JP3065821B2 - Object detection device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an object detecting device, and more particularly, to an object detecting device for detecting a moving object such as a running vehicle.

[0002]

2. Description of the Related Art Conventionally, there has been known an apparatus which recognizes a preceding vehicle using radars and travels following the preceding vehicle. In recognition of the preceding vehicle by these devices, there are problems such as a temporary detection error of the preceding vehicle due to uncertainty of the sensor, and a false recognition of the preceding vehicle due to mixing of noise.

[0003] JP-A-2-36500, JP-A-1-36500
No. 213593, JP-A-61-20877,
JP-A-61-145474, JP-A-61-278
Japanese Patent Application Laid-Open No. 775, Japanese Patent Application Laid-Open No. 62-36581 and Japanese Patent Application Laid-Open No. 2-40800 disclose techniques for recognizing a preceding vehicle. However, in these techniques, measures are taken to prevent the preceding vehicle from being lost. Absent.

In Japanese Patent Application Laid-Open No. 1-114550, when a preceding vehicle is lost at an entrance or exit of a curve or the like, a countermeasure is taken to maintain the control vehicle speed of the own vehicle for a fixed time w1. As a similar method, a method of holding the preceding vehicle data for a certain time or a method of holding the inter-vehicle distance for a certain time can be considered, but has the following problems.

[0005] In order to cope with a long loss of the preceding vehicle (about several seconds), if the holding time is set to be longer, control is delayed by the holding time when the preceding vehicle is erroneously recognized due to noise or the like. Causes of the noise include a case where an obstacle other than the preceding vehicle is erroneously recognized as the preceding vehicle, noise mixing into the sensor system, and the like.

[0006] Conversely, if the holding time is shortened,
Can't cope with a long loss. Causes of long sight loss include driving in and out of curves, up and down hills, bad environment such as rainy weather, radar malfunction, and the like.

Japanese Patent Application Laid-Open No. 61-272673 discloses a technique for increasing the S / N ratio of a radar signal by accumulating a plurality of observation results in a memory. However, depending on the mechanism of the radar device, the preceding vehicle may temporarily go out of the radar measurement range when the nose of the vehicle bounces or crosses an inflection point on a slope. In this case, since the radar signal itself cannot be obtained, even if the S / N ratio is increased by this method, the loss cannot be solved.

[0008]

SUMMARY OF THE INVENTION The present invention has been made in order to solve the above-mentioned problems, and is intended to solve the problem when a moving object such as a preceding vehicle or a vehicle running on another lane is lost. The holding time is long, and when noise or the like is mistaken for the preceding vehicle, the influence of the misidentification is reduced, and the detected object is
It is an object of the present invention to provide an object detection device capable of coping with a loss when the light beam temporarily deviates from a measurement area such as a laser.

[0009]

To achieve the above object, the present invention provides an object detecting means for detecting a current position of an object, and a storing means for storing a past position of the detected object detected by the object detecting means. And whether the detected object whose current position is detected based on the current position of the detected object and the position of the storage object stored in the storage means corresponds to the storage object whose position is stored in the storage means Determining means for determining whether the detected object and the stored object correspond to each other.
If it is determined that the
Detected object and the memory object do not correspond
Calculating means for calculating a certainty factor that is smaller than the previous value, and changing the storage content of the storage object having the certainty factor equal to or greater than the predetermined value based on the detected current position and storing the memory object having the certainty factor less than the predetermined value. And a storage content changing means for erasing the storage content of the detected object and storing the position with respect to the detected object determined not to correspond to the storage object by the determination means.

[0010]

The object detecting means of the present invention detects the current position of a moving object such as a traveling vehicle or a stopped object such as a stopped vehicle, and the storage means stores the detected object detected by the object detecting means. Of the past is stored. The determination means determines whether the detected object whose current position is detected based on the current position of the object and the position of the storage object stored in the storage means corresponds to the storage object whose position is stored in the storage means Determine whether or not. For example, if the detected object whose current position is detected exists in an area within a predetermined range including the stored past position, the detected object whose current position is detected and the position are stored in the storage unit. It can be determined that the storage object corresponds. The calculating means calculates a certainty factor indicating the degree of existence of the storage object based on the determination result of the determining means. At this time, calculation is performed so that the certainty factor is increased for the storage object determined to correspond to the detected object by the determination means, and the reliability is determined for the storage object determined to not correspond to the detected object by the determination means. The calculation is performed so that the degree becomes small.

The storage content changing means changes the storage content of the storage object having the certainty factor equal to or higher than the predetermined value based on the detected current position. As a result, the stored content of the storage means is changed to the latest content for a stored object having a certainty factor equal to or higher than a predetermined value, that is, an object that has not been lost or has been lost within a predetermined time, so that a short lost Can be dealt with. Further, the storage content changing means deletes the storage content of the storage object having the certainty factor less than the predetermined value. As a result, for an object whose certainty factor is less than the predetermined value, that is, for an object whose misplacement exceeds a predetermined time, the storage content of the storage unit is deleted. Further, the storage content changing means includes:
The position with respect to the detected object, which is detected by the detection means but is determined not to correspond to the storage object of the determination means, is stored in the storage means. This makes it possible to cope with a case where a lost object is detected again after a long loss.

[0012]

Embodiments of the present invention will be described below. First, this embodiment is represented by functional blocks as shown in FIG. As shown in the figure, the present embodiment includes an object detection unit 10 for detecting the current position of a moving object, and a storage unit 12 for storing a past position of the detected object detected by the object detection unit 10. Have. The object detection means 10 and the storage means 12
The detected object whose current position is detected based on the current position of the detected object and the past position of the storage object stored in the storage unit 12 corresponds to the storage object whose position is stored in the storage unit 12. It is connected to the judging means 14 for judging whether or not there is. The judging means 14 is connected to a calculating means 16 for calculating a certainty factor indicating the degree of existence of the storage object based on the judgment result of the judging means 14. The calculating means 16 changes the storage content of the storage object having the certainty factor equal to or more than the predetermined value based on the detected current position,
It is connected to storage content changing means 18 for erasing the storage content of the storage object having a certainty factor less than a predetermined value and storing the position of the detected object determined by the determination means 14 not to correspond to the storage object.

The object position detecting means 10 is a sensor capable of detecting the position of a moving object, and is capable of distinguishing a plurality of objects, for example, two-dimensional coordinates such as a scanning laser and an image type. A sensor or a sensor that can obtain one-dimensional coordinates using a microwave or the like can be used. Alternatively, sensors that can obtain relative speed at the same time as the position may be used. By selecting these sensors, an effect corresponding to the performance of the selected sensor can be expected. Hereinafter, a description will be given using a sensor that can obtain two-dimensional coordinates.

The object information storage means 12 uses, for example, a RAM, and detects the position of the detected object (preceding vehicle) (two-dimensional information corresponding to the sensor), certainty factor, and other information (for example, speed).
Are collectively stored for each detected object as a storage object, for example.

As a storage method, information may be collectively stored for each storage object as in Example 1 below, or each information may be stored in an array form as in Example 2. Since the number of objects to be detected, that is, the number of preceding vehicles, is about five at the time of general follow-up traveling, it is sufficient to secure an area capable of storing about 10 to 20 vehicles with a margin.

Example 1 of storage method Storage object 1: x coordinate at previous measurement y coordinate at previous measurement Confidence at previous measurement x speed at previous measurement y direction speed at previous measurement Storage object 2: previous measurement X-coordinate at the time y-coordinate at the previous measurement y-coordinate at the previous measurement x-speed at the previous measurement y-direction speed at the last measurement ・ ・ ・ Example 2 of storage method x-coordinate of storage object 1 x-coordinate of storage object 2 The y-coordinate of the storage object 1 The y-coordinate of the storage object 2 The certainty of the storage object 1 The certainty of the storage object 2 The determination unit 14 stores the detected object (for example, a car) and information as follows. Memory objects (for example,
) Is determined. First, it is assumed that the positions of the storage objects 1 to 3 as shown in FIG. The fan shape 20 is the measurement range of the sensor, and the vehicle is located at the left end, which is the corner of the fan shape. Then, it is assumed that the detected objects 1 to 3 are detected at the positions shown in FIG. 3 in the next measurement. Detected objects 1-3 and memory objects 1-3
When it is determined whether or not they correspond to each other, ranges 1 to 3 to be associated with each storage object are determined as shown in FIG.
In these ranges 1 to 3, the detected object closest to the storage object is associated with the storage object existing in this range. In this example, since the storage object 1 and the detection object 1 exist in the range 1, the storage object 1 and the detection object 1 correspond to each other, and the storage object 2 and the detection object 2 exist in the range 2. Therefore, the storage object 2 and the detection object 2 correspond to each other, and no detection object exists in the range 3 so that no detection object corresponding to the storage object 3 exists. That is, the storage object corresponding to the detection object 3 exists. Will not do. It should be noted that a storage area is provided in the storage means for a detected object having no corresponding storage object.

The range to be associated is determined in consideration of the distance that the object may relatively move within one measurement cycle, the magnitude of an error included in the obtained measurement result, the certainty of existence, and the like. To decide. Of course, other factors can be considered, and the range to be associated can be arbitrarily determined. The shape of the range to be associated may be selected according to the method of actually performing the association.

When the relative speed is calculated by the object detecting means 10 , the position obtained by the current measurement is predicted,
By obtaining the difference between the predicted position and the detected position and correcting the relative speed according to the difference, it is possible to increase the reliability. Furthermore, if a sensor capable of obtaining a relative speed is used, it is possible to determine whether or not the detected object corresponds to the stored object using the relative speed in addition to the position. Further, the position obtained by the current measurement can be predicted from the relative speed and the calculated relative acceleration.

Calculation means 1 for calculating the certainty of the existence of an object
In step 6, based on the correspondence result determined by the determination means 14, a certainty factor indicating the degree of existence of the storage object is calculated. Specifically, the degree of certainty is increased or decreased as follows.

That is, for the stored object corresponding to the detected object, the value of the certainty factor in the stored information is increased. On the other hand, for a storage object that does not correspond to the detected object, the value of the certainty factor in the stored information is reduced. In addition, "confidence"
Instead of, for example, it is also possible to use “uncertainty” in which the increase / decrease of the certainty factor is reversed.

As a method of increasing or decreasing the value of the certainty factor, for example, a probability that the sensor detects an object is experimentally obtained, a method using this probability (method 1), a missing time and a period to be associated are set, There is a method (method 2) for obtaining the increase / decrease amount of the certainty factor from the set missing time or period.

The certainty factor P (k + 1) in the method 1 can be increased or decreased by the following equations (1) and (2).

When there is a detected object corresponding to the stored object P (k + 1) = P (k) + (1−P (k)) · α (1) There is no detected object corresponding to the stored object Case P (k + 1) = P (k) −P (k) · (1−α) = P (k) · α (2) where P (k) is a memory object at a certain point at time k. Α (0 << α <1) is the probability that the sensor will detect an object at that point, and P (k + 1) is the degree of certainty when a stored object exists at that point at time k + 1. .

The certainty factor of the method 1 will be described with a specific example. It is assumed that the probability α is set for each range as shown in FIG. 5 (here, the probability α is set for each range divided by distance, but the range α is set according to the detection characteristics of the sensor. It is also possible to set the probability α for each time, to express the probability α as a function of the position, etc.). Also, certainty factors P1 (k), P2 (k), P
3 (k) are respectively stored object 1: certainty P1 (k) = 0.95 storage object 2: certainty P2 (k) = 0.80 storage object 3: certainty P3 (k) = 0.60 , The threshold value of the certainty factor for deleting the storage content of the storage object is θ
= 0.5, the confidence levels P1 (k + 1), P2 (k +
1), P3 (k + 1) is updated as follows.

Storage object 1: confidence factor P1 (k + 1) = P1 (k) + (1
−P1 (k)) · α1 = 0.95 + (1−0.95) ·
0.95 = 0.9975 Storage object 2: certainty factor P2 (k + 1) = P2 (k) + (1
−P1 (k)) · α2 = 0.80 + (1−0.80) ·
0.90 = 0.98 Memory object 3: Confidence P3 (k + 1) = P3 (k) · α3
= 0.60 · 0.80 = 0.48 The degree of certainty P (k + 1) in the method 2 is defined as a blind period, for example, assuming that an object having a certainty of 1 exists at least during n times of missing. If there is a request to handle, α in the above equations (1) and (2) is replaced by ⁿ √
(1−θ). Here, θ is a threshold of certainty as a reference for deleting the storage object.

In this case, since the detection result has no probabilistic meaning, in the following equations (3) and (4), simply α = (1−θ) / n ( ^{ｎ n} √ (1−θ) The expected effect can be obtained even with the simplified calculation of)).

When there is a detected object corresponding to the stored object P (k + 1) = P (k) + α (3) When there is no detected object corresponding to the stored object P (k + 1) = P (k) -Α (4) The storage content changing means 18 performs the following processing. First, as a result of changing the certainty by the calculating means 16 for calculating the certainty of the existence of the storage object, the storage contents stored in the storage means 12 are erased for the storage object whose certainty is smaller than the threshold. . For example, in the above specific example, the certainty factor P3 (k + 1) = 0.48 of the storage object 3 and the threshold value θ = 0.5
Because of the following, the storage contents of the storage object 3 are deleted.

Next, for a storage object having a certainty factor equal to or higher than the threshold, the storage contents such as the position coordinates and the relative speed of the storage object are updated. Lastly, for a detected object that does not correspond to a storage object, an area for storing information of a new storage object is created in the storage unit 12, and an initial value of the certainty factor, position coordinates, relative The initial values such as the speed are stored.

Next, the operation of this embodiment will be described.
FIG. 6 shows how the certainty factor changes when an object is lost in a situation where an object actually exists. The probability of correctly recognizing an object is much greater than the probability of missing, so that the confidence will take a value close to 1 during long observations. If the object is lost and the object cannot be detected in this state, the confidence decreases, but the probability of correctly recognizing the object is much larger than the probability of missing, and the object is detected again before the threshold is reduced. , Confidence increases again. On the other hand, when an object is erroneously recognized due to noise as shown in FIG. 7, since the period of erroneous recognition is temporary, the certainty does not increase so much, and the certainty quickly falls below the threshold. Immediately erases the stored content.

Thus, a car that has been observed many times is
Even if a certain amount of time (the number of times) is lost, its existence can be determined, and the erroneous recognition time becomes shorter for erroneous recognition due to noise or the like. Further, as already shown in FIG. 6, the degree of certainty fluctuates greatly when a loss of sight occurs, and when correctly detected, the certainty level becomes approximately 1 and has a constant value. From this, the rate at which the sensor system can detect the object can be confirmed.

A specific example for realizing an embodiment of the present invention will be described below with reference to the drawings. FIG. 8 is a block diagram illustrating a vehicle detection device in which the object detection device according to the present embodiment is implemented using a computer. As shown in FIG. 8, the vehicle detection device includes a detection device 21 that detects distance information, angle information, and the like of another vehicle with respect to the own vehicle, and a computer 3 that performs data processing on an output signal of the detection device 21. Have been. As the detection device 21, for example, a scan-type laser radar described in the embodiment of Japanese Patent Application Laid-Open No. 1-213593 can be used. The detection device 21 outputs a detection signal including a distance signal, a scan signal, and an angle signal.

The computer 3 takes in the detection signal output from the detection device 21 and calculates information such as the distance to another vehicle. The computer 3 performs a CP for performing arithmetic processing.
U31, an input / output interface 33 for converting an external signal into a digital signal that can be processed internally, and an internal digital signal to an external signal, a RAM 35 as temporary storage means, ROM 37 storing the control programs and data of the above, and a bus such as a data bus and a control bus for connecting these.

The computer 3 executes processing according to the flowcharts shown in FIGS. First, FIG.
In step 100, the position of the vehicle measured from the detection signal obtained from the detection device 21 is obtained. The position of the vehicle can be determined, for example, using the method described in Japanese Patent Application Laid-Open No. 1-213593. At this time, as shown in FIG. 9, the measured positions of the other vehicles are defined as vehicle 1, vehicle 2,..., Vehicle j,. On the other hand, RA
A model in which position information and the like are stored in M is referred to as a model, and the models are referred to as model 1, model 2,..., Model j,.

In the next step 200, the measured vehicle is
It is determined whether or not the vehicle (model) stored in the RAM corresponds to the vehicle. The details of this determination are executed by the routine shown in the flowchart of FIG. This routine will be described. First, in order to execute a loop for examining the car measured in step 202, the number j of the car is set to 1, which is an initial value. Next, in step 204, a model number i is set to 1, which is an initial value, to execute a loop for sequentially examining models. Next, step 206
It is determined whether or not the corresponding flag of the car with the number j (vehicle j) is reset. When the flag is reset, there is no car corresponding to the model i (the i-th model).
In step 208, the correspondence between the car j and the model i is checked, and in step 210, it is determined whether or not the car j and the model i correspond. Whether or not the vehicle is compatible is determined based on whether or not the vehicle j and the model i can be regarded as being the same. This determination is carried out following good Unishi.

That is, a search range for determining whether or not the model i is supported is determined, and the car j having the highest evaluation value within the range is determined to be the corresponding car. If no car exists within the range, it is determined that the corresponding car j could not be measured. The range of the search for judging whether or not it is compatible takes into account the distance that the car may move relatively within one measurement cycle, the position error of the model, the position error of the car, and the like. It is only necessary to determine, and it is also possible to automatically change the range of the search for determining whether or not they correspond according to the calculation result of the error.

The evaluation values are determined by the model i and the car j
A method that gives a higher evaluation value as the distance from the vehicle is shorter, obtains a stochastic matching degree from the variance of an error between the model i and the car j,
A method of using the degree of coincidence as an evaluation value is conceivable. Also,
In this embodiment, the evaluation value is determined only by the position information,
It is also possible to determine whether information other than the position, for example, the speed, the color and shape of the vehicle, and the like are supported by adding them to the evaluation value.

At step 208 and step 210, the vehicle j
And the model i is determined. If the vehicle j and the model i correspond, the process proceeds to step 212, and if not, the process proceeds to step 226.

At step 212, it is determined whether or not the corresponding flag of the model i has been reset.
It is determined whether a car corresponding to the model i has existed before. If the corresponding flag of the model i has been reset and there is no vehicle corresponding to the model i, step 2
Proceeding to 22, the model number i corresponding to the corresponding flag of the car j is set. On the other hand, if a corresponding car already exists, the process proceeds to step 214.

In step 214, the model i and the car determined to have already corresponded are examined. Here, it is assumed that it corresponds to the car j '.

At step 216, the distance L (i, j), which is the evaluation value of the correspondence between the car j and the model i, and the distance L (i, j '), which is the evaluation value of the correspondence between the car j' and the model i Is calculated. In step 218, the evaluation values are compared. if,
If the distance L (i, j ′) is equal to or less than the distance L (i, j), the process proceeds to step 226. On the other hand, the distance L (i, j ′) is the distance L
If it is larger than (i, j), the process proceeds to step 220.

In step 220, since the distance L (i, j ') is determined to be larger than the distance L (i, j) in step 218, the car j is more likely to be the model i than the car j'. Since they correspond (the distance is short), the corresponding flag of the car j 'is reset, and in step 222, the model number i corresponding to the corresponding flag of the car j is set. Then, in the next step 224, the number j of the corresponding car is set in the corresponding flag of the model i.

In step 226, i = i + 1 is set to check the next model. At step 228, it is determined whether or not the model loop has been completed. If the model loop has ended, the process proceeds to step 230, and if the model loop has not ended, the process proceeds to step 206 to repeat the above processing.

At step 230, j = j + 1 is set in order to check the next measured vehicle. At step 232, it is determined whether or not the vehicle loop has been completed. If the car loop has ended, the process is terminated. If the car loop has not ended, the process proceeds to step 204 to repeat the above process.

After the correspondence between the car and the model is determined in step 200 in FIG. 10 as described above, a certainty factor is calculated in step 300 according to the correspondence state. The calculation processing of the certainty factor will be described in detail with reference to the flowchart of FIG.

First, in order to start execution of a loop for sequentially examining models, the model number is set to i = 1 (step 302). Next, in step 304, it is determined whether or not the corresponding flag of the model i has been reset, thereby determining whether or not it is determined in step 200 that a vehicle corresponding to the model i exists. When it is determined that the corresponding flag of the model i is reset, that is, when it is determined that the corresponding vehicle does not exist, the process proceeds to step 306 to decrease the certainty factor, and the corresponding flag of the model i is set. When it is determined that the vehicle is present, that is, when it is determined that the corresponding vehicle exists, the process proceeds to step 308 to increase the certainty factor.

The increase / decrease of the certainty factor will be described with an example. As shown in FIG. 17, the measurement range 20 of the detection device 21;
It is assumed that the detection probability α has been set. Here, the probability α is set by dividing the range by the distance, but it is also possible to delimit the range according to the detection characteristics of the sensor, to express α as a function of the position, and the like. Further, as shown in FIG. 14, the confidence factor of each model at time t is as follows.

Model 1: certainty P1 (t) = 0.95 Model 2: certainty P2 (t) = 0.80 Model 3: certainty P3 (t) = 0.60 The position of the vehicle obtained by the calculation is as shown in FIG. From the correspondence result of FIG. 16, the certainty factor at time t + 1 is updated as follows.

Model 1: increase in confidence P1 (t + 1) = P1 (t) + (1−P1 (t)) · α
1 = 0.95 + (1−0.95) · 0.95 = 0.975 Model 2: Increased confidence P2 (t + 1) = P2 (t) + (1−P1 (t)) · α
2 = 0.80 + (1−0.80) · 0.90 = 0.98 Model 3: Decrease in confidence P3 (k + t) = P3 (t) · α3 = 0.60 · 0.80 = 0.48 Step At 310, to find out about the next model,
Let i = i + 1. In step 312, it is determined whether or not the model loop has ended. If the model loop has ended, the process ends. If the model loop has not ended, the process proceeds to step 304.

In the next step 400, the model of the car in the RAM is updated. This model updating routine will be described with reference to FIG. In this routine, unnecessary models are deleted from the storage area, and a storage area for a new model is created. In order to smoothly delete and create in this routine, for example, the entire model may be linked by a bidirectional list.

In step 402, i = 1 is set in order to start a model loop. In step 404, it is determined whether or not the degree of certainty of the model is equal to or smaller than a threshold value. If the confidence of the model is equal to or less than the threshold, the process proceeds to step 406 to delete the storage contents of the corresponding model. On the other hand, if the certainty factor of the model is larger than the threshold, the process proceeds to step 408. In step 408, to determine about the next model, i = i
+1.

In the next step 410, it is determined whether or not the model loop has been completed. If the model loop has ended, the flow advances to step 412; if not, the flow advances to step 404 to repeat the above processing.

In step 412, j = 1 is set in order to start a car loop. In step 414, it is determined whether the model corresponding to the vehicle j exists by determining whether the corresponding flag of the vehicle j is reset. If the corresponding model does not exist, the corresponding flag is set to 0, so the process proceeds to step 416 to store the position of the car where the corresponding model did not exist, thereby storing the new model in the RAM. Create within. On the other hand, if a corresponding model exists, a model number other than 0 is set in the corresponding flag, so that the process proceeds to step 418, and j =
j + 1.

In step 420, it is determined whether or not the vehicle loop has been completed. If the loop has not ended, the process proceeds to step 414, and if the loop has ended, the process ends.

[0054]

As described above, according to the present invention, the degree of certainty representing the degree of existence of a storage object is calculated, and the stored contents are changed, deleted, and newly stored according to the degree of certainty. Since the contents are stored, it is possible to cope with a long loss of the object and to reduce the influence of erroneous detection due to noise or the like.

[Brief description of the drawings]

FIG. 1 is a block diagram showing one embodiment of the present invention.

FIG. 2 is a diagram showing a position of a storage object.

FIG. 3 is a diagram illustrating a position of a detection object.

FIG. 4 is a diagram showing a range for determining whether a storage object and a detected object correspond to each other.

FIG. 5 is a diagram showing an existence probability in a detection range.

FIG. 6 is a diagram showing a change in the certainty factor of the existence of a moving object with respect to the missing of the moving object.

FIG. 7 is a diagram showing a change in certainty of the presence of a moving object when an erroneous detection is made due to noise;

FIG. 8 is a block diagram showing a specific example of this embodiment.

FIG. 9 is a diagram showing a model stored in a RAM and a storage area of a vehicle.

FIG. 10 is a flowchart illustrating a vehicle detection processing routine according to the present embodiment.

FIG. 11 is a flowchart showing details of step 200 in FIG. 10;

FIG. 12 is a flowchart showing details of step 300 in FIG. 10;

FIG. 13 is a flowchart showing details of step 400 in FIG. 10;

FIG. 14 is a diagram showing the position of a model.

FIG. 15 is a diagram showing a detected car position.

FIG. 16 is a diagram showing a range for determining whether a model and a detected car correspond to each other.

FIG. 17 is a diagram showing the existence probability in the detection range.

[Explanation of symbols]

 3 Computer 21 Detector

Continued on the front page (72) Inventor Takero Hongo 41-41, Yokomichi, Nagakute-machi, Aichi-gun Inside (72) Inventor Yuichi Kubota 1 Toyota Town, Toyota City, Aichi Prefecture Inside Toyota Motor Co., Ltd. (56) References JP-A-4-2433491 (JP, A) JP-A-62-36581 (JP, A) (58) Field surveyed (Int.Cl. ⁷ , DB name) G08G 1/16 G01S 13/93

Claims (6)

(57) [Claims] 1. An object detection means for detecting a current position of an object, a storage means for storing a past position of the detection object detected by the object detection means, and a current position of the detection object and stored in the storage means Judgment means for judging whether or not the detected object whose current position is detected based on the position of the storage object and the storage object whose position is stored in the storage means correspond to the detection object, and the detection object Corresponding to the memory object
If it is determined that the
It is determined that the detected object and the memory object do not correspond
A calculating means for calculating a certainty factor that is smaller than the previous value; and a method for changing the storage content of the storage object having the certainty factor equal to or more than the predetermined value based on the detected current position and for the storage object having the certainty factor less than the predetermined value An object detection device, comprising: storage content changing means for erasing storage content and storing a position for a detected object determined not to correspond to a storage object by the determination means. 2. The method according to claim 1, wherein the determining means determines that the detected object whose current position has been detected exists in an area within a predetermined range including the stored past position. 2. The object detection device according to claim 1, wherein it is determined that the storage object whose position is stored in the means corresponds to the storage object. 3. The method according to claim 1, wherein the determining unit determines that the detected object and the storage object correspond to each other when the detected object and the storage object can be considered to be the same. The object detection device according to claim. 4. The calculation means calculates a certainty factor P (k) when a stored object is present at a certain point at a time k, and sets a probability that the object detection means detects an object at the point at α (0 <0).
<Α <1), the certainty factor P at time k + 1
The object detection device according to any one of claims 1 to 3, wherein (k + 1) is calculated based on the following expression. When it is determined that the detected object and the storage object correspond to each other P (k + 1) = P (k) + (1−P (k)) · α It is determined that the detection object and the storage object do not correspond to each other. P (k + 1) = P (k) · α Wherein when the predetermined value for erasing the stored contents to the storage object theta, the number of times losing sight be treated as what the object is present is n ^{times, α = n √ (1-} θ)
The object detection device according to claim 4, wherein 6. The computing means calculates a certainty factor P (k) when a stored object is present at a certain point at a time k, and calculates a probability that the object detecting means detects an object at the point at α = (1
−θ) / n (where θ is the predetermined value for erasing the stored contents of the stored object, and n is the number of times the object is lost as if the object is present), and the confidence at the time k + 1 is obtained. The object detection device according to claim 1, wherein the degree P (k + 1) is calculated based on the following equation. P (k + 1) = P (k) + α When it is determined that the detected object and the storage object do not correspond P (k + 1) = P (k) ) -Α