Driving assistance systems

Abstract

To appropriately prompt a driver of a vehicle 10 to perform safety confirmation.SOLUTION: A driving support device (11) performs notification to a user of an own vehicle (10). The driving support device comprises notification means (15) that performs notification according to a possibility of collision between the own vehicle and a peripheral target based on predicted trajectories of the own vehicle and the peripheral target, based on position potential information indicating a relation between a plurality of future positions at a second time after a minute time from a first time, and a probability that the own vehicle will be present at each of the plurality of future positions at the second time.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to the technical field of driving assistance devices.

Background Art

[0002] As this type of device, for example, based on the risk level by direction obtained based on data of the surrounding situation obtained by a surrounding monitoring sensor that monitors the surrounding situation of a vehicle, and the driver's fixation ratio by direction based on an image captured by an in-vehicle camera, a device for detecting an oversight of risk has been proposed (see Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] There is room for improvement in the technology described in Patent Document 1. The present invention has been made in view of this situation, and an object thereof is to provide a driving assistance device that can appropriately prompt a vehicle driver to confirm safety.

Means for Solving the Problems

[0005] A driving assistance device according to one aspect of the present invention is a driving assistance device that provides notification to the user of the vehicle, and based on first position potential information showing the relationship between a plurality of first future positions where the vehicle may be at a second time a small time interval after a first time interval and the probability that the vehicle is at each of the plurality of first future positions at the second time interval, the device calculates second position potential information showing the relationship between a plurality of second future positions where the vehicle may be at a third time a small time interval after a second time interval and the probability that the vehicle is at each of the plurality of second future positions at the third time interval, assuming that the vehicle is at one of the plurality of first future positions with a relatively high probability at the second time interval, and predicts the trajectory of the vehicle based on the calculation of second position potential information showing the relationship between a plurality of second future positions where the vehicle may be at a third time a small time interval after a second time interval and the probability that the vehicle is at each of the plurality of second future positions at the third time interval, and the vehicle The system includes a notification means that provides notification based on the predicted trajectory of the surrounding object, which is determined by calculating a fourth position potential information that shows the relationship between a plurality of third future positions in which the surrounding object may be located at the third time and the probability that the surrounding object is located at each of the plurality of third future positions at the second time, assuming that the surrounding object is located at one of the plurality of third future positions in which the probability is relatively high at the second time. [Brief explanation of the drawing]

[0006] [Figure 1] This is a block diagram showing the configuration of the driver assistance system according to the embodiment. [Figure 2] This is a conceptual diagram showing an example of an computational model. [Figure 3] This is a conceptual diagram illustrating an example of a position potential. [Figure 4] This is a conceptual diagram showing an example of a predicted trajectory. [Figure 5] This is a diagram illustrating the concept of contact possibility. [Figure 6]This is a flowchart showing the operation of the driver assistance system according to the embodiment. [Modes for carrying out the invention]

[0007] Embodiments relating to the driver assistance device will be described with reference to Figures 1 to 6. In the following embodiments, driver assistance system 1 is given as an example of a driver assistance device.

[0008] In Figure 1, the driver assistance system 1 comprises a vehicle 10 and a server 20. The vehicle 10 includes an ECU (Electronic Control Unit) 11, an external sensor 12, an internal sensor 13, a communication device 14, an HMI (Human Machine Interface) 15, and an actuator 16. The server 20 includes a computing device 21, a storage device 22, and a communication device 23. The vehicle 10 and the server 20 are configured to communicate with each other via the communication devices 14 and 23.

[0009] The external sensor 12 is a sensor that detects the conditions around the vehicle 10. The external sensor 12 may include a front camera positioned to capture images in front of the vehicle 10, an omnidirectional camera having multiple cameras positioned to capture images to the sides and rear of the vehicle 10, a millimeter-wave radar capable of detecting objects in front of the vehicle 10, and a clearance sonar capable of detecting objects in the vicinity of and around the vehicle 10. The external sensor 12 may also include a LiDAR (Light Detection and Ranging).

[0010] The internal sensor 13 is a sensor that detects the motion state of the vehicle 10. The internal sensor 13 may include a wheel speed sensor, an acceleration / deceleration sensor, a yaw rate sensor, a steering angle sensor, an accelerator sensor, and a brake sensor. The motion state of the vehicle 10 may also be referred to as the behavior of the vehicle 10.

[0011] The HMI 15 is an interface for inputting and outputting information between the driver of the vehicle 10 and the vehicle's system (e.g., the ECU 11). The HMI 15 may have a display, a buzzer, and a speaker. The display of the HMI 15 may be a HUD (Head Up Display) or a MID (Multi Information Display) provided on the instrument panel.

[0012] The actuator 16 is a device used to control the vehicle 10. The actuator 16 may include a drive actuator, a brake actuator, and a steering actuator.

[0013] If vehicle 10 is equipped with an engine, the drive actuator may be an actuator for controlling the amount of air supplied to the engine (e.g., throttle opening). If vehicle 10 is equipped with a motor as a power source, the motor as a power source may be referred to as the drive actuator. The brake actuator is an actuator for controlling the braking force applied to the wheels of vehicle 10. The steering actuator is an actuator for controlling the steering torque.

[0014] The ECU 11 may be configured to perform autonomous driving by automatically controlling at least one of the acceleration / deceleration and steering of the vehicle 10 based on the surrounding conditions of the vehicle 10 detected by the external sensor 12 and the motion state of the vehicle 10 detected by the internal sensor 13. In other words, the ECU 11 may have an autonomous driving function.

[0015] The ECU 11 may be configured to perform driver assistance, such as issuing a warning to the driver, activating safety devices of the vehicle 10, or controlling at least one of the acceleration / deceleration and steering of the vehicle 10, based on the surrounding conditions of the vehicle 10 detected by the external sensor 12 and the motion state of the vehicle 10 detected by the internal sensor 13, in order to avoid a collision between the vehicle 10 and an object. In this case, the ECU 11 may constitute part of an Advanced Driver-Assistance System (ADAS).

[0016] The ECU 11 may recognize objects present around the vehicle 10 based on the detection results from the external sensor 12. These recognized objects may include moving objects such as vehicles and pedestrians, as well as stationary objects such as structures and fallen objects. The ECU 11 may also recognize regulatory information such as traffic lights, road signs, and road markings based on the detection results from the external sensor 12. Since various existing methods can be applied to the technology for recognizing objects based on the detection results from the external sensor 12, a detailed explanation is omitted. The objects present around the vehicle 10 recognized by the ECU 11 may represent just one specific example of the conditions surrounding the vehicle 10.

[0017] The ECU 11 transmits to the server 20 via the communication device 14 information indicating objects present around the recognized vehicle 10 and vehicle environment information indicating the motion state of the vehicle 10 detected by the internal sensor 13. The information indicating the objects may include information indicating the attributes of the objects and information indicating the location of the objects. The object attributes may include at least one of pedestrians, passenger cars, large vehicles, motorcycles, mopeds, bicycles, structures, and fallen objects. The information indicating the location of the objects may include the orientation of the objects as seen from the vehicle 10 and the distance from the vehicle 10 to the objects. The vehicle environment information may also include information indicating the attributes of the road on which the vehicle 10 is traveling (e.g., expressway, non-expressway, number of lanes).

[0018] The arithmetic unit 21 of the server 20 may store the vehicle environment information transmitted from the vehicle 10 in the storage device 22. Note that the arithmetic unit 21 may acquire (in other words, receive) the vehicle environment information from each of a plurality of vehicles (including the vehicle 10) configured to be communicable with the server 20. The acquired vehicle environment information may be stored in the storage device 22. As a result, a large-scale database (so-called big data) related to the vehicle environment information may be constructed in the storage device 22.

[0019] The arithmetic unit 21 may use the vehicle environment information transmitted from the vehicle 10 for the learning process of an arithmetic model for predicting the trajectories of objects including the vehicle 10. The arithmetic unit 21 may also use the vehicle environment information transmitted from the vehicle 10 for the prediction process of predicting the trajectories of objects including the vehicle 10.

[0020] The learning process of the arithmetic model performed by the arithmetic unit 21 will be described. Note that the arithmetic model for predicting the trajectory of an object may mean an arithmetic model that outputs a prediction result of the trajectory of an object when vehicle environment information is input. As an example of such an arithmetic model, an arithmetic model using a neural network (for example, Convolutional Neural Network: CNN) can be mentioned.

[0021] The arithmetic unit 21 first classifies a plurality of pieces of vehicle environment information stored in the storage device 22. For example, the arithmetic unit 21 may classify the plurality of pieces of vehicle environment information based on at least one classification item such as whether it is an exclusive road for automobiles, the number of lanes, the attributes of the object, the congestion conditions in front of and behind the host vehicle (that is, the vehicle that transmitted the vehicle environment information to the server 20), and the direction in which the object exists as seen from the host vehicle. At this time, the arithmetic unit 21 may obtain the similarity between one piece of vehicle environment information and another piece of vehicle environment information for the above classification items, and classify the plurality of pieces of vehicle environment information based on the obtained similarity.

[0022] In the classification described above, for example, multiple vehicle environment information is classified (in other words, categorized) for each case, such as when the vehicle (i.e., the vehicle that transmitted the vehicle environment information to the server 20) is traveling on an expressway without pedestrians or bicycles, when the vehicle is traveling near an intersection with traffic lights, when the vehicle is traveling on a curved section, when the vehicle is traveling on a road without a sidewalk and there are pedestrians around the vehicle, etc. In other words, in the classification described above, multiple vehicle environment information may be classified based on road conditions (e.g., road shape and gradient, road attributes, operating speed, degree of pedestrian-vehicle separation, traffic volume, etc.). In addition to road conditions, multiple vehicle environment information may also be classified based on environmental conditions (e.g., time of day, weather, etc.).

[0023] The arithmetic unit 21 may further classify the multiple vehicle environment information classified by the above-described process. For example, the arithmetic unit 21 may further classify and subdivide the multiple vehicle environment information classified into Category 1 by the above-described process. In this case, the arithmetic unit 21 may further classify the multiple vehicle environment information classified into Category 1 based on at least one of the classification items: the positional relationship between the vehicle and objects present in the vicinity of the vehicle, and the relationship between the attributes of the vehicle and the attributes of objects present in the vicinity of the vehicle. In this case, the arithmetic unit 21 may determine the similarity between one piece of vehicle environment information and other pieces of vehicle environment information with respect to the above classification item, and classify the multiple pieces of vehicle environment information based on the determined similarity. The arithmetic unit 21 may perform the same process for multiple pieces of vehicle environment information classified into Category n (where n is a natural number of 2 or more).

[0024] In the further classification described above, for example, when the vehicle itself (i.e., the vehicle that sent the vehicle environment information to the server 20) is traveling on an expressway, multiple vehicle environment information may be classified (in other words, categorized into cases) for each of the following situations: (i) when there is another vehicle ahead of the lane in which the vehicle is traveling; (ii) when there is another vehicle ahead of the vehicle and in a lane adjacent to the lane in which the vehicle is traveling; (iii) when there is no other vehicle ahead of the vehicle; (iv) when there is another vehicle behind the vehicle and in a lane adjacent to the lane in which the vehicle is traveling. In other words, in the further classification described above, multiple vehicle environment information may be further classified based on the relationship between the vehicle and objects present in its vicinity.

[0025] As described above, the computing device 21 may broadly categorize multiple vehicle environment information into multiple vehicle environment information groups based on, for example, road conditions. The computing device 21 may then further classify the multiple vehicle environment information belonging to each of the multiple vehicle environment information groups based on the relationship between the vehicle itself and objects present in its surroundings.

[0026] The computing device 21 may perform machine learning (e.g., deep learning) using the vehicle environment information classified as described above as a learning process for the computational model. In this case, the computing device 21 may perform machine learning using multiple vehicle environment information belonging to each of the multiple vehicle environment information groups. The computing device 21 may perform machine learning using multiple vehicle environment information belonging to each of the multiple vehicle information groups generated by further classifying the multiple vehicle information groups belonging to each of the multiple vehicle environment information groups. As a result of the learning process for the computational model performed by the computing device 21, the computing device 21 may generate (or construct) multiple computational models corresponding to the classification of the vehicle environment information.

[0027] Next, we will explain the prediction process performed by the computing device 21 to predict the trajectory of objects, including the vehicle 10. The outline of the prediction process performed by the computing device 21 is as follows: The computing device 21 selects one computing model from among the multiple computing models generated by the learning process described above that is suitable for the vehicle environment information transmitted from the vehicle 10. Then, the computing device 21 inputs the vehicle environment information transmitted from the vehicle 10 into the selected computing model to predict the trajectory of the vehicle 10 and objects in its vicinity.

[0028] Specifically, the computing device 21 may identify the group of vehicle environment information to which the vehicle environment information transmitted from vehicle 10 belongs, based on at least one of the following: whether or not it is an expressway, the number of lanes, the attributes of the object, the degree of congestion in front of and behind the vehicle 10, and the direction in which the object is located as seen from vehicle 10. The term "group of vehicle environment information" refers to the group of vehicle environment information generated in the learning process of the computing model described above by classifying multiple pieces of vehicle environment information based on, for example, road conditions.

[0029] Next, the arithmetic unit 21 may compare the positional relationship between the vehicle (i.e., the vehicle corresponding to vehicle 10) and objects in its vicinity, as indicated by each of the multiple vehicle environment information pieces that correspond to at least a portion of the multiple vehicle environment information pieces that belong to the group of vehicle environment information pieces to which the vehicle environment information transmitted from vehicle 10 belongs (i.e., the multiple vehicle environment information pieces stored in the storage device 22), with the positional relationship between vehicle 10 and objects in its vicinity, as indicated by the vehicle environment information transmitted from vehicle 10, to determine the degree of similarity between the two.

[0030] Specifically, the arithmetic unit 21 may associate (or link) objects located around its own vehicle, indicated by each of the multiple vehicle environment information items corresponding to at least some of the above, with objects located around the vehicle 10, indicated by the vehicle environment information transmitted from the vehicle 10, based on the position of the objects. Next, the arithmetic unit 21 may determine whether the attributes of the linked objects match, based on the attributes of the linked objects (e.g., automobile, motorcycle, bicycle, pedestrian, etc.). At this time, the arithmetic unit 21 may calculate a Jackard coefficient for the linked objects. The arithmetic unit 21 may determine that the attributes of the linked objects match if the calculated Jackard coefficient is equal to or greater than a first predetermined value.

[0031] The arithmetic unit 21 may determine the degree of similarity of each of the multiple vehicle environment information items corresponding to at least a portion of the above-mentioned vehicle environment information transmitted from the vehicle 10, based on the proportion of matching attributes of the linked objects. In this case, the arithmetic unit 21 may set the degree of similarity higher the higher the proportion of matching attributes of the linked objects.

[0032] Next, the arithmetic unit 21 may extract one or more vehicle environment information from a plurality of vehicle environment information corresponding to at least a part of the above, the degree of similarity being higher than the second predetermined value. The arithmetic unit 21 may exclude from the extracted one or more vehicle environment information information any vehicle environment information that includes an object indicated by the one or more vehicle environment information, to which the object is linked, and in at least one of its position and attributes differs significantly from the object indicated by the vehicle environment information transmitted from the vehicle 10.

[0033] In this case, the calculation unit 21 may determine the vehicle environment information to be excluded based on the Euclidean distance and Jackard coefficient for each of the multiple objects indicated by the extracted vehicle environment information. For example, let "n" be the number of multiple objects indicated by the vehicle environment information, "x" be the Euclidean distance, and "y" be the Jackard coefficient. The calculation unit 21 will then use "L=Σ(x i 2 +y i2 L may be calculated using the formula ) / n. The calculated value of L increases as the number of objects that deviate significantly increases. In other words, the calculated value of L decreases as the number of objects that deviate significantly decreases. The calculation device 21 may exclude vehicle environment information in which the calculated value of L is greater than the third predetermined value.

[0034] The computing device 21 may select a computing model corresponding to at least one of the vehicle environment information groups and groups to which the vehicle environment information that was not excluded from the extracted vehicle environment information belongs, as a computing model for predicting the trajectory of an object including the vehicle 10. The selected computing model may be a neural network having multiple hidden layers, as shown in Figure 2. The selected computing model may also be a neural network having only one hidden layer.

[0035] The computing device 21 may input the vehicle environment information transmitted from the vehicle 10 into the selected computing model to obtain information relating to an object including the vehicle 10 after a small time interval Δt (i.e., one or more objects indicated by the vehicle environment information transmitted from the vehicle 10).

[0036] For example, as shown in Figure 2, the calculation model may be input to information relating to the vehicle 10 (e.g., speed, driving route information), information relating to multiple objects ob#1, ob#2, ..., ob#n (e.g., attributes, lane, orientation of the object as seen from the vehicle 10, relative distance from the vehicle 10, relative speed from the vehicle 10, direction of travel of the object as seen from the vehicle 10, status of lights), information relating to the road (e.g., lateral distance to the white line, distance to the stop line, distance to the traffic light, distance to the pedestrian crossing), and information relating to regulations (speed limit, traffic light color, lane restrictions). Note that the information relating to the vehicle 10, the information relating to multiple objects ob#1, ob#2, ..., ob#n, the information relating to the road, and the information relating to regulations are information indicated by the vehicle environment information transmitted from the vehicle 10.

[0037] For example, as shown in Figure 2, the calculation model may output information relating to the vehicle 10 after a small time interval Δt, and information relating to multiple objects ob#1, ob#2, ..., ob#n after a small time interval Δt. The information relating to the vehicle 10 after a small time interval Δt may be position potential information showing the relationship between multiple possible future positions where the vehicle 10 may exist after Δt and the probability that the vehicle 10 will exist in each of these multiple future positions after Δt.

[0038] The above position potential information can be visually represented as shown in Figure 3(a). Of the position potential shown in Figure 3(a), part a shows the range of future positions where vehicle 10 may be located after Δt if vehicle 10 is traveling in a straight line. Part b shows the range of future positions where vehicle 10 may be located after Δt if vehicle 10 is turning to the right. Part c shows the range of future positions where vehicle 10 may be located after Δt if vehicle 10 is turning to the left. In the position potential shown in Figure 3(a), darker colors indicate a higher probability of vehicle 10 being located after Δt compared to lighter colors.

[0039] Similarly, information relating to multiple objects ob#1 after a small time interval Δt may be positional potential information showing the relationship between multiple possible future locations where object ob#1 may exist after Δt and the probability that object ob#1 may exist in each of those future locations after Δt. Information relating to multiple objects ob#2 after a small time interval Δt may be positional potential information showing the relationship between multiple possible future locations where object ob#2 may exist after Δt and the probability that object ob#2 may exist in each of those future locations after Δt. Information relating to multiple objects ob#n after a small time interval Δt may be positional potential information showing the relationship between multiple possible future locations where object ob#n may exist after Δt and the probability that object ob#n may exist in each of those future locations after Δt.

[0040] Here, the vehicle environment information transmitted from vehicle 10 represents information relating to vehicle 10 at time t1. In this case, when the vehicle environment information transmitted from vehicle 10 is input into the calculation model, the information relating to vehicle 10 after Δt output from the calculation model represents information relating to vehicle 10 at time t2, which is Δt after time t1.

[0041] The calculation device 21 may input information relating to the vehicle 10 at time t2 (i.e., information output from the calculation model) into the calculation model to obtain information relating to the vehicle 10 at time t3, Δt after time t2. At this time, the calculation device 21 may set the position of the vehicle 10 at time t2 to the position with the highest probability indicated by the position potential information relating to the vehicle 10 at time t2. Similarly, the calculation device 21 may set the positions of each of the multiple objects ob#1, ob#2, ..., ob#n at time t2 to the position with the highest probability indicated by the position potential information relating to each of the multiple objects ob#1, ob#2, ..., ob#n at time t2.

[0042] The calculation device 21 may further input information relating to the vehicle 10 at time t3 into the calculation model to obtain information relating to the vehicle 10 at time t4, Δt after time t3. At this time, the calculation device 21 may set the position of the vehicle 10 at time t3 to the position with the highest probability indicated by the position potential information relating to the vehicle 10 at time t3. Similarly, the calculation device 21 may set the positions of each of the multiple objects ob#1, ob#2, ..., ob#n at time t3 to the position with the highest probability indicated by the position potential information relating to each of the multiple objects ob#1, ob#2, ..., ob#n at time t3.

[0043] The arithmetic unit 21 can predict the positional transitions of the vehicle 10 and each of the multiple objects ob#1, ob#2, ..., ob#n through the processing described above. For example, the positional potential information for the vehicle 10 at times t2, t3, and t4 can be visually represented as shown in Figure 3(b). In Figure 3(b), "P1" indicates the position of the vehicle 10 at time t1, "P2" shows an example of the position of the vehicle 10 at time t2, "P3" shows an example of the position of the vehicle 10 at time t3, and "P4" shows an example of the position of the vehicle 10 at time t4. The arithmetic unit 21 may predict the trajectory of the vehicle 10 by connecting positions P1, P2, P3, and P4 in Figure 3(b), for example.

[0044] As a result of this processing, the computing unit 21 may predict the trajectory Tr1 of the vehicle 10, the trajectory Tr2 of object Ob#1, and the trajectory Tr3 of object Ob#2, for example, as shown in Figure 4. For convenience, only the trajectories of three objects (i.e., vehicle 10, object Ob#1, and object Ob#2) are shown in Figure 4, but the computing unit 21 may predict the trajectories of four or more objects. Also, to avoid complexity, the drawing of the position potential shown in Figure 3 is omitted in Figure 4.

[0045] Next, the computing unit 21 makes a determination regarding the possibility of contact between the vehicle 10 and the multiple objects ob#1, ob#2, ..., ob#n based on the predicted trajectory. As mentioned above, the position potential is not shown in Figure 4, but the trajectory predicted by the computing unit 21 is represented as a series of position potentials, for example, as shown in Figure 3(b).

[0046] When the computing unit 21 determines the possibility of contact between the vehicle 10 and an object in the vicinity of the vehicle 10 (for example, object Ob#1), it may determine the possibility of contact between the vehicle 10 and the object based on a plurality of future positions indicated by the position potential information of the vehicle 10 and a plurality of future positions indicated by the position potential information of the object at the same time (for example, one of times t2, t3, and t4).

[0047] In Figure 5(a), let's assume that vehicle 30 is an example of the above-mentioned object. As shown in Figure 5(a), if a portion of the position potential of vehicle 10 and a portion of the position potential of vehicle 30 overlap (see dotted circle C1), the computing device 21 may determine that there is a possibility of contact between vehicle 10 and vehicle 30.

[0048] When the computing device 21 determines the possibility of contact between one object (e.g., object Ob#1) and another object (e.g., object Ob#2) in the vicinity of the vehicle 10, it may determine the possibility of contact between the one object and the other object based on a plurality of future positions indicated by the position potential information of the one object and a plurality of future positions indicated by the position potential information of the other object at the same time (e.g., one of times t2, t3, and t4).

[0049] In Figure 5(b), let vehicle 30 be an example of the first object, and vehicle 40 be an example of the other object. As shown in Figure 5(b), if a portion of the position potential of vehicle 30 and a portion of the position potential of vehicle 40 overlap (see dotted circle C2), the computing device 21 may determine that there is a possibility of contact between vehicle 30 and vehicle 40.

[0050] The computing device 21 determines, based on the result of the collision possibility determination, whether or not the driver of vehicle 10 needs to take safety check actions. If the collision possibility determination indicates at least one of the following: there is a possibility of collision between vehicle 10 and an object in the vicinity of vehicle 10, and there is a possibility of collision between one object in the vicinity of vehicle 10 and another object, the computing device 21 may determine that the driver of vehicle 10 needs to take safety check actions.

[0051] The computing unit 21 transmits safety confirmation necessity information, which indicates whether or not the driver of the vehicle 10 needs to perform a safety confirmation action, to the vehicle 10 via the communication device 23. If the safety confirmation necessity information indicates that a safety confirmation action is necessary, the ECU 11 controls the HMI 15 to notify the driver of the vehicle 10 to prompt them to perform a safety confirmation.

[0052] The operation of the driver assistance system 1 configured as described above will be explained with reference to the flowchart in Figure 6. In Figure 6, the ECU 11 of the vehicle 10 acquires detection results (e.g., sensor values) from the external sensor 12 and the internal sensor 13 (step S111). Based on the detection results from the external sensor 12, the ECU 11 may recognize objects present around the vehicle 10. In other words, the ECU 11 may analyze the traffic environment around the vehicle 10 (step S112).

[0053] The ECU 11 transmits vehicle environment information, which includes information indicating objects present around the vehicle 10 recognized in step S112 and the motion state of the vehicle 10 detected by the internal sensor 13, to the server 20 via the communication device 14 (step S113). The server 20 receives the vehicle environment information via the communication device 23 (step S121).

[0054] The arithmetic unit 21 of the server 20 calculates the degree of similarity between the vehicle environment information received in step S121 (i.e., the vehicle environment information transmitted from the vehicle 10) and at least some of the multiple vehicle environment information stored in the storage device 22 (step S122). Based on the result of the processing in step S122, the arithmetic unit 21 selects one arithmetic model from the multiple arithmetic models (step S123).

[0055] The computing device 21 inputs the vehicle environment information received in step S121 into a selected computing model in step S123 to generate predicted trajectories for each of the multiple objects, including the vehicle 10 (step S124). Based on the predicted trajectories generated in step S124, the computing device 21 makes a determination regarding the possibility of contact (step S125).

[0056] Based on the result of the processing in step S125, the arithmetic unit 21 determines whether or not the driver of the vehicle 10 needs to perform a safety check action (step S126). The arithmetic unit 21 transmits safety check necessity information, indicating the result of the processing in step S126, to the vehicle 10 via the communication device 23 (step S127). The vehicle 10 receives the safety check necessity information via the communication device 14 (step S114). The ECU 11 of the vehicle 10 may control the HMI 15 to provide notification to the driver to prompt them to perform a safety check based on the safety check necessity information (step S115).

[0057] (Technical effects) To ensure safety while driving a vehicle, it is not enough to simply detect or sense an imminent danger. It is necessary to understand the surrounding environment of the vehicle, predict potential hazards, and then drive in a preventative manner. However, depending on the driver's circumstances, they may not be able to adequately perform safety checks. In such cases, prompting the driver to perform safety checks from the vehicle itself is a possible solution. However, it takes some time for a driver, prompted by the vehicle, to actually perform the checks and understand the situation. Therefore, when prompting the driver to perform safety checks from the vehicle, it is necessary to ensure that the prompting allows sufficient time for the driver to understand the situation.

[0058] In the driver assistance system 1, the computing unit 21 of the server 20 predicts the trajectory of each of several objects, including the vehicle 10, over a period of n times the infinitesimal time Δt (where "n" is a natural number greater than or equal to 2). Therefore, compared to the comparative example which predicts the trajectory of objects up to an infinitesimal time Δt, the driver assistance system 1 can make determinations regarding the possibility of contact between objects not only at a relatively near future point in time, but also at a relatively distant future point in time. Accordingly, the driver assistance system 1 can provide notifications to encourage safety checks so that the driver of the vehicle 10 has enough time to recognize the situation. In other words, the driver assistance system 1 can appropriately encourage the driver of the vehicle 10 to check for safety.

[0059] In the driver assistance system 1, for example, multiple computational models are provided to correspond to multiple classified environments (or situations) based on road conditions and the relationship between the vehicle and objects in its surroundings. When the trajectory of each of the multiple objects, including the vehicle 10, is predicted, one computational model is selected from the multiple computational models based on the vehicle environment information transmitted from the vehicle 10. In other words, the driver assistance system 1 selects the optimal computational model for predicting the trajectory of each of the multiple objects, including the vehicle 10. Therefore, the driver assistance system 1 can predict the trajectory of each of the multiple objects, including the vehicle 10, with high accuracy.

[0060] As described above, if the system indicates at least one of the following: that there is a possibility of contact between vehicle 10 and an object in its vicinity, or that there is a possibility of contact between one object in the vicinity of vehicle 10 and another object, the computing unit 21 may determine that the driver of vehicle 10 needs to take safety check actions. In other words, the driver assistance system 1 notifies the driver of vehicle 10 to take safety checks not only when there is a possibility of contact between the vehicle (e.g., vehicle 10) and an object in its vicinity, but also when there is a possibility of contact between objects in the vicinity of vehicle 10. Therefore, the driver assistance system 1 can prompt the driver to take safety checks not only for direct risks related to vehicle 10, but also for indirect risks.

[0061] The embodiments of the invention derived from the above-described embodiments are described below.

[0062] A driving assistance device according to one aspect of the invention is a driving assistance device that provides notification to the user of the vehicle, and based on first position potential information showing the relationship between a plurality of first future positions where the vehicle may be at a second time a small time interval after a first time interval and the probability that the vehicle is at each of the plurality of first future positions at the second time interval, the device calculates second position potential information showing the relationship between a plurality of second future positions where the vehicle may be at a third time a small time interval after a second time interval and the probability that the vehicle is at each of the plurality of second future positions at the third time interval, assuming that the vehicle is at one of the plurality of first future positions with a relatively high probability at the second time interval, and predicts the trajectory of the vehicle, and the vehicle The system includes a notification means that provides notification based on the predicted trajectory of the surrounding object, which is determined by calculating a fourth position potential information that shows the relationship between a plurality of third future positions in which the surrounding object may be located at the third time and the probability that the surrounding object is located at each of the plurality of third future positions at the second time, assuming that the surrounding object is located at one of the plurality of third future positions with a relatively high probability at the second time.

[0063] In the said driving support device, if a first surrounding object and a second surrounding object exist as surrounding objects, the notification means may provide notification in addition to the possibility of contact between the vehicle and the surrounding objects, based on the possibility of contact between the first surrounding object and the second surrounding object, which is determined by the predicted trajectory of the first surrounding object and the predicted trajectory of the second surrounding object.

[0064] The present invention is not limited to the embodiments described above, and can be modified as appropriate without contradicting the gist or idea of ​​the invention as can be read from the claims and specification as a whole. Driving assistance devices with such modifications are also included within the technical scope of the present invention. [Explanation of symbols]

[0065] 1…Driver assistance system, 10…Vehicle, 11…ECU, 12…External sensor, 13…Internal sensor, 14, 23…Communication device, 15…HMI, 20…Server, 21…Computer unit, 22…Memory device

Claims

1. A driver assistance device that provides notifications to the user of the vehicle, Based on first position potential information showing the relationship between a plurality of first future positions where the vehicle may be at a second time a small time interval after the first time interval, and the probability that the vehicle is at each of the plurality of first future positions at the second time interval, the trajectory of the vehicle is predicted by calculating second position potential information showing the relationship between a plurality of second future positions where the vehicle may be at a third time a small time interval after the second time interval, and the probability that the vehicle is at each of the plurality of second future positions at the third time interval. Based on third position potential information showing the relationship between a plurality of third future positions in which a surrounding object present in the vicinity of the vehicle may exist at the second time, and the probability that the surrounding object is present at each of the plurality of third future positions at the second time, the trajectory of the surrounding object is predicted by calculating fourth position potential information showing the relationship between a plurality of fourth future positions in which the surrounding object may exist at the third time, assuming that the surrounding object is present at one of the plurality of third future positions with a relatively high probability at the second time, and a plurality of fourth future positions in which the surrounding object may exist at the third time, The vehicle is equipped with notification means that provides notification based on the possibility of contact between the vehicle and the surrounding object. A driving assistance device characterized by the following features.

2. If a first surrounding object and a second surrounding object exist as the surrounding objects, the notification means provides notification based on the possibility of contact between the vehicle and the surrounding objects, as well as the possibility of contact between the first surrounding object and the second surrounding object based on the predicted trajectory of the first surrounding object and the predicted trajectory of the second surrounding object. The driving support device according to feature 1.

Images