JP4924208B2 - Vehicle travel support device - Google Patents

Description

  The present invention relates to a vehicle travel support apparatus that supports travel of a host vehicle based on the risk of collision in the host vehicle.

There is a travel support device that calculates the route of the host vehicle and performs the travel support of the host vehicle based on the route when the host vehicle is supported. Conventionally, for example, there is a route search device disclosed in Patent Document 1 as a device for obtaining a route in such a driving support device. In this route search device, a fatigue level or a risk level when passing through the road section is assigned to a road section in the road network data, and a route search start point is used by using the fatigue level or the risk level. It searches for the optimal route from to the end point.
JP 2004-12247 A

  However, the route search device disclosed in Patent Document 1 searches for a route (route) in consideration of the degree of fatigue or risk in a road section based on a map or statistical data. The risk level is not taken into consideration. For this reason, there is a difference in skills such as driving skills and driving behavior of the driver.For example, even when there is a difference in the risk of driving by each driver, the route search is performed uniformly, The route according to the skill difference such as driving behavior cannot be obtained. Therefore, when this route search device is used, there is a problem that it is difficult to obtain a route according to the skill of the driver, and it is difficult to perform appropriate driving support.

  Accordingly, an object of the present invention is to provide a vehicle travel support device that can obtain a route according to the skill of the driver and can perform travel support according to the skill of the driver by using this route.

The vehicle travel support apparatus according to the present invention that has solved the above problems is a travel support apparatus that assists the travel of the host vehicle, wherein the risk of a route that is a part of the travel path on which the host vehicle travels is determined as an obstacle. A risk evaluation means for evaluating based on the possibility of collision, and a route acquisition means for acquiring a traveling route of the host vehicle formed by connecting a plurality of routes based on the evaluation result of the risk, The route acquisition means acquires a traveling route of the host vehicle by connecting a plurality of routes selected by cutting a route having an estimated degree of risk exceeding a predetermined upper limit value among the plurality of routes .

  In the vehicle travel support device according to the present invention, the travel route of the host vehicle is obtained by cutting a route in which the risk of the route on the travel route on which the host vehicle travels exceeds a predetermined upper limit value. For this reason, since the route can be obtained by avoiding a route that is dangerous for the driver, a route corresponding to the skill of the driver can be obtained, and traveling according to the skill of the driver by using this route Can provide support.

Further, the vehicle travel support device according to the present invention which has solved the above problems, in the driving support apparatus for supporting running of the vehicle, the risk of root vehicle is part of the traveling path of travel, the vehicle failure A risk evaluation means for evaluating based on the possibility of collision with an object, and a route acquisition means for acquiring a traveling route of the host vehicle formed by connecting a plurality of routes based on the evaluation result of the risk. The route acquisition means acquires a travel route of the host vehicle by connecting a plurality of routes having a common risk degree of a route having an estimated risk level equal to or lower than a predetermined lower limit value among the plurality of routes. Is.

In the vehicle travel support apparatus according to the present invention, the route risk level in which the route risk level on the travel path on which the host vehicle travels is equal to or lower than a predetermined lower limit value is shared. For this reason, the route can be acquired without making a difference within the range of the risk level acceptable as the driver. Therefore, it is possible to obtain a route according to the skill of the driver and use this route to provide driving support according to the skill of the driver. In addition, since it is possible to acquire a route according to other factors such as distance and time, for example, within a range that can be permitted by the driver, it is possible to perform driving support with improved efficiency as a distance and time.
Also, the risk level of the route can be evaluated based on the risk level on the route or at a point closest to the route.
Furthermore, it is possible to further comprise a risk storage means for storing the evaluated route risk, and use the route risk stored in the stored risk storage means as the route risk. it can.

  According to the vehicle travel support device of the present invention, a route corresponding to the skill of the driver can be obtained, and travel support corresponding to the skill of the driver can be performed by using this route.

  Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the description of the drawings, the same elements are denoted by the same reference numerals, and redundant description is omitted. For the convenience of illustration, the dimensional ratios in the drawings do not necessarily match those described.

  FIG. 1 is a block diagram showing the configuration of the driving support apparatus according to the first embodiment of the present invention, FIG. 2 is a block diagram of a risk map creating unit, and FIG. 3 is a block diagram of a collision probability calculating unit. is there.

  As shown in FIG. 1, the driving support device according to the present embodiment includes a risk map creation unit 1, an obstacle sensor 2, an obstacle extraction unit 3, a host vehicle sensor 4, a position sensor 5, a timer 6, and a map database. 7 is provided. In addition, the driving support device includes a destination setting unit 8, a database reading unit 9, a normal risk calculation unit 10, a filter unit 11, an optimum route search unit 12, and a driving support unit 13.

  The risk map creation unit 1 is a computer of an electronically controlled automobile device, and includes a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), an input / output interface, and the like. Yes. As shown in FIG. 2, the risk map creation unit 1 includes a collision probability calculation unit 20, a database reading unit 21, a statistical processing unit 22, and a database writing unit 23. In addition, an obstacle sensor 2 is connected to the danger map creation unit 1 via an obstacle extraction unit 3, and a host vehicle sensor 4 is connected. Further, a position sensor 5, a timer 6, and a map database 7 are connected to the risk map creating unit 1.

  The obstacle sensor 2 includes a millimeter wave radar sensor, a laser radar sensor, an image sensor, and the like, and detects obstacles such as other vehicles and passers-by around the host vehicle. The obstacle sensor 2 transmits obstacle-related information including information on the detected obstacle to the obstacle extraction unit 3.

  The obstacle extraction unit 3 extracts obstacles from the obstacle-related information transmitted from the obstacle sensor 2, and the collision probability calculation unit in the risk map creation unit 1 as obstacle information such as the position and movement speed of the obstacles. 20 is output. For example, when the obstacle sensor 2 is a millimeter wave radar sensor or a laser radar sensor, the obstacle extraction unit 3 detects an obstacle based on the wavelength of a reflected wave reflected from the obstacle. When the obstacle sensor 2 is an image sensor, for example, another vehicle is extracted from the captured image as an obstacle by a technique such as pattern matching.

  The own vehicle sensor 4 includes a speed sensor, a yaw rate sensor, and the like, and detects information related to the traveling state of the own vehicle. The own vehicle sensor 4 transmits the detected vehicle traveling state information regarding the traveling state of the own vehicle to the collision probability calculating unit 20 in the risk map creating unit 1. The traveling state information of the host vehicle here includes, for example, the speed and yaw rate of the host vehicle.

  The position sensor 5 detects the position of the host vehicle and transmits the detected current position (x, y) of the host vehicle to the database reading unit 9 and the database reading unit 21 in the risk map creating unit 1. As the position sensor 5, a GPS (Global Positioning System) device, an odometry device, or the like can be used. When an odometry device is used as the position sensor 5, the origin is fixed at, for example, a driver's home parking lot.

  The timer 6 measures the time, and transmits the measured time t to the database reading unit 21 in the risk map creating unit 1. Here, when a GPS device is used as the position sensor 5, this GPS device can be used as the timer 6.

  As shown in FIG. 3, the collision probability calculation unit 20 in the risk map creation unit 1 includes an obstacle information temporary storage unit 31, an obstacle possible route calculation unit 32, a host vehicle route recording unit 33, and a host vehicle route reading unit 34. A host vehicle position reading unit 35, a host vehicle possible course calculation unit 36, an actual course collision probability calculation unit 37, a best host vehicle course collision probability calculation unit 38, and a host vehicle risk calculation unit 39.

  The obstacle information temporary storage unit 31 stores the obstacle information transmitted from the obstacle extraction unit 3 for a predetermined time, for example, 5 seconds. The obstacle possible course calculation unit 32 reads the obstacle information for the past 5 seconds stored in the obstacle information temporary storage unit 31, and based on the obstacle information for 5 seconds, the obstacle for a certain period of time thereafter. A plurality of courses predicted to move are calculated and acquired. Note that, instead of the past 5 seconds, an appropriate time (for example, a time between 3 seconds and 10 seconds) can be set. The obstacle possible course calculation unit 32 outputs the obstacle course information regarding the calculated course of the obstacle to the realized course collision probability calculation unit 37 and the best own vehicle course collision probability calculation unit 38.

  The host vehicle course recording unit 33 records the course history of the host vehicle based on the running state information of the host vehicle transmitted from the host vehicle sensor 4. The own vehicle route reading unit 34 reads the history of the own vehicle track for a predetermined time, for example, 5 seconds, which is recorded in the own vehicle route recording unit 33. The predetermined time set in advance here is in common with the time of the obstacle information stored in the obstacle information temporary storage unit 31. The own vehicle course reading unit 34 outputs, to the realized course collision probability calculation unit 37, the actual course information related to the actual course that the host vehicle has actually taken based on the read course history of the host vehicle.

  The own vehicle position reading unit 35 reads the position of the own vehicle for a predetermined time, for example, 5 seconds, from the history of the own vehicle recorded in the own vehicle course recording unit 33, and the own vehicle position related to the position of the own vehicle. The information is output to the own vehicle possible route calculation unit 36. The own vehicle possible route calculating unit 36 can move the own vehicle from the position where the own vehicle is recorded based on the own vehicle position information output from the own vehicle position reading unit 35. Calculate and obtain multiple courses. The own vehicle possible route calculation unit 36 outputs the own vehicle possible route information regarding the calculated possible route of the own vehicle to the best own vehicle route collision probability calculation unit 38.

  Based on the obstacle route information output from the obstacle possible route calculation unit 32 and the realized route information output from the own vehicle route reading unit 34, the actual route collision probability calculation unit 37 is subject to the own vehicle during the past 5 seconds. Calculates and acquires the actual course collision probability that may have collided with an obstacle in the actual course. The realized course collision probability calculation unit 37 outputs the realized course collision probability information related to the calculated realized course collision probability to the own vehicle risk calculation unit 39.

  Based on the obstacle route information output from the obstacle possible route calculation unit 32 and the own vehicle possible route information output from the own vehicle possible route calculation unit 36, the best own vehicle route collision probability calculation unit 38 The best own vehicle course that minimizes the probability of collision with another vehicle is calculated. Further, the best own vehicle course collision probability calculation unit 38, based on the calculated best own vehicle course, has the possibility that the own vehicle may collide with an obstacle in the best own vehicle course during the past 5 seconds. Calculate and obtain the course collision probability. The best own vehicle course collision probability calculation unit 38 outputs the best own vehicle course collision probability information based on the calculated best own vehicle course collision probability to the own vehicle risk calculation unit 39.

  The own vehicle risk calculating unit 39 is a degree of divergence between the realized course collision probability information output from the realized course collision probability calculating unit 37 and the best own vehicle course collision probability information output from the best own vehicle course collision probability calculating unit 38. Based on the above, the risk level of the host vehicle is calculated. The degree of risk here can be, for example, the degree of divergence between the realized course collision probability based on the realized course collision probability information and the best own vehicle course collision probability based on the best own vehicle course collision probability information. The degree of divergence can be the ratio of the two. The own vehicle risk level calculation unit 39 outputs the own vehicle risk level Rcur based on the calculated risk level to the statistical processing unit 22.

  Further, the map database 7 shown in FIG. 1 and FIG. 2 stores road information related to roads that serve as roads for automobiles. Further, the map database 7 can write the risk obtained at time t at any point (x, y) in the road information as the memory risk Rdb (x, y, t). The map database 7 stores the storage risk Rdb (x, y, t) corresponding to a plurality of points and times. The storage risk Rdb (x, y, z) stored in the map database 7 may be set in advance, or may be configured to be automatically constructed or updated from the risk of the host vehicle.

  One aspect of automatic construction and update of the memory risk Rdb (x, y, z) is as follows. The database reading unit 21 shown in FIG. 2 refers to the map information stored in the map database 7 with respect to the position (x, y) and time t transmitted from the position sensor 5, and stores the current and time storage risk of the host vehicle. Rdb (x, y, t) is acquired. The database reading unit 21 outputs the memory risk Rdb (x, y, t) based on the current position (x, y) of the host vehicle and the risk at the current time t to the statistical processing unit 22.

  The statistical processing unit 22 includes a storage risk based on the storage risk Rdb (x, y, t) output from the database reading unit 21 and a host vehicle output from the host vehicle risk calculation unit 39 in the collision probability calculation unit 20. Statistical processing is performed using the own vehicle risk level Rcur based on the risk level information. As the statistical processing here, the statistical processing unit 22 outputs the host vehicle risk output from the collision probability calculation unit 20 to the storage risk Rdb (x, y, t) of the host vehicle output from the database reading unit 21. The degree is integrated, and the integrated risk Rnew (x, y, t) is calculated. The statistical processing unit 22 records the calculated storage risk Rdb (x, y, t), the current position (x, y) of the host vehicle, and the accumulated risk R new (x, y, t) based on the current time t. Output to the insertion unit 23.

  The database writing unit 23 writes the cumulative risk Rnew (x, y, t) output from the statistical processing unit 22 in the map database 7. The map database 7 stores the cumulative risk Rnew (x, y, t) written by the database writing unit 23 as the storage risk Rdb at the position and time.

  The destination setting unit 8 sets a destination based on a driver's operation on a switch (not shown) provided in the driver's seat. The destination setting unit 8 transmits the set destination position (xg, yg) to the database reading unit 9.

  The database reading unit 9 stores the current position (x, y) of the host vehicle transmitted from the position sensor 5 and the position (xg, yg) of the destination transmitted from the destination setting unit 8 in the map database 7. Refer to the map information. The database reading unit 9 stores the risk Rall (xg, yg, x, y) of all positions on the route where the destination (xg, yg) can be reached from the current position (x, y) of the host vehicle. Search from the risk level Rdb. The database reading unit 9 outputs the searched risk level Rall (xg, yg, x, y) and the route for this risk level to the filter unit 11.

  The normal risk level calculation unit 10 reads all the storage risk levels Rdb stored in the map database 7, performs statistical processing on all the storage risk levels Rdb, and calculates the normal risk level θ. The normal risk level calculation unit 10 outputs the calculated normal risk level θ to the filter unit 11.

  The filter unit 11 compares all the risk levels Rall output from the database reading unit 9 with the normal risk level θ, and extracts only the risk levels below the normal risk level θ from the risk level Rall. The filter unit 11 outputs the extracted risk level and a route corresponding to the risk level to the optimum route search unit 12.

  The optimum route search unit 12 is based on the danger levels output from the filter unit 11 and the routes corresponding to these danger levels, and the optimum route route (optimum route) from the current position of the host vehicle to the destination. Explore. The optimum route searching unit 12 outputs the searched optimum route to the driving support unit 13.

  The driving support unit 13 performs driving support for the driver based on the optimum route output from the optimum route searching unit 12. As the driving support, a display for navigating the driving along the optimum route or voice guidance can be performed. In the case of automatic driving, driving along the optimum route can be performed.

  Next, a processing procedure of the driving support device according to the present embodiment will be described. Here, FIG. 4 is a flowchart showing a processing procedure of the driving support apparatus.

  As shown in FIG. 4, in the driving assistance apparatus according to the present embodiment, the risk map creating unit 1 creates a risk map (S1). Creation of the risk map is performed according to the flowchart shown in FIG. As shown in FIG. 5, when creating the risk map, first, the current position (x, y) of the host vehicle transmitted from the position sensor 5 is acquired (S11). Next, the current time t transmitted from the timer 6 is acquired (S11). When the current position (x, y) and the current time t of the own vehicle are acquired, the acquired current position / time (x, y, t) is referred to the map database 7 to correspond to the current position and the current time of the own vehicle. The memory risk level Rdb (x, y, t) to be read is read (S12). Here, when referring to the current position / time (x, y, t) in the map information, it is rare that the storage risk Rdb (x, y, t) corresponding to the current position / time of the host vehicle exists. is there. For this reason, when the memory risk Rdb (x, y, t) at the current position of the host vehicle does not exist, it is stored at the point closest to the current position / time (x, y, t) of the host vehicle. The memory risk level Rdb is acquired. Further, when the distance between the closest point from the current position / time (x, y, t) of the own vehicle and the current position / time (x, y, t) of the own vehicle exceeds a predetermined threshold Reads a predetermined number of memory risks stored at a predetermined number of points close to the current position of the host vehicle, for example, an arbitrary number (n = 2 to 5) of memory risks of 2 to 5 .

  When the storage risk Rdb (x, y, t) corresponding to the current position and the current time of the host vehicle is read, the host vehicle risk Rcur is calculated (S13). The calculation of the own vehicle risk level Rcur is performed according to the procedure shown in FIG. FIG. 6 is a flowchart showing a procedure for calculating the own vehicle risk level. As shown in FIG. 6, when calculating the vehicle risk Rcur, the obstacle extraction unit 3 determines obstacles around the host vehicle based on the obstacle related information transmitted from the obstacle sensor 2. Extract (S21). Here, another vehicle is extracted as an obstacle. When a plurality of other vehicles are included, all of the plurality of other vehicles are extracted.

  If the other vehicle as an obstacle is extracted, the other vehicle information regarding the extracted other vehicle is stored in the obstacle information temporary storage unit 31, and based on the other vehicle information for the past 5 seconds stored in the obstacle information temporary storage unit 31. Thus, the possible course where the other vehicle can move is calculated by the obstacle possible course calculation unit 32 as a trajectory on time and space constituted by time and space for each other vehicle (S22). Here, as a possible course in which the other vehicle can move, a certain destination point is not defined and the possible course to the destination point is not calculated, but until a predetermined movement time in which the other vehicle moves passes. Find the course of In general, there is no place where safety is guaranteed in advance on the road on which the host vehicle travels, so in order to determine the possibility of collision between the host vehicle and the other vehicle, the arrival point between the host vehicle and the other vehicle However, it cannot be said that the collision can be surely avoided.

  For example, as shown in FIG. 7, on a three-lane road R, the host vehicle M travels in the first lane r1, the first other vehicle H1 travels in the second lane r2, and the third lane travels in the second lane. Assume that H2 is traveling. At this time, in order to avoid collision with the other vehicles H1 and H2 that the vehicle M travels in the second and third lanes r2 and r3, respectively, the vehicle M reaches positions Q1, Q2, and Q3, respectively. It is considered preferable to travel. However, when the second other vehicle H2 takes the route B3 so as to change the route to the second lane r2, the first other vehicle H1 takes the route B2 in order to avoid a collision with the second other vehicle H2. It is conceivable that the vehicle enters the first lane r1. In this case, if the host vehicle M travels so as to reach the positions Q1, Q2, and Q3, there is a risk of collision with the first other vehicle H1.

  Therefore, the positions of the own vehicle and other vehicles are not determined in advance, but the courses of the own vehicle and other vehicles are predicted each time. By predicting the course of the host vehicle and the other vehicle each time, for example, the course B1 as shown in FIG. 8 can be used as the course of the host vehicle, so that the danger of the host vehicle M traveling can be avoided accurately. Safety.

  In addition, instead of prescribing until a predetermined travel time for the other vehicle to move has elapsed, an aspect in which the possible course of the other vehicle is obtained until the travel distance traveled by the other vehicle reaches a predetermined distance may be adopted. it can. In this case, the predetermined distance can be appropriately changed according to the speed of the other vehicle (or the speed of the host vehicle).

  The possible routes of other vehicles are calculated for each other vehicle as follows. An initialization process is performed in which the value of the counter k for identifying another vehicle is set to 1, and the value of the counter n indicating the number of possible course generations for the same other vehicle is set to 1. Subsequently, the position and moving state (speed and moving direction) of the other vehicle based on the other vehicle information transmitted from the obstacle sensor 2 and extracted from the other vehicle related information are set as the initial state.

  Subsequently, one behavior is selected from a plurality of selectable behaviors according to a behavior selection probability given in advance to each behavior as a behavior of the other vehicle assumed during the subsequent fixed time Δt. The behavior selection probability when selecting one behavior is defined, for example, by associating elements of a selectable behavior set with a predetermined random number. In this sense, a different behavior selection probability may be given for each behavior, or an equal probability may be given to all elements of the behavior set. Moreover, it is also possible to adopt a mode in which the behavior selection probability depends on the position and traveling state of another vehicle and the surrounding road environment.

  The selection of the behavior of the other vehicle assumed during a certain time Δt based on the behavior selection probability is repeatedly performed, and the behavior of the other vehicle is selected up to a time that is a predetermined movement time for the other vehicle to move. One possible course of the other vehicle can be calculated based on the behavior of the other vehicle thus selected.

  When one possible route of another vehicle is calculated, a plurality (N) of possible routes of the other vehicle are calculated by the same procedure. Even when a similar procedure is used, since one behavior is selected according to a behavior selection probability given in advance to each behavior, different possible routes are calculated in most cases. The number of possible routes calculated here is determined in advance and can be set to 1000 (N = 1000), for example. Of course, it can also be set as the aspect which calculates another some possible course, for example, it can be set as the number between hundreds-tens of thousands. The possible course calculated in this way is set as the predicted course of the other vehicle.

  Furthermore, when there are a plurality of other vehicles extracted, possible routes are calculated for each of the plurality of other vehicles.

  When the possible course of the other vehicle is calculated, the own vehicle course reading unit 34 reads the course of the host vehicle recorded in the own vehicle course recording unit 33 for the past 5 seconds (S23). The own vehicle course reading unit 34 outputs the read actual course information related to the realized course of the host vehicle for the past 5 seconds to the realized course collision probability calculation unit 37.

  Subsequently, the own vehicle position reading unit 35 reads the position of the own vehicle in the past 5 seconds recorded in the own vehicle course recording unit 33. Then, based on the position of the host vehicle in the past 5 seconds, the host vehicle possible path calculation unit 36 calculates a plurality of possible paths where the host vehicle can move from that position (S24). The possible course of the host vehicle is calculated as a trajectory on time and space composed of time and space by the same calculation procedure as the possible course of the other vehicle. When the own vehicle possible route calculation unit 36 acquires the possible route of the own vehicle, the own vehicle possible route calculation unit 36 outputs the own vehicle possible route information to the best own vehicle route collision probability calculation unit 38.

  Thus, after the possible course of the host vehicle has been calculated, the actual course collision probability calculation unit 37 calculates the actual course that the host vehicle has moved based on the actual course information output from the host vehicle course reading unit 34 ( S25). Then, the actual course collision probability calculation unit 37 calculates the actual course collision probability that the host vehicle allowed to collide with another vehicle (S26). Here, the predicted courses of a plurality of other vehicles in the past 5 seconds based on the obstacle course information output from the obstacle possible course calculation unit 32 are compared with the actual course obtained in step S25, and in the past 5 seconds, The collision probability allowed by the host vehicle is calculated.

  Now, an example of the predicted course of the other vehicle and the actual course of the host vehicle obtained in step S22 and step S23 is represented by the three-dimensional space shown in FIG. In the three-dimensional space in FIG. 9, the position of the vehicle is shown on the xy plane indicated by the x axis and the y axis, and the t axis is set as the time axis. Accordingly, the predicted courses of the other vehicle and the host vehicle can be indicated by (x, y, t) coordinates, and the trajectory obtained by projecting each course of the host vehicle and the other vehicle on the xy plane This is a travel locus on the road of the travel route and the predicted travel route predicted by the other vehicle to travel.

  In this way, by expressing the predicted courses of the subject vehicle and other vehicles in the space shown in FIG. 9, a plurality of vehicles (own vehicle and other vehicles) existing within a predetermined range of the three-dimensional space-time are taken. A spatiotemporal environment consisting of a set of possible prediction paths is formed. The spatiotemporal environment Env (M, H) shown in FIG. 9 is a set of the actual course of the own vehicle M and the predicted course of the other vehicle H, and the actual course {M (n1)} of the own vehicle M and the other vehicle H. It consists of a predicted course set {H (n2)}. More specifically, the spatiotemporal environment (M, H) is when the host vehicle M and the other vehicle H are moving in a + y-axis direction on a flat and straight road R such as an expressway. It shows the spatial environment. Here, when the predicted course of the other vehicle is obtained, the predicted course is independently obtained for each of the own vehicle M and the other vehicle H without considering the correlation between the own vehicle M and the other vehicle H. Paths may intersect in space and time.

Thus, when the actual course of the own vehicle M and the predicted course of the other vehicle H are obtained, the probability that the own vehicle M has allowed the own vehicle M and the other vehicle H to collide is obtained. Now, when the actual course of the own vehicle M and the predicted course of the other vehicle H intersect, the own vehicle M and the other vehicle H will collide, but the predicted course of the own vehicle M and the other vehicle H is predetermined. It is obtained based on the behavior selection probability. Therefore, the collision probability between the own vehicle M and the other vehicle H can be determined by the number of the predicted courses of the other vehicle H that intersect the predicted course of the own vehicle M. For example, the case of 1000 calculates a predicted route of the vehicle H, if five of which intersects the predicted course of the vehicle M is 0.5% chance of a collision (collision possibility) P _A Can be calculated. In other words, the remaining 99.5% can be a probability that the own vehicle M and the other vehicle H do not collide (non-collision possibility).

Further, as the other vehicle H, when a plurality of other vehicles are extracted, the collision probability P _A that allows to collide with at least one of the plurality of other vehicles be determined by the following equation (1) it can.

Here, k: number of other vehicles extracted P _A k: probability of collision with the kth vehicle In this way, a plurality of predicted routes of the other vehicle H are calculated, and the own vehicle is calculated using the plurality of predicted routes. By predicting the possibility of collision between M and the other vehicle H, the routes that the other vehicle can take are widely calculated. Accordingly, the collision probability can be calculated taking into account the case where there is a large change in the course of another vehicle, such as when an accident occurs at a place where there is a branch such as an intersection.

  When the collision probability between the own vehicle and the other vehicle is obtained in this way, the best own vehicle route collision probability calculation unit 38 has the lowest collision probability with the other vehicle among the possible routes in which the own vehicle can move. The best own vehicle course is selected (S27). In the best own vehicle course collision probability calculation unit 38, a predicted course of a plurality of other vehicles based on the obstacle course information output from the obstacle possible course calculation unit 32, and the own vehicle output from the own vehicle possible course calculation unit 36. The best own vehicle route is selected by comparing with possible routes of a plurality of own vehicles based on the possible route information.

  Now, an example of the predicted course of the other vehicle obtained in step S22 and the actual course of the subject vehicle obtained in step S24 are represented by the three-dimensional space shown in FIG. The spatiotemporal environment Env (M, H) shown in FIG. 10 is a set of possible courses of the own vehicle M and predicted courses of the other vehicle H. The possible course set {M (n1)} of the own vehicle M and the other vehicle H Of the predicted course set {H (n2)}.

From the possible course set of the own vehicle M and the predicted course set of the other vehicle H, it is confirmed that the own vehicle M collided with the other vehicle H for each time the own vehicle M moved along the possible course of each own vehicle The probability that the host vehicle M has allowed is obtained. Collision probability P _A of that vehicle M had collided with another vehicle H vehicle M had acceptable can be determined by the above equation (1).

Thus, after each determined collision probability P _A plurality of possible tracks of the vehicle M, to select the most collision probability it can lower course as the best vehicle track. The best own vehicle course collision probability calculating unit 38 calculates the collision probability in the selected best own vehicle course as the best own vehicle course collision probability, and outputs the best own vehicle course collision probability information to the own vehicle risk calculating unit 39. To do.

  When the actual course collision probability is calculated, the host vehicle risk calculation unit 39 calculates the host vehicle risk (S28). The own vehicle risk calculation unit 39 realizes the actual course collision probability based on the actual course collision probability information output from the actual course collision probability calculation unit 37 and the best own vehicle course output from the best own vehicle course collision probability calculation unit 38. Based on the degree of deviation from the collision probability, the host vehicle risk Rcur is calculated. The host vehicle risk Rcur here is determined by the ratio of the actual course collision probability and the best host vehicle path collision probability. The closer the ratio of the actual course collision probability and the best host vehicle path collision probability is to 1, the more the host vehicle risk The degree Rcur will be low. Thus, the own vehicle risk level Rcur is calculated.

  Returning to the flow shown in FIG. 5, when the collision probability calculation unit 20 calculates the own vehicle risk level Rcur, the statistical processing unit 22 calculates the integrated risk level Rnew (x, y, t) (S14). The cumulative risk Rnew (x, y, t) is the vehicle risk Rcur output from the collision probability calculation unit 20 with respect to the memory risk Rdb (x, y, t) output from the database reading unit 21. Is calculated using the following equation (2).

Rnew = αRcur + (1-α) Rdb (2)
Here, α: Arbitrary constant satisfying α <1. Further, when a plurality of storage risk levels Rdb are acquired in step S12, the closest point from the current position / time (x, y, t) of the host vehicle. Then, the distance from the current position / time (x, y, t) of the host vehicle exceeds a predetermined threshold value. In this case, the own vehicle risk level Rcur is directly used as the integrated risk level Rnew (x, y, t). In this way, the integrated risk Rnew (x, y, t) is calculated, and the risk is evaluated.

  In addition, the distance between the current location / time (x, y, t) of the host vehicle and the current location / time (x, y, t) of the host vehicle exceeds a predetermined threshold. For example, when four storage risk levels Rdb1 to Rdb4 are read, the own vehicle risk level Rcur is added, and at the same time, the four storage risk levels Rdb1 to Rdb4 are integrated. The integrated amount at the time of this integration is obtained according to the absolute difference between the respective separation distances of the own vehicle risk level Rcur and the storage risk levels Rdb1 to Rdb4.

  For example, as shown in FIG. 11, the second memory risk level Rdb2 and the third memory risk level Rdb3 are located near the own vehicle risk level Rcur, and the first memory risk level Rdb1 is located farther than the second memory risk level Rdb2. Suppose that there is a fourth memory risk level Rdb4 at a position farther than the third memory risk level Rdb3. In this case, for example, the first to fourth cumulative risk levels Rnew1 to Rnew4 can be calculated based on the following equation (3).

Rnewn = βRcur + (1-β) Rdbn (3)
Where n is an integer from 1 to 4
β: a coefficient less than 1 that decreases with increasing distance from the current position / time (x, y, t) of the host vehicle. Therefore, as shown in FIG. 11, in this example, the second cumulative risk from the second memory risk Rdb2 The rate of increase to Rnew2 is increased from the first memory risk Rdb1 to the first cumulative risk Rnew1. Similarly, the rate of increase from the third memory risk Rdb3 to the third cumulative risk Rnew3 is increased, and the rate of increase from the fourth memory risk Rdb4 to the fourth cumulative risk Rnew4 is increased. Thus, when a plurality of storage risk levels Rdb are acquired, the nearest point from the current position / time (x, y, t) of the own vehicle and the current position / time (x, y, t) of the own vehicle When the separation distance exceeds the predetermined threshold value, the risk levels are evaluated by calculating the integrated risk levels Rnew1 to Rnew4 for the surrounding memory risk levels Rdb1 to Rdb4.

  After calculating the integrated risk Rnew or the integrated risk Rnew1 to Rnew4 in this way, the database writing unit 23 writes the calculated integrated risk Rnew or the integrated risk Rnew1 to Rnew4 in the map database 7 (S15). By writing such integrated risk Rnew, the evaluation of the storage risk Rdb in the map database 7 is accumulated, and a risk map is created. Note that the method of creating the storage risk Rdb is not limited to this, and it is only necessary to express the risk of the position on each road.

The memory risk Rall can be calculated as follows, for example. A plurality ( _nm ) of positions on a plurality (m) of routes from the current position (x, y) of the host vehicle to the destination position (xg, yg) are obtained. For example, the position (x _1i , x _1i ) on the route of the route 1 and the position (x _mi , x _mi ) on the route of the route m are as follows.
(X _1i , x _1i ) = {(x ₁₁ , x ₁₁ ) (x ₁₂ , y ₁₂ )... (X _1n1 , y _1n1 )}
(X _mi , x _mi ) = {(x _m1 , x _m1 ) (x _m2 , y _m2 ) ... (x _mnm , y _mnm )}
Hereinafter, subscripts representing routes for simplifying the explanation are omitted. For example, the position on the route (x _i, y _i) to added the transit time t _i with reference _{(x i, y i, t} i) to the map database 7, corresponding to the position and passing time on the route The memory risk Rdb (x _i , y _i , t _i ) to be read is read. For example, passage time _{t i} is the current position (x, y) position on the route (x _i, y _i) a constant speed with the journey distance to (e.g. general road 30 km / h, such as highway 80 km / h It can be determined by dividing by a fixed value or the court speed set for the road.

Here, when there is rarely a memory risk Rdb (x _i , y _i , t _i ) corresponding to the position / passing time on the route, for example, there may be a case where the traffic volume is low or in an initial state. . For this reason, when there is no memory risk Rdb (x _i , y _i , t _i ) corresponding to the position / passing time on the route, it is stored at the point closest to the position / passing time on the route. The memory risk Rdb is acquired. Further, when the distance between the closest point from the position / passing time (x, y, t) on the route and the position / passing time (x, y, t) on the route exceeds a predetermined threshold Reads a predetermined number (for example, 2 to 5) of risk stored in each of a predetermined number of points close to the position / passing time on the route. When the storage risk Rdb (x _i , y _i , t _i ) corresponding to the position / passing time on the route is read, the storage risk Rall of all routes is calculated. The calculation of the memory risk Rall for all routes is performed according to the following procedure. For each route (1, 2,..., M), an average value of all memory risks Rdb (x _i , y _i , t _i ) on the route is obtained, and the memory risk of the route is calculated based on the average value. To do. The average value mentioned here is an example of a realization means, and the median value, mode value, minimum value, maximum value, average value + 2σ can be used as the storage risk of the route. By calculating the memory risk for all routes by the above-described method, the calculated memory risk can be set as the memory risk Rall for all routes.

  Returning to the flow shown in FIG. 4, when the risk map is created, the database reading unit 9 transmits the current position (x, y) of the host vehicle transmitted from the position sensor 5 and the destination setting unit 8. The destination position (xg, yg) obtained is acquired (S2). Subsequently, in the database reading unit 9, the current position (x, y) of the own vehicle and the position (xg, yg) of the destination are referred to the map database 7, and the current position (x, y) of the own vehicle is determined. The memory risk Rall of all routes to the destination position (xg, yg) is read (S3).

When the storage risk level is read, the normal risk level calculation unit 10 reads the total risk level stored in the map database 7, and calculates the normal risk level θ based on the total risk level (S4). The normal risk degree θ is calculated as follows, for example. First, an average and a variance σ ² are obtained from all the risks in the map database 7. Obtained here, an average + 3 [sigma] from the average and variance sigma ^2, it is possible to make this value as the normal risk theta.

  After calculating the normal risk level, the filter unit 11 compares the memory risk level Rall with the normal risk level θ, and extracts a memory risk level that is equal to or lower than the normal risk level from among the memory risk levels (S5). For example, as shown in FIG. 12, let us consider a case where there are ten routes whose risk levels are evaluated on the travel route from the starting point PS to the end point PG of the route through which the vehicle passes. Here, the risk of the route Ru1 between PS-P1 is “1”, the risk of the route Ru2 between P1 and P2 is “6”, the risk of the route Ru3 between P2 and PG is “1”, PS− The risk of the route Ru4 between P3 is “4”, the risk of the route Ru5 between P1 and P3 is “3”, the risk of the route Ru6 between P3 and P4 is “4”, and the route Ru7 between P4 and P2 The risk of route Ru8 between P1 and P4 is "2", the risk of route Ru9 between P3 and P2 is "6", and the risk of route Ru10 between P4 and PG is "6" ”. Further, it is assumed that the normal risk degree θ is “4”. In this case, as shown in FIG. 13, the route Ru2, Ru9, Ru10 has a risk level of “6”, which is a value exceeding the normal risk level of “4”.

  The filter unit 11 sets the normal risk degree θ as an upper limit value, excludes a route having a memory risk level exceeding the normal risk level from the memory risk level Rall, and a route having a memory risk level equal to or lower than the normal risk level and its route The route storage risk level is output to the optimum route search unit 12. In the example shown in FIG. 12, the route for the route Ru1, Ru3, Ru4, Ru5, Ru6, Ru7, Ru8 and the risk level of the route are output.

  The optimum route search unit 12 searches for the optimum route based on the route output from the filter unit 11 and the storage risk of the route (S6). Since the route used here is a route whose risk level is less than or equal to the normal risk level, a route that has a risk level exceeding the normal risk level is searched for a disconnected route. In the example shown in FIG. 12, the routes Ru2, Ru9, Ru10 having a degree of risk exceeding “4” are cut off from the starting point PS to the end point PG, and the route is searched. As the optimum route, for example, the shortest distance route can be obtained. As a method of calculating the shortest distance route, for example, the distance of an arbitrary route can be obtained and the Dijkstra method can be used. Thus, the shortest distance path is calculated. It is also possible to calculate the minimum time route and the minimum fee route by the same method. When the optimum route is obtained by calculating the shortest distance route, driving support based on the optimum route is performed (S7). Thus, the process by the driving support device is finished.

  As described above, in the travel support device according to the present embodiment, the travel route of the host vehicle is acquired by cutting the route in which the risk of the route on the travel route on which the host vehicle travels exceeds a predetermined upper limit value. For this reason, the route can be acquired while avoiding a route that is dangerous for the driver. Therefore, a route corresponding to the skill of the driver is obtained, and the driving support according to the skill of the driver is performed by using this route. be able to. In addition, for a route having a risk level exceeding the normal risk level, a cut route is searched. For this reason, the search time for searching for the optimum route can be shortened.

  Next, a second embodiment of the present invention will be described. Compared with the first embodiment, the driving support device according to the present embodiment has the same device configuration as that of the first embodiment, and the processing in the filter unit 11 is mainly different. Hereinafter, the present embodiment will be described focusing on the processing in the filter unit 11.

  The filter unit 11 in the driving support apparatus according to the present embodiment compares all the risk levels Rall output from the database reading unit 9 with the normal risk level θ, and the risk less than the normal risk level θ from the risk level Rall. The degree Rall is extracted. For the risk level extracted here, a risk level common to the normal risk level θ is given. The filter unit 11 outputs the extracted risk level and a route corresponding to the risk level to the optimum route search unit 12. Other points are the same as those in the first embodiment.

  In the filter unit 11 in the driving support apparatus according to the present embodiment, the normal risk degree θ is set as a lower limit value, and a risk degree common to the normal risk degree θ is assigned to the risk degree Rall below the normal risk degree θ. For example, in the case of the example shown in FIG. 12, as shown in FIG. 14 (a), the risk of the routes Ru2, Ru9, Ru10 exceeds the normal risk “4”, and the other routes For Ru1, Ru3, Ru4, Ru5, Ru6, Ru7, and Ru8, the risk is less than the normal risk. The risk levels for the routes Ru1, Ru3, Ru4, Ru5, Ru6, Ru7, and Ru8 are made common to the normal risk level θ as shown in FIG.

  In this way, by sharing the risk level in the route of the risk level that is less than or equal to the normal risk level, the route can be acquired without making a difference within the range of the risk level that can be accepted by the driver. it can. Therefore, it is possible to obtain a route according to the skill of the driver and use this route to provide driving support according to the skill of the driver. In addition, since it is possible to acquire a route according to other factors such as distance and time, for example, within a range that can be permitted by the driver, it is possible to perform driving support with improved efficiency as a distance and time.

  The preferred embodiment of the present invention has been described above, but the present invention is not limited to the above embodiment. For example, in the first embodiment, the normal risk level is set as the upper limit value, and in the second embodiment, the normal risk level is set as the lower limit value. However, the upper limit value and the lower limit value in both embodiments are set to other values. You can also. Further, in each of the above embodiments, the normal risk level θ is obtained as the threshold value, and the route of the risk level exceeding the normal risk level θ is cut, or the risk level equal to or lower than the normal risk level is shared. Further, it is possible to adopt a mode in which the driver sets a risk level that becomes a threshold value instead of the normal risk level. In this case, for example, a risk setting knob or the like may be provided at a position where the driver can operate. By making it possible to set the threshold value of the danger level in this way, it is possible to search for the optimum route with the risk level corresponding to the driver.

  Moreover, in the said embodiment, although the own vehicle risk level is calculated | required in the collision probability calculation part 20, the own vehicle risk level can also be calculated | required by another aspect. Furthermore, in the said embodiment, in the database reading part 21, the separation distance of the nearest point from the present position (x, y) of the own vehicle and the present position (x, y) of the own vehicle is below a predetermined threshold value. In this case, the memory risk level Rdb at the closest point from the current position (x, y) of the host vehicle is acquired, but a plurality of memory risk levels Rdb around it can also be acquired. In this case, as in the case where the distance between the closest point from the current position (x, y) of the host vehicle and the current position (x, y) of the host vehicle exceeds a predetermined threshold, a plurality of surrounding memories are stored. Regarding the risk level, the integrated risk level can be obtained and rewritten.

  Moreover, in the said embodiment, although the example suitable when the driver | operator who drives the own vehicle was specified was demonstrated, the map database according to a driver | operator is used about the vehicle by which a several driver | operator is assumed, for example. It can also be created. Furthermore, in the above-described embodiment, the map database is provided in the host vehicle. For example, the map database is installed in a base station or other vehicle away from the host vehicle, and the host vehicle and the other vehicle are It is also possible to connect the base station and acquire the risk level. In this case, the degree of danger can be calculated by exchanging information with other vehicles. Moreover, in each said embodiment, although other vehicles are assumed as an obstruction, living organisms, such as a passerby, can also be assumed, for example.

It is a block block diagram which shows the structure of a driving assistance device. It is a block block diagram which shows the structure of a risk map preparation part. It is a block block diagram which shows the structure of a collision probability calculation part. It is a flowchart which shows the process sequence of a traveling control apparatus. It is a flowchart which shows the process sequence of a risk map preparation part. It is a flowchart which shows the process sequence of a collision probability calculation part. It is a schematic diagram which shows typically the driving state of the own vehicle and another vehicle. It is a schematic diagram which shows typically the driving | running route which the own vehicle can take. It is a graph which shows the structure of the spatiotemporal environment where the prediction course of the own vehicle is one. It is a graph which shows the structure of the spatiotemporal environment where the prediction course of the own vehicle is multiple. It is a graph which shows the relationship between the own vehicle risk in the position of the own vehicle, and the memory risk in the surrounding position. It is explanatory drawing which shows the example of the path | route from the starting point PS to the end point PG. It is a graph which shows the relationship between normal risk and memory risk. (A) is a graph which shows the relationship between normal risk and memory | storage risk, (b) is a graph which shows the state which wrote in common the risk below normal risk.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Risk map creation part, 2 ... Obstacle sensor, 3 ... Obstacle extraction part, 4 ... Own vehicle sensor, 5 ... Position sensor, 6 ... Timer, 7 ... Map database, 8 ... Destination setting part, 9 ... Database reading unit, 10 ... Normal risk calculation unit, 11 ... Filter unit, 12 ... Optimal route search unit, 13 ... Driving support unit, 21 ... Database reading unit, 22 ... Statistical processing unit, 23 ... Database writing unit, 31 ... obstacle information temporary storage unit, 32 ... obstacle possible route calculation unit, 33 ... own vehicle route recording unit, 34 ... own vehicle route reading unit, 35 ... own vehicle position reading unit, 36 ... own vehicle possible route calculation unit, 37 ... Realized track collision probability calculation unit, 38 ... Best vehicle track collision probability calculation unit, 39 ... Self vehicle risk calculation unit, Rcur ... Self vehicle risk level, Rdb ... Memory risk level, Rnew ... Total risk level, Env ... Spatio-temporal environment, θ ... Normal danger , H, H1, H2 ... other vehicles, M ... vehicle.

Claims (4)

In a driving support device that supports driving of the host vehicle,
A risk evaluation means for evaluating the risk of a route that is a part of a travel path on which the host vehicle travels based on a possibility that the vehicle collides with an obstacle ;
Route acquisition means for acquiring a travel route of the host vehicle formed by connecting a plurality of the routes based on the evaluation result of the degree of risk,
The route acquiring unit, among the plurality of routes, the degree of risk is evaluated by connecting a plurality of routes selected by cutting a route exceeds a predetermined upper limit value, and acquires the travel route of the vehicle A vehicle travel support device characterized by the above. In a driving support device that supports driving of the host vehicle,
A risk evaluation means for evaluating the risk of a route that is a part of a travel path on which the host vehicle travels based on a possibility that the vehicle collides with an obstacle ;
Route acquisition means for acquiring a travel route of the host vehicle formed by connecting a plurality of the routes based on the evaluation result of the degree of risk,
The route acquiring unit, among the plurality of routes, by connecting a plurality of routes risk evaluated was a common risk of routes equal to or less than a predetermined lower limit value, obtains the travel route of the vehicle A vehicle travel support device. The vehicle travel support apparatus according to claim 1, wherein the risk degree of the route is evaluated based on a risk degree on the route or a point closest to the route. A risk storage means for storing the estimated risk of the route;
The vehicle travel support apparatus according to any one of claims 1 to 3, wherein the risk degree of the route stored in the stored risk degree storage unit is used as the risk degree of the route.

Images