JP7193202B2 - Motion prediction method and motion prediction device - Google Patents

Description

The present invention relates to a motion prediction method and a motion prediction device.

Patent Literature 1 discloses a vehicle travel control device that issues an alarm and/or controls the own vehicle in accordance with the operation of another vehicle. The vehicle travel control device is a relative speed at which the adjacent vehicle moves away from the own vehicle even if the inter-vehicle distance between the adjacent vehicle and the own vehicle is shorter than the set inter-vehicle distance when the adjacent vehicle changes lanes into the travel lane of the own vehicle. and/or limiting means for limiting the control for decelerating the own vehicle.

Japanese Patent Application Laid-Open No. 2005-199930

The behavior of a vehicle differs depending on the driving characteristics of the vehicle, i.e., how the driver tends to drive, and even if the vehicle is driven by the same driver, the driving characteristics change according to changes in the surrounding environment. Sometimes.
For this reason, even if the alarm and/or the own vehicle is controlled based on the relative speed between the other vehicle and the own vehicle as in Patent Document 1, for example, the difference in behavior of the other vehicle due to the difference in driving characteristics and the change in the surrounding environment. Occupants of one's own vehicle may feel uncomfortable due to a change in the behavior of another vehicle that accompanies a change in driving characteristics due to the change.
SUMMARY OF THE INVENTION An object of the present invention is to improve the accuracy of predicting the behavior of another vehicle based on the driving characteristics of the other vehicle.

According to one aspect of the present invention, there is provided a motion prediction method for predicting motions of other vehicles in the vicinity of the host vehicle. In this motion prediction method, a first driving characteristic of another vehicle is estimated based on a specific first driving environment, and based on the first driving characteristic, a first driving characteristic of the other vehicle is estimated in a second driving environment different from the first driving environment. 2) Driving characteristics are calculated, and when the driving environment of the other vehicle changes from the first driving environment to the second driving environment, the behavior of the other vehicle is predicted based on the second driving characteristics.

According to the present invention, it is possible to improve the accuracy of predicting the behavior of another vehicle based on the driving characteristics of the other vehicle.

It is a figure which shows an example of the operation|movement of another vehicle predicted by the operation|movement prediction method of embodiment. 1 is a block diagram showing an example of a schematic configuration of a driving support device according to an embodiment; FIG. FIG. 5 is a diagram showing another example of the motion of another vehicle predicted by the motion prediction method of the embodiment; FIG. 10 is an explanatory diagram of an example of a difference between a basic trajectory predicted as a motion candidate and an effective trajectory predicted based on actual behavior; FIG. 11 is an explanatory diagram of another example of the difference between the basic trajectory predicted as a motion candidate and the effective trajectory predicted based on the actual behavior; 4 is a flow chart showing an example of the operation of the driving assistance device of the embodiment; 4 is a flowchart showing details of an example of another vehicle motion prediction step; 4 is a flow chart showing details of an example of a motion candidate prediction step; 4 is a flow chart showing details of an example of a driving characteristic estimation step; 4 is a flowchart showing details of another example of the driving characteristic estimation step; 4 is a flow chart showing details of an example of a driving characteristic modification step;

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described with reference to the drawings. Note that each drawing is schematic and may differ from the actual one. Further, the embodiments of the present invention shown below are examples of apparatuses and methods for embodying the technical idea of the present invention. are not specific to the following: Various modifications can be made to the technical idea of the present invention within the technical scope defined by the claims.

Please refer to FIG. The driving support device of the embodiment predicts the motion of the other vehicle 2 and supports the own vehicle 1 based on the prediction result. Here, driving assistance includes automatic driving control that does not involve driving operation by the driver who is a passenger, and driving assistance that assists the driver's driving of the own vehicle 1 by intervening in the driving by the driver.
For example, in the driving scene shown in FIG. 1, the own vehicle 1 is traveling in the right lane, and the other vehicle 2 is traveling in the left lane diagonally in front of the own vehicle 1 . Furthermore, the preceding vehicle 3 of the other vehicle 2 is traveling ahead of the other vehicle 2 in the same driving lane as the other vehicle 2 .

When the speed of the preceding vehicle 3 is slower than the speed of the other vehicle 2, the inter-vehicle distance between the preceding vehicle 3 and the other vehicle 2 gradually becomes shorter. In such a driving scene, as an action of the other vehicle 2, trajectories t2 and t3 in which the other vehicle 2 changes lanes in front of the own vehicle 1 are conceivable.
On the other hand, the driver of the other vehicle 2 continues traveling in the left lane while decelerating in consideration of the own vehicle 1, and does not change lanes to the right lane until the own vehicle 1 overtakes the other vehicle 2. t1 is also conceivable.

Which of these trajectories t1 to t3 the operation of the other vehicle 2 follows depends on the "driving characteristics" of the other vehicle 2, that is, what kind of driving operation the driver of the other vehicle 2 tends to perform.
For example, when the driver tends to drive selfishly (for example, relatively abrupt driving), the trajectory t3 that cuts in front of the vehicle 1 in a relatively short lane change time is selected. It can be determined that there is a high possibility that

On the other hand, if the driver has a tendency to drive cautiously (for example, if the driver is a beginner), it can be determined that there is a high possibility that the trajectory t1 that gives way to the own vehicle 1 will be selected. In addition, an average driver, who is between a driver who drives selfishly and a driver who drives cautiously, chooses a trajectory t2 that changes lanes in front of own vehicle 1 in a relatively long lane change time. You can decide by choosing.
By predicting the trajectory of the other vehicle 2 in consideration of the driving characteristics of the other vehicle 2 in this way, the accuracy of prediction of the motion of the other vehicle 2 is enhanced. As a result, it is possible to predict the movement of the other vehicle 2 that will stray into the own lane of the own vehicle 1, so that the sudden behavior of the own vehicle 1 can be prevented, and the sense of discomfort given to the occupants of the own vehicle 1 can be reduced.

(composition)
Please refer to FIG. A driving support device 10 according to the embodiment includes an object detection device 11 , an own vehicle position estimation device 12 , a map acquisition device 13 and a controller 14 .
The object detection device 11 includes a plurality of different types of object detection sensors such as a laser radar, a millimeter wave radar, and a camera mounted on the vehicle 1 that detect objects around the vehicle 1 . The object detection device 11 detects objects around the vehicle 1 using a plurality of object detection sensors.

The object detection device 11 detects moving objects including other vehicles, motorcycles, bicycles and pedestrians, and stationary objects including parked vehicles. For example, it detects the position, attitude, size, speed, acceleration, deceleration, and yaw rate of a moving object and a stationary object with respect to the own vehicle 1 .
In the following description, the position, orientation (yaw angle), size, velocity, acceleration, deceleration, and yaw rate of an object are collectively referred to as "behavior" of the object. The object detection device 11 outputs the behavior of a two-dimensional object as a detection result, for example, in a zenith view (also referred to as a plan view) viewed from above the own vehicle 1 in the air.

The vehicle position estimation device 12 includes a GPS (Global Positioning System) receiver mounted on the vehicle 1 and a position detection sensor such as odometry that measures the absolute position of the vehicle 1 . The own vehicle position estimation device 12 uses a position detection sensor to measure the absolute position of the own vehicle 1, that is, the position, attitude and speed of the own vehicle 1 with respect to a predetermined reference point.

The map acquisition device 13 acquires map information indicating the structure of the road on which the vehicle 1 travels. The map acquisition device 13 may own a map database storing map information, or may acquire map information from an external map data server by cloud computing. The map information acquired by the map acquisition device 13 includes road structure information such as the absolute position of lanes, the connection relationship between lanes, and the relative positional relationship.

Further, the map acquisition device 13 acquires frequently updated map information (for example, information embedded in a dynamic map). Specifically, the map acquisition device 13 provides dynamic information updated at a frequency of one second or less, semi-dynamic information updated at a frequency of one minute or less, semi-static information updated at a frequency of one hour or less. Information is obtained from the outside of the own vehicle 1 by wireless communication.
For example, dynamic information includes information on surrounding vehicles, pedestrians, and traffic lights; semi-dynamic information includes accident information, congestion information, and narrow-area weather information; Includes regulatory information, road construction information, and regional weather information. On the other hand, the "map information indicating the structure of roads" described above corresponds to static information that is updated at a frequency of one hour or less.

The controller 14 predicts the movement of the other vehicle 2 based on the detection results by the object detection device 11 and the vehicle position estimation device 12 and the information acquired by the map acquisition device 13, and calculates the route of the vehicle 1 from the movement of the other vehicle 2. and control the own vehicle 1 according to the generated route.
Controller 14 may be a computer that includes a processor and peripheral components such as storage devices. The processor may be, for example, a CPU (Central Processing Unit) or an MPU (Micro-Processing Unit).

The storage device may comprise any one of a semiconductor storage device, a magnetic storage device and an optical storage device. The storage device may include memories such as registers, cache memories, ROMs (Read Only Memories) used as main storages, and RAMs (Random Access Memories).
Note that the controller 14 may be formed of dedicated hardware for executing each information processing described below.
For example, the controller 14 may be realized by a functional logic circuit set in a general-purpose semiconductor integrated circuit. For example, the controller 14 may comprise a programmable logic device (PLD), such as a Field-Programmable Gate Array (FPGA), or the like.

In addition, in the embodiment, the controller 14 will be described as an example of a driving support device that controls the own vehicle 1 . However, the invention is not so limited. For example, the controller 14 can be implemented as a motion prediction device that predicts the motion of other vehicles. In other words, the controller 14 may not generate the route for the own vehicle 1 and assist the traveling along the route, and may finally output the predicted motion of the other vehicle. The controller 14 is an example of the motion prediction device described in the claims.

The controller 14 includes a detection integration unit 20 , an object tracking unit 21 , an intra-map position calculation unit 22 , a motion prediction unit 23 , an own vehicle route generation unit 24 , and a vehicle control unit 25 . The controller 14 executes a computer program stored in a predetermined storage device with a processor to perform a detection integration unit 20, an object tracking unit 21, a map position calculation unit 22, a motion prediction unit 23, and an own vehicle route generation unit 24. , and the functions of the vehicle control unit 25 may be realized.

The detection integration unit 20 integrates a plurality of detection results obtained from each of the plurality of object detection sensors included in the object detection device 11, and outputs one detection result for each object.
Specifically, from the behavior of the object obtained from each of the object detection sensors, the most rational behavior of the object with the least error is calculated after considering the error characteristics of each of the object detection sensors.
Specifically, by using a known sensor fusion technique, detection results obtained by a plurality of types of sensors are comprehensively evaluated to obtain more accurate detection results.

The object tracking section 21 tracks the object detected by the object detection device 11 . Specifically, from the detection results integrated by the detection integrating unit 20, the identity of the object between different times is verified (associated) from the behavior of the object output at different times, and the correspondence is performed. based on the prediction of the behavior of the object. The behavior of the object output at different times is stored in the memory within the controller 14 and used for trajectory prediction, which will be described later.

The in-map position calculator 22 estimates the position and orientation of the vehicle 1 on the map from the absolute position of the vehicle 1 obtained by the vehicle position estimation device 12 and the map data obtained by the map acquisition device 13. do.
Further, the in-map position calculation unit 22 identifies the road on which the vehicle 1 is traveling and the lane on which the vehicle 1 is traveling on the road.

The motion prediction unit 23 predicts the motion of a moving object around the vehicle 1 based on the detection result obtained by the detection integration unit 20 and the position of the vehicle 1 specified by the intra-map position calculation unit 22. .
The motion prediction unit 23 includes a behavior determination unit 30, a motion candidate prediction unit 31, a motion candidate correction unit 32, a trajectory prediction unit 33, a driving characteristic estimation unit 34, a driving characteristic correction unit 35, and a likelihood estimation unit. 36.

The behavior determination unit 30 identifies the position and behavior of the object on the map from the position of the vehicle 1 on the map and the behavior of the object obtained by the detection integration unit 20 .
Furthermore, when the position of an object on the map changes over time, the behavior determination unit 30 determines that the object is a “moving object”, and determines the attributes of the moving object from the size and speed of the moving object. (Other vehicles and pedestrians in motion).

When determining that the moving object is the "other vehicle" that is traveling, the behavior determining unit 30 determines the road and lane on which the other vehicle is traveling.
On the other hand, if the position of the object on the map does not change with the passage of time, the behavior determination unit 30 determines that this object is a stationary object, and based on the position, posture, and size of the stationary object on the map, the stationary object attributes (such as other stopped vehicles, parked vehicles, and pedestrians).

The motion candidate prediction unit 31 predicts motion candidates of other vehicles based on the map. The motion candidate prediction unit 31 predicts the "motion intention" of how the other vehicle will travel next from the road structure and the lane information to which the other vehicle belongs, which is included in the map information, and predicts the "motion intention" based on the motion intention. The "basic trajectory" of other vehicles is calculated based on the road structure.
“Motion candidate” is a higher concept including motion intention and basic trajectory. The basic trajectory shows not only the profile of the other vehicle's position at different times, but also the profile of the other vehicle's velocity at each position.

For example, when another vehicle travels on a single-lane road and a curved road, the motion candidate prediction unit 31 predicts the motion intention (straight ahead) to travel along the shape of the lane, and uses the lane on the map as the basic trajectory. Calculate the trajectory along .
Further, when the other vehicle travels on a single road and a curved road with multiple lanes, the motion candidate prediction unit 31 predicts the motion intention (straight ahead) and the motion intention to change lanes to the right or left (lane change). The basic trajectory of the other vehicle in the action intent (lane change) is the trajectory to change lanes based on road structure and predetermined lane change times.

Furthermore, when driving through an intersection, the motion candidate prediction unit 31 predicts motion intentions of going straight, turning right, and turning left, and calculates the straight trajectory, right-turning trajectory, and left-turning trajectory based on the road structure at the intersection on the map as basic trajectories. .
Although the road structure is considered in the calculation of the "basic trajectory", the moving and stationary objects integrated by the detection integration unit 20 are not considered.

For example, in the driving scene shown in FIG. 1 , the motion candidate prediction unit 31 calculates the motion intention (straight ahead) of traveling along the left lane and the basic trajectory t1 as motion candidates for the other vehicle 2 .
However, the motion candidate prediction unit 31 does not calculate the motion intention (lane change) caused by the preceding vehicle 3, which is a moving object around the own vehicle 1, and the lane change trajectories t2 and t3. The same is true for motion candidates caused by stationary objects.

The motion candidate correction unit 32 corrects the motion candidate predicted by the motion candidate prediction unit 31 in consideration of the stationary object detected by the object detection device 11 .
Specifically, the motion candidate correction unit 32 determines whether or not the base trajectory of the other vehicle interferes with the position of the stationary object. In the case of interference, the motion intention and basic trajectory of the other vehicle 2 to avoid the stationary object are newly added.

When a moving object other than the other vehicle 2 shown in FIG. 1 (for example, the preceding vehicle 3) is detected by the object detection device 11, the motion candidate correction unit 32 causes the motion candidate prediction unit 31 to Modify predicted motion candidates.
Specifically, the motion candidate correction unit 32 determines over time whether or not there is interference between another moving object and the other vehicle 2 . When there is interference between moving objects, the motion intention and basic trajectory of the other vehicle 2 for avoiding interference with other moving objects are newly added.

At this time, the motion candidate correction unit 32 generates a plurality of different basic trajectories as motion candidates for the other vehicle 2 according to the driving characteristics of the other vehicle 2 .
For example, a plurality of different basic trajectories may be generated depending on whether the driving characteristics of the other vehicle 2 tend to drive selfishly or to drive cautiously.
In this embodiment, three types of driving characteristics including "Aggressive", "Ordinary" and "Cautious" are illustrated as examples of the driving characteristics of the other vehicle 2. FIG.

Aggressive is a driving characteristic with a tendency to drive egocentrically, and has a tendency to perform relatively abrupt driving maneuvers. In other words, the propensity for safe driving is relatively weak and the predictability of behavior is relatively low.
Conversely, Cautious is a driving characteristic that tends to drive cautiously, with a tendency to perform relatively sluggish driving maneuvers and drive at speeds below average surrounding vehicles and legal speeds. In other words, the propensity for safe driving is relatively strong and the predictability of behavior is relatively high.

Ordinary is a driving characteristic that tends to perform an average driving operation, and has a tendency to perform a driving operation that is slower than Aggressive and sharper than Cautious. In addition, the tendency of safe driving is stronger than Aggressive, and the tendency of safe driving is weaker than Cautious. It also has higher predictability than Aggressive.
When the driving characteristics of the other vehicle 2 have not yet been estimated, "Unknown" is used as the state indicating that the driving characteristics have not yet been estimated.

The driving characteristics used for predicting the behavior of the other vehicle 2 are not limited to the driving characteristics illustrated above, and other driving characteristics useful for predicting the behavior of the other vehicle 2 may be employed.
Aggressive is an example of a first property recited in the claims, and Cautious and Ordinary are examples of a second property recited in the claims. Further, when Ordinary is taken as an example of the first characteristic, Cautious may be taken as an example of the second characteristic.

FIG. 3 shows an example of a basic trajectory of a motion intention (lane change) to overtake the preceding vehicle 3 as a motion candidate of the other vehicle 2 . The motion candidate correction unit 32 generates basic trajectories t11, t12, and t13 with different lane change times when the driving characteristics of the other vehicle 2 are Aggressive, Ordinary, and Cautious.
For example, the lane change time when the driving characteristic of the other vehicle 2 is Aggressive may be 3 seconds or less, when it is Cautious, it may be 5 seconds or more, and when it is Ordinary, it may be longer than 3 seconds. It may be long and shorter than 5 seconds.

For example, the motion candidate correction unit 32 determines the lane change time according to the driving characteristic based on the center of the driving lane of the other vehicle 2 before the lane change indicated by the map information and the center of the driving lane of the other vehicle 2 after the lane change. The basic trajectories t11 to t13 may be generated using the clothoid curve with . The curve for generating the basic trajectories t11 to t13 does not necessarily have to be the clothoid curve, and other curves may be used.
Note that the basic trajectory when the driving characteristics are Ordinary (for example, the basic trajectory t12) may also be used as the basic trajectory when the driving characteristics are not estimated (that is, in the case of Unknown).

Please refer to FIG. For example, the motion candidate correction unit 32 generates basic trajectories t3 and t2 with different lane change times when the driving characteristics of the other vehicle 2 are Aggressive and Ordinary, respectively, and when the driving characteristics of the other vehicle 2 are Cautious. may generate a trajectory t1 that gives way to the own vehicle 1.
In addition to or instead of the lane change time, a plurality of basic trajectories with different lane change start timings may be generated as the plurality of basic trajectories that differ according to the driving characteristics.

Note that the motion candidate prediction unit 31 may also generate a plurality of different basic trajectories as motion candidates for the other vehicle 2 according to the driving characteristics of the other vehicle 2 . That is, a plurality of different trajectories may be generated according to the driving characteristics of the other vehicle 2 even for the basic trajectory calculated based on the road structure regardless of whether the moving or stationary object other than the other vehicle 2 is present.
For example, the motion candidate prediction unit 31 predicts the movement of the other vehicle 2 regardless of moving or stationary objects other than the other vehicle 2, even when going straight on the travel road, when changing lanes, when going straight at an intersection, and when turning right or left. A basic trajectory may be generated for each of the cases in which the driving characteristics are Aggressive, Ordinary, and Cautious.

Please refer to FIG. The trajectory prediction unit 33 predicts a trajectory (effective trajectory) taken by the other vehicle 2 based on the actual behavior of the other vehicle 2 detected by the behavior determination unit 30 .
Specifically, the trajectory prediction unit 33 calculates the effective trajectory of the other vehicle 2 when operating according to the predicted motion intention, using a known state estimation technique such as a Kalman filter, for example.

The "effective trajectory" not only indicates the position of the other vehicle 2 at different times, but also indicates the profile of the speed of the other vehicle 2 at each position in the same manner as the basic trajectory. Both the effective trajectory and the basic trajectory are common in that they are trajectories taken by the other vehicle 2, but while the effective trajectory is calculated considering the behavior of the other vehicle 2, the basic trajectory is calculated without considering the behavior of the other vehicle. Both are different in that they are calculated.

The basic trajectories t21 and t22 shown in FIGS. 4A and 4B are examples of trajectories of the other vehicle 2 derived based on the motion intention and the road structure, and the behavior of the other vehicle 2 is not considered. Therefore, for example, since the current attitude (yaw angle) of the other vehicle 2 is not taken into consideration, a plurality of basic trajectories t21 and t22 extend from the current position of the other vehicle 2 in different directions. On the other hand, the trajectory prediction unit 33 calculates a trajectory (effective trajectory) in line with the above-described motion intention in consideration of the behavior of the other vehicle 2 . In other words, the effective trajectory of the other vehicle 2 is calculated when the vehicle 2 moves in accordance with the motion intention described above.
The trajectories t1 to t3 and t11 to t13 shown in FIGS. 1 and 3 are examples of basic trajectories of the other vehicle 2 derived based on the motion intention of the other vehicle 2 and the road structure.

4A and 4B, the attitude (yaw angle) of the other vehicle 2 is inclined to the left of the basic trajectory t21 of traveling along the shape of the road, and the speed of the other vehicle 2 consists only of the speed component in the traveling direction. , the velocity component in the vehicle width direction is zero. In other words, the other vehicle 2 is in a straight-ahead state.
Therefore, when the other vehicle 2 travels according to the motion intention of traveling along the shape of the road with this posture and speed as the starting point, as shown in FIG. , the effective trajectory t31 coincides.

In other words, it is expected that a corrected trajectory (overshoot trajectory) will be drawn to correct the departure from the driving lane. The trajectory prediction unit 33 predicts an effective trajectory t31 to travel according to the motion intention (straight ahead) of travel along the shape of the road, starting from the attitude (yaw angle) and speed of the other vehicle 2 .
Next, when the other vehicle 2 travels according to the intention of lane change with the same posture and speed as the starting point, as shown in FIG. 4B, it starts turning to the left, moves to the left lane, and then turns to the right. Then, the effective trajectory t32 is corrected to the trajectory along the left lane.

In other words, an effective trajectory t32 is drawn which consists of a left-turning clothoid curve and a right-turning clothoid curve starting from the state where the steering angle is in the neutral position. Therefore, the effective trajectory t32 becomes a trajectory in which the lane change is completed in approximately the same time as the "predetermined lane change time" when calculating the lane change trajectory t22. The curve for drawing the effective trajectory does not necessarily have to be the clothoid curve, and other curves may be used for drawing. In the example of FIGS. 4A and 4B, the effective trajectory t32 is substantially the same trajectory as the basic trajectory t22 in lane change.

4A and 4B, for the basic trajectories t1 to t3 and t11 to t13 shown in FIGS. 1 and 3, the trajectory prediction unit 33 considers the behavior of the other vehicle 2, Calculate the trajectory (effective trajectory).
For example, in the driving scenes shown in FIGS. 1 and 3, starting from the position, attitude (yaw angle), and speed of the other vehicle 2, there is an intention to change lanes in front of the own vehicle 1, or until the own vehicle 1 overtakes. The effective trajectory of the other vehicle 2 traveling according to the intention of continuing to travel in the left lane while decelerating without changing lanes to the right lane is calculated.
Here, as an example of the behavior of the other vehicle 2, the position, attitude, and speed are taken into consideration, but the acceleration and deceleration of the other vehicle 2 may be considered to calculate the effective trajectory. For example, it can be predicted that the deceleration will be greater when changing lanes than when traveling straight ahead.

The predicted trajectory is an example of the first predicted trajectory predicted based on the driving characteristics described in the claims. The effective trajectory is an example of the second predicted trajectory predicted based on the actual behavior of the other vehicle described in the claims.

Next, the driving characteristic estimator 34 estimates the driving characteristic of the other vehicle 2 based on the specific driving environment around the other vehicle 2 . The driving characteristic estimated by the driving characteristic estimating unit 34 determines which of the motion candidates predicted by the motion predicting unit 23 and the motion candidate correcting unit 32 (that is, a plurality of motion candidates that differ according to the driving characteristics). Used to determine what to predict as behavior.
For example, the driving characteristic estimator 34 may estimate which one of the Aggressive, Ordinary, and Cautious driving characteristics of the other vehicle 2 is.

For example, the driving characteristic estimator 34 may estimate the driving characteristic of the other vehicle 2 from the vehicle behavior of the other vehicle 2 in this driving environment based on a specific driving environment. For example, the driving characteristic estimator 34 may estimate the driving characteristic of the other vehicle 2 based on the turn signal operation of the other vehicle 2 when changing lanes.
For example, assume a driving scene in which another vehicle 2 changes lanes as shown in FIG. The driving characteristic estimator 34 calculates an intersection area 4 where the predicted route t4 of the own vehicle 1 and the predicted trajectory t3 of the other vehicle 2 intersect when the other vehicle 2 changes lanes.

Then, the arrival time at which the other vehicle 2 reaches the intersection area 4 is estimated, and the time difference between the arrival time at the intersection area 4 and the turn signal lighting start time is calculated. When the calculated time difference is equal to or greater than the first threshold value X1, the driving characteristic of the other vehicle 2 is estimated to be cautious, and the driving characteristic of the other vehicle 2 is less than the first threshold value X1 and equal to or greater than the second threshold value X2 (second threshold value X2<first threshold value X1). If it is less than the second threshold value X2, it may be estimated to be Aggressive.

Alternatively, the driving characteristic estimation unit 34 may estimate the likelihood that the driving characteristic of the other vehicle 2 is Aggressive, the likelihood that it is Ordinary, and the likelihood that it is Cautious.
Hereinafter, the likelihood that the driving characteristic of the other vehicle 2 is Aggressive, the likelihood that it is Ordinary, and the likelihood that it is Cautious are referred to as "aggressive likelihood,""ordinarylikelihood," and "cautious likelihood," respectively.

The driving characteristic estimating unit 34 calculates the likelihoods of the plurality of motion candidates predicted by the motion predicting unit 23 and the motion candidate correcting unit 32 (i.e., motion candidates that differ according to the driving characteristics). (that is, the likelihood of being selected as the action of the other vehicle 2) is stored in the storage device.
For example, the aggressive likelihood is the likelihood that the other vehicle 2 will pass through the basic trajectory t11 in FIG. be.
The likelihood that each motion candidate predicted based on the estimation of the driving characteristics of the other vehicle 2 will be the motion of the other vehicle 2 is denoted as "first likelihood α1".

The driving characteristic estimation unit 34 sequentially increases or decreases these likelihoods according to the vehicle behavior of the other vehicle 2 .
For example, when the time difference between the arrival time at the intersection area 4 and the turn signal lighting time is equal to or greater than the first threshold value X1, the Cautious likelihood is increased and the Ordinary likelihood and Aggressive likelihood are decreased. In the case of two thresholds X2 or more, the ordinary likelihood is increased and the cautious likelihood and the aggressive likelihood are decreased, and in the case of less than the second threshold X2, the aggressive likelihood is increased and the cautious likelihood and the ordinary likelihood are decreased. you can

Further, for example, the driving characteristic estimator 34 may estimate the driving characteristic of the other vehicle 2 based on the lane change time.
Specifically, when the lane change time is equal to or greater than the third threshold value X3, the driving characteristic of the other vehicle 2 is estimated to be cautious, and when less than the third threshold value X3, the fourth threshold value X4 or greater (fourth threshold value X4<third If the threshold value X3), it may be estimated to be Ordinary, and if it is less than the fourth threshold value X4, it may be estimated to be Aggressive.
Alternatively, when the lane change time is greater than or equal to the third threshold X3, the Cautious likelihood is increased and the Ordinary likelihood and Aggressive likelihood are decreased, and when the time is less than the third threshold X3 and is greater than or equal to the fourth threshold X4, the Ordinary likelihood While increasing, the cautious likelihood and the aggressive likelihood may be decreased, and in the case of less than the fourth threshold value X4, the aggressive likelihood may be increased and the cautious likelihood and the ordinary likelihood may be decreased.

In addition, for example, when the speed of the other vehicle 2 is lower than the average speed of the surrounding vehicles or the legal speed as a specific driving environment, the driving characteristics of the other vehicle 2 are It may be estimated to be cautious, or the cautious likelihood may be increased and the ordinary likelihood and aggressive likelihood may be decreased.
Alternatively, when the other vehicle 2 overtakes the preceding vehicle 3 in a lane other than the overtaking lane, or when the relative speed to the preceding vehicle 3 is equal to or greater than the threshold and the inter-vehicle distance is equal to or less than the threshold, the driving characteristic of the other vehicle 2 is determined to be aggressive. Alternatively, the aggressive likelihood may be increased and the cautious likelihood and the ordinary likelihood may be decreased.

Based on the driving characteristics of the other vehicle 2 estimated by the driving characteristic estimating unit 34, one of the plurality of motion candidates (i.e., motion candidates that differ according to the driving characteristics) predicted by the motion predicting unit 23 and the motion candidate correcting unit 32 is selected. By predicting whether or not as the motion of the other vehicle 2, it becomes possible to predict the motion of the other vehicle 2 in consideration of the driving characteristics of the other vehicle 2, so that the prediction accuracy increases.

However, the driving characteristics of the other vehicle 2 may change due to changes in the surrounding environment.
For example, when the preceding vehicle 3 of the other vehicle 2 continues to travel at a speed lower than the average traveling speed of the vehicle or the legal speed (for example, when the driving characteristic of the preceding vehicle 3 is Cautious), the usual driving characteristics of the other vehicle 2 Even if is Ordinary or Cautious, the vehicle may drive more aggressively than usual in order to overtake the preceding vehicle 3, and may behave in the same manner as a vehicle with aggressive driving characteristics.
Further, even if the normal driving characteristics of the other vehicle 2 are Aggressive, when the road surface is low μ or the forward visibility is poor, the other vehicle 2 behaves in the same manner as the vehicle whose driving characteristics are Ordinary or Cautious. Sometimes.

For this reason, the driving characteristic correction unit 35 changes when the surrounding environment of the other vehicle 2 changes (that is, when the driving environment surrounding the other vehicle 2 changes from a specific first driving environment to a second driving environment). Then, the driving characteristics of the other vehicle 2 estimated by the driving characteristics estimation unit 34 are corrected. That is, based on the driving characteristics estimated by the driving characteristics estimating unit 34, the driving characteristics obtained by modifying the driving characteristics are calculated.
The driving characteristics estimated by the driving characteristics estimation unit 34 and the driving characteristics corrected by the driving characteristics correction unit 35 are examples of the first driving characteristics and the second driving characteristics, respectively.
For example, when the preceding vehicle 3 continues to travel at a speed lower than the average vehicle speed or legal speed (for example, when the driving characteristic of the preceding vehicle 3 is Cautious), the driving characteristic correction unit 35 The characteristic may be changed from Cautious or Ordinary to Aggressive. Alternatively, it may be changed from Cautious to Ordinary.

Hereinafter, changing the driving characteristic to Aggressive includes increasing the Aggressive likelihood and decreasing the Cautious likelihood and the Ordinary likelihood.
Changing the driving characteristic to Ordinary includes increasing the Ordinary likelihood and decreasing the Cautious and Aggressive likelihoods.
Changing the driving characteristic to Cautious includes increasing the Cautious likelihood and decreasing the Ordinary and Aggressive likelihoods.

Further, for example, when the other vehicle 2 and the preceding vehicle 3 approach each other, the driving characteristic of the other vehicle 2 may be changed from Cautious or Ordinary to Aggressive. Alternatively, it may be changed from Cautious to Ordinary.
Further, for example, when an obstacle such as a falling object or road construction appears in front of the other vehicle 2, the driving characteristic of the other vehicle 2 may be changed from Cautious or Ordinary to Aggressive. Alternatively, it may be changed from Cautious to Ordinary.

Further, for example, when the vehicle at the end of the traffic jam becomes the preceding vehicle 3 of the other vehicle 2, the driving characteristic of the other vehicle 2 may be changed from Cautious or Ordinary to Aggressive. Alternatively, it may be changed from Cautious to Ordinary.
Further, for example, when the preceding vehicle 3 of the other vehicle 2 is a bus and is predicted to stop, the driving characteristic of the other vehicle 2 may be changed from Cautious or Ordinary to Aggressive. Alternatively, it may be changed from Cautious to Ordinary.

On the other hand, when the road surface in front of the other vehicle 2 changes to low μ due to rain or snow, for example, the driving characteristic of the other vehicle 2 may be changed from Aggressive to Ordinary or Cautious. Alternatively, it may be changed from Ordinary to Cautious.
Further, for example, when visibility in front of the other vehicle 2 deteriorates due to snow, rain, fog, or the like, the driving characteristic of the other vehicle 2 may be changed from Aggressive to Ordinary or Cautious. Alternatively, it may be changed from Ordinary to Cautious.

Further, for example, when the other vehicle 2 is expected to travel in a place with many passers-by (for example, a school road of an elementary school or a road near a downtown area) based on the map information, the driving characteristic of the other vehicle 2 is set to Aggressive. to Ordinary or Cautious. Alternatively, it may be changed from Ordinary to Cautious.
Further, for example, based on the map information, it is determined that there is a police-related facility (for example, an automatic speed control device, a police station, a police station, or other police facility) in front of the other vehicle 2 and the other vehicle 2 passes by the police-related facility. If so, the driving characteristic of the other vehicle 2 may be changed from Aggressive to Ordinary or Cautious. Alternatively, it may be changed from Ordinary to Cautious.

Further, for example, when police vehicles are present around the other vehicle 2, the driving characteristic of the other vehicle 2 may be changed from Aggressive to Ordinary or Cautious. Alternatively, it may be changed from Ordinary to Cautious.
Hereinafter, the driving characteristics of the other vehicle 2 corrected by the driving characteristic correcting unit 35 (the driving characteristics estimated by the driving characteristic estimating unit 34 when there is no correction by the driving characteristic correcting unit 35) are simply referred to as "estimated driving It is sometimes referred to as “characteristics”.

The likelihood estimating unit 36 predicts the other vehicle based on the estimated driving characteristics and the result of comparison between the motion candidates predicted by the motion predicting unit 23 and the motion candidate correcting unit 32 and the effective trajectory predicted by the trajectory predicting unit 33. Predict 2 actions.
Specifically, the likelihood estimation unit 36 compares the basic trajectory and the effective trajectory for each motion candidate predicted by the motion candidate prediction unit 31 and the motion candidate correction unit 32 . Then, from the difference between the basic trajectory and the effective trajectory, the likelihood that each motion candidate will be the motion of the other vehicle 2 is obtained.

The likelihood that each motion candidate will be the motion of the other vehicle 2, which is predicted based on the difference between the basic trajectory and the effective trajectory, is denoted as "second likelihood α2". The likelihood estimator 36 calculates the second likelihood α2 such that the smaller the difference between the basic trajectory and the effective trajectory, the higher the second likelihood α2.
Next, the likelihood estimator 36 predicts the motion of the other vehicle 2 based on the estimated driving characteristics and the second likelihood α2.

For example, when a first likelihood α1 (for example, Aggressive likelihood, Ordinary likelihood, or Cautious likelihood) is estimated as the driving characteristic of the other vehicle 2, the first likelihood α1 of each motion candidate is estimated based on the first likelihood α1. 2 The likelihood α2 is weighted.
For example, the final likelihood (final likelihood α) is calculated by multiplying the second likelihood α2 of each motion candidate by using the first likelihood α1 as a coefficient. As a result, the final likelihood α that is combined with the first likelihood α1 predicted by the driving characteristic estimation unit 34 and the driving characteristics correction unit 35 and the second likelihood α2 estimated by the likelihood estimation unit 36 can be obtained.

For example, when the driving characteristic of the other vehicle 2 is Aggressive in the driving scene shown in FIG. The second likelihood α2 of t11 is multiplied by a larger coefficient (first likelihood α1).
The motion candidate for which the highest likelihood is calculated can be determined as the most likely motion candidate in consideration of the driving characteristics and behavior of the other vehicle 2 . Therefore, the likelihood estimation unit 36 determines the motion candidate with the highest final likelihood α as the motion of the other vehicle 2 .

The difference between the basic trajectory and the effective trajectory is calculated, for example, based on the sum of the differences in the position and velocity profiles between the two trajectories. Areas S1 and S2 shown in FIGS. 4A and 4B are an example of the sum of the integrated differences in the positions of the basic trajectory and the effective trajectory. Since it can be determined that the smaller the area, the smaller the positional difference, a higher likelihood is calculated. As another example, a low likelihood can be computed if the velocity profile is significantly different, even if the position difference is small. Note that the first likelihood α1, the second likelihood α2, and the final likelihood α are examples of indicators representing the possibility of the motion candidate actually occurring, and expressions other than the likelihood may be used. .

As described above, the motion prediction unit 23 predicts the motion of the other vehicle 2 based on the likelihood of each motion candidate assumed by the likelihood estimation unit 36 . Note that the "movement of other vehicle" includes the trajectory and speed profile of the other vehicle. The trajectory of the other vehicle 2 indicates the profile of the position of the other vehicle 2 at different times.

The own vehicle route generation unit 24 generates the route of the own vehicle 1 based on the motion of the other vehicle 2 predicted by the motion prediction unit 23 .
For example, when the motion prediction unit 23 predicts the motion t2 or t3 shown in FIG. 1, the lane deviation of the other vehicle 2 can be predicted and the route of the own vehicle 1 can be generated. The route of the host vehicle 1 is a route that does not interfere with the lane change. Alternatively, if the lane width (lane width) is sufficiently wide, it may be a route in which the vehicle travels on the right side of the right lane. Furthermore, if there is an adjacent lane on the right side, the route may be a lane change route in advance.

Therefore, it is possible to generate a smooth route for the own vehicle 1 that does not come into contact with the other vehicle 2 and that the own vehicle 1 does not suddenly decelerate or turn sharply due to the behavior of the other vehicle 2 . "Path of own vehicle 1" indicates not only the profile of the position of the own vehicle 1 at different times, but also the profile of the speed of the own vehicle 1 at each position.

Here, based on the behavior of the other vehicle 2 on the map, the motion of the other vehicle including the trajectory of the other vehicle 2 is predicted. Therefore, generating the route of the own vehicle 1 based on the trajectory of the other vehicle 2 generates the route of the own vehicle 1 based on the change in the relative distance from the other vehicle 2, the difference in the acceleration/deceleration, or the attitude angle. It means that

For example, in the driving scene shown in FIG. 1, when the other vehicle 2 decelerates and continues to run in the left lane, the behavior of the other vehicle 2 is to let the own vehicle 1 go first, and then the other vehicle 2 follows. It can be interpreted as indicating the intention of the operation to change lanes to the right lane.
In this case, by forming the route of the own vehicle 1 or controlling the own vehicle 1 in consideration of the motion intention of the other vehicle 2 , the own vehicle 1 does not decelerate or accelerates to get ahead of the preceding vehicle 3 . can pass through. As a result, it is possible to avoid a situation in which both the other vehicle 2 and the own vehicle 1 yield to each other, thereby realizing smooth traffic flow.

The vehicle control unit 25 controls the steering actuator, accelerator pedal actuator, and a brake pedal actuator.
In addition, although the case where control is performed according to the route of the own vehicle 1 is shown in the embodiment, the own vehicle 1 may be controlled without generating the route of the own vehicle 1 . In this case, it is also possible to perform control based on the relative distance to the other vehicle 52 or the difference in attitude angle between the other vehicle 2 and the host vehicle 1 .

(motion)
Next, an example of the operation of the driving support device 10 according to the embodiment will be described with reference to FIG.
In step S1, the object detection device 11 detects the behavior of objects around the vehicle 1 using a plurality of object detection sensors.
In step S2, the detection integration unit 20 integrates a plurality of detection results obtained from each of the plurality of object detection sensors and outputs one detection result for each object. Then, the object tracking unit 21 tracks each detected and integrated object.

In step S3, the own vehicle position estimation device 12 measures the position, attitude and speed of the own vehicle 1 with respect to a predetermined reference point using a position detection sensor.
In step S4, the map acquisition device 13 acquires map information indicating the structure of the road on which the vehicle 1 travels.
In step S5, the in-map position calculator 22 estimates the position and orientation of the vehicle 1 on the map from the position of the vehicle 1 measured in step S3 and the map data acquired in step S4.

In step S6, the motion prediction unit 23 predicts the other vehicle 2 around the own vehicle 1 based on the detection result (behavior of the other vehicle 2) obtained in step S2 and the position of the own vehicle 1 specified in step S5. Predict the behavior of The other vehicle motion prediction step in step S6 will be described later with reference to FIGS. 6 to 10. FIG.
In step S7, the own vehicle route generation unit 24 generates the route of the own vehicle 1 based on the motion of the other vehicle predicted in step S6.
In step S8, the vehicle control unit 25 controls the vehicle 1 so that the vehicle 1 travels along the route generated in step S7.

Next, step S6 (another vehicle motion prediction step) in FIG. 5 will be described with reference to FIG. The other vehicle motion prediction step is an example of the motion prediction method described in the claims.
In step S10, the behavior determination unit 30 determines the road and lane on which the other vehicle 2 travels from the position of the own vehicle 1 on the map and the behavior of the object obtained in step S2.
In step S11, the motion candidate prediction unit 31 predicts motion candidates of the other vehicle 2 based on the map. For example, motion intention is predicted from the road structure.

FIG. 7 is a flow chart showing an example of the motion candidate prediction step by the motion candidate prediction unit 31 in step S11.
In step S<b>30 , the motion candidate prediction unit 31 selects one of various motion intentions predicted for the other vehicle 2 (eg, going straight or changing lanes, or going straight or turning left or right at an intersection).

In step S31, the motion candidate prediction unit 31 calculates the basic trajectory of the other vehicle 2 for the selected motion intention, and generates the selected motion intention and the corresponding basic trajectory as motion candidates. At this time, the motion candidate prediction unit 31 calculates a plurality of basic trajectories that differ according to the plurality of driving characteristics. That is, the motion candidate prediction unit 31 predicts different motion candidates for each driving characteristic.

In step S<b>32 , the motion candidate prediction unit 31 determines whether or not the process of step S<b>31 has been performed for all of the motion intentions predicted for the other vehicle 2 . When all motion intentions have been processed (step S32: Y), the motion candidate prediction step ends, and the process proceeds to step S12 in FIG.
If any motion intention has not been processed (step S32: N), the process returns to step S30.

Please refer to FIG. In step S12, the controller 14 determines whether the processes of steps S10 and S11 have been performed for all the other vehicles 2 detected in step S1. If all the other vehicles 2 have been processed (step S12: Y), the process proceeds to step S13. If any of the other vehicles 2 has not been processed (step S12: N), the process returns to step S10.

In step S13, the motion candidate modification unit 32 modifies the motion candidate predicted in step S11 in consideration of the stationary object detected at the same time as the other vehicle 2 in step S1, and generates a modified motion candidate. The motion candidate correction unit 32 predicts different motion candidates for each driving characteristic in the same manner as in the motion candidate prediction step shown in FIG.
In step S14, the motion candidate correction unit 32 modifies the motion candidate predicted in step S11 in consideration of the moving object other than the other vehicle 2 detected in step S1, and generates a revised motion candidate. The motion candidate correction unit 32 predicts different motion candidates for each driving characteristic in the same manner as in the motion candidate prediction step shown in FIG.

In step S15, the driving characteristic estimator 34 estimates the driving characteristic of the other vehicle 2 from the vehicle behavior of the other vehicle 2 in this driving environment based on the specific driving environment. For example, the driving characteristic estimation unit 34 estimates the aggressive likelihood, the ordinary likelihood, and the cautious likelihood as the driving characteristics of the other vehicle 2 .
FIG. 8 shows an example of the driving characteristic estimation step in step S15. For example, the driving characteristic estimation unit 34 may estimate the aggressive likelihood, the ordinary likelihood, and the cautious likelihood based on the turn signal operation of the other vehicle 2 when changing lanes.

In step S40, the driving characteristic estimator 34 selects one of the motion candidates of the other vehicle 2 predicted in steps S11, S13, and S14 of FIG.
In step S<b>41 , the driving characteristic estimation unit 34 acquires the predicted route of the own vehicle 1 .
In step S<b>42 , the driving characteristic estimation unit 34 calculates an intersection area where the selected basic trajectory of the motion candidate of the other vehicle 2 intersects with the predicted route of the own vehicle 1 .

If the basic trajectory of the other vehicle 2 and the predicted route of the own vehicle 1 do not intersect, the subsequent steps S43 to S50 are skipped. If none of the basic trajectories of the other vehicle 2 intersects with the predicted route of the own vehicle 1, the driving characteristics of the other vehicle 2 are not estimated, and the driving characteristics of the other vehicle 2 are unknown (unestimated).

In step S<b>43 , the driving characteristic estimator 34 estimates the arrival time at which the other vehicle 2 reaches the intersection area 4 .
In step S44, the driving characteristic estimator 34 determines whether or not the arrival time is less than a threshold. That is, it is determined whether another vehicle 2 is approaching the intersection area 4 or not. If the arrival time is less than the threshold (step S44: Y), the process proceeds to step S45. If the arrival time is not less than the threshold (step S44: N), the process proceeds to step S51.

In step S45, the driving characteristic estimator 34 determines the time when the other vehicle 2 started turning on the blinkers. Then, the time difference between the time when the vehicle reaches the crossing area 4 and the time when the blinker starts lighting is calculated.
In step S46, the driving characteristic estimation unit 34 determines whether or not the time difference is equal to or greater than the first threshold value X1. If the time difference is greater than or equal to the first threshold value X1 (step S46: Y), the process proceeds to step S47. If the time difference is not greater than or equal to the first threshold value X1 (step S46: N), the process proceeds to step S48.

In step S47, the driving characteristic estimation unit 34 increases the cautious likelihood and decreases the ordinary likelihood and the aggressive likelihood. After that, the process proceeds to S51.
In step S48, the driving characteristic estimation unit 34 determines whether or not the time difference is equal to or greater than the second threshold value X2. If the time difference is greater than or equal to the second threshold value X2 (step S48: Y), the process proceeds to step S49. If the time difference is not greater than or equal to the second threshold value X2 (step S48: N), the process proceeds to step S50.

In step S49, the driving characteristic estimation unit 34 increases the ordinary likelihood and decreases the cautious likelihood and the aggressive likelihood. After that, the process proceeds to S51.
In step S50, the driving characteristic estimation unit 34 increases the aggressive likelihood and decreases the cautious likelihood and the ordinary likelihood. After that, the process proceeds to S51.

In step S51, the driving characteristic estimator 34 determines whether or not all of the predicted motion candidates of the other vehicle 2 have been selected in step S40. When all motion candidates have been selected in step S40 (step S51: Y), the driving characteristic estimation step ends, and the process proceeds to step S16 in FIG.
If any motion candidate has not been selected in step S40 (step S51: N), the process returns to step S40 in FIG.

Another example of the driving characteristic estimation step will be described with reference to FIG. For example, the driving characteristic estimation unit 34 may estimate the aggressive likelihood, the ordinary likelihood, and the cautious likelihood based on the lane change time.
In step S60, the driving characteristic estimator 34 records the start timing when the other vehicle 2 starts changing lanes.
In step S61, the driving characteristic estimator 34 records the end timing when the other vehicle 2 ends the lane change.

In step S62, the driving characteristic estimation unit 34 calculates the lane change time by the other vehicle 2 based on the start timing recorded in step S60 and the end timing recorded in step S61.
In step S63, the driving characteristic estimation unit 34 determines whether or not the lane change time is equal to or greater than the third threshold value X3. If the lane change time is equal to or greater than the third threshold value X3 (step S63: Y), the process proceeds to step S64. If the lane change time is less than the third threshold value X3 (step S63: N), the process proceeds to step S65.

In step S64, the driving characteristic estimation unit 34 increases the cautious likelihood and decreases the ordinary likelihood and the aggressive likelihood. After that, the driving characteristic estimation step ends, and the process proceeds to step S16 in FIG.
In step S65, the driving characteristic estimation unit 34 determines whether or not the lane change time is equal to or greater than the fourth threshold value X4. If the lane change time is equal to or greater than the fourth threshold value X4 (step S65: Y), the process proceeds to step S66. If the lane change time is less than the fourth threshold value X4 (step S65: N), the process proceeds to step S67.

In step S66, the driving characteristic estimation unit 34 increases the ordinary likelihood and decreases the cautious likelihood and the aggressive likelihood. After that, the driving characteristic estimation step ends, and the process proceeds to step S16 in FIG.
In step S67, the driving characteristic estimation unit 34 increases the aggressive likelihood and decreases the cautious likelihood and the ordinary likelihood. After that, the driving characteristic estimation step ends, and the process proceeds to step S16 in FIG.

Please refer to FIG. In step S16, the driving characteristic correction section 35 corrects the driving characteristic of the other vehicle 2 estimated by the driving characteristic estimation section 34 when the surrounding environment of the other vehicle 2 changes.
FIG. 10 shows an example of the driving characteristic correction step in step S16. For example, the driving characteristic correction unit 35 may change the driving characteristic of the other vehicle 2 from Ordinary to Aggressive when the preceding vehicle 3 continues to travel at a speed lower than the average vehicle traveling speed or legal speed.

Specifically, in step S70, the driving characteristic correction unit 35 determines whether or not the driving characteristic of the other vehicle 2 has already been estimated. If the driving characteristics have been estimated (step S70: Y), the process proceeds to step S71. If not (step S70: N), the driving characteristic correction step ends, and the process proceeds to step S17 in FIG.
In step S<b>71 , the driving characteristic correction unit 35 attempts to detect the preceding vehicle 3 ahead of the other vehicle 2 . If the preceding vehicle 3 exists ahead of the other vehicle 2 (step S72: Y), the process proceeds to step S73. If the preceding vehicle 3 does not exist (step S72: N), the driving characteristics correction step ends without modifying the driving characteristics, and the process proceeds to step S17 in FIG.

In step S<b>73 , the driving characteristic correction unit 35 estimates the speed of the preceding vehicle 3 .
In step S74, the driving characteristic correction unit 35 determines whether or not the speed of the preceding vehicle 3 is lower than the average speed of surrounding vehicles. If the speed of the preceding vehicle 3 is lower than the average speed of the surrounding vehicles (step S74: Y), the process proceeds to step S75. If the speed of the preceding vehicle 3 is not lower than the average vehicle speed of the surrounding vehicle (step S74: N), the driving characteristics correction step ends without modifying the driving characteristics, and the process proceeds to step S17 in FIG.

In step S75, the driving characteristic correction unit 35 determines whether or not the driving characteristic of the other vehicle 2 is Ordinary. When the likelihood is estimated as the driving characteristic, for example, the driving characteristic correction unit 35 may determine whether the driving characteristic is Ordinary based on whether the Ordinary likelihood is equal to or greater than a predetermined value. Further, for example, it may be determined whether or not the driving characteristic is Ordinary depending on whether or not the Ordinary likelihood is higher than other likelihoods (that is, Aggressive likelihood and Cautious likelihood).
If the driving characteristic is Ordinary (step S75: Y), the process proceeds to step S76. If the driving characteristic is not Ordinary (step S75: N), the driving characteristic correction step ends without modifying the driving characteristic, and the process proceeds to step S17 in FIG.

In step S76, the driving characteristic correction unit 35 changes the driving characteristic of the other vehicle 2 from Ordinary to Aggressive. For example, the driving characteristic modification unit 35 may increase the aggressive likelihood and decrease the cautious likelihood and the ordinary likelihood. After that, the driving characteristic correction step ends, and the process proceeds to step S17 in FIG.
The driving characteristic correction step of FIG. 10 is an example of the driving characteristic correction processing by the driving characteristic correction unit 35. Properties may be modified.

For example, when the other vehicle 2 approaches the preceding vehicle 3 as described above, or when an obstacle such as a fallen object or road construction appears in front of the other vehicle 2, the driving characteristic correction unit 35 may adjust the vehicle at the tail end of the traffic jam. becomes the preceding vehicle 3, if the preceding vehicle 3 is a bus and is predicted to stop, if the road ahead changes to a low μ, if the visibility ahead deteriorates, a place with many passers-by The driving characteristics of the other vehicle 2 may be modified when the other vehicle 2 is traveling, when there is a police-related facility in front of the other vehicle 2, and when there are police vehicles around the other vehicle 2.

Please refer to FIG. In step S17, the motion prediction unit 23 determines the cautious likelihood, the ordinary likelihood, and the aggressive likelihood corrected by the driving characteristic correction unit 35 (or estimated by the driving characteristic estimation unit 34 when there is no correction by the driving characteristic correction unit 35). These likelihoods) are the respective first likelihoods that a plurality of motion candidates predicted by the motion candidate prediction unit 31 and the motion candidate correction unit 32 (that is, motion candidates that differ according to driving characteristics) will be motions of the other vehicle 2. Determined as α1.

In step S18, the controller 14 determines whether or not the processes of steps S13 to S17 have been performed for all the other vehicles 2 detected in step S1. If all the other vehicles 2 have been processed (step S18: Y), the process proceeds to step S19. If any of the other vehicles 2 has not been processed (step S18: N), the process returns to step S13.

In step S19, the trajectory prediction unit 33 calculates the effective trajectory of the other vehicle 2 when the other vehicle 2 maintains its behavior and operates according to the predicted motion intention, using a known state estimation technique such as a Kalman filter. to calculate.
In step S<b>20 , the likelihood estimation unit 36 compares the basic trajectory of the motion candidate predicted by the motion candidate prediction unit 31 and the motion candidate correction unit 32 with the effective trajectory. Based on the difference between the basic trajectory and the effective trajectory, the likelihood estimating unit 36 calculates a second likelihood α2 that the motion candidate predicted by the motion candidate predicting unit 31 and the motion candidate correcting unit 32 is the motion of the other vehicle 2. to estimate

In step S21, the likelihood estimator 36 weights the second likelihood α2 of each motion candidate based on the first likelihood α1, and finalizes the likelihood that each motion candidate can perform a motion with the other vehicle 2. Each likelihood (final likelihood α) is determined.
In step S22, the controller 14 determines whether or not the processes of steps S19 to S21 have been performed for all motion candidates predicted by the motion candidate prediction unit 31 and the motion candidate correction unit 32. FIG. If any motion candidate has not been processed (step S22: N), the process returns to step S19.

If all motion candidates have been processed (step S22: Y), the process proceeds to step S23. When the final likelihood α is determined for all motion candidates, the likelihood estimation unit 36 determines the motion candidate with the highest final likelihood α as the motion of the other vehicle 2 .
At step S23, the controller 14 determines whether or not the processes of steps S19 to S22 have been performed for all the other vehicles 2 detected at step S1. When all the other vehicles 2 have been processed (step S23: Y), the other vehicle motion prediction step ends, and the process proceeds to step S7 in FIG. If any other vehicle 2 has not been processed (step S23: N), the process returns to step S19.

(Effect of Embodiment)
(1) The controller 14 of the driving support device 10 executes a motion prediction method for predicting the motion of the other vehicle 2 around the own vehicle 1 . In this motion prediction method, the driving characteristic estimator 34 estimates the first driving characteristic of the other vehicle 2 based on the specific first driving environment. The driving characteristic correction unit 35 calculates second driving characteristics of the other vehicle 2 in a second driving environment different from the first driving environment, based on the first driving characteristics. When the driving environment of the other vehicle 2 changes from the first driving environment to the second driving environment, the motion candidate predicting unit 31, the motion candidate correcting unit 32, the trajectory predicting unit 33, and the likelihood estimating unit 36 change the second driving characteristic. Based on this, the motion of the other vehicle 2 is predicted.

As a result, when the motion of the other vehicle 2 is predicted based on the driving characteristics of the other vehicle 2, the prediction accuracy can be improved. For example, by correcting the driving characteristics of the other vehicle 2 when there is a change in the surrounding environment of the other vehicle 2, the movement of the other vehicle can be predicted with high accuracy, and the trajectory of the own vehicle 1 can be corrected at an early stage, resulting in greater safety. can run smoothly.

(2) When the environment ahead of the other vehicle 2 changes, the driving characteristic correction unit 35 changes the environment ahead of the other vehicle 2 from the first driving environment to the second driving environment. Compute properties.
As a result, by correcting the driving characteristics of the other vehicle 2 when the environment ahead of the other vehicle 2 changes, the movement of the other vehicle can be accurately predicted, and the trajectory of the own vehicle 1 can be corrected early. This allows for a safer, smoother ride.

(3) The likelihood estimating unit 36 predicts each first predicted trajectory predicted as the behavior of the other vehicle 2 based on the first driving characteristic and the second driving characteristic, respectively, and the actual behavior of the other vehicle 2. Either the first predicted trajectory based on the first driving characteristic or the first predicted trajectory based on the second driving characteristic is selected based on each comparison result between the second predicted trajectory of the other vehicle 2 and the first predicted trajectory based on the second driving characteristic.
By estimating the trajectory of the other vehicle 2 based on the actual behavior of the other vehicle 2 in addition to the driving characteristics, the motion of the other vehicle 2 can be estimated with higher accuracy, and the trajectory of the own vehicle 1 can be corrected more appropriately.

(4) The first driving characteristic is either a first characteristic in which the driving operation is relatively abrupt (for example, Aggressive) or a second characteristic in which the driving operation is relatively slow (for example, Ordinary), and the second driving characteristic is the second characteristic. It is the other of the 1st property and the 2nd property.
As a result, the motion of the other vehicle 2 can be predicted according to whether the driving operation of the other vehicle 2 is abrupt or slow. It can be corrected to a safer and smoother ride.

(5) The driving characteristic correction unit 35 determines that the surrounding environment of the other vehicle 2 has changed from the first driving environment to the second driving environment, and the vehicle traveling at a speed lower than the average vehicle speed or the legal speed A second driving characteristic, which is a first characteristic, is calculated based on the fact that the vehicle has become the preceding vehicle 3 of the other vehicle 2 .
As a result, when the preceding vehicle 3 travels at a low speed, the driving characteristic of the other vehicle 2 is corrected to the first characteristic, so that the lane change trajectory of the other vehicle 2 can be estimated with higher accuracy. The trajectory of the own vehicle 1 can be made safer and smoother.

(6) The driving characteristic correction unit 35 determines that the surrounding environment of the other vehicle 2 has changed from the first driving environment to the second driving environment, based on the fact that the preceding vehicle 3 of the other vehicle 2 and the other vehicle 2 are approaching each other. Then, the second driving characteristic, which is the first characteristic, is calculated.
As a result, when the speed of the preceding vehicle 3 is low, the relative speed is high, and there is no change in the acceleration of the preceding vehicle 3, the driving characteristic of the other vehicle 2 is corrected to the first characteristic. The lane change trajectory can be estimated with higher accuracy, and as a result, the trajectory of the own vehicle 1 can be made safer and smoother.

(7) The driving characteristic correction unit 35 determines that the surrounding environment of the other vehicle 2 has changed from the first driving environment to the second driving environment, and based on the appearance of an obstacle in front of the other vehicle 2, the first The second driving characteristic, which is a characteristic, is calculated.
As a result, by correcting the driving characteristic of the other vehicle 2 to the first characteristic due to the presence of an obstacle in front of the other vehicle 2, the lane change trajectory of the other vehicle 2 can be estimated with higher accuracy. , the trajectory of the own vehicle 1 can be made safer and smoother.

(8) The driving characteristic correction unit 35 determines that the vehicle at the end of the traffic congestion has become the preceding vehicle 3 of the other vehicle 2, assuming that the surrounding environment of the other vehicle 2 has changed from the first driving environment to the second driving environment. Based on this, the second driving characteristic, which is the first characteristic, is calculated.
As a result, when the vehicle at the end of the traffic congestion becomes the preceding vehicle 3 of the other vehicle 2, the driving characteristic of the other vehicle 2 is corrected to the first characteristic, so that the trajectory of the lane change of the other vehicle 2 is made higher. It can be estimated with accuracy, and as a result, the trajectory of the own vehicle 1 can be made safer and smoother.

(9) The driving characteristic correction unit 35 detects that the surrounding environment of the other vehicle 2 has changed from the first driving environment to the second driving environment, and that the bus that is predicted to stop has become the preceding vehicle 3 of the other vehicle. Based on this, the second driving characteristic, which is the first characteristic, is calculated.
As a result, when the preceding vehicle 3 is a bus and is likely to stop, by correcting the driving characteristic of the other vehicle 2 to the first characteristic, the lane change trajectory of the other vehicle 2 can be estimated with higher accuracy. As a result, the trajectory of the own vehicle 1 can be made safer and smoother.

(10) Based on the fact that the surrounding environment of the other vehicle 2 has changed from the first driving environment to the second driving environment and that the road surface in front of the other vehicle 2 has changed to a low μ, the driving characteristic correction unit 35 A second driving characteristic, which is a second characteristic, is calculated.
As a result, when the road surface changes to a low μ due to rain, snow, etc., the driving characteristic of the other vehicle 2 is corrected to the second characteristic. You can run without hindering the amount.

(11) The driving characteristic correction unit 35 determines that the surrounding environment of the other vehicle 2 has changed from the first driving environment to the second driving environment, and based on the fact that the visibility in front of the other vehicle 2 has deteriorated, the second characteristic A second driving characteristic is calculated.
As a result, when the forward visibility deteriorates due to snow, rain, fog, etc., the driving characteristic of the other vehicle 2 is corrected to the second characteristic, thereby preventing the own vehicle 1 from unnecessarily decelerating. It is possible to run without disturbing the traffic volume.

(12) The driving characteristic correction unit 35 assumes that the surrounding environment of the other vehicle 2 has changed from the first driving environment to the second driving environment, and that the other vehicle 2 is expected to travel in a place with many passers-by. A second driving characteristic, which is a second characteristic, is calculated based on.
As a result, when there are many pedestrians and passers-by, by correcting the driving characteristics of the other vehicle 2 to the second characteristics, the own vehicle 1 does not unnecessarily decelerate and hinders the overall traffic volume. You can run without it.

(13) The driving characteristic correction unit 35 determines that the surrounding environment of the other vehicle 2 has changed from the first driving environment to the second driving environment, and based on the existence of the police-related facility in front of the other vehicle 2, the second A second driving characteristic, which is two characteristics, is calculated.
As a result, when there is a police-related facility, by correcting the driving characteristics of the other vehicle 2 to the second characteristics, the own vehicle 1 does not decelerate unnecessarily and travels without interfering with the overall traffic volume. can.

(14) The driving characteristic correction unit 35 determines that the surrounding environment of the other vehicle 2 has changed from the first driving environment to the second driving environment, based on the existence of the police vehicle around the other vehicle 2, the second driving environment. A second driving characteristic, which is a characteristic, is calculated.
As a result, when a police vehicle is present, by correcting the driving characteristic of the other vehicle 2 to the second characteristic, the own vehicle 1 does not needlessly decelerate and can travel without interfering with the overall traffic volume. .

1 Own vehicle 2 Other vehicle 3 Leading vehicle 10 Driving support device 11 Object detection device 12 Own vehicle position estimation device 13 Map acquisition device 14 Controller 20 Detection integration unit , 21... Object tracking unit, 22... Intra-map position calculation unit, 23... Motion prediction unit, 24... Vehicle route generation unit, 25... Vehicle control unit, 30... Behavior determination unit, 31... Motion candidate prediction unit, 32... Motion candidate correcting unit 33 Trajectory predicting unit 34 Driving characteristic estimating unit 35 Driving characteristic correcting unit 36 Likelihood estimating unit

Claims (13)

A motion prediction method for predicting the motion of another vehicle in the vicinity of the own vehicle,
Based on the vehicle behavior of the other vehicle when the driving environment around the other vehicle is the first driving environment, the driving characteristic of the other vehicle is compared with the first characteristic in which the driving operation is relatively abrupt or the driving operation. estimated to be one of the slow second characteristics ,
When the driving environment around the other vehicle changes from the first driving environment to the second driving environment, the driving characteristic of the other vehicle estimated based on the second driving environment is changed to the first characteristic. or modifying one of the second characteristics to the other,
A motion prediction method, comprising predicting the motion of the other vehicle based on the corrected driving characteristics . The behavior of the other vehicle is predicted based on the first driving characteristic that is the driving characteristic estimated based on the vehicle behavior and the second driving characteristic that is the driving characteristic corrected based on the second driving environment . The first prediction based on the first driving characteristic based on each comparison result between each first predicted trajectory and the second predicted trajectory of the other vehicle predicted based on the actual behavior of the other vehicle. 2. The motion prediction method according to claim 1 , wherein one of a trajectory and said first predicted trajectory based on said second driving characteristic is selected. A vehicle running at a speed lower than an average vehicle speed or a legal speed becomes the preceding vehicle of the other vehicle, assuming that the surrounding environment of the other vehicle has changed from the first driving environment to the second driving environment. 3. The motion prediction method according to claim 1 , wherein the driving characteristic of the other vehicle, which is estimated based on the vehicle behavior, is corrected from the second characteristic to the first characteristic based on the fact. estimating based on the behavior of the vehicle based on the fact that the preceding vehicle of the other vehicle and the other vehicle approach each other as the surrounding environment of the other vehicle has changed from the first driving environment to the second driving environment; 4. The motion prediction method according to any one of claims 1 to 3, wherein the driving characteristic of the other vehicle that has been detected is corrected from the second characteristic to the first characteristic . Another vehicle estimated based on the behavior of the other vehicle based on the appearance of an obstacle in front of the other vehicle as a change in the surrounding environment of the other vehicle from the first driving environment to the second driving environment. 5. The motion prediction method according to claim 1 , wherein the driving characteristic of the vehicle is modified from the second characteristic to the first characteristic . Based on the behavior of the vehicle , based on the fact that the surrounding environment of the other vehicle has changed from the first driving environment to the second driving environment, that the vehicle at the end of the traffic congestion has become the preceding vehicle of the other vehicle. The motion prediction method according to any one of claims 1 to 5 , wherein the estimated driving characteristic of the other vehicle is corrected from the second characteristic to the first characteristic . based on the fact that the surrounding environment of the other vehicle has changed from the first driving environment to the second driving environment, and based on the fact that the bus predicted to stop has become the preceding vehicle of the other vehicle, based on the vehicle behavior ; 7. The motion prediction method according to any one of claims 1 to 6 , wherein the driving characteristic of the other vehicle estimated by the second characteristic is corrected from the second characteristic to the first characteristic . Assuming that the surrounding environment of the other vehicle has changed from the first driving environment to the second driving environment, it is estimated based on the behavior of the vehicle based on the fact that the road surface in front of the other vehicle has changed to a low μ. The motion prediction method according to any one of claims 1 to 7 , wherein the driving characteristic of the other vehicle is modified from the first characteristic to the second characteristic . The other vehicle estimated based on the behavior of the other vehicle based on deterioration of the forward visibility of the other vehicle as a change in the surrounding environment of the other vehicle from the first driving environment to the second driving environment. The motion prediction method according to any one of claims 1 to 8 , wherein the driving characteristic is modified from the first characteristic to the second characteristic . Based on the fact that the surrounding environment of the other vehicle has changed from the first driving environment to the second driving environment and that the other vehicle is expected to travel in a place with many passers-by, the behavior of the vehicle is changed. The motion prediction method according to any one of claims 1 to 9 , wherein the driving characteristic of the other vehicle estimated based on the first characteristic is modified from the first characteristic to the second characteristic . Estimated based on the behavior of the vehicle based on the presence of a police-related facility in front of the other vehicle as a change in the surrounding environment of the other vehicle from the first driving environment to the second driving environment The motion prediction method according to any one of claims 1 to 10 , wherein the driving characteristic of the vehicle is modified from the first characteristic to the second characteristic . Another vehicle estimated based on the behavior of the other vehicle based on the presence of a police vehicle around the other vehicle as a change in the surrounding environment of the other vehicle from the first driving environment to the second driving environment. The motion prediction method according to any one of claims 1 to 11 , wherein the driving characteristic of the vehicle is modified from the first characteristic to the second characteristic . A motion prediction device comprising a controller that predicts the motion of another vehicle in the surroundings of the own vehicle,
The controller is
Based on the vehicle behavior of the other vehicle when the driving environment around the other vehicle is the first driving environment, the driving characteristic of the other vehicle is compared with the first characteristic in which the driving operation is relatively abrupt or the driving operation. estimated to be one of the slow second characteristics,
When the driving environment around the other vehicle changes from the first driving environment to the second driving environment, the driving characteristic of the other vehicle estimated based on the second driving environment is changed to the first characteristic or Modifying one of the second characteristics to the other;
A motion prediction device that predicts the motion of the other vehicle based on the corrected driving characteristics .

Images