MVa is traveling, and it is necessary to make an accurate judgment in various situations. A plurality of rules are stored in the rule memory 53a of the unit 53 (see FIG. 14). Therefore, since the IF part of a plurality of rules may be established for one situation, a plurality of rules for which the IF part is established are competed to output an appropriate degree of approach.

For the competing method, for example, a method performed in the field of artificial intelligence may be used. For example,
I. Take the sum of the degrees of approach output from each of the rules for which the IF part is established. Prioritize rules, and output the degree of approach of rules with higher priority c. A method such as a combination of these two methods can be mentioned. For a, a weighted average of the degree of approach may be taken. An example of rule conflict will be described later in detail.

(6) Approach degree output In step S80 of FIG. 4, the approach degree is output as follows. When all the IF parts of the rule stored in the rule memory 53a are not established, a default value of 0 is output as the approach degree. The output degree of approach is notified to the driver by an alarm device such as a lamp or buzzer provided in the vehicle MVa, or displayed on a display or the like provided in the vehicle MVa using characters. An example of controlling the travel of the vehicle MVa based on the output degree of approach will be described later.

  Therefore, according to the present embodiment, in addition to information such as the position and speed of the forward vehicle MVb existing in front of the vehicle MVa, travel guidance information such as the shape of a white line (unit direction vector) in the forward field of view of the vehicle MVa is also considered. Since the approach degree is calculated, the approach degree of the preceding vehicle can be accurately determined and predicted according to the road condition. For example, in FIGS. 2 and 15, the position B of the forward vehicle MVb and the direction of the velocity vector Vb are almost the same, and the conventional approach prediction device outputs an alarm as a possibility of approach in any case. In this embodiment, an approach degree (= 1) indicating that there is a possibility of approach only for those that may actually change lanes as shown in FIG. 2 is output, and the forward vehicle MVb is simply curved as shown in FIG. For those that are only traveling on the road and are less likely to change lanes, an approach degree (= 0) indicating that the possibility of approach is low is output, and an accurate approach degree determination is performed according to the road condition.

-Modification of the first embodiment-
In the first embodiment described above, the unit direction vector of the white line is used as the travel guide information. However, the present invention is not limited to this, and travel guide information that can accurately grasp the road condition can be used. FIG. 16 is a flowchart showing a modification of the first embodiment, and is for explaining details of step S40 in FIG. Except for step S40 and step S50, which will be described later, the contents of the other steps in FIG. 4 are the same as those in the first embodiment described above, and a description thereof will be omitted.

After performing the same processing as steps S41 to S44 in FIG. 11, in step S45A, as shown in FIG. 2, the position B of the front vehicle MVb, the lane R ₀ in which the vehicle MVa is traveling, and the front vehicle MVb are A distance C (t) (t: time) between the white line WLb separating the traveling lane R _{+ 1} is calculated. The distance C (t) is stored in the data extraction unit 51 together with the elapsed time t from the predetermined position described above.

ここに、ｋ₈：定数（０＞ｋ₈） In step S50 of FIG. 4, the behavior of the preceding vehicle, that is, whether the preceding vehicle MVb changes the lane to the traveling lane of the vehicle MVa is predicted by the following equation.

Where k _{8 is a} constant (0> k ₈ )

(dC (t) / dt) is the speed of the direction of the forward vehicle MVb orthogonal to the white line WLb, k ₈ corresponds to the threshold of the speed, running forward vehicle MVb is slightly meandering lane R ₊₁ Even if the lane changes, the value is set so that it is not determined that the lane may be changed. Therefore, if (dC (t) / dt) is determined to be smaller than the predetermined threshold value (<k _8), the forward vehicle MVb is trying to change lanes in the lane R ₀ being the vehicle MVa is traveling from the lane R ₊₁ The fact RC (+1,0) is true.

  Thereafter, based on the fact RC (+1,0) predicted in step S50 and the velocity vectors Va and Vb and the coordinate values A and B calculated in steps S41 to S44, the degree of approach is determined in steps S60 to S80. Calculated and output. Therefore, also by this modification, the same operation effect as the above-mentioned 1st example can be obtained.

-Second Example-
In the above-described first embodiment and its modified example, the unit direction vector of the white line is used as the travel guide information, but the result of pattern recognition of the pattern shape other than the white line drawn on the lane as in the second embodiment. Can also be used as travel guidance information.

For example, as shown in FIG. 17, when a forward vehicle MVb exists in front of the vehicle MVa, the lane in which the forward vehicle MVb is traveling is any type of lane, for example, the lanes R ₀ and R ₊₁ . Whether the vehicle is a traveling lane such as the above or a merging lane such as the lane R- ₁ is recognized by the shape of the pattern drawn in the lane, and the recognized lane shape is determined based on the position A, Using the speed vector Va, the position B of the forward vehicle MVb, and the speed vector Vb, whether the forward vehicle MVb changes the lane to the lane in which the vehicle MVa travels, and is safe for the vehicle MVa after the lane change Judge whether or not the distance is less than the correct inter-vehicle distance. As shown in FIG. 17, the merge lane has a shape that merges with the right lane in the middle and the left roadside separation zone merges with the white line on the right, and usually the merge lane has the meaning of merging at the merge position. Therefore, it can be determined that the roadside separation zone crosses the lane and that the lane R- ₁ is a merging lane if the pattern MP is recognized.

  The second embodiment is different from the first embodiment in that the data extraction executed in step S40 in FIG. 4, the behavior prediction of the preceding vehicle in step S50, the IF check of the rule in step S60, the rule in step S70 This is a specific procedure for contention and the degree of approach in step S80, and the rest is the same, including the configuration of FIG. Therefore, in the following description, it demonstrates centering on difference with 1st Example, and abbreviate | omits description about a common part.

FIG. 18 shows a data extraction subroutine in the second embodiment, which corresponds to FIG. 11 of the first embodiment. After performing the same processing as steps S41 to S44 in FIG. 11, in step S45B, it is determined whether or not the lane R- ₁ in which the preceding vehicle MVb is traveling is a merged lane. Details of the merge lane determination method are shown in the flowchart of FIG.

In step S4501 in FIG. 19, among the lanes separated by the white line detected in step S13 in FIG. 5, the lane adjacent to the lane in which the vehicle MVa is traveling (in which the lane in which the forward vehicle MVb is traveling is included). R- ₁ is included), and an edge extending in the direction crossing the white line (referred to as a lateral edge) is extracted from the detected edges.

FIG. 20 is a diagram showing an example of the horizontal edge detection process. When an observation image as shown in FIG. 20A is obtained, the result of the viewpoint conversion operation for this observation image is shown in FIG. The result of performing the white line detection operation on the viewpoint conversion image is as shown in FIG. The lane in which the vehicle MVa is traveling is the center lane R ₀ , and the forward vehicle MVb is traveling in the lane R _-1 on the left side of the lane R ₀ . Therefore, when the horizontal edge is extracted from the viewpoint conversion image corresponding to the lanes R ₊₁ and R _-1 adjacent to the lane R ₀ , a plurality of horizontal edges (represented in the figure are representative). Only horizontal edges b ₁ , b ₂ and b ₃ are labeled). Note that the edge detection method is well known, and for example, a method of multiplying the viewpoint conversion image by an edge detection filter such as a differential filter can be cited.

In step S4502, a labeling process is performed on each horizontal edge extracted in step S4501, and the horizontal length of each labeled horizontal edge is measured. In step S4503, it is determined whether the horizontal length of the horizontal edge is substantially equal to the width of the adjacent lanes R _{+ 1} , R- ₁ , and if the determination is affirmative, the process proceeds to step S4504, where the horizontal length equal to the adjacent lane width is determined. It is determined that the lane R- ₁ where the edge exists is a merged lane. This fact
R ₋₁ = RJ (11)
It expresses. On the other hand, if the determination is negative, the process proceeds to step S4505, and it is determined that the adjacent lane is not a merged lane.

As shown in FIG. 17, the pattern MP drawn in the adjacent lane R ₋₁ that is the merging lane and the roadside separation band at the end of the merging lane are the horizontal edges in FIG. Extracted as b ₁ , b ₂ and b ₃ . The lateral lengths of the lateral edges b ₁ and b ₂ are close to the lane width of the adjacent lane R ₋₁ , and the lateral length of the lateral edge b ₃ is equal to the lane width of the adjacent lane R ₋₁ . Therefore, if the horizontal length of the horizontal edge extracted in step S4501 is substantially equal to the lane width of the adjacent lane R- ₁ (input in advance), it is determined that the adjacent lane R- ₁ is a merged lane. it can.

ここに、ＲＣ₄(-1,0)：前方車両が車線Ｒ_-1からＲ₀に車線変更するという事実
Ｂ∈Ｒ-1：座標Ｂが車線Ｒ_-1内にあるという事実 In step S50 of FIG. 4, the behavior of the preceding vehicle, that is, whether the preceding vehicle MVb changes the lane to the traveling lane of the vehicle MVa is predicted by the following equation.

Where RC ₄ (-1,0): the fact that the vehicle ahead changes lanes from R _-1 to R ₀
B∈R-1: the fact that the coordinate B is in the lane R- ₁

Accordingly, in a state where the vehicle MVa is traveling on the lane R _0, the preceding vehicle MVb is traveling lane R _-1, if it is determined that the lane R _-1 is the merging lane, the vehicle ahead MVb lane R _- It is predicted from ₁ that the vehicle MVa is traveling to the lane R ₀ in which the vehicle MVa is traveling, and the fact RC (−1,0) becomes true.

  In step S60 of FIG. 4, various data are input for the rule selected from the various rules in the rule memory 53a of the proximity determination unit 53, and the IF unit is verified. The related rule in the present embodiment is the rule of rule number 4 shown in FIG.

In the example shown in FIG. 17, the forward vehicle MVb is traveling on a merging lane R _-1, always changing lane to the running lane R0 of the vehicle MVa in to the front vehicle MVb reaches the end of the merging lane R _-1 Therefore, the left side of Equation (12) is determined to be true. Further, when the distance y _b -ya between the forward vehicle MVb and the vehicle MVa along the Y-axis direction has _a positive value smaller than the safe inter-vehicle distance f (Va, Vb), the result of the lane change is that When the vehicle MVb enters the lane R ₀ , the distance between the vehicles may be narrow and the vehicle MVb may approach sufficiently. The IF part of the rule number 4 is determined to be true, the rule is established, and the approach degree = 2 is output. The

In step S70, the rule in which the IF unit is collated in step S60 is conflicted. For example, as shown in FIG. 21, a case where the forward vehicles MVb and MVc are traveling in the merging lane R _{-1 on} the left side and the traveling lane R ₊₁ on the right side of the lane R _{0 on} which the vehicle MVa travels, respectively. Think. In the example shown in FIG. 21, it is the rules of rule numbers 2 and 4 shown in FIG. 14 that the IF part is determined to be true in step S60. Therefore, the degree of approach = 1 is output from the rule of rule number 2, and the degree of approach = 2 is output from the rule of rule number 4.

According to the above-mentioned method (a), the sum of the approach degrees = 1 + 2 = 3. In this case, the approach degree is evaluated in three or more stages. Further, according to the above-mentioned method of B, when comparing the forward vehicle MVb and the forward vehicle MVc, the forward vehicle MVb always changes the lane to the lane R ₀ , so that the rule of rule number 4 has a higher priority. I can say that. Therefore, for example, if the priority order is 10 steps as a whole, the priority order of the rule of rule number 2 is 5, and the priority order of the rule of rule number 4 is priority 7, the priority order of the rule of rule number 4 is higher. , Proximity = 2 is output.

  Therefore, according to the present embodiment, in addition to information such as the position and speed of the vehicle ahead of the forward vehicle MVb existing in front of the vehicle MVa, a pattern indicating whether the vehicle is a merge lane in the forward field of view of the vehicle MVa, etc. Since the approach degree is calculated in consideration of the travel guide information, the approach degree of the preceding vehicle can be accurately determined and predicted according to the road condition.

-Modification of the second embodiment-
In the second embodiment described above, pattern recognition is performed on the screen imaged from the front view of the vehicle MVa to determine the merge lane. However, the present invention is not limited to this, and the adjacent lane is a merge lane by various methods. It can be determined whether or not. A description will be given with reference to FIG. 22, which is another example of the merge lane determination routine.

  In this example, it is assumed that a sign post (not shown) is installed at every predetermined interval of the road, and a road information signal indicating a road condition is transmitted from the sign post. Therefore, as shown in FIG. 22B, the vehicle MVa has the configuration shown in FIG. 3 except that the vehicle MVa includes an antenna 21 for receiving the electromagnetic wave and a receiver 22 for receiving the electromagnetic wave via the antenna 21. is there.

  In step S4511 of FIG. 22A, the road information signal from the sign post is received by the antenna attached to the vehicle MVa, and the received signal is taken into the control device 5. In step S4512, it is determined whether there is a signal indicating a merged lane in the received road information signal.

  FIG. 23A is a flowchart showing still another example of the merging lane determination routine. In this example, the vehicle MVa is equipped with a GPS (Global Positioning System) navigation device, and this navigation device includes an antenna 31 for receiving GPS signals from a satellite, as shown in FIG. A receiver 32 that receives a GPS signal via an antenna 31, a storage device 33 that stores road map data whose position is specified by latitude and longitude, a display device 34, and a GPS signal received by the receiver 32; Based on the road map data stored in the storage device 33, the position of the own vehicle on the road map is specified, road information representing the state of the road is extracted, and the map and the own vehicle position are displayed on the display device 34. And a control circuit 35.

  In step S4521 of FIG. 23A, the navigation device recognizes the current position of the vehicle MVa from the GPS signal from the satellite, and the current vehicle MVa travels from the position and road map data provided in the navigation device. Find the location on the map. Then, road information given to the road on which it is traveling is extracted and sent to the control device 5. In step S4522, the control device 5 determines whether the road information includes information indicating a merged road.

  Therefore, also by these modified examples, the same effects as those of the second embodiment described above can be obtained.

-Third Example-
The third embodiment is another example in which the pattern drawn in the lane where the vehicle ahead is present is used as the travel guidance information. As shown in FIG. 24, the road shoulder is in front of the vehicle MVa traveling in the lane R _0. It is assumed that there is a forward vehicle MVb traveling in a certain lane R- ₁ . Therefore, what kind of lane the front vehicle MVb is traveling on is, for example, a traveling lane such as the lanes R ₀ and R ₊₁ , or a shoulder such as the lane R _−1. Is recognized by the shape of the pattern drawn on the road surface. Then, using the position A, the speed vector Va of the vehicle MVa, the position B, the speed vector Vb of the forward vehicle MVb, whether the forward vehicle MVb changes the lane to the lane in which the vehicle MVa travels, and the lane After the change, it is determined whether or not the vehicle MVa approaches a safe inter-vehicle distance or less.

  The third embodiment is different from the first embodiment in that the data extraction executed in step S40 in FIG. 4, the behavior prediction of the preceding vehicle in step S50, the IF check of the rule in step S60, the rule in step S70 This is a specific procedure for contention and the degree of approach in step S80, and the rest is the same including the configuration of FIG. Therefore, in the following description, it demonstrates centering on difference with 1st Example, and abbreviate | omits description about a common part.

FIG. 25 shows a data extraction subroutine in the third embodiment, which corresponds to FIG. 11 for explaining details of step S40 in FIG. After processing similar to steps S41 to S44 in FIG. 11 is performed, in step S45C, it is determined whether or not the lane R- ₁ in which the forward vehicle MVb is traveling is a road shoulder. Details of the road shoulder determination method are shown in the flowchart of FIG.

In step S4531, the roadside reflector RF (see FIG. 24) is detected from the viewpoint-converted image, and the distance between the roadside reflector RF and the white line WLa adjacent to the right is detected. In step S4532, the distance between the roadside reflector detected in step S4531 and the white line is the width of a normal traveling lane (for example, the width of the lane R ₀ or R ₊₁ , and a known value is input in advance). If the determination is affirmative, the process proceeds to step S4533, and it is determined that the lane R- ₁ having a lane width sufficiently smaller than the travel lane width is a road shoulder. This fact
R _-1 = RC (13)
It expresses. On the other hand, if the determination is negative, the process proceeds to step S4534, where it is determined that the adjacent lane is not a shoulder.

ここに、ｋ₃：定数（ｋ₃＞０でありかつ０に近い値）
ｋ₃は符号判定のための値であり、測定誤差等を考慮して０に近い正の定数に設定してある。 In step S50 of FIG. 4, the behavior of the preceding vehicle, that is, whether the preceding vehicle MVb changes the lane to the traveling lane of the vehicle MVa is predicted by the following equation.

Where k _{3 is a} constant (k ₃ > 0 and close to 0)
k ₃ is a value for sign determination, and is set to a positive constant close to 0 in consideration of measurement error and the like.

Therefore, the vehicle MVa is traveling on the lane R _0, in a state where the front vehicle MVb is traveling lane R _-1, if it is determined that the lane R _-1 is shoulder, the forward vehicle MVb lane R _-1 From this, it is predicted that the vehicle MVa is going to change to the lane R ₀ where the vehicle MVa is traveling, and the fact RC (−1,0) becomes true.

  In step S60 of FIG. 4, various data are input for the rule selected from the various rules in the rule memory 53a of the proximity determination unit 53, and the IF unit is verified. The related rule in this embodiment is the rule with rule number 6 shown in FIG.

In the example shown in FIG. 24, if the preceding vehicle MVb is moving on the road shoulder R ₋₁ (| Vb |> k ₃ : If it is stopped, it is a broken car and the possibility of changing lanes is low). Since it can be predicted that there is a high possibility of changing to the lane R ₀ of the vehicle MVa, the left side of the equation (14) is determined to be true. Further, when the distance y _b -ya between the forward vehicle MVb and the vehicle MVa along the Y-axis direction has _a positive value smaller than the safe inter-vehicle distance f (Va, Vb), the result of the lane change is that When the vehicle MVb enters the lane R ₀ , the distance between the vehicles may be narrow and the vehicle MVb may approach sufficiently. The IF part of the rule number 6 is determined to be true, the rule is established, and the approach degree = 1 is output. The

  Therefore, also according to the present embodiment, it is possible to obtain the same effect as the above-described embodiment.

-Fourth embodiment-
In the fourth embodiment, the vehicle receives road information indicating road conditions transmitted from sign posts installed at predetermined intervals on the road side of the road, and predicts the behavior of the preceding vehicle using the received road information. To do.

For example, as shown in FIG. 27, when the forward vehicle MVb exists in front of the vehicle MVa, the lane in which the forward vehicle MVb is traveling is any type of lane, for example, the lanes R ₀ and R ₊₁ . Whether the vehicle is a traveling lane such as the above or a departure road such as the lane R- ₁ is recognized from the road information from the sign post, and the recognized road information is determined based on the position A of the vehicle MVa and the speed vector Va. And the position B of the forward vehicle MVb and the speed vector Vb, whether the forward vehicle MVb changes the lane in which the vehicle MVa travels, and the safe inter-vehicle distance with respect to the vehicle MVa after the lane change Determine if you are approaching:

  The fourth embodiment is different from the first embodiment in that, in addition to the configuration shown in FIG. 2, both the antenna for receiving the electromagnetic wave from the sign post and the receiver (see FIG. 22B) are omitted. 4), data extraction executed in step S40 of FIG. 4, prediction of the behavior of the preceding vehicle in step S50, IF section verification of the rule in step S60, rule conflict in step S70, and approach in step S80 It is the point of the concrete procedure of degree, and others are the same. Therefore, in the following description, it demonstrates centering on difference with 1st Example, and abbreviate | omits description about a common part.

FIG. 28 shows a data processing routine showing the fourth embodiment, which corresponds to the flowchart of FIG. 11 of the first embodiment. After performing the same processing as steps S41 to S44 in FIG. 11, in step S45D, it is determined whether or not the lane R- ₁ on which the preceding vehicle MVb is traveling is a departure road. Details of the departure lane determination method are shown in the flowchart of FIG.

In step S4541 in FIG. 29, the road information signal from the sign post is received by the antenna attached to the vehicle MVa, and the received signal is taken into the control device 5 via the receiver. In step S4542, it is determined based on the received road information signal whether adjacent lane R- ₁ is a departure road. The fact that the adjacent lane R- ₁ is a departure road,
R ₋₁ = RX (15)
It expresses.

ここに、ｋ₆：定数（０＞ｋ₆＞−１であり、かつ、ｋ₆は０に近い値）
ｋ₇：定数（ｋ₇＞０であり、かつ、ｋ₇は０に近い値）
(16)式は上述の(６)式と類似しており、ただし、図２７に示すように前方車両ＭＶｂが左隣の車線（離脱路）Ｒ_-1を走行しているため、符号の判定が逆になっている点のみ異なる。 In step S50 of FIG. 4, the behavior of the preceding vehicle, that is, whether the preceding vehicle MVb changes the lane to the traveling lane of the vehicle MVa is predicted by the following equation.

Here, k ₆ : constant (0> k ₆ > −1, and k ₆ is a value close to 0)
k ₇ : constant (k ₇ > 0 and k ₇ is a value close to 0)
The expression (16) is similar to the above expression (6), except that the forward vehicle MVb is traveling on the left lane (leaving road) R ₋₁ as shown in FIG. The only difference is that is reversed.

Therefore, in a state where the vehicle MVa is traveling on the lane R _{0 and the} preceding vehicle MVb is traveling on the lane R ₋₁ , (Db × Vb) _z / | Vb | takes a negative value (<k ₆ ). If | (Db × Vb) _z | is determined to be larger than a predetermined threshold (> k ₇ ), the forward vehicle MVb is changed from the lane R _-1 to the lane R ₀ on which the vehicle MVa is traveling. Predicting what is going to happen, the fact RC (-1,0) becomes true.

  In step S60 of FIG. 4, various data are input for the rule selected from the various rules in the rule memory 53a of the proximity determination unit 53, and the IF unit is verified. The rule related in the present embodiment is the rule of rule number 8 shown in FIG.

In the example shown in FIG. 27, the forward vehicle MVb is going to change the lane to the travel lane R ₀ of the vehicle MVa, and the speed vector Vb of the forward vehicle MVb is oriented in the direction of the lane R ₀ . The left side is determined to be true. Furthermore, since the lane in which the forward vehicle MVb is traveling is a departure road, R ₋₁ = RX is established, and as a result, it is predicted that the forward vehicle will be forcibly changed to the lane R _0. The IF part of number 8 is determined to be true, the rule is established, and the degree of approach = 2 is output.

  Therefore, also according to the present embodiment, it is possible to obtain the same effect as the above-described embodiment.

  Alternatively, in the above-described fourth embodiment, when the vehicle MVa is equipped with a navigation device (see FIG. 23B), the road information from the navigation device is used as in the modification of the second embodiment. Also good. This will be described with reference to FIG.

  FIG. 30 is a flowchart illustrating another example of the departure path determination routine. First, in the same manner as described above in step S4521, the position of the vehicle on the map is recognized based on the GPS signal, and the road information given to the road is extracted and transmitted to the control device 5. In step S4522A, it is determined based on the received road information whether the adjacent lane is a departure road.

-Fifth embodiment-
In the fifth embodiment, when there are two forward vehicles, which are forward vehicles traveling in front of the vehicle, the behavior is predicted from the positions and speeds of these forward vehicles. For example, as shown in FIG. 31, when there are two forward vehicles MVb and MVc in front of the vehicle MVa, the forward vehicle MVb following the preceding forward vehicle MVc changes the lane to the travel lane of the vehicle MVa. Whether or not the vehicle MVc is overtaken and whether or not the vehicle MVa approaches a safe inter-vehicle distance after changing the lane, the position A of the vehicle MVa, the speed vector Va, and Judgment is made based on positions B and C (x _c , y _c , 0) and speed vectors Vb and Vc of the preceding vehicles MVb and MVc.

  The fifth embodiment is different from the first embodiment in that the data extraction executed in step S40 in FIG. 4, the behavior prediction of the forward vehicle in step S50, the IF check of the rule in step S60, and the rule in step S70 This is a specific procedure for contention and the degree of approach in step S80, and the rest is the same including the configuration of FIG. Therefore, in the following description, it demonstrates centering on difference with 1st Example, and abbreviate | omits description about a common part.

ここに、ｋ₄：定数（ｋ₄＞０）
ｋ₅：定数（ｋ₅＜０）
ｙ_a＜ｙ_b＜ｙ_cなる条件は、前から順に前方車両ＭＶｃ、前方車両ＭＶｂ、車両ＭＶａの順に並んでいることを示しており、図３１に示す配置状態に対応している。また、ｙ_c−ｙ_bは、Ｙ軸方向に沿った前方車両ＭＶｃと前方車両ＭＶｂとの間の距離であり、｜Ｖｃ｜−｜Ｖｂ｜は、前方車両ＭＶｃと前方車両ＭＶｂとの間の速度差である。 In step S40 of FIG. 4, necessary data regarding the two forward vehicles is extracted in the same manner as in the first embodiment in which there is one forward vehicle. In step S50 of FIG. 4, the behavior of the preceding vehicle, that is, whether the preceding vehicle MVb changes the lane to the traveling lane of the vehicle MVa is predicted by the following equation.

Where k _{4 is a} constant (k ₄ > 0)
k ₅ : constant (k ₅ <0)
The condition y _a <y _b <y _c indicates that the front vehicle MVc, the front vehicle MVb, and the vehicle MVa are arranged in order from the front, and corresponds to the arrangement state shown in FIG. Further, y _c −y _b is a distance between the forward vehicle MVc and the forward vehicle MVb along the Y-axis direction, and | Vc | − | Vb | is between the forward vehicle MVc and the forward vehicle MVb. It is a speed difference.

Therefore, in a state where the vehicle MVa travels on the lane R ₀ and both the front vehicle MVb and the front vehicle MVc travel on the lane R ₊₁ , the distance y _c −y _b between the front vehicles MVb and MVc is a predetermined value. When the inter-vehicle distance k ₄ or less and the speed difference | Vc | − | Vb | between the front vehicles MVb and MVc are a predetermined speed difference k ₅ or more, the front vehicle MVb gradually approaches and approaches the front vehicle MVc. Then, it is predicted that the lane will be changed to the lane R ₀ in order to overtake the preceding vehicle MVc, and the fact RC ₇ (+1,0) becomes true.

  In step S60 of FIG. 4, various data are input for the rule selected from the various rules in the rule memory 53a of the proximity determination unit 53, and the IF unit is verified. The related rule in this embodiment is the rule of rule number 7 shown in FIG.

In the example shown in FIG. 31, since the vehicle MVa is traveling in the lane R ₀ and the preceding vehicles MVb and MVc are both traveling in the lane R ₊₁ , the distance y _c −y _b between the preceding vehicles MVb and MVc is predetermined. If the inter-vehicle distance k ₄ or less and the speed difference | Vc | − | Vb | between the preceding vehicles MVb and MVc is greater than or equal to a predetermined speed difference k ₅ , the left side of Expression (17) is determined to be true. Further, when the distance y _b -ya between the forward vehicle MVb and the vehicle MVa along the Y-axis direction has _a positive value smaller than the safe inter-vehicle distance f (Va, Vb), the result of the lane change is that When the vehicle MVb enters the lane R ₀ , there is a possibility that the distance between the vehicles MVb is narrow and the vehicle MVb may approach sufficiently. The IF part of the rule number 7 is determined to be true, the rule is established, and the approach degree = 2 is output. The

  Therefore, also according to the present embodiment, the same operational effects as those of the above-described embodiments can be obtained.

-Sixth Example-
In the first to fifth embodiments described above, the approach to the vehicle is detected. However, the present invention is not limited to this. For example, the approach to a pedestrian crossing a pedestrian crossing may be detected. The sixth embodiment is an example in which the shape of the pattern drawn on the road surface is recognized to determine the presence or absence of a pedestrian crossing. For example, as shown in FIG. 32, when a pedestrian WP is standing on the side of a pedestrian crossing CR in the forward view of the vehicle MVa, the pattern shape recognizes whether the pedestrian crossing CR exists in the forward view of the vehicle MVa. The pedestrian WP crosses the pedestrian crossing CR using the recognition result together with the position B of the pedestrian WP, the position A of the vehicle MVa, and the velocity vector Va, and the pedestrian WP crosses the pedestrian crossing CR. In this case, it is determined whether the vehicle MVa can be stopped safely. As shown in FIG. 32, the pedestrian crossing CR has two white lines WLd and WLe in the horizontal direction extending over the entire width of the lane width, and therefore, two white lines WLd and WLe extending in the horizontal direction over the entire width of the lane width. Can be recognized as a pedestrian crossing CR.

  The sixth embodiment is different from the first embodiment as follows. First, FIG. 33 shows the overall configuration of the approach prediction apparatus in the sixth embodiment. The difference from the approach prediction apparatus of FIG. 3A of the first embodiment is that an infrared detection apparatus 10 is provided. The infrared detection device 10 is constituted by an infrared camera, for example, and detects infrared rays coming from the front of the vehicle MVa as an image. The pre-processing unit 7 analyzes the shape of the infrared ray detected by the infrared detection device 10 if there is a region higher than the ambient temperature to determine whether or not the human is imaged. If so, the direction of the person relative to the vehicle MVa is detected.

  In addition, the image processing in step S10 in FIG. 4, the data extraction in step S40, the behavior prediction of the preceding vehicle in step S50, the IF part matching of the rule in step S60, the rule conflict in step S70, and the degree of approach in step S80 It is the same except that the general procedure is different. Therefore, in the following description, it demonstrates centering on difference with 1st Example, and abbreviate | omits description about a common part.

  FIG. 34 is a flowchart showing an image processing routine according to the sixth embodiment, and is for explaining details of step S10 in FIG. After performing the same processing as steps S11 to S13 in FIG. 5, in step S14, it is determined whether or not a person is present in front of the vehicle MVa. If a person is present, the relative coordinates with respect to the vehicle MVa are determined. Ask.

  That is, an infrared ray coming from the front of the vehicle MVa is imaged by the infrared detection device 10 described above, and a region having a temperature higher than the ambient temperature is extracted. Next, the shape of the high temperature region is analyzed to determine whether the high temperature region is caused by infrared rays radiated from humans, that is, whether or not a human is present in the high temperature region. And if it is judged that a person exists in a front view, the direction of a person to vehicles MVa will be detected.

Next, FIG. 35 is a flowchart showing the data extraction routine of the fourth embodiment, and is for explaining details of step S40 of FIG. Calculating the absolute coordinate values of the vehicle as in step S41 in FIG. 11, calculates absolute coordinates and human absolute coordinate value B of the white line area (x _b, y _b) to a step S 42 A, the vehicle at step S43A The velocity vector is calculated. In this embodiment, the human velocity vector is not obtained. In step S46, it is determined whether or not there is a pedestrian crossing CR in the forward field of view of the vehicle MVa. Details of the pedestrian crossing determination method are shown in the flowchart of FIG.

In FIG. 36, in step S461, white line detection is performed on the forward view image of the vehicle MVa whose viewpoint has been changed, and edge detection is performed in the lane R ₀ in which the vehicle MVa is traveling among the lanes separated by the detected white line. Then, a horizontal edge extending in a direction crossing the white line is extracted from the detected edges. Details of the horizontal edge extraction have already been performed in the description of the operation in step S4501 in FIG.

In step S462, similarly to step S4502 in FIG. 19, a labeling process is performed on each of the horizontal edges extracted in step S461, and the horizontal lengths of the labeled horizontal edges are measured. In step S4503, the length of the transverse direction of the transverse edge substantially equal to the width of the travel lane R _0, and whether the lateral edges of two to the traffic lane R ₀ is present is determined, the determination is affirmative In step S464, it is determined that there is a pedestrian crossing in the travel lane _R0 . This fact
Kb = Human. . . (18)
It expresses. On the other hand, if the determination is negative, the process proceeds to step S465, and it is determined that there is no pedestrian crossing in the travel lane _R0 .

ここに、ｋ₁₀：定数（ｋ₁₀＞０）
ＲＥ：歩行者ＷＰが横断歩道を横断しようとしている事実
ｙ_p ：横断歩道ＣＲの中央点Ｐ（ｘ_p，ｙ_p）のＹ座標 Next, in step S50 of FIG. 4, the behavior of the front obstacle, that is, whether or not the pedestrian WP is going to cross the pedestrian crossing CR is predicted by the following equation.

Where k _{10 is a} constant (k ₁₀ > 0)
RE: The fact that pedestrian WP is about to cross a pedestrian crossing
y _p : Y coordinate of the center point P (x _p , y _p ) of the pedestrian crossing CR

| Y _p −y _b | is a distance between the pedestrian WP and the pedestrian crossing CR along the Y-axis direction. Therefore, if it is determined that the pedestrian WP is present in the forward field of view of the vehicle MVa and the distance between the pedestrian WP and the pedestrian crossing CR is smaller than a predetermined safety distance (<k ₁₀ ), the pedestrian WP Predicts that he is going to cross the pedestrian crossing CR, and the fact RE becomes true.

ここに、α：車両ＭＶａの減速度 In step S60 of FIG. 4, various data are input for the rule selected from the various rules in the rule memory 53a of the proximity determination unit 53, and the IF unit is verified. The related rule in this embodiment is the rule of rule number 11 shown in FIG. Here, f ′ (Va) is a function indicating the stop distance of the vehicle MVa, and is defined by the following equation.

Where α: deceleration of vehicle MVa

In the example shown in FIG. 32, since the pedestrian WP exists in the forward field of view of the vehicle MVa, the distance y _p -y _b between the pedestrian WP and the pedestrian crossing CR is smaller than a predetermined safety distance ( <K ₁₀ ), that is, when the pedestrian WP is sufficiently close to the pedestrian crossing CR, the left side of the equation (19) is determined to be true. Furthermore, when the distance y _b -ya between the pedestrian WP and the vehicle MVa along the Y-axis direction has _a positive value smaller than the stop distance f ′ (Va), the braking distance is even when the brake is applied. When the pedestrian WP crosses the pedestrian crossing CR for a long time, the vehicle MVa may be approaching the pedestrian crossing CR, and the IF part of the rule number 11 is determined to be true, the rule is established, and the approach degree = 2 is set. Is output.

  Therefore, according to the present embodiment, the presence of a human being is detected by the infrared detection device, and the degree of approach is calculated by predicting the human behavior in consideration of the shape of the white line. It is possible to accurately judge and predict the degree of approach with humans.

-Modification of the sixth embodiment-
In the sixth embodiment described above, pattern recognition is performed on the screen that captures the forward field of view of the vehicle MVa to determine the pedestrian crossing CR. However, as in the case of the merging lane determination described above, various methods are used. Thus, it can be determined whether or not there is a pedestrian crossing CR in the front view.

  FIG. 37 is a flowchart illustrating another example of the pedestrian crossing determination routine. In this example, it is assumed that a sign post (not shown) is installed at every predetermined interval of the road, and a road information signal indicating a road condition is transmitted from the sign post. Therefore, the vehicle MVa includes an antenna and a receiver (see FIG. 22B) for receiving this electromagnetic wave. The other structure is shown in FIG.

  In step S4601 of FIG. 37, the road information signal from the sign post is received by the antenna attached to the vehicle MVa, and the received signal is taken into the control device 5. In step S4602, it is determined based on the received road information signal whether a crosswalk CR exists in the forward field of view.

  FIG. 38 is a flowchart illustrating an example in which a pedestrian crossing is determined based on road information input from the navigation device described with reference to FIG. In step S4611, the navigation apparatus recognizes the position of the vehicle on the map based on the GPS signal in the same manner as described above, extracts road information given to the road, and transmits the road information to the control apparatus 5. In step S4612, it is determined whether or not there is a pedestrian crossing in the forward view based on the received road information.

  Therefore, also by these modified examples, it is possible to obtain the same operational effects as those of the sixth embodiment described above.

Next, seventh and eighth embodiments in which the approach degree output from the approach degree determination unit 53 is used for controlling the vehicle MVa will be described.
-Seventh Example-
FIG. 39 is a block diagram showing a schematic configuration of a seventh embodiment in which the approach degree output from the above approach predicting apparatus is used for controlling the so-called constant speed traveling device. In this figure, 50 is an approach prediction apparatus according to the first to sixth embodiments described above. Reference numeral 61 denotes a desired vehicle speed setting unit, in which a vehicle occupant inputs a desired vehicle speed. The desired vehicle speed setting unit 61 includes, for example, a push button for instructing UP / DOWN. When the UP button is continuously pressed, a signal for increasing the set vehicle speed is transmitted, while the DOWN button is continuously pressed. When it is moved, a signal for lowering the set vehicle speed is transmitted. Reference numeral 62 denotes an operation ON / OFF switch which instructs the start and end of a constant speed running operation by the control unit 63.

Reference numeral 64 denotes an actuator group that controls each part of the vehicle by a signal from the control unit 63 and performs constant speed running. The actuator group 64 includes a throttle actuator that controls the opening of the throttle, a brake actuator that controls the brake device, and a transmission actuator that controls the transmission. The control unit 63 determines a target speed based on signals input from the desired vehicle speed setting unit 61 and the approach prediction device 50, compares the target speed with the current vehicle speed, and the current speed is the target speed. A control signal is output to the actuator group 64 so that Further, the actuator group 64 is controlled based on a signal from the approach prediction device 50 to perform traveling control according to the degree of approach. For constant speed control and traveling control according to the degree of approach,
(1) Drives only the throttle actuator
(2) Drives the throttle actuator and variable speed actuator
(3) The throttle actuator and brake actuator can be driven.

  FIG. 40 is a flowchart for explaining the operation of this embodiment. In addition, since the operation | movement of the approach prediction apparatus 50 is the same as that of the above-mentioned 1st-6th Example, the description is abbreviate | omitted.

  The program shown in FIG. 40 starts when the operation ON / OFF switch 62 is turned on. First, in step S91, the desired vehicle speed set by the desired vehicle speed setting unit 61 is taken into the control unit 63. Next, in step S92, the absolute coordinate values, speed vectors, and travel lanes of the vehicle MVa and the forward vehicle MVb are read from the approach prediction device 50.

  In step S93, a safe inter-vehicle distance that can ensure safety is calculated based on the absolute coordinate values and speed vectors of the vehicle MVa and the forward vehicle MVb input from the proximity determination unit 53. The safe inter-vehicle distance is, for example, as defined by the above formula (7). In step S94, an inter-vehicle distance between the vehicle MVa and the preceding vehicle MVb is calculated, and the actual inter-vehicle distance is equal to or greater than the safe inter-vehicle distance calculated in step S93, or on the travel lane on which the vehicle MVa travels. It is determined whether there is no vehicle ahead. As a result, if the determination is positive, the process proceeds to step S95, and if the determination is negative, the process proceeds to step S96.

  The affirmative determination is made in step S94 because it is determined that safety can be ensured even if the vehicle speed of the vehicle MVa is set to the desired vehicle speed. In step S95, the actuator group 64 is set so that the actual vehicle speed becomes the desired vehicle speed. Control. On the other hand, in step S96, it is determined that there is a possibility that safety cannot be secured if the vehicle speed of the vehicle MVa is set to the desired vehicle speed, and the actuator group 64 is controlled so that the actual inter-vehicle distance becomes the safe inter-vehicle distance.

  On the other hand, when the approach degree is output from the approach prediction device 50 while the program shown in the flowchart of FIG. 40 is being executed, an interrupting operation is executed and the constant speed running operation is ended as shown in FIG.

  In this way, the behavior of the vehicle MVa is controlled so that the vehicle MVa does not approach the front vehicle MVb using the position, speed, lane, and the degree of proximity to the front vehicle MVb detected by the approach prediction device 50. can do.

-Modification of the seventh embodiment-
In the seventh embodiment described above, when the approach degree is output from the approach predicting device 50, the constant speed traveling operation is uniformly terminated. However, it may be controlled according to the level of the approach degree. FIG. 42 is a flowchart showing an example of such a control procedure. When the approach degree is output from the approach prediction device 50 while the program shown in the flowchart of FIG. 40 is being executed, the subroutine flowchart of FIG. The interrupt operation shown in FIG. In step S901, the target speed targeted by the constant speed traveling operation is decelerated in accordance with the level of the degree of approach output from the approach prediction device 50. In particular, the higher the approach level, the lower the target speed may be set. For example, if the approach degree = 1, the throttle actuator is simply returned and the engine brake is decelerated. The actuator may be controlled to decelerate.

  By controlling the constant speed travel operation stepwise according to the level of approach, fine control according to the travel situation becomes possible.

-Eighth embodiment-
FIG. 43 is a block diagram showing a schematic configuration of an approach avoiding device for avoiding the host vehicle from approaching a preceding vehicle based on a signal output from the approach predicting device according to an eighth embodiment of the present invention. In the following description, the same components as those in the seventh embodiment are denoted by the same reference numerals, and the description thereof is omitted. In FIG. 43, reference numeral 65 denotes a rear and side vehicle detection unit that detects the presence of vehicles behind and behind the vehicle MVa. The rear and side vehicle detection units include, for example, a distance sensor, an optical switch, and the like, and detect the presence of the vehicle based on detection results of the distance sensor, the optical switch, and the like. The actuator group 64 includes a steering actuator, a throttle actuator, a brake actuator, and a speed change actuator.

  The normal operation of the present embodiment is substantially the same as the operation of the seventh embodiment described with reference to FIG. 40, so the description thereof is omitted, and only the interrupt operation is illustrated in the flowchart. When the approach degree is output from the approach prediction device 50 while the program shown in the flowchart of FIG. 40 is being executed, a subroutine flowchart of the interrupt operation shown in FIG. 44 is executed.

  In step S911 in FIG. 44, the lane of the preceding vehicle that causes the approach degree and the approach degree output is captured from the approach prediction device 50. The lane of the vehicle ahead can be detected by searching for the IF part of the rule that has contributed to the proximity output. In step S912, it is determined whether or not approach to the preceding vehicle can be prevented by simply decelerating the vehicle MVa. If the determination is affirmative, the process proceeds to step S913, and if the determination is negative, the process proceeds to step S914. The determination in step S912 may be made by affirming the determination when the degree of approach is high (the degree of approach is 2 or more, for example), and denied when the degree of approach is low.

  In step S913, similarly to step S901 in FIG. 42, the target speed targeted for the constant speed traveling operation is decelerated in accordance with the level of approach output from the approach prediction device 50. In particular, the higher the proximity level, the lower the target speed may be set. On the other hand, in step S914, a search is made for a lane in the opposite direction to the lane in which the preceding vehicle that has caused the proximity output is present (that is, the next lane on the left if the preceding vehicle is in the right lane). If no vehicle is detected behind and to the side of the vehicle MVa in this lane, it is determined that the lane is an avoidable lane, and the lane is changed to an avoidable lane. That is, the vehicle MVa is moved to this lane toward the lane where the steering of the vehicle MVa can be avoided by the steering actuator constituting the actuator group 64.

  Therefore, also according to the present embodiment, it is possible to obtain the same operational effects as those of the seventh embodiment described above.

  In the correspondence between the embodiment described above and the claims, the CCD camera 3 and the data extraction unit 51 use the travel guidance information detection unit 201, the wheel speed sensor 1 and the data extraction unit 51 use the vehicle behavior detection unit 202, and infrared rays. The detection device 10 and the data extraction unit 51 constitute a pedestrian crossing detection means 201, and the approach degree determination part 53 constitutes an approach degree prediction means 204. The details of the approach prediction device of the present invention are not limited to the above-described embodiments, and various modifications can be made.

It is a functional block diagram which shows the approach prediction apparatus of this invention. It is a figure explaining operation | movement of the approach prediction apparatus of 1st Example. (A) is a whole block diagram which shows one Example of an approach prediction apparatus, (b) is a figure which shows the detail of the control apparatus. It is a main flowchart for demonstrating operation | movement of an approach prediction apparatus. It is a flowchart which shows the image processing routine of 1st Example-5th Example. (a) is a figure which shows an example of the front view image of the vehicle imaged with the imaging device, (b) is a figure which shows the viewpoint conversion image corresponding to (a). It is a flowchart which shows a white line detection routine. It is a figure which shows an example of the filter for white line detection. (a)-(f) is a figure for demonstrating the connection method of a white line candidate area | region, respectively. (a) is a figure which shows an example of a white line candidate area | region, (b) is a template obtained from the white line candidate area | region of (a), (c) is a figure which shows the white line area | region detected using the template of (b). It is. It is a flowchart which shows the data extraction routine of 1st Example. It is a flowchart which shows the direction vector detection routine of a white line. (a)-(c) is a figure for demonstrating the calculation method of the unit direction vector of a white line, respectively. It is a figure which shows an example of the rule memorize | stored in the rule memory. It is a figure explaining operation | movement of the approach prediction apparatus of 1st Example. It is a flowchart which shows the data extraction routine of the modification of 1st Example. It is a figure explaining operation | movement of the approach prediction apparatus of 2nd Example. It is a flowchart which shows the data extraction routine of 2nd Example. It is a flowchart which shows the merge lane determination routine of 2nd Example. FIG. 6 is a diagram illustrating a second embodiment, where (a) is a diagram illustrating an example of a front view image of a vehicle imaged by an imaging device, (b) is a diagram illustrating a viewpoint conversion image corresponding to (a), () Is a diagram showing a white line area detected for the viewpoint converted image of (b), and (d) is a diagram showing horizontal edges extracted from the viewpoint converted image of (b). It is a figure explaining operation | movement of the approach prediction apparatus of 2nd Example. (A) is a flowchart which shows an example of the merge lane determination routine of the modification of 2nd Example, (b) is a figure which shows the structural example of the system which receives the signal from a sign post. (A) is a flowchart which shows the other example of the merge lane determination routine of the modification of 2nd Example, (b) is a figure which shows the structural example of a navigation apparatus. It is a figure explaining operation | movement of the approach prediction apparatus of 3rd Example. It is a flowchart which shows the data extraction routine of 3rd Example. It is a flowchart which shows the road shoulder determination routine of 3rd Example. It is a figure explaining operation | movement of the approach prediction apparatus of 4th Example. It is a flowchart which shows the data extraction routine of 4th Example. It is a flowchart which shows the departure lane determination routine of 4th Example. It is a flowchart which shows the departure lane determination routine of the modification of 4th Example. It is a figure explaining operation | movement of the approach prediction apparatus of 5th Example. It is a figure explaining operation | movement of the approach prediction apparatus of 6th Example. It is a figure which shows the whole structure of the approach prediction apparatus in 6th Example. It is a flowchart which shows the image processing routine of 6th Example. It is a flowchart which shows the data extraction routine of 6th Example. It is a flowchart which shows the pedestrian crossing determination routine of 6th Example. It is a flowchart which shows an example of the pedestrian crossing determination routine of the modification of 6th Example. It is a flowchart which shows the other example of the pedestrian crossing determination routine of the modification of 6th Example. It is a block diagram which shows the constant speed travel apparatus with which the approach prediction apparatus which is 7th Example is applied. It is a flowchart for demonstrating operation | movement of 7th Example. It is a flowchart for demonstrating the interruption operation | movement of 7th Example. It is a flowchart for demonstrating the interruption operation | movement of the modification of 7th Example. It is a block diagram which shows the approach avoidance apparatus with which the approach prediction apparatus which is 8th Example is applied. It is a flowchart for demonstrating the operation | movement of 8th Example.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Wheel speed sensor 2 Tire 3 CCD camera 4,7 Pre-processing part 5 Control apparatus place 6 Radar apparatus 10 Infrared detector 50 Approach prediction apparatus 51 Data extraction part 52 Prediction behavior part 53 Approach degree judgment part 53a Rule memory 53b IF section collation section 53c Rule selection section 53d Rule competition section MVa Vehicle MVb, MVc Front vehicle WLa-WLc White line WLRa-WLRc White line area MP Merge lane pattern RF Roadside reflector CR Crosswalk 201 Crosswalk detection means 202 Self-vehicle behavior detection means 203 Front vehicle behavior detection means 204 Approach degree prediction means

Claims (5)

An apparatus for predicting the degree of proximity of a moving object such as a human to a host vehicle traveling on a road surface,
A pedestrian crossing detection means for detecting a pedestrian crossing;
Own vehicle behavior detecting means for detecting the behavior of the own vehicle;
Object behavior detecting means for detecting the behavior of the moving object present in the traveling direction of the host vehicle;
An approach degree predicting means for predicting an approach degree of the moving object to the own vehicle based on the detected pedestrian crossing, the behavior of the moving object and the behavior of the own vehicle ;
The own vehicle behavior detecting means calculates the position and speed vector of the own vehicle,
The object behavior detection means calculates the position of the moving object, and determines that the moving object crosses the pedestrian crossing when a distance between the moving object and the pedestrian crossing is a predetermined value or less,
The approach degree predicting means outputs an approach degree based on the position of the moving object and the own vehicle and the speed vector of the own vehicle when it is determined that the moving object crosses the pedestrian crossing. An approach prediction device. In the approach prediction apparatus according to claim 1,
The approach prediction apparatus, wherein the pedestrian crossing detection means detects a pedestrian crossing from the shape of a white line drawn on the road surface. In the approach prediction apparatus according to claim 1,
The approach prediction device, wherein the pedestrian crossing detecting means includes receiving means for receiving road information transmitted from a sign post installed on a road side, and detects the pedestrian crossing from the received road information. In the approach prediction apparatus according to claim 1,
The pedestrian crossing detection means includes a reception means for receiving a GPS signal from a satellite and a storage means for storing road map data, and detects the pedestrian crossing from the received GPS signal and the road map data. An approach prediction device characterized by: In the approach prediction apparatus according to any one of claims 1 to 4,
The approach predicting apparatus according to claim 1, wherein the object behavior detecting means includes infrared detecting means for detecting infrared rays emitted from the moving object.

Images