Vehicle control device

Abstract

Provided is a vehicle control device capable of ensuring control margin time required for avoidance control and performing collision avoidance control. The present invention provides a means and a device that: store recognition information on a two-wheel vehicle at the time of overtaking the two-wheel vehicle; on the occasion of turning left by a host vehicle at an intersection, read recognition information on two-wheel vehicles recognized in the past, judge the possibility of collision with the host vehicle, and calculate prediction recognition information; and when an obstacle is actually detected at the time of turning left by the host vehicle, immediately confirm that the obstacle detection is a re-detection of a moving body which was recognized in the past, by using the prediction recognition information calculated in advance; and perform avoidance control. Accordingly, control margin time for avoiding collision with an obstacle which comes from behind the host vehicle at the time of turning left and which is difficult to avoid can be ensured, and collision can be prevented.

Description

[Technical Field]

[0001] The present invention relates to a vehicle control device. [Background technology]

[0002] 2. Description of the Related Art Conventionally, there have been proposed vehicle control devices and vehicle control methods that perform collision avoidance control to avoid a collision between the host vehicle and an object in front of the host vehicle.

[0003] A conventional vehicle control device is known that uses the detection results of an object in front of the vehicle to determine the possibility of a collision between the object and the vehicle, and if it determines that there is a high possibility of a collision between the object in front of the vehicle and the vehicle, performs collision avoidance control to avoid a collision with the object.

[0004] In addition, when a vehicle turns right or left at an intersection, the detection area changes as the direction of the vehicle changes, and if an object that was located in a blind spot enters the detection area, the object and the vehicle are close to each other at the time the object is detected, which causes a delay in the timing of activation of the collision avoidance control.In response to this problem, a technology has been proposed in which the timing of activation of the collision avoidance control is advanced when it is determined that the vehicle has overtaken the object and then will turn right or left, compared to when it is not determined that the vehicle has overtaken the object, and when it is determined that the vehicle has overtaken the object and then will not turn right or left (see Patent Document 1). [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Patent No. 6617696 Summary of the Invention [Problem to be solved by the invention]

[0006] Normally, to execute avoidance control, an object is detected by an external recognition sensor such as a camera sensor attached to the vehicle, and tracking is performed for a certain period of time to determine and identify whether the object is an object to be avoided. Then, a flow is followed to execute avoidance control for that object. Generally, to identify an object as an object to be avoided, a certain period of tracking processing time is required to prevent erroneous recognition.

[0007] Also, consider the case where the host vehicle performs avoidance control against an object (bicycle, motorcycle, etc.) approaching from the crossing direction or from behind when turning right or left at an intersection. When performing avoidance control against an object approaching from the crossing direction, the position at which avoidance control starts is far from the host vehicle, making it easy to ensure sufficient time for avoidance control. On the other hand, when performing avoidance control against an object approaching from behind, the position at which avoidance control starts is closer to the host vehicle, making it difficult to ensure sufficient time for avoidance control. In reality, the host vehicle also moves, so it is even more difficult to ensure sufficient time for avoidance control against an object approaching from behind. In this way, the direction from which the object is approaching relative to the host vehicle determines whether or not avoidance control can be performed in time (see Figure 1).

[0008] Therefore, when a fast-moving obstacle such as a bicycle approaches from behind the vehicle, time is similarly required to determine it as an object to be avoided, but even if avoidance control is initiated after determining it as an object to be avoided, there is a high possibility that avoidance will not be possible in time.

[0009] The collision avoidance control proposed in the method of Patent Document 1 attempts to deal with objects in blind spots by determining that an object is to be avoided and then accelerating the activation timing based on additional conditions, but does not take into consideration the time it takes to determine that an object is to be avoided. Therefore, even if Patent Document 1 is applied, if a fast-moving object such as a bicycle is approaching from behind, it will still require time to determine that the object is to be avoided, so there is a high possibility that a collision cannot be avoided even if the activation timing is accelerated.

[0010] The present invention has been made in consideration of the above circumstances, and has an object to provide a vehicle control device that can secure the control margin time required for avoidance control and perform collision avoidance control. [Means for solving the problem]

[0011] The vehicle control device according to the present invention comprises an external environment recognition unit that acquires external environment information at least ahead of the host vehicle, a control unit that controls the host vehicle based on the external environment information acquired by the external environment recognition unit, and a memory unit that stores detection information of the moving body acquired by the external environment recognition unit when the host vehicle overtakes a moving body, wherein when a lane change of the host vehicle is predicted, the external environment recognition unit reads out past detection information of the moving body stored in the memory unit, determines the possibility of a collision between the host vehicle and the moving body, and based on the collision possibility, sets as predicted recognition information a moving body that is likely to collide with the host vehicle when the host vehicle changes lane, and determines whether the detection information of the moving body detected by the external environment recognition unit when the host vehicle changes lane is identical to the past detection information of the moving body included in the predicted recognition information. [Effects of the Invention]

[0012] According to the present invention, it is possible to ensure a control margin time for avoidance control for an obstacle approaching from behind the vehicle when turning right or left at an intersection, for example, and perform collision avoidance control.

[0013] The present invention makes it possible to prevent accidents involving motorcycles when turning right or left at intersections, which has been difficult to address in the past.

[0014] Problems, configurations, and effects other than those described above will become apparent from the following description of the embodiments. [Brief explanation of the drawings]

[0015] [Figure 1] 1 is a schematic diagram of a problem to be solved by the present invention; [Figure 2] 1 is a schematic diagram of an embodiment of the present invention; [Figure 3]1 is a schematic block diagram of a vehicle control device according to an embodiment of the present invention; [Figure 4] 1 is an overall flowchart in an embodiment of the present invention. [Figure 5] 4 is a flowchart illustrating a collision possibility determination process according to an embodiment of the present invention. [Figure 6] 10 is a flowchart illustrating a process for setting a prediction region according to an embodiment of the present invention. [Figure 7] 10 is a flowchart illustrating a method for shortening a tracking process using predicted recognition information according to an embodiment of the present invention. [Figure 8] 4 is a flowchart relating to a right / left turn determination in an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0016] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0017] Prior to describing the embodiments of the present invention, the problem to be solved by the present invention will be described with reference to Fig. 1. Fig. 1 shows a situation in which a vehicle is turning left at an intersection, and is composed of a host vehicle 50 with an avoidance control function traveling on a general road 55 that has an intersection, a sensor detection area 51 used for the avoidance control function, a period 54 required to identify an object to be avoided when performing avoidance control, a bicycle 53 approaching the host vehicle 50 from behind, and a bicycle 52 approaching the host vehicle 50 from a crossing direction.

[0018] Consider a case where a vehicle 50 turns left at an intersection and performs avoidance control against a bicycle 52 approaching from the crossing direction and a bicycle 53 approaching from behind. The avoidance control function detects an obstacle in the sensor detection area 51, identifies it as an object to be avoided, and then actually performs avoidance control to prevent a collision.

[0019] First, in the case of avoidance control for a bicycle 52 coming from the crossing direction, the avoidance control actually starts after a period 54 has passed since the bicycle 52 entered the sensor detection area 51 and is identified as an object to be avoided, but because the position at which the avoidance control starts is far from the vehicle 50, it is possible to ensure a margin of time for avoidance control.

[0020] In contrast, in the case of avoidance control for a bicycle 53 approaching from behind, avoidance control actually begins after a period 54 has passed since the bicycle 53 entered the sensor detection area 51 and is identified as an obstacle to be avoided, but because the position at which avoidance control begins is close to the vehicle 50, it is not possible to ensure sufficient time for control to avoid the obstacle. In reality, the vehicle 50 also moves, so it is even more difficult to ensure sufficient time for control to avoid an obstacle approaching from behind.

[0021] In this way, whether or not avoidance control can be implemented in time depends on the direction from which bicycles 52 and 53 are approaching relative to vehicle 50. Even if the above-mentioned prior art (Patent Document 1) is applied to address this issue, it is difficult to respond to cases such as bicycle 53 approaching from behind, because the period 54 specified as an object to be avoided is not clearly stated.

[0022] Furthermore, the idea that avoidance control should simply be initiated when the vehicle enters the sensor detection area 51 is difficult to implement from the standpoint of preventing malfunctions, and it is not realistic to simply shorten the period 54 during which a detected object is identified as an object to be avoided.

[0023] The present invention addresses the issue of how to shorten the period required to identify an obstacle as an object to be avoided for performing avoidance control while preventing malfunctions not only for obstacles coming from the crossing direction or the front, but also for obstacles coming from behind.

[0024] In an embodiment of the present invention, to address this problem, a means and device are provided that perform the above three operations: (1) when overtaking a moving object such as a bicycle while approaching an intersection, the recognition information of the moving object is stored, and when it is predicted that the vehicle will turn left at the intersection, the previously stored recognition information is read out to determine the possibility of a collision between the vehicle and the previously stored moving object; (2) using information on the moving object identified as having a high collision probability as a result of determining the possibility of collision, the detection timing and detection area at which the moving object is predicted to be detected in the sensor detection range of the vehicle when the vehicle turns left is calculated as predicted recognition information, and the area at which the moving object is predicted to be detected is set in the sensor detection area at least in front of the vehicle; and (3) when the vehicle turns left at an intersection, if an object is detected in the predicted detection timing or detection area, the previously stored moving object is determined to have been re-detected, and the period for identifying it as an avoidance target for which avoidance control is to be performed is shortened, thereby ensuring the control margin time required for avoidance control.

[0025] First, the basic concept of this proposal will be explained using Fig. 2. A vehicle control device 100 (Fig. 3) of this embodiment is mounted on a host vehicle 3, and performs collision avoidance control to avoid collision between the host vehicle and an object ahead of the host vehicle. This embodiment assumes that the collision avoidance control is performed on a general road 1 including traffic lights 2 and intersections.

[0026] FIG. 2 is a conceptual diagram showing a case where the host vehicle 3 performs collision avoidance control against a two-wheeled vehicle 6 traveling from behind when turning left at an intersection, and the operation will be explained by dividing it into Step 1, Step 2, and Step 3 showing the operation flow of this embodiment.

[0027] Of the three operations in the embodiment of the present invention, (1) when overtaking a moving object such as a bicycle while approaching an intersection, the operation of storing the recognition information of the moving object, and when it is predicted that the vehicle will turn left at the intersection, reading out the recognition information stored in the past, and determining the possibility of a collision between the vehicle and the previously stored moving object will be explained using Step 1 and Step 2.

[0028] First, as shown in Step 1 of Fig. 2, the host vehicle 3 and motorcycle 6 are traveling on a public road 1. When the motorcycle 6 is detected in the sensor detection area 4 of the host vehicle 3, recognition information of the motorcycle 6, such as its speed, relative position to the host vehicle, and detection time, is acquired, and when the host vehicle 3 overtakes the motorcycle 6 and the motorcycle 6 leaves the sensor detection area 4, this information is stored in the memory unit 5 as reusable information.

[0029] Next, as shown in Step 2, the travel route of the host vehicle 3 is obtained from map information such as a navigation system, and it is determined whether the host vehicle 3 is likely to pass through an intersection. When it is determined that the host vehicle 3 will pass through an intersection, the recognition information of the motorcycle 6 stored in the memory unit 5 is read, and the likelihood of a collision between the motorcycle 6 and the host vehicle 3 when turning left at the intersection is determined based on the read past recognition information. Specifically, the distance from the host vehicle 3 to the intersection is obtained, and the intersection arrival time is calculated. Next, based on the recognition information of the motorcycle 6 read from the memory unit 5, the intersection arrival time of the motorcycle 6 is calculated, assuming that the motorcycle 6 moves at a constant speed in a straight line from the time of detection. The calculated intersection arrival times of the host vehicle 3 and the motorcycle 6 are compared, and the possibility that the motorcycle 6 will approach from behind when the host vehicle 3 turns left at the intersection is determined as a possibility of a collision. If it is determined that there is a high likelihood of a collision, the recognition information (past recognition information) stored in the memory unit 5 as predicted recognition information is identified as an obstacle that is likely to be re-detected when turning left at the intersection. This makes it possible to store the recognition information of a moving body when passing a moving body such as a bicycle while traveling before an intersection, and to read out the previously stored recognition information when it is predicted that the vehicle will turn left at the intersection, thereby determining the possibility of collision between the vehicle and the previously stored moving body.

[0030] Next, the remaining two operations will be explained using Step 3: (2) using information on a moving object identified as having a high collision probability as a result of the collision probability determination, calculate the detection timing and detection area at which the moving object is predicted to be detected in the sensor detection range of the host vehicle when the host vehicle makes a left turn as predicted recognition information, and set the area at which the moving object is predicted to be detected in the sensor detection area at least in front of the host vehicle; and (3) when the host vehicle makes a left turn at an intersection and an object is detected in the predicted detection timing or detection area, determine that a previously stored moving object has been re-detected, and shorten the period for identifying the moving object as an avoidance target for which avoidance control is to be performed, thereby ensuring the control margin time required for avoidance control.

[0031] Step 3 illustrates a situation in which a motorcycle 6 is approaching from behind at the exact time that the host vehicle 3 is actually about to turn left at an intersection. It is determined from control information of the turn signal or the like that the host vehicle 3 is about to turn left at the intersection. The fact that the host vehicle 3 is about to turn left at the intersection is used as a trigger to determine whether the predictive recognition information calculated in Step 2 exists. If the predictive recognition information exists, the obstacle detection timing (also referred to as predicted detection timing) and detection area (also referred to as predicted detection area) when turning left at the intersection are predicted from the predictive recognition information, and a predicted area 8 corresponding to the predicted detection area is set for the image 7 generated by the image processing unit. As a result, using information about a moving object identified as having a high collision risk, the detection timing and detection area at which the moving object is predicted to be detected within the sensor detection range of the host vehicle when the host vehicle makes a left turn can be calculated as predictive recognition information, and the area at which the moving object is predicted to be detected can be set within the sensor detection area at least ahead of the host vehicle.

[0032] When an object is detected in the sensor detection area 4 while the host vehicle 3 is turning left at an intersection, it is possible to immediately determine whether or not it is an obstacle predicted in advance by making a conditional judgment in conjunction with the predicted recognition information, and it is possible to move to avoidance control (or a warning) by shortening the period required to identify it as an object to be avoided in normal collision avoidance control. Specifically, when an object is detected in the sensor detection area 4, it is immediately identified as the stored motorcycle 6 based on conditions such as whether the detection timing matches the predicted timing and whether the detection position was detected in the predicted area 8 (corresponding to the predicted detection area), thereby shortening the time required to identify it as an object to be avoided in collision avoidance control.

[0033] According to this proposal, the time required to execute collision avoidance control, which is difficult to shorten simply, from detecting an object to recognizing it for a certain period of time until it is identified as an object to be avoided, can be shortened by adding predictive recognition information. This time reduction makes it possible to advance the timing to start avoidance control, thereby ensuring sufficient control time to avoid the obstacle.

[0034] A configuration diagram of a vehicle control device according to an embodiment of the present invention is shown in Fig. 3. The vehicle control device 100 according to this embodiment includes an external environment recognition unit 101 that recognizes obstacles and traffic lights, a control unit 102 that actually issues instructions for avoidance control, and a storage unit 106 that stores (preserves) the recognition information from the external environment recognition unit 101, and is a vehicle control device that realizes collision avoidance control.

[0035] The external environment recognition unit 101 acquires external environment information ahead of the vehicle, and includes an image processing unit 103 for generating images to be used in the recognition process, a recognition unit 104 for detecting vehicles and lanes and tracking the time series of detected objects, and an obstacle prediction recognition unit 105 for determining the possibility of a collision when the vehicle turns left at an intersection from past recognition information stored in a memory unit 106 that stores information recognized by the recognition unit 104, and generating predicted recognition information.

[0036] The obstacle prediction and recognition unit 105 is configured to acquire the position information of the vehicle via an MPU having high-precision map information such as data on road structure, such as intersection width information, intersection stop line positions, and number of lanes, and information on three-dimensional objects on the road, such as road signs and traffic light positions, or via a GNSS (Global Navigation Satellite System) positioning satellite. an intersection arrival time calculation unit 115 that calculates the intersection arrival time when an obstacle travels to the intersection at a constant speed linear motion and the intersection arrival time when the vehicle arrives at the intersection, based on obstacle information (past recognition information) stored in the storage unit 106; a collision possibility determination unit 116 that determines the possibility of a collision from a comparison between the intersection arrival times of the obstacle and the vehicle calculated by the intersection arrival time calculation unit 115; and a prediction reliability setting unit 117 that adds reliability to the prediction recognition information calculated by the collision possibility determination unit 116 based on external factors that reduce detection accuracy, such as the time elapsed since the obstacle was detected and the weather conditions at the time of detection. The reliability set by the prediction reliability setting unit 117 is used to determine the possibility that an obstacle detected by the prediction area detection determination unit 112 (when the vehicle turns left at an intersection) is an obstacle previously detected and stored in the memory unit 106 (to determine whether the obstacle detection information matches or is identical to the previous detection information).

[0037] The recognition unit 104 has a moving object detection unit 111 that identifies moving objects, such as motorcycles, traveling on public roads, a predicted area detection and determination unit 112, and a tracking processing unit 113. The predicted area detection and determination unit 112 receives predicted recognition information from the obstacle prediction and recognition unit 105, calculates an area (predicted area) in which an obstacle included in the predicted recognition information is predicted to be detected in an image captured by the imaging unit 126, and passes the predicted area information to a predicted area setting unit 110 of the image processing unit 103. The predicted area setting unit 110 sets a prediction area for the image captured by the imaging unit 126 (more specifically, for at least a part of the image) (see FIG. 2). When a moving object is actually detected by the moving object detection unit 111, the predicted area detection and determination unit 112 determines whether the moving object has been detected in the predicted area. The predicted area detection determination unit 112 determines whether the detected obstacle is an obstacle detected in the past based on the conditions of whether a moving object has been detected in the predicted area and whether the time at which the moving object was detected is the predicted time, and passes the result to the tracking processing unit 113. In other words, the predicted area detection determination unit 112 determines whether the detection information of the detected moving object is identical to the past detection information of the moving object based on the predicted detection timing and predicted detection area, and passes the result to the tracking processing unit 113. The tracking processing unit 113 normally detects vehicles and lanes, tracks detected objects over time, etc., but in the case of the collision avoidance control of this embodiment, it is necessary to perform tracking processing for a certain period of time after the obstacle is first detected, in order to identify the obstacle as an object to be avoided for which avoidance control is to be performed, and to determine whether the obstacle is an obstacle with a possibility of collision. In addition to the normal tracking process, when the tracking processing unit 113 of this embodiment receives predictive recognition information indicating that a detected obstacle is likely to be an obstacle detected in the past (in other words, when it is determined that the detection information of the detected moving object and the past detection information of the moving object included in the predictive recognition information are the same), it adds the information recognized in advance, making it possible to shorten the tracking processing time for identifying the detected obstacle as an obstacle to be avoided compared to the normal tracking process. This makes it possible to shorten the period required to identify the detected obstacle as an obstacle to be avoided for which collision avoidance control is to be performed.

[0038] The control unit 102 controls the host vehicle based on the external environment information acquired by the external environment recognition unit 101, and includes a right / left turn determination unit 120, a collision risk calculation unit 121, a route planning unit 122, and a control instruction unit 123. The right / left turn determination unit 120 determines whether the host vehicle is turning right or left based on information from a yaw rate sensor 125, a steering angle sensor 127, and a direction indicator 130. The collision risk calculation unit 121 calculates a time to collision (TTC) for an obstacle with a high collision probability determined by the external environment recognition unit 101 (in other words, identified as an obstacle to be avoided by avoidance control), determines the collision risk, and determines whether avoidance control needs to be executed. The route planning unit 122 calculates a driving route for collision avoidance. The control instruction unit 123 calculates a control value for collision avoidance control and issues an instruction to a braking force / driving force control ECU 124.

[0039] The braking force / driving force control ECU 124 issues appropriate control instructions to each actuator of the vehicle in accordance with the control instruction value of the control instruction unit 123 .

[0040] The warning instruction unit 128 plays a role of issuing a warning to the driver that a collision risk is imminent based on information from the collision risk calculation unit 121, or notifying the driver that avoidance control is in progress by displaying a display or making a sound when collision avoidance control is in progress. In addition, the above components are connected to each other by a communication network 131.

[0041] In this embodiment, the external environment recognition unit 101 is exemplified by an imaging unit (camera sensor) mounted at the front of the vehicle that captures images of at least the area in front of the vehicle as a means for acquiring external environment information at least in front of the vehicle, but it can also be applied to not only a camera sensor alone, but also a combination of a camera sensor and a radar sensor or a combination of other sensors.

[0042] In this embodiment, the external environment recognition unit 101 and the control unit 102 operate independently, but a configuration in which the external environment recognition unit 101 and the control unit 102 are integrated is also applicable.

[0043] When determining the possibility of a collision, the obstacle prediction and recognition unit 105 acquires detection information from the memory unit 106. In order to reduce the data calculation load, the obstacle prediction and recognition unit 105 may limit the data to be acquired by filtering whether the detected obstacle is a two-wheeled vehicle or not based on width information of the detected obstacle contained in the detection information.

[0044] The storage unit 106 may be shared with a normal storage unit used in the external environment recognition unit 101, or may be independently arranged as a storage unit dedicated to this embodiment.

[0045] The prediction region setting unit 110 may set a plurality of prediction regions, not just one, and may change the method of setting the prediction region depending on the reliability included in the prediction recognition information.

[0046] When the setting information is not detected in the prediction area set by the prediction area setting unit 110 even after the predicted detection time has passed based on the detection timing included in the predicted recognition information, the setting information can be deleted from the storage unit 106. In other words, when the moving object included in the predicted recognition information is not detected in the prediction area set by the prediction area setting unit 110 even after the predicted detection timing has passed, the predictive recognition information can be deleted from the storage unit 106.

[0047] As an example of determining the possibility of a collision in the collision possibility determination unit 116, a simple method of determining the possibility of a collision by comparing the intersection arrival times of the vehicle 3 and the two-wheeled vehicle 6 has been described, but this is not the only method for determining the possibility of a collision. For example, methods for improving the accuracy of determining the possibility of a collision, such as predicting the future route of the two-wheeled vehicle 6 and determining the possibility of a collision, are not limited to this embodiment.

[0048] Since the time it takes for the vehicle 3 to reach the intersection varies depending on the driver's control from the time the collision possibility judgment unit 116 makes a collision possibility judgment until the vehicle 3 reaches the intersection, the collision possibility judgment may be performed at any time (at a cycle time) until the vehicle 3 reaches the intersection.

[0049] The right / left turn determining unit 120 may obtain a steering angle control value to determine whether to turn right or left.

[0050] In this embodiment, the intersection distance calculation unit 114 calculates the intersection distance from the map information 129, but it may also calculate the intersection distance from the recognition result of a sensor (such as a camera sensor) attached to the host vehicle, or may calculate the intersection distance based on road-to-vehicle communication information received from a roadside communication device. In other words, as a trigger for calculating the intersection distance and reading out obstacle information (past detection information) from the storage unit 106, the host vehicle's travel route may be calculated from the map information 129 and the past detection information may be read out by determining that there is a high possibility of turning right or left at the intersection, or the past detection information may be read out by determining that there is a high possibility of turning right or left at the intersection from the intersection recognition result by the external environment recognition unit 101, or the past detection information may be read out by determining that there is a high possibility of turning right or left at the intersection using intersection information acquired by road-to-vehicle communication.

[0051] In this embodiment, an example has been described in which collision avoidance control is performed against an obstacle approaching from behind when the vehicle is turning left, but this is not limited to left turns, and can also be applied to general lane changes such as right turns.

[0052] Based on the above configuration, the overall flow of this embodiment is shown in FIG. 4 , (1) when overtaking a moving body such as a bicycle while approaching an intersection, recognition information of the moving body is stored, and when it is predicted that the host vehicle will turn left at the intersection, previously stored recognition information is read out to determine the possibility of a collision between the host vehicle and the previously stored moving body, FIG. 5 , (2) using information on the moving body identified as having a high collision possibility as a result of determining the possibility of collision, an operation of calculating, as predicted recognition information, a detection timing and a detection area at which the moving body is predicted to be detected in the sensor detection range of the host vehicle when the host vehicle turns left, and setting an area at least in front of the host vehicle in the sensor detection area, FIG. 6 , and (3) when the host vehicle turns left at an intersection, an operation of determining that a previously stored moving body has been re-detected and shortening the period for identifying the moving body as an avoidance target for which avoidance control is to be performed, thereby ensuring the control margin time required for avoidance control, FIG. 7 , and a method of determining whether the host vehicle will turn left by the right / left turn determination unit 120 will be described using the flowchart in FIG. 8 .

[0053] First, we will explain the flowchart in Fig. 4. Note that Steps 1 to 3 in Fig. 4 correspond to Steps 1 to 3 in Fig. 2.

[0054] First, the external environment recognition unit 101 attached to the vehicle acquires external environment information ahead of the vehicle (S300). It is determined whether a motorcycle is included in the acquired external environment information (S301). If No, the process returns to S300, and if Yes, the process proceeds to S302. The detection information and detection time of the motorcycle recognized by the recognition unit 104 are stored in the storage unit 106 (S302).

[0055] The obstacle prediction / recognition unit 105 acquires driving plan information for the host vehicle from the map information 129, the location information of the host vehicle, etc. (S303). The obstacle prediction / recognition unit 105 determines the possibility of collision when the host vehicle turns right or left for moving object information among the obstacle information (past recognition information) stored in the storage unit 106, and extracts obstacles determined to have a high collision possibility (S304).

[0056] Then, the right / left turn determination unit 120 determines that the host vehicle will turn right or left (S305). If No, the process returns to S300; if Yes, the process proceeds to S306. A moving object with a high probability of collision when the host vehicle turns right or left is identified, and the detection timing (predicted detection timing) and detection area (predicted detection area) predicted to be detected in the host vehicle's detection area when turning right or left are acquired as predicted recognition information (S306). It is determined whether an obstacle has been detected in the host vehicle's detection area (S307). If No, the process returns to S306; if Yes, the process proceeds to S308. Using the predicted recognition information acquired in S306, the process compares the detection timing and detection area actually detected in S307 with the predicted recognition information, and determines whether the obstacle actually detected in S307 is the same as the obstacle identified in the predicted recognition information (S308). In other words, it is determined whether the detection information of the moving object actually detected in S307 when the host vehicle turns right or left is the same as the past detection information of the moving object included in the predicted recognition information. If it is determined that the possibility is high, the process proceeds to S309; ​​if it is determined that the possibility is low, the process proceeds to S310. If it is determined in S308 that the re-detected obstacle is highly likely to be the same as the obstacle detected in the predictive recognition information previously detected, the period (number of detections) required for identifying the obstacle as an avoidance target required for normal collision avoidance control is reduced, and the obstacle in the predictive recognition information is immediately determined to have been detected (S309). If it is determined that the re-detected obstacle is highly likely to be the same as the obstacle detected in the predictive recognition information previously detected, a tracking process is performed for a certain period (i.e., the normal number of detections) to identify the obstacle as an avoidance target, as in normal collision avoidance control, and the obstacle is identified as an avoidance target (S310). Avoidance control is implemented or an alarm is issued for the obstacle identified in the previous procedure (S311).

[0057] Next, (1) when passing a moving object such as a bicycle while approaching an intersection, the recognition information of the moving object is stored, and when it is predicted that the vehicle will turn left at the intersection, the recognition information stored in the past is read out and the possibility of a collision between the vehicle and the moving object stored in the past is determined. Of these, the operation of reading out the recognition information stored in the past and determining the possibility of a collision between the vehicle and the moving object stored in the past will be explained using Figure 5.

[0058] The obstacle prediction / recognition unit 105 acquires driving plan information for the host vehicle from the map information 129, the host vehicle's position information, etc. (S400). The obstacle prediction / recognition unit 105 determines whether the host vehicle is scheduled to pass through an intersection from the map information 129, the host vehicle's position information, etc. (S401). If No, the process returns to S400, and if Yes, the process proceeds to S402. The moving object information is read out from the past recognition information stored in the storage unit 106 (S402). The prediction reliability setting unit 117 calculates and adds a reliability to the moving object information read out from the storage unit 106 based on the time when the moving object was detected, the detection accuracy, etc. (S403). The intersection distance calculation unit 114 calculates the distance to the intersection from the current position of the host vehicle and information in the map information 129 (S404). The intersection arrival time calculation unit 115 calculates the intersection arrival time predicted when the host vehicle will turn right or left after arriving at the intersection from the intersection distance information calculated in S404 (S405). The collision possibility determination unit 116 calculates the intersection turn time (intersection arrival time) for each piece of moving object information read out in S402, and determines the collision possibility for each piece of moving object information when the host vehicle turns right or left (S406). The prediction reliability setting unit 117 evaluates and sets the reliability for the determined collision possibility (S407). From the collision possibility evaluated in S406 and the reliability evaluated in S407, the unit identifies moving object information that is likely to collide when the host vehicle turns right or left at the intersection, and passes the information to the recognition unit 104 as predicted recognition information (S408).

[0059] (2) Using information on a moving object identified as having a high collision probability as a result of determining the possibility of collision, the operation of calculating, as predictive recognition information, the detection timing and detection area in which the moving object is predicted to be detected within the sensor detection range of the vehicle when the vehicle makes a left turn, and setting the area in which the moving object is predicted to be detected within the sensor detection area at least in front of the vehicle is explained using Figure 6.

[0060] First, the right / left turn determination unit 120 detects that the host vehicle is turning right or left (S500). If No, the process returns to S500; if Yes, the process proceeds to S501. The recognition unit 104 determines whether it has received the predicted recognition information (determined by the collision possibility determination unit 116) described in FIG. 5 from the obstacle prediction / recognition unit 105 (S501). If No, the process ends; if Yes, the process proceeds to S502. The recognition unit 104 calculates and predicts the timing of collision between the host vehicle and a moving object that is likely to collide with the host vehicle when the host vehicle turns right or left, and is included in the received predicted recognition information (S502). The recognition unit 104 also calculates a sensor detection area that is predicted to be detected when the host vehicle turns right or left, based on relative position information of the moving object with respect to the host vehicle, which is included in the predicted recognition information (S503). The prediction area setting unit 110 of the image processing unit 103 sets the sensor detection area as a prediction area (see Figure 2) in (at least a part of) the sensor detection area at least in front of the vehicle based on the calculated collision timing and predicted recognition information including sensor detection area information (S504).

[0061] (3) When the vehicle is turning left at an intersection, if an object is detected at the predicted detection timing or in the detection area, it is determined that a previously stored moving object has been detected again. This operation is explained with reference to FIG. 7, in which the control margin time required for avoidance control is secured by shortening the period for identifying the moving object as an object to be avoided for avoidance control.

[0062] The right / left turn determination unit 120 determines whether the host vehicle is currently under right / left turn control (S600). If No, the process ends; if Yes, the process proceeds to S601. The recognition unit 104 determines whether a moving object has been detected in a sensor detection area at least ahead of the host vehicle (S601). If No, the process returns to S600; if Yes, the process proceeds to S602. The prediction area detection determination unit 112 of the recognition unit 104 compares the collision timing included in the predicted recognition information calculated by the operation of FIG. 6 for the moving object detected in S601 with the timing at which the moving object was actually detected (S602). In other words, it determines whether the moving object was detected at the predicted collision timing. If No, the process proceeds to S604; if Yes, the process proceeds to S603. The prediction area detection determination unit 112 also determines whether the moving object detected in S601 was detected within the prediction area set by the operation of FIG. 6 (S603). In other words, it determines whether a moving object was detected in the prediction area. If No, proceed to S604, and if Yes, proceed to S606. That is, for the detection information of the moving object detected in S601 when the host vehicle is turning right or left, it is determined whether or not the information is identical to the past detection information of the moving object included in the predicted recognition information, using the predicted detection timing and predicted detection area as conditions.

[0063] In S604, the prediction area detection and determination unit 112 determines that the moving object detected in S601 is unlikely to be the same moving object that is included in the predicted recognition information and that has been recognized in the past and determined to have a high collision possibility, and cancels the setting of the prediction area set in Fig. 6 (S604). Then, the tracking process that is performed in normal collision avoidance control is performed (S605).

[0064] If S603 returns Yes, the prediction area detection determination unit 112 determines that the moving object detected in S601 is likely to be the same moving object included in the prediction recognition information that was previously recognized and determined to have a high collision risk. That is, it determines that the detection information for the moving object detected in S601 when the host vehicle turns right or left is the same as the past detection information for the moving object included in the prediction recognition information. In this case, tracking processing is performed only on moving objects detected in the prediction area. Using the prediction recognition information, the number of detections required for tracking processing (tracking processing time) can be reduced, enabling identification of an object to be avoided. This simplifies the tracking processing method while maintaining detection accuracy. As a result, the period required to identify an object to be avoided, which is necessary for normal collision avoidance control, can be shortened (S606). The control unit 102 then performs collision avoidance control or issues a warning for the object to be avoided identified by the recognition processing in S605 or S606 (S607).

[0065] Finally, how the right / left turn determining unit 120 of the control unit 102 determines whether the host vehicle is turning right or left will be described with reference to FIG.

[0066] The right / left turn determination unit 120 acquires control information for the direction indicator 130 (S700). The right / left turn determination unit 120 acquires information about the vehicle's position from information from the yaw rate sensor 125 and map information 129, etc. (S701). From the information acquired in S700, it is determined whether the direction indicator is being operated for a left or right turn (S702). From the yaw rate sensor information acquired in S701, it is determined whether the direction is changing in the same direction as the direction indicator direction indicated by the vehicle (S703). If the determination results in S702 and S703 are Yes, the right / left turn determination unit 120 determines that the vehicle is in a left or right turn state (S704).

[0067] In this embodiment, (1) when overtaking a moving body such as a bicycle while approaching an intersection, recognition information for the moving body is stored, and when it is predicted that the host vehicle will turn left at the intersection, the previously stored recognition information is read out to determine the possibility of a collision between the host vehicle and the previously stored moving body; (2) using information on the moving body identified as having a high collision possibility as a result of determining the possibility of collision, the detection timing and detection area at which the moving body is predicted to be detected in the sensor detection range of the host vehicle when the host vehicle turns left is calculated as predicted recognition information, and the area at which the moving body is predicted to be detected is set in the sensor detection area at least in front of the host vehicle; (3) when an object is detected in the predicted detection timing or detection area when the host vehicle turns left at an intersection, it is determined that the previously stored moving body has been re-detected, and the period for identifying it as an avoidance target for which avoidance control is to be performed is shortened, thereby ensuring the control margin time required for avoidance control.By using these three methods, it is possible to shorten the period from object detection, to recognizing the object for a certain period of time, and identifying it as an avoidance target, which is necessary before collision avoidance control can be performed, by adding predictive recognition information. This shortening of the period allows the timing from the first detection of an obstacle to the start of avoidance control to be accelerated, making it possible to ensure sufficient control time to avoid the obstacle.

[0068] The issue defined at the beginning of this paper was to ensure the control margin time for avoidance control against obstacles approaching from behind the vehicle when turning right or left at an intersection. By shortening the period in which obstacles are identified as targets for avoidance using predictive recognition information, it is possible to ensure the control margin time for avoidance control against obstacles approaching from behind and perform collision avoidance control.

[0069] As described above, the vehicle control device 100 of this embodiment includes an external environment recognition unit 101 that acquires external environment information at least ahead of the host vehicle, a control unit 102 that controls the host vehicle based on the external environment information acquired by the external environment recognition unit 101, and a memory unit 106 that stores detection information of the moving body acquired (detected) by the external environment recognition unit 101 when the host vehicle overtakes a moving body. When a lane change of the host vehicle is predicted, the external environment recognition unit 101 reads out past detection information of the moving body stored in the memory unit 106, determines the possibility of a collision between the host vehicle and the moving body, and sets (the predicted detection timing and predicted detection area of) a moving body that is likely to collide with the host vehicle when the host vehicle changes lane as predicted recognition information based on the collision possibility (obstacle prediction recognition unit 105), and determines whether the detection information of the moving body detected by the external environment recognition unit when the host vehicle changes lane is the same as the past detection information of the moving body included in the predicted recognition information (recognition unit 104).

[0070] When the vehicle control device 100 determines that the detection information of a moving object detected by the external environment recognition unit 101 when the vehicle changes course is identical to the past detection information of the moving object contained in the predictive recognition information, it shortens the period for identifying the moving object as an object to be avoided in collision avoidance control, identifies the moving object as an object to be avoided, and transitions to avoidance control or an alert for the moving object (recognition unit 104).

[0071] The predicted recognition information includes a predicted detection timing and a predicted detection area of ​​the moving body, and the vehicle control device 100 sets a predicted detection area (predicting the detection area) in which the moving body is predicted to be detected by the external environment recognition unit 101 when the host vehicle changes course, in at least a part of the external environment information of the host vehicle acquired by the external environment recognition unit 101, sets a predicted detection timing (predicting the detection timing) in which the moving body is predicted to be detected by the external environment recognition unit 101 when the host vehicle changes course, and determines whether the detection information of the moving body detected by the external environment recognition unit 101 when the host vehicle changes course is identical to past detection information of the moving body, based on the predicted detection timing and the predicted detection area (recognition unit 104).

[0072] The vehicle control device 100 compares the intersection arrival time of the vehicle with the intersection arrival time of the moving object calculated from past detection information of the moving object, and determines the possibility of collision (obstacle prediction recognition unit 105).

[0073] When the vehicle control device 100 determines that the detection information of a moving object detected by the external environment recognition unit 101 when the vehicle changes course is identical to the past detection information of the moving object contained in the predictive recognition information, it shortens the tracking processing time for identifying the moving object as an object to be avoided in collision avoidance control and identifies the moving object as an object to be avoided (recognition unit 104).

[0074] The vehicle control device 100 sets the predicted detection area in a part of an image captured by a camera sensor that captures an image of at least the area ahead of the host vehicle (recognition unit 104).

[0075] The vehicle control device 100 sets a reliability for the predicted recognition information based on the time when the moving object was detected and the detection accuracy (weather conditions, etc.), and uses the reliability to determine the consistency (identity) of the detection information and the past detection information.

[0076] In other words, the vehicle control device 100 according to this embodiment includes a control unit 102 that performs collision avoidance control to avoid a collision between the host vehicle and an object ahead of the host vehicle, an external environment recognition unit 101 that recognizes obstacles, a memory unit 106 that stores obstacle recognition information, an obstacle prediction and recognition unit 105 that calculates the distance to the intersection of the host vehicle and the time it takes to reach the intersection, and determines the possibility of a collision when the host vehicle turns left at the intersection for each piece of obstacle information stored in the memory unit 106, and generates predicted recognition information, and a prediction area setting unit that identifies, from the predictive recognition information, obstacles that are likely to collide when the host vehicle turns left at the intersection, and sets an area where the obstacle is predicted to be detected. When the host vehicle turns left at the intersection, the vehicle control device 100 determines that a previously stored moving object has been re-detected, and shortens the time required to identify the obstacle as an avoidance target for which avoidance control is to be performed, thereby ensuring the control margin time required for avoidance control.

[0077] That is, the vehicle control device 100 according to this embodiment stores recognition information when overtaking a two-wheeled vehicle, reads out the recognition information of previously recognized two-wheeled vehicles when turning right or left at an intersection, determines the possibility of a collision with the vehicle, calculates predicted recognition information, and when an obstacle is actually detected when the vehicle turns left, immediately determines that it is a re-detection of a moving object that was previously recognized by using the predicted recognition information, and provides means and a device for performing avoidance control.

[0078] According to the present invention, when a vehicle is traveling on a road before an intersection, (1) when passing a moving object such as a bicycle while traveling before an intersection, recognition information of the moving object is stored, and when it is predicted that the vehicle will turn left at the intersection, the previously stored recognition information is read out and a collision possibility between the vehicle and the previously stored moving object is determined; (2) using information on the moving object identified as having a high collision possibility as a result of determining the collision possibility, a detection timing and a detection area in which the moving object is predicted to be detected in the sensor detection range of the vehicle when the vehicle turns left are calculated as predicted recognition information, and the moving object is predicted to be detected in the sensor detection area at least ahead of the vehicle. (3) When the vehicle turns left at an intersection, if an object is detected at the predicted detection timing or in the detection area, it is determined that a previously stored moving object has been re-detected, and the period for identifying it as an object to be avoided for avoidance control is shortened, thereby ensuring the control margin time required for avoidance control.Through these three operations, the current challenge of ensuring the margin time for avoidance control for obstacles coming from behind the vehicle when turning right or left at an intersection is addressed.By shortening the period for identifying an object to be avoided using predictive recognition information, it is possible to ensure the control margin time for avoidance control for obstacles coming from behind and perform collision avoidance control.

[0079] This embodiment makes it possible to prevent accidents involving motorcycles when turning right or left at an intersection, which has been difficult to address in the past.

[0080] It should be noted that the above-described embodiments are merely examples, and the present invention is not limited to these details as long as the features of the invention are not impaired. Also, although various embodiments have been described above, the present invention is not limited to these details. Other aspects conceivable within the scope of the technical idea of ​​the present invention are also included within the scope of the present invention.

[0081] Furthermore, the present invention is not limited to the above-described embodiments and includes various modifications. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and are not necessarily limited to those including all of the described configurations. Furthermore, some or all of the above-described configurations, functions, processing units, processing means, etc. may be implemented in hardware, for example, by designing them as integrated circuits. Furthermore, the above-described configurations, functions, etc. may be implemented in software by a processor interpreting and executing a program that realizes each function. Information such as programs, tables, and files that realize each function can be stored in a storage device such as a memory, a hard disk, or an SSD (Solid State Drive), or in a recording medium such as an IC card, an SD card, or a DVD. [Explanation of symbols]

[0082] 1. General roads including intersections 2. Traffic lights 3. Vehicle 4. Sensor detection area 5...Storage section 6...Motorcycle 7. Images generated by the image processing unit 8. Prediction region 50... Your vehicle 51 Sensor detection area 52 Bicycle 53 Bicycle 54. Period to be identified as an object of avoidance 55 General roads with intersections 100 Vehicle control device 101...External world recognition department 102 Control unit 103 Image processing unit 104...Recognition section 105 Obstacle prediction and recognition unit 106...Storage section 110... Prediction area setting unit 111 Moving object detection unit 112: Prediction area detection and judgment unit 113 Tracking processing unit 114 Intersection distance calculation unit 115 Intersection arrival time calculation unit 116...Collision possibility determination unit 117...Prediction reliability setting section 120 Right / Left Turn Judgment Unit 121 Collision risk calculation unit 122 Route planning section 123 Control instruction section 124 Braking force / driving force control ECU 125···Yaw rate sensor 126 Imaging unit 127 Steering angle sensor 128...Alarm instruction section 129···Map Information 130...turn signal 131···Communication Network

Claims

1. an external environment recognition unit that acquires external environment information at least ahead of the host vehicle; a control unit that controls the host vehicle based on the external environment information acquired by the external environment recognition unit; a storage unit that stores detection information of the moving object acquired by the external environment recognition unit when the host vehicle overtakes the moving object, When a lane change of the host vehicle is predicted, the external environment recognition unit reads out past detection information of the moving body stored in the memory unit, determines a possibility of a collision between the host vehicle and the moving body, and sets a moving body that is likely to collide with the host vehicle when the host vehicle changes lane as predicted recognition information based on the collision possibility, determines whether or not detection information of the moving body detected by the external environment recognition unit when the host vehicle changes lane is identical to past detection information of the moving body included in the predicted recognition information, and if it is determined that the detection information of the moving body detected by the external environment recognition unit when the host vehicle changes lane is identical to the past detection information of the moving body included in the predicted recognition information, shortens a period for identifying the moving body as an avoidance target in collision avoidance control, identifies the moving body as an avoidance target, and transitions to avoidance control or an alert for the moving body.

2. an external environment recognition unit that acquires external environment information at least ahead of the host vehicle; a control unit that controls the host vehicle based on the external environment information acquired by the external environment recognition unit; a storage unit that stores detection information of the moving object acquired by the external environment recognition unit when the host vehicle overtakes the moving object, When a lane change of the host vehicle is predicted, the external environment recognition unit reads out past detection information of the moving body stored in the memory unit, determines a possibility of a collision between the host vehicle and the moving body, and sets a moving body that is likely to collide with the host vehicle when the host vehicle changes lane as predicted recognition information based on the collision possibility, determines whether detection information of the moving body detected by the external environment recognition unit when the host vehicle changes lane is identical to past detection information of the moving body included in the predicted recognition information, and if it is determined that the detection information of the moving body detected by the external environment recognition unit when the host vehicle changes lane is identical to the past detection information of the moving body included in the predicted recognition information, shortens a tracking processing time for identifying the moving body as an avoidance target for collision avoidance control and identifies the moving body as an avoidance target.

3. the predicted recognition information includes a predicted detection timing and a predicted detection area of ​​the moving object; 3. The vehicle control device according to claim 1 or 2, wherein the vehicle control device sets a predicted detection area in which the moving body is predicted to be detected by the external environment recognition unit when the vehicle changes course, in at least a part of the external environment information of the vehicle acquired by the external environment recognition unit, sets a predicted detection timing in which the moving body is predicted to be detected by the external environment recognition unit when the vehicle changes course, and determines whether the detection information of the moving body detected by the external environment recognition unit when the vehicle changes course is identical to past detection information of the moving body, using the predicted detection timing and the predicted detection area as conditions.

4. 3. The vehicle control device according to claim 1, wherein the vehicle control device compares an intersection arrival time of the vehicle with an intersection arrival time of the moving object calculated from past detection information of the moving object to determine the possibility of a collision.

5. The vehicle control device according to claim 3, wherein the predicted detection area is set in a part of an image captured by a camera sensor that captures an image of at least a front area of ​​the host vehicle.

6. 3. The vehicle control device according to claim 1, wherein the vehicle control device performs the collision possibility determination as needed until the host vehicle reaches the intersection.

7. 4. The vehicle control device according to claim 3, characterized in that, if the moving object included in the predictive recognition information is not detected in the predictive detection area even after the predictive detection timing has passed, the vehicle control device deletes the predictive recognition information from the storage unit.

8. 3. The vehicle control device according to claim 1, wherein the vehicle control device sets a reliability for the predicted recognition information based on the time when the moving object is detected and the detection accuracy, and uses the reliability to determine whether the detection information matches the past detection information.

9. 3. The vehicle control device according to claim 1, wherein the vehicle control device calculates a driving route of the vehicle from map information and determines that there is a high possibility of turning right or left at an intersection as a trigger for reading out the past detection information, thereby reading out the past detection information.

10. 3. The vehicle control device according to claim 1, wherein the vehicle control device reads the past detection information by determining, from an intersection recognition result by the external environment recognition unit, that there is a high possibility of turning right or left at the intersection, as a trigger for reading the past detection information.

11. 3. The vehicle control device according to claim 1, wherein the vehicle control device uses intersection information acquired through road-to-vehicle communication as a trigger for reading out the past detection information, and reads out the past detection information by determining that there is a high possibility of turning right or left at the intersection.

12. 4. The vehicle control device according to claim 3, wherein the vehicle control device cancels the setting of the predicted detection area when it is determined that the detection information of the moving body detected by the external environment recognition unit when the vehicle changes course is not identical to the past detection information of the moving body included in the predicted recognition information.

Images