JP4849000B2 - Traffic environment prediction device - Google Patents

Description

  The present invention relates to a traffic environment predicting apparatus that predicts a future course of another vehicle or the like existing at an arbitrary position.

As a conventional traffic environment prediction device, for example, as described in Patent Document 1, a travel prediction distance of another vehicle is obtained based on other vehicle data such as a current position and vehicle speed starting from another vehicle, and the other vehicle There is known a predicted circular area centered on a data reception position and having a radius of a predicted movement distance as a movable range of another vehicle.
JP 2000-242898 A

  However, in the above-described prior art, since a wide range of a circular area centered on the reception position of the other vehicle data and having a radius of the predicted movement distance is used as the route prediction range of the other vehicle, accurate route prediction cannot be performed. .

  An object of the present invention is to provide a traffic environment prediction apparatus capable of improving the prediction accuracy of a future course of a moving object.

The traffic environment prediction apparatus of the present invention stores travel history storage means for storing positions where a plurality of moving objects have traveled as travel history, position setting means for setting the position of a moving object to be predicted, and travel history storage means. From the travel histories of the plurality of moving objects, a travel history close to the position of the prediction target moving object set by the position setting means is extracted, and a course of the prediction target moving object is predicted based on the travel history A route prediction means , the route prediction means divides the prediction area including the position of the moving object to be predicted set by the position setting means into a plurality of small areas, and travels close to the position of the moving object to be predicted The frequency with which the history exists in each small area is obtained, and the course of the moving object to be predicted is predicted based on the frequency .

  In such a traffic environment prediction apparatus, when the position of the moving object to be predicted is set by the position setting unit, the setting position of the moving object to be predicted is calculated from a plurality of traveling histories stored in the traveling history storage unit. A travel history close to is extracted. Then, the travel route corresponding to the extracted travel history is predicted as a route on which the predicted moving object may travel in the future. As described above, by using the traveling histories of a plurality of moving objects, it is possible to improve the prediction accuracy of the future course of the moving object to be predicted.

At this time , a travel route that passes through a small area where a travel history close to the set position of the prediction target moving object exists is predicted to be a route in which the prediction target moving object is likely to travel in the future. The Thereby, the prediction accuracy of the future course of the moving object to be predicted can be further improved.

  Preferably, the course predicting means approaches a plurality of temporally continuous positions of the moving object to be predicted set by the position setting means from the traveling histories of the plurality of moving objects stored in the traveling history storage means. The travel history is extracted, the similarity between the history of a plurality of temporally continuous positions of the moving object to be predicted and the travel history close to the history of the plurality of positions is determined, and the prediction is performed based on the similarity Predict the path of the target moving object.

  In this case, when there is a high similarity between the history of a plurality of set positions that are temporally continuous for the moving object to be predicted and the travel history that is close to the plurality of set positions, the travel route corresponding to the travel history is selected. It is predicted that the moving object to be predicted is a route that is likely to travel in the future. Thereby, the prediction accuracy of the future course of the moving object to be predicted can be further improved.

In addition, the traffic environment prediction apparatus of the present invention includes a travel history storage unit that stores a position where a plurality of moving objects have traveled as a travel history, a position setting unit that sets the position of a moving object to be predicted, and a travel history storage unit. Is extracted from the travel histories of the plurality of moving objects stored in the position setting means, the travel history close to the position of the prediction target moving object set by the position setting means, and the path of the prediction target moving object is determined based on the travel history. A route predicting unit that predicts, and the route predicting unit includes a plurality of temporally continuous moving objects to be predicted set by the position setting unit from the traveling histories of the plurality of moving objects stored in the traveling history storing unit. The travel history close to the position of the vehicle is extracted, the similarity between the history of a plurality of temporally continuous positions of the moving object to be predicted and the travel history close to the history of the plurality of positions is determined, and the similarity Based on Is characterized in that to predict the path of a moving object of the prediction target are.

In such a traffic environment prediction apparatus, when the position of the moving object to be predicted is set by the position setting unit, the setting position of the moving object to be predicted is calculated from a plurality of traveling histories stored in the traveling history storage unit. A travel history close to is extracted. Then, the travel route corresponding to the extracted travel history is predicted as a route on which the predicted moving object may travel in the future. As described above, by using the traveling histories of a plurality of moving objects, it is possible to improve the prediction accuracy of the future course of the moving object to be predicted.
  At this time, when the similarity between the history of the plurality of set positions that are temporally continuous for the moving object to be predicted and the travel history close to the plurality of set positions is high, the travel route corresponding to the travel history is predicted. It is predicted that the target moving object is likely to travel in the future. Thereby, the prediction accuracy of the future course of the moving object to be predicted can be further improved.

  ADVANTAGE OF THE INVENTION According to this invention, the prediction precision of the future course of a moving object can be improved. This makes it possible to effectively use the prediction result of the future course of the moving object in the vehicle control system or the like.

  DESCRIPTION OF EMBODIMENTS Hereinafter, a preferred embodiment of a traffic environment prediction apparatus according to the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a schematic configuration diagram showing a first embodiment of a traffic environment prediction apparatus according to the present invention. In the figure, a traffic environment prediction apparatus 1 according to the present embodiment is mounted on a vehicle such as an automobile, and predicts a future course of another vehicle existing at an arbitrary position.

  The traffic environment prediction apparatus 1 includes a route prediction ECU (Electronic Control Unit) 2, a periphery monitoring sensor 3, and a position setting device 4. A travel control ECU 5 is connected to the route prediction ECU 2.

  The perimeter monitoring sensor 3 is a sensor that detects the position of other vehicles existing around the host vehicle, and is composed of, for example, a radar, a camera, or inter-vehicle communication means and road-to-vehicle communication means. The position setting device 4 is a device for setting and inputting the absolute position of another vehicle for which a future traveling route is to be predicted.

  The course prediction ECU 2 includes an other vehicle travel history storage unit 6 and an other vehicle course prediction unit 7. The other vehicle travel history storage unit 6 chronologically records the absolute position (travel position) where any other other vehicle has traveled in the past based on the information on the other vehicle acquired by the periphery monitoring sensor 3. As a memory save. As shown in FIG. 2, the travel history of the other vehicle has a graph structure that plots and expresses the travel position (node) of the other vehicle and the temporal connection relationship (edge) of each node on the XY coordinates. Yes. As the coordinates of the traveling position of the other vehicle, for example, coordinates of latitude and longitude may be used, or Euclidean coordinates representing the distance from a certain reference point may be used.

  The other vehicle course prediction unit 7 is based on the position data of the other vehicle to be predicted set and input by the position setter 4 and the travel history data stored in the other vehicle travel history storage unit 6. The future course is predicted, and the prediction result is sent to the travel control ECU 5.

  FIG. 3 is a flowchart showing details of the prediction processing procedure executed by the other vehicle course prediction unit 7. In the figure, first, it is determined whether or not the position of the other vehicle to be predicted has been set and inputted by the position setter 4 (step S51). Note that the position of the other vehicle to be predicted is set and input as a position in the same coordinate system as the XY coordinate system expressing the travel history stored in the other vehicle travel history storage unit 6.

  When the position of the other vehicle to be predicted is set and input by the position setter 4, the prediction target that is set and input from the traveling positions of a plurality of other vehicles stored as the traveling history in the other vehicle traveling history storage unit 6. A travel position close to the position (set position) of the other vehicle is extracted (step S52).

Specifically, as shown in FIG. 4A, from the travel position Qi (i = 1 to N; N is the number of history data) stored as the travel history, the set position P of the other vehicle to be predicted and The traveling position Qs (s = 1 to M: M <N) is extracted such that the distance (see the following formula) is smaller than a separately defined threshold value θ. Here, the distance between the set position P and the travel position Qi is not necessarily the Euclidean distance, and may be anything as long as it satisfies the distance axiom.
√ (P _x −Qi _x ) ² + (P _y −Qi _y ) ²

  Subsequently, in the prediction region that includes the travel position Qs extracted in step S52 and that has a distance from the set position P that is less than or equal to a separately determined threshold π in the travel history that is temporally later than the travel position Qs. A travel history (hereinafter referred to as a partial travel history) Rs (s = 1 to M) is extracted (procedure S53).

Subsequently, as shown in FIG. 4B, the prediction area having the radius π set in step S52 is divided into a plurality of block-shaped small areas at intervals δ separately defined, and each small area S _ij (i, j is an index of the small area), and the partial traveling history Rs passing through the small area S _ij is integrated, and the frequency H at which the partial traveling history Rs passes through each small area S _{ij is} calculated for each small area S _ij by the following formula. Then, a probability distribution (predicted probability distribution) in which another vehicle to be predicted at the set position P is present at the center position of each small region _{Sij in} the future is obtained (step S54).
Frequency H = number of passage histories (pieces) / number of partial travel histories Rs (M pieces)

In addition, by storing time in the accumulated travel history at the same time, the existence position of the travel history may be counted for each time, and the probability distribution may be obtained from the following formula.
Number of small areas S _ij at time t / total number of small areas ΣS _ij with other vehicles at time t

  In addition, in order to obtain a continuous probability distribution with respect to space, the probability distribution may be calculated using only the number of peaks of each integrated number as a mixture distribution of normal distributions. Further, as a method of dividing the prediction region having the radius π, not only the above method but also, for example, concentric circles may be used. Furthermore, it is not always necessary to obtain the probability distribution, and all possible future routes may be output, and by adding the time and vehicle type together with the position, a part limited to a specific time and specific vehicle type It may be output as a possible future course.

  Subsequently, the predicted probability distribution data obtained in step S54 is sent to the travel control ECU 5 (step S55). Thereafter, the traveling control ECU 5 executes various applications related to traveling control of the host vehicle using the probability distribution data.

In such a traffic environment prediction apparatus 1, when predicting the future course of another vehicle, first, the position setting device 4 sets and inputs the position (current position) of the other vehicle whose future course is to be predicted. Then, a travel position Qs existing in the vicinity of the set position P is extracted from the travel history data stored in the other vehicle travel history storage unit 6, and a travel route along the partial travel history Rs including the travel position Qs is obtained. , Predicted as a possible future course. For example, in the case shown in FIG. 4, two partial travel histories Rs existing in the vicinity of the set position P are extracted. Then, a prediction probability distribution is obtained according to the frequency H at which the partial travel history Rs passes through each small area S _ij of the prediction area. At this time, when the prediction probability of the partial travel history Rs is high, it is predicted that the other vehicle to be predicted is likely to travel along the same route as the partial travel history Rs.

  As described above, in the present embodiment, the travel history of a plurality of other vehicles that have traveled in the past is stored in advance as a database, and the future course of the other vehicle is predicted using this travel history data. It is also possible to predict a course that cannot be predicted at the current position. Thereby, the future course of other vehicles can be predicted with high accuracy.

  FIG. 5 is a flowchart showing details of a prediction processing procedure executed by another vehicle route prediction unit of the route prediction ECU in the second embodiment of the traffic environment prediction device according to the present invention. In addition, about the schematic structure of the traffic environment prediction apparatus of this embodiment, it is the same as that of what is shown in FIG.

  In FIG. 5, it is first determined whether or not the position of the other vehicle to be predicted has been set and inputted by the position setter 4 (step S61). Note that the position of the other vehicle to be predicted is set and input as a position in the same coordinate system as the XY coordinate system expressing the travel history stored in the other vehicle travel history storage unit 6.

When the position of the other vehicle to be predicted is set and input by the position setter 4, the number of positions (set positions) of the other vehicle to be predicted set and input by the position setter 4 is used to predict the future course. It is determined whether the number of minutes corresponds to the required specified time T (step S62). That is, as shown in FIG. 6, it is determined whether there is an input from the most recent (current) temporally consecutive multiple set position of the setting position by an amount corresponding to a specified time T back from the time of P _t.

When a plurality of time-sequential set positions are input, the set positions Pj of the other vehicles to be predicted between the current set position P _t and the position that is back by the amount corresponding to the specified time T (J = t to t−N + 1) is extracted (procedure S63). For example, in the case shown in FIG. 6, set positions P _t , P _t−1 , and P _t−2 are extracted.

  Subsequently, a travel position close to the set position Pj of the other vehicle to be predicted is extracted from the travel positions of a plurality of other vehicles stored as travel histories in the other vehicle travel history storage unit 6 (step S64).

Specifically, as shown in FIG. 6, the distance from the travel position Qi (i = 1 to N; N is the number of history data) stored as the travel history to the set position Pj of the other vehicle to be predicted. A traveling position Qs (s = 1 to M: M <N) is extracted such that the sum (see the following formula) is smaller than a separately defined threshold value Θ. Here, the distance between the set position Pj and the travel position Qi is not necessarily the Euclidean distance, and may be anything as long as it satisfies the distance axiom.
Σ√ (P _x −Qi _x ) ² + (P _y −Qi _y ) ²

  Subsequently, as shown in FIG. 6, a threshold π that includes the travel position Qs extracted in step S <b> 64, and in which the distance from the set position P in the travel history after the travel position Qs is separately determined. A travel history (hereinafter referred to as a partial travel history) Rs (s = 1 to M) in the prediction region below is extracted (step S65). As a result, the similarity between the history of the set position Pj and the partial travel history Rs that are temporally continuous for the other vehicle to be predicted is determined, and the partial travel history Rs that is highly similar to the history of the set position Pj is extracted. The Rukoto.

Subsequently, similarly to step S54 shown in FIG. 3, the prediction region having the radius π set in step S65 is divided into a plurality of block-like small regions, and each small region S _ij (i, j is an index of the small region). ) in integrating the partial travel history Rs passing, by calculating the frequency with which partial driving history Rs passes in each small area S _ij, other vehicles future each small prediction target in the set position Pj A probability distribution (predicted probability distribution) existing at the center position of the region S _ij is obtained (procedure S66). Then, the probability distribution data is sent to the travel control ECU 5 (step S67).

In the present embodiment as described above, when the future course of another vehicle is predicted, the position setter 4 first sets and inputs the position (current position) of the other vehicle for which the future course is to be predicted and the past position. Then, the travel position Qs existing in the vicinity of the set position Pj of the other vehicle to be predicted is extracted from the travel history data stored in the other vehicle travel history storage unit 6, and the partial travel history Rs including the travel position Qs is extracted. Is determined to be similar to the history of the set position Pj, a travel route along the partial travel history Rs is predicted as a possible future course. Then, depending on the frequency of partial travel history Rs passes in each small area S _ij of the prediction area, the prediction probability distribution is determined.

  Thus, since the future course of the other vehicle is predicted using the similarity between the history of the set position Pj of the other vehicle and the travel history stored as the database, more accurate prediction can be performed.

  FIG. 7 is a schematic configuration diagram showing a third embodiment of the traffic environment prediction apparatus according to the present invention. In the figure, the same or equivalent members as those in the first and second embodiments are denoted by the same reference numerals, and the description thereof is omitted.

  In the figure, the traffic environment prediction apparatus 1 of the present embodiment further includes an alarm device 11 such as a speaker in addition to the components of the first and second embodiments. The other vehicle course prediction unit 7 of the course prediction ECU 2 predicts the future course of the other vehicle, sends the prediction result to the travel control ECU 5, and controls the alarm device 11 according to the travel state of the other vehicle to be predicted. .

  FIG. 8 is a flowchart showing details of the prediction processing procedure executed by the other vehicle course prediction unit 7. In the figure, steps S61 to S65 are the same as those in the flowchart shown in FIG.

  After the partial travel history Rs in the prediction area is extracted in step S65, the number of partial travel histories Rs similar to the history of the set position Pj that is temporally continuous for the other vehicle to be predicted is equal to or less than a separately specified number. Whether or not (step S70). At this time, when the number of partial traveling histories Rs similar to the history of the set position Pj is equal to or less than the specified number, it is determined that the traveling state of the other vehicle to be predicted is abnormal traveling, and an alarm is issued from the alarm device 11. Thus, the alarm device 11 is controlled (step S71).

  On the other hand, when the number of partial histories Rs similar to the history of the set position Pj is larger than the specified number, it is determined that the running state of the other vehicle to be predicted is normal running, and the same procedure as the flowchart shown in FIG. S66 and S67 are executed.

  As described above, when it is determined that the other vehicle to be predicted is traveling abnormally, an alarm is generated from the alarm device 11 to notify the driver that the traveling of the other vehicle is in an abnormal traveling state. As a result, the driver can perform a driving operation for avoiding a collision with another vehicle with careful attention.

  FIG. 9 is a schematic configuration diagram showing a fourth embodiment of the traffic environment prediction apparatus according to the present invention. In the figure, the same or equivalent members as those in the first to third embodiments are denoted by the same reference numerals, and the description thereof is omitted.

  In the figure, the traffic environment prediction device 1 of this embodiment includes a route prediction ECU 21 instead of the route prediction ECU 2 in the first to third embodiments. In addition to the components of the first and second embodiments, the traffic environment prediction device 1 displays a GPS receiver 22 that acquires position information using GPS (Global Positioning System), and a route of the host vehicle. A route indicator 23 for displaying is further provided.

  The course prediction ECU 21 further includes a host vehicle position calculation unit 24, a host vehicle travel history storage unit 25, and a planned course generation unit 26 in addition to the other vehicle travel history storage unit 6 and the other vehicle course prediction unit 7. is doing.

  The other vehicle course prediction unit 7 executes the processing procedure shown in FIG. 8 described above, and when it is determined that the traveling state of the other vehicle to be predicted is abnormal traveling, Along with the history information of the setting position Pj of the vehicle, the information is sent to the planned route generation unit 26.

  The own vehicle position calculation unit 24 calculates the current position of the own vehicle from the position information acquired by the GPS receiver 22. The own vehicle travel history storage unit 25 stores and saves the absolute position of the travel of the host vehicle as travel history data in a time series based on the calculation result of the own vehicle position calculation unit 24.

  When the other vehicle route prediction unit 7 determines that the traveling state of the other vehicle to be predicted is abnormal traveling, the planned route generation unit 26 and the history of the set position Pj of the other vehicle and the own vehicle traveling history storage unit A course for avoiding the other vehicle is determined using the traveling history of the host vehicle stored in the vehicle 25, and the course is displayed on the course indicator 23.

  As a result, the driver performs a driving operation according to the screen display of the course indicator 23, so that a collision with another vehicle traveling abnormally can be avoided.

  FIG. 10 is a schematic configuration diagram illustrating a modification of the traffic environment prediction apparatus illustrated in FIG. 9. In the present modified embodiment, when the other vehicle course prediction unit 7 determines that the running state of the other vehicle to be predicted is abnormal running, the planned course generation unit 26 automatically records the history of the set position Pj of the other vehicle. Using the travel history of the host vehicle stored in the vehicle travel history storage unit 25, a guidance route for avoiding the other vehicle is determined, and the guidance route information is sent to the travel control ECU 5. Then, the travel control ECU 5 controls the drive unit 27 such as a wheel drive motor so that the host vehicle automatically travels in accordance with the guidance route.

  The present invention is not limited to the above embodiment. For example, the above embodiment is for predicting the future course of other vehicles around the own vehicle, but the present invention is not limited to such vehicles, but various moving objects that run on the road such as motorcycles and bicycles. You may predict the future course.

It is a schematic block diagram which shows 1st Embodiment of the traffic environment prediction apparatus concerning this invention. It is a figure which shows the expression structure of the travel history of the other vehicle memorize | stored and preserve | saved at the other vehicle travel history preservation | save part shown in FIG. It is a flowchart which shows the detail of the prediction process procedure performed by the other vehicle course prediction part shown in FIG. It is a conceptual diagram which shows the method of estimating the future course of another vehicle according to the process sequence shown in FIG. It is a flowchart which shows the detail of the prediction process procedure performed by the other vehicle course prediction part in 2nd Embodiment of the traffic environment prediction apparatus concerning this invention. It is a conceptual diagram which shows the method of estimating the future course of another vehicle according to the process sequence shown in FIG. It is a schematic block diagram which shows 3rd Embodiment of the traffic environment prediction apparatus concerning this invention. It is a flowchart which shows the detail of the prediction process procedure performed by the other vehicle course prediction part shown in FIG. It is a schematic block diagram which shows 4th Embodiment of the traffic environment prediction apparatus concerning this invention. It is a schematic block diagram which shows the modification of the traffic environment prediction apparatus shown in FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Traffic environment prediction apparatus, 2 ... Course prediction ECU, 4 ... Position setter (position setting means), 6 ... Other vehicle travel history storage part (travel history storage means), 7 ... Other vehicle course prediction part (course prediction means) 11 ... alarm device, 21 ... route prediction ECU, 23 ... route indicator, 24 ... own vehicle position calculation unit, 25 ... own vehicle travel history storage unit, 26 ... planned route generation unit.

Claims (3)

Travel history storage means for storing the positions traveled by a plurality of moving objects as a travel history;
Position setting means for setting the position of the moving object to be predicted;
A travel history close to the position of the moving object to be predicted set by the position setting means is extracted from the travel histories of the plurality of moving objects stored in the travel history saving means, and based on the travel history A route prediction means for predicting the route of the moving object to be predicted ;
The course prediction unit divides a prediction area including the position of the moving object to be predicted set by the position setting unit into a plurality of small areas, and a travel history close to the position of the moving object to be predicted is A traffic environment prediction device characterized in that a frequency existing in each small area is obtained and a course of the prediction target moving object is predicted based on the frequency . The route predicting means approaches a plurality of temporally continuous positions of the prediction target moving object set by the position setting means from the traveling histories of the plurality of moving objects stored in the traveling history storage means. The travel history is extracted, and the similarity between the history of a plurality of temporally continuous positions of the moving object to be predicted and the travel history close to the history of the plurality of positions is determined, and based on the similarity The traffic environment prediction apparatus according to claim 1, wherein a route of the prediction target moving object is predicted. Travel history storage means for storing the positions traveled by a plurality of moving objects as a travel history;
Position setting means for setting the position of the moving object to be predicted;
A travel history close to the position of the moving object to be predicted set by the position setting means is extracted from the travel histories of the plurality of moving objects stored in the travel history saving means, and based on the travel history A route prediction means for predicting the route of the moving object to be predicted;
The route predicting means approaches a plurality of temporally continuous positions of the prediction target moving object set by the position setting means from the traveling histories of the plurality of moving objects stored in the traveling history storage means. The travel history is extracted, and the similarity between the history of a plurality of temporally continuous positions of the moving object to be predicted and the travel history close to the history of the plurality of positions is determined, and based on the similarity wherein it characterized by predicting the course of the moving object to be predicted transportation environment prediction device.

Images