Other vehicle action prediction apparatus, method, and program product

Abstract

The present disclosure relates to other-vehicle-action-prediction devices, methods, and program products. An other-vehicle-action-prediction device predicts whether a driver is present in an other vehicle located in the periphery of a host vehicle based on a detection result of a periphery-condition sensor that detects a periphery condition of the host vehicle, and predicts an action of the other vehicle based on a prediction result of whether a driver is present in the other vehicle and a time-series detection result of the other vehicle obtained by the periphery-condition sensor.

Description

Technical Field

[0001] This disclosure relates to other vehicle motion prediction devices, other vehicle motion prediction methods, and program products. Background Technology

[0002] Patent document 1 (Japanese Patent Publication No. 2023-508986) describes a technology for predicting the intentions of users who share roads with vehicles.

[0003] While previous predictions of other vehicles' actions have been conducted, these predictions did not consider the presence of drivers in those vehicles. When other vehicles are manually driven, such as in situations without intersections or leading vehicles traveling at low speeds, lane changes, acceleration, and deceleration are frequently observed. In contrast, when other vehicles are assumed to be autonomously driven without drivers, such as in situations without intersections or leading vehicles traveling at low speeds, lane changes, acceleration, and deceleration are almost never observed. Nevertheless, because previous predictions of other vehicles' actions did not consider the presence of drivers, high-precision predictions of other vehicles' actions were not possible. Summary of the Invention

[0004] In view of the above points, the purpose of this disclosure is to provide a device, method, and program product for predicting the actions of other vehicles with high accuracy.

[0005] (1) One aspect of this disclosure is an other vehicle action prediction device having a processor, wherein the processor is configured to: predict whether there is a driver among other vehicles located in the vicinity of the vehicle based on the detection results of a surrounding condition sensor that detects the surrounding condition of the vehicle; and predict the action of the other vehicles based on the prediction result of whether there is a driver among the other vehicles and the detection results of the time series of the other vehicles obtained by the surrounding condition sensor.

[0006] (2) In other vehicle action prediction devices of scheme (1), the action of the other vehicle predicted when the driver is predicted to be present in the other vehicle may be different from the action of the other vehicle predicted when the driver is predicted not to be present in the other vehicle.

[0007] (3) In other vehicle action prediction devices of scheme (1) or (2), the processor may use a prediction model to predict the likelihood of the other vehicle changing lanes based on the prediction result of whether the driver exists in the other vehicle and the detection result of the time series of the other vehicle obtained by the surrounding condition sensor. The prediction model is a model obtained by learning using training data. The training data is a dataset of detection results and labels of the time series of the learning other vehicle from the first time point to the second time point obtained by the learning surrounding condition sensor mounted on the learning vehicle. The labels are labels indicating whether the driver exists in the learning other vehicle and whether the learning other vehicle changed lanes at the third time point after the second time point.

[0008] (4) In any of the other vehicle action prediction devices in schemes (1) to (3), the processor may predict whether the driver is present in the other vehicles based on the behavior of the other vehicles detected by the surrounding situation sensor.

[0009] (5) In any of the other vehicle movement prediction devices in schemes (1) to (4), the processor may predict whether the driver is present in the other vehicle based on images of the other vehicle captured by a camera that serves as the surrounding situation sensor.

[0010] (6) In any of the other vehicle movement prediction devices in schemes (1) to (5), the processor may predict whether the driver is in the other vehicle based on information obtained by a wireless communication device that serves as the surrounding condition sensor via wireless communication with the outside of the vehicle, indicating whether the driver is in the other vehicle.

[0011] (7) One aspect of this disclosure is a method for predicting the actions of other vehicles, the method comprising: predicting whether a driver exists in other vehicles located around the vehicle based on the detection results of a surrounding condition sensor that detects the surrounding conditions of the vehicle; and predicting the actions of the other vehicles based on the prediction results of whether the driver exists in the other vehicles and the detection results of the time series of the other vehicles obtained by the surrounding condition sensor.

[0012] (8) One aspect of this disclosure is a program product containing a computer program that causes a processor to perform the following actions: predicting the presence of a driver in other vehicles located in the vicinity of the vehicle based on detection results from a surrounding condition sensor that detects the surrounding conditions of the vehicle; and predicting the actions of the other vehicles based on the prediction of the presence of the driver in the other vehicles and the detection results of the time series of the other vehicles obtained by the surrounding condition sensor.

[0013] According to this disclosure, the actions of other vehicles can be predicted with high accuracy. Attached Figure Description

[0014] Figure 1 This is a diagram showing an example of the vehicle 1 that applies the other vehicle movement prediction device 15 of the first embodiment.

[0015] Figure 2 A is a diagram representing an example of the prediction results of the actions of other vehicle OV obtained by other vehicle action prediction unit 3D when it is predicted by other vehicle driver prediction unit 3C that there is a driver in other vehicle OV.

[0016] Figure 2 B is a diagram representing an example of the prediction result of the action of the other vehicle OV obtained by the other vehicle action prediction unit 3D when the other vehicle driver prediction unit 3C predicts that there is no driver in the other vehicle OV.

[0017] Figure 3 This is a flowchart illustrating an example of a process performed by the processor 153 of another vehicle movement prediction device 15 of the first embodiment.

[0018] Figure 4 A is a diagram representing an example of the prediction results of the actions of other vehicle OV obtained by other vehicle action prediction unit 3D when it is predicted by other vehicle driver prediction unit 3C that there is a driver in other vehicle OV.

[0019] Figure 4 B is a diagram representing an example of the prediction result of the action of the other vehicle OV obtained by the other vehicle action prediction unit 3D when the other vehicle driver prediction unit 3C predicts that there is no driver in the other vehicle OV. Detailed Implementation

[0020] Hereinafter, embodiments of other vehicle movement prediction devices, other vehicle movement prediction methods, and program products of this disclosure will be described with reference to the accompanying drawings.

[0021] <First Implementation> Figure 1 This is a diagram showing an example of the vehicle 1 that applies the other vehicle movement prediction device 15 of the first embodiment.

[0022] exist Figure 1 In the example shown, the vehicle 1 includes a surrounding condition sensor 11, a vehicle status sensor 12, an HMI (Human Machine Interface) 13, a vehicle control device 14, a steering actuator 14A, a brake actuator 14B, a drive actuator 14C, and other vehicle movement prediction devices 15.

[0023] The surrounding conditions sensor 11 monitors the surrounding conditions of vehicle 1 (e.g., other vehicles (OV) located around vehicle 1). Figure 2 A and Figure 2 The surrounding environment sensor 11 detects obstacles (such as B) located around the vehicle 1 and sends the detection results to the vehicle control unit 14 and other vehicle movement prediction devices 15. The surrounding environment sensor 11 includes, for example, a camera, LiDAR (Light Detection and Ranging), radar, and a wireless communication device that acquires information about the external conditions of the vehicle 1 from the outside of the vehicle 1 via wireless communication.

[0024] The vehicle status sensor 12 performs functions such as detecting the status of the vehicle 1 and measuring the position of the vehicle 1, and sends the detection results and position measurement results of the vehicle 1 to the vehicle control device 14 and other vehicle movement prediction devices 15. The vehicle status sensor 12 includes, for example, a vehicle speed sensor, an acceleration sensor, a yaw rate sensor, a gyroscope sensor, and a GPS (Global Positioning System) receiver.

[0025] HMI13 has functions such as accepting various operations of the driver of vehicle 1 and sending signals indicating the operation of the driver of vehicle 1 to vehicle control device 14.

[0026] The vehicle control unit 14 controls the steering actuator 14A, brake actuator 14B, and drive actuator 14C, for example, based on information (data, signals) sent from the surrounding condition sensor 11, the vehicle status sensor 12, and the HMI 13. Specifically, the vehicle control unit 14 has an autonomous driving function that allows the vehicle 1 to drive autonomously without requiring operation from the driver of the vehicle 1 to control the steering actuator 14A, brake actuator 14B, and drive actuator 14C. Specifically, the vehicle control unit 14 generates a driving plan for the vehicle 1 to reach its destination, for example, based on map information, the location information of the vehicle 1, and information indicating the destination of the vehicle 1. Furthermore, the vehicle control unit 14 drives the vehicle 1 autonomously according to this driving plan. Specifically, the vehicle control unit 14 uses the detection results of the surrounding condition sensor 11 and other vehicle movement prediction devices (OVs) obtained by the other vehicle movement prediction device 15 (described later) as referenced in the following section. Figure 2 A and Figure 2 Based on the predicted results of B), the driving plan is modified to avoid collisions between the vehicle 1 and other vehicles, such as OV, so that the vehicle 1 can drive autonomously.

[0027] Other vehicle movement prediction device 15 consists of a microcomputer with a communication interface (I / F) 151, a memory 152 and a processor 153.

[0028] The communication interface 151 has interface circuitry for connecting the other vehicle movement prediction device 15 to the surrounding condition sensor 11, the vehicle status sensor 12, the HMI 13, and the vehicle control device 14. The memory 152 stores programs and various data used in the processing executed by the processor 153. The processor 153 functions as an acquisition unit 3A, an object lane recognition unit 3B, an other vehicle driver prediction unit 3C, and an other vehicle movement prediction unit 3D.

[0029] The acquisition unit 3A acquires the detection results of the surrounding conditions of the vehicle 1 and the measurement results of the position of the vehicle 1.

[0030] The object lane recognition unit 3B performs OV (Operational Vehicle View) for other vehicles located around the vehicle 1 based on the detection results of the surrounding conditions of the vehicle 1 obtained by the acquisition unit 3A. Figure 2 A and Figure 2 The identification of objects such as B) and the identification of lanes located around the vehicle 1.

[0031] The other vehicle driver prediction unit 3C predicts whether there is a driver among the other vehicles (OVs) located around the vehicle 1 based on the detection results of the surrounding conditions of the vehicle 1 obtained by the acquisition unit 3A (sensor data of the surrounding conditions sensor 11).

[0032] exist Figure 1 In the example shown, the other vehicle driver prediction unit 3C predicts whether a driver is present in the other vehicle OV based on the behavior of the other vehicle OV detected by a camera, which is a surrounding condition sensor 11 (time-series sensor data of the surrounding condition sensor 11). Specifically, for example, if the surrounding condition sensor 11 detects human-specific driving behaviors of the other vehicle OV, such as sudden acceleration, sudden braking, or frequent lane changes, the other vehicle driver prediction unit 3C predicts that a driver is present in the other vehicle OV.

[0033] In another example, the other vehicle driver prediction unit 3C predicts other vehicle drivers based on images of other vehicle OVs captured by a camera acting as a surrounding condition sensor 11 (see reference). Figure 2 A and Figure 2 (B) Whether a driver exists. For example, if the other vehicle driver prediction unit 3C predicts that a driver exists in the other vehicle OV when the image captured by the camera, which is a peripheral situation sensor 11, contains the driver of the other vehicle OV in the rearview mirror or side mirror of the other vehicle OV (that is, the driver of the other vehicle OV is reflected in the rearview mirror or side mirror of the other vehicle OV).

[0034] In another example, the other vehicle driver prediction unit 3C predicts whether a driver is in another vehicle (OV) based on information obtained from outside the vehicle 1 via a wireless communication device that functions as a surrounding condition sensor 11, indicating whether a driver is in another vehicle (OV). Specifically, the wireless communication device that functions as the surrounding condition sensor 11 obtains information indicating whether a driver is in another vehicle (OV) via, for example, V2I (Vehicle-to-roadside-Infrastructure) or V2V (Vehicle-to-Vehicle) communication, and the other vehicle driver prediction unit 3C predicts whether a driver is in another vehicle (OV) based on this information.

[0035] exist Figure 1 In the example shown, the Other Vehicle Action Prediction Unit 3D is based on the prediction result of whether a driver exists in other vehicles (OV) obtained by the Other Vehicle Driver Prediction Unit 3C and the detection result DR of the time series of other vehicle OVs obtained by the surrounding situation sensor 11 (see reference). Figure 2 A and Figure 2 (B) predicts the actions of other vehicles (OV).

[0036] Figure 2 A and Figure 2 Figure B is an example of the detection result DR of the time series of other vehicle OVs obtained by the surrounding situation sensor 11 and the prediction result of the action of other vehicle OVs obtained by the other vehicle action prediction unit 3D. In detail, Figure 2 A represents an example of the prediction result of the action of the other vehicle OV obtained by the other vehicle action prediction unit 3D, in the case where the other vehicle driver prediction unit 3C predicts that there is a driver in the other vehicle OV. Figure 2 B represents an example of the prediction result of the action of the other vehicle OV obtained by the other vehicle action prediction unit 3D when the other vehicle driver prediction unit 3C predicts that there is no driver in the other vehicle OV.

[0037] exist Figure 2 In the example shown in A, vehicle 1 is traveling in lane L1, and another vehicle OV is traveling in lane L2. Specifically, the other vehicle OV passes from position P2 in lane L2. To ensure vehicle 1 can travel safely without colliding with other vehicles OV, the other vehicle movement prediction device 15... Figure 2 The actions of other vehicles (OV) after the time point shown in A are predicted.

[0038] Specifically, in Figure 2 In the example shown in A, the other vehicle driver prediction unit 3C predicts, based on the detection results of the surrounding situation sensor 11, that there is a driver in the other vehicle OV (that is, the other vehicle OV may be driving manually).

[0039] The Other Vehicle Action Prediction Unit 3D predicts the actions of other vehicle OVs based on the prediction results obtained by the Other Vehicle Driver Prediction Unit 3C, which states that there is a driver in the other vehicle OV, and the detection results DR (specifically, the trajectory of the other vehicle OV's position from the time point when the other vehicle OV passes through position P1 to the time point when the other vehicle OV passes through position P2) obtained by the Surrounding Condition Sensor 11.

[0040] Specifically, the Other Vehicle Action Prediction 3D predicts that there is a 30% probability that another vehicle (OV) will change lanes from lane L2 to lane L3, a 60% probability that another vehicle (OV) will not change lanes and will continue to travel in lane L2, and a 10% probability that another vehicle (OV) will change lanes from lane L2 to lane L1.

[0041] exist Figure 2In the example shown in B, vehicle 1 is traveling in lane L1, and another vehicle OV is traveling in lane L2. Specifically, the other vehicle OV passes from position P2 in lane L2. To ensure vehicle 1 can travel safely without colliding with other vehicles OV, the other vehicle movement prediction device 15... Figure 2 The actions of other vehicles (OV) after the time point shown in B are predicted.

[0042] Specifically, in Figure 2 In the example shown in B, the other vehicle driver prediction unit 3C predicts, based on the detection results of the surrounding situation sensor 11, that there is no driver in the other vehicle OV (that is, the other vehicle OV is driving through autonomous driving).

[0043] The Other Vehicle Action Prediction Unit 3D predicts the actions of other vehicle OVs based on the prediction results obtained by the Other Vehicle Driver Prediction Unit 3C, which states that there is no driver in the other vehicle OVs, and the detection results DR (specifically, the trajectory of the other vehicle OVs' positions from the time point when the other vehicle OV passes through position P1 to the time point when the other vehicle OV passes through position P2) obtained by the Surrounding Condition Sensor 11.

[0044] In detail, the Other Vehicle Action Prediction 3D predicts that there is a 10% probability that another vehicle (OV) will change lanes from lane L2 to lane L3, an 80% probability that another vehicle (OV) will not change lanes and will continue to travel in lane L2, and a 10% probability that another vehicle (OV) will change lanes from lane L2 to lane L1.

[0045] exist Figure 2 A and Figure 2 In the example shown in B, the actions of the other vehicle OV predicted by the other vehicle action prediction unit 3D when the other vehicle driver prediction unit 3C predicts that there is a driver in the other vehicle OV (the probability that the other vehicle OV will change lanes from lane L2 to lane L3 is 30%, the probability that the other vehicle OV will not change lanes and continue to drive in lane L2 is 60%, and the probability that the other vehicle OV will change lanes from lane L2 to lane L1 is 10%) are different from the actions of the other vehicle OV predicted by the other vehicle action prediction unit 3D when the other vehicle driver prediction unit 3C predicts that there is no driver in the other vehicle OV (the probability that the other vehicle OV will change lanes from lane L2 to lane L3 is 10%, the probability that the other vehicle OV will not change lanes and continue to drive in lane L2 is 80%, and the probability that the other vehicle OV will change lanes from lane L2 to lane L1 is 10%).

[0046] exist Figure 2 A and Figure 2 In the example shown in B, the detection result DR of the time series of other vehicle OVs obtained by the surrounding condition sensor 11 uses the trajectory of the positions of other vehicle OVs from the time point when other vehicle OVs pass through position P1 to the time point when other vehicle OVs pass through position P2. However, in another example, the detection result DR of the time series of other vehicle OVs obtained by the surrounding condition sensor 11 can also use the detection result of the orientation of other vehicle OVs from the time point when other vehicle OVs pass through position P1 to the time point when other vehicle OVs pass through position P2.

[0047] In another example, as the detection result DR of the time series of other vehicle OVs obtained by the surrounding conditions sensor 11, the detection results of the speed, acceleration, braking mode, etc. of other vehicle OVs from the time point when other vehicle OVs pass through position P1 to the time point when other vehicle OVs pass through position P2 can also be used.

[0048] exist Figure 1 In the example shown, the Other Vehicle Action Prediction Unit 3D uses a prediction model to predict the likelihood of other vehicles (OVs) changing lanes, based on the prediction results of whether a driver exists in other vehicles (OVs) obtained by the Other Vehicle Driver Prediction Unit 3C and the detection results of the time series of other vehicles (OVs) obtained by the Surrounding Condition Sensor 11. The prediction model is obtained by learning using training data, which is a dataset of detection results and labels of the time series of other learning vehicles (not shown) from a first time point to a second time point obtained by the learning Surrounding Condition Sensor (not shown) mounted on the learning vehicle (not shown). The labels are labels indicating whether a driver exists in the other learning vehicles and whether the other learning vehicles have changed lanes at a third time point after the second time point.

[0049] In another example, it could also be that other vehicle motion prediction units 3D use [methods] through [interaction / methods]. Figure 1 The example shown uses different methods to obtain prediction models, which predict the likelihood of other vehicles changing lanes based on the prediction results of whether there is a driver in other vehicles (OVs) obtained by the other vehicle driver prediction unit 3C and the detection results of the time series of other vehicles (OVs) obtained by the surrounding conditions sensor 11.

[0050] Figure 3 This is a flowchart illustrating an example of a process performed by the processor 153 of another vehicle movement prediction device 15 of the first embodiment.

[0051] exist Figure 3 In the example shown, in step S10, the acquisition unit 3A acquires the detection results of the surrounding conditions of the vehicle 1 and the measurement results of the position of the vehicle 1.

[0052] In step S11, the object lane recognition unit 3B performs the recognition of other vehicles (OVs) and other objects located around the vehicle 1 and the recognition of lanes located around the vehicle 1 based on the detection results of the surrounding conditions of the vehicle 1 obtained in step S10.

[0053] In step S12, the other vehicle driver prediction unit 3C predicts whether there is a driver among the other vehicles (OVs) located around the vehicle 1 based on the detection results of the surrounding conditions of the vehicle 1 obtained in step S10. If yes, proceed to step S13; otherwise, proceed to step S14.

[0054] In step S13, the Other Vehicle Action Prediction Unit 3D predicts the actions of other vehicles (OVs) based on the detection results DR of the time series of other vehicles (OVs) obtained by the surrounding situation sensor 11 and the unique driving behavior of the human (the driver in the other vehicle OV) (irregular driving behavior based on human emotions, attention, experience, etc.).

[0055] In step S14, the Other Vehicle Action Prediction Unit 3D predicts the actions of other vehicles (driving actions that are safe and in accordance with the rules of the programmed algorithm) based on the detection results DR of the time series of other vehicles (OV) obtained by the surrounding condition sensor 11 and the properties (algorithms) of AI (Artificial Intelligence) applied to other vehicles (OVs).

[0056] As described above, in the other vehicle action prediction device 15 of the first embodiment, unlike the prior art which predicts the actions of other vehicles (OVs) under the premise that they are being driven manually by their drivers, when the other vehicles (OVs) are autonomous vehicles, it can predict the actions of autonomous vehicles (OVs) that do not conform to predictions based on human reactions and driving tendencies with high accuracy. As a result, the safety and reliability of the autonomous driving of the vehicle 1 can be improved.

[0057] <Second Implementation> The vehicle 1 that applies the other vehicle movement prediction device 15 of the second embodiment is configured in the same way as the vehicle 1 that applies the other vehicle movement prediction device 15 of the first embodiment, except for the points described later.

[0058] Figure 4 A and Figure 4 Figure B is an example of the detection result DR of the time series of other vehicle OVs obtained by the surrounding situation sensor 11 and the prediction result of the action of other vehicle OVs obtained by the other vehicle action prediction unit 3D of the other vehicle action prediction device 15 of the second embodiment. In detail, Figure 4 A represents an example of the prediction result of the other vehicle's (OV) actions obtained by the other vehicle action prediction unit 3D when the other vehicle driver prediction unit 3C predicts the presence of a driver in the other vehicle (OV). Figure 4 B represents an example of the prediction result of the other vehicle's actions obtained by the other vehicle's actions prediction unit 3D in the case where there is no driver in the other vehicle's OV, as predicted by the other vehicle driver prediction unit 3C.

[0059] exist Figure 4 In the example shown in A, vehicle 1 is traveling in lane L2, and another vehicle OV is traveling in front of vehicle 1. Specifically, the other vehicle OV passes from position P2 in lane L2. To ensure vehicle 1 can travel safely without colliding with other vehicles OV, the other vehicle movement prediction device 15... Figure 4 The actions of other vehicles (OV) after the time point shown in A are predicted.

[0060] Specifically, in Figure 4 In the example shown in A, the other vehicle driver prediction unit 3C predicts, based on the detection results of the surrounding situation sensor 11, that there is a driver in the other vehicle OV (that is, the other vehicle OV may be driving manually).

[0061] The Other Vehicle Action Prediction Unit 3D predicts the actions of other vehicle OVs based on the prediction results obtained by the Other Vehicle Driver Prediction Unit 3C, which states that there is a driver in the other vehicle OV, and the detection results DR of the time series of other vehicle OVs obtained by the Surrounding Condition Sensor 11 (specifically, the speed, acceleration, braking mode, etc. of other vehicle OVs from the time point when other vehicle OVs pass through position P1 to the time point when other vehicle OVs pass through position P2).

[0062] Specifically, the Other Vehicle Action Prediction Department (3D) predicts a 20% probability that the Other Vehicle (OV) will accelerate, a 60% probability that the Other Vehicle (OV) will not accelerate or decelerate and will continue to maintain its speed, and a 20% probability that the Other Vehicle (OV) will decelerate.

[0063] exist Figure 4In the example shown in B, vehicle 1 is traveling in lane L2, and another vehicle OV is traveling in front of vehicle 1. Specifically, the other vehicle OV passes from position P2 in lane L2. To ensure vehicle 1 can travel safely without colliding with other vehicles OV, the other vehicle movement prediction device 15... Figure 4 The actions of other vehicles (OV) after the time point shown in B are predicted.

[0064] Specifically, in Figure 4 In the example shown in B, the other vehicle driver prediction unit 3C predicts, based on the detection results of the surrounding situation sensor 11, that there is no driver in the other vehicle OV (that is, the other vehicle OV is driving through autonomous driving).

[0065] The Other Vehicle Action Prediction Unit 3D predicts the actions of other vehicle OVs based on the prediction results obtained by the Other Vehicle Driver Prediction Unit 3C, which states that there is no driver in the other vehicle OVs, and the detection results DR of the time series of other vehicle OVs obtained by the Surrounding Condition Sensor 11 (specifically, the speed, acceleration, braking mode, etc. of other vehicle OVs from the time point when other vehicle OVs pass through position P1 to the time point when other vehicle OVs pass through position P2).

[0066] Specifically, the Other Vehicle Action Prediction 3D predicts that there is a 10% probability that the Other Vehicle (OV) will accelerate, an 80% probability that the Other Vehicle (OV) will not accelerate or decelerate and will continue to maintain its speed, and a 10% probability that the Other Vehicle (OV) will decelerate.

[0067] In one example of the vehicle 1 using the other vehicle action prediction device 15 of the second embodiment, the other vehicle action prediction unit 3D predicts the likelihood of other vehicle OVs accelerating or decelerating by using a prediction model based on the prediction results of whether a driver exists in other vehicle OVs obtained by the other vehicle driver prediction unit 3C and the detection results of the time series of other vehicle OVs obtained by the surrounding condition sensor 11. The prediction model is a model obtained by learning using training data, which is a dataset of detection results and labels of the time series of learning other vehicles (not shown) from a first time point to a second time point obtained by the learning surrounding condition sensor (not shown) mounted on the learning vehicle (not shown). The labels are labels indicating whether a driver exists in the learning other vehicles and whether the learning other vehicles accelerated or decelerated at a third time point after the second time point.

[0068] In another example, the other vehicle action prediction unit 3D may use a prediction model obtained by a different method than the example above to predict the likelihood of other vehicle OVs accelerating or decelerating, based on the prediction results of whether a driver exists in other vehicle OVs obtained by the other vehicle driver prediction unit 3C and the detection results of the time series of other vehicle OVs obtained by the surrounding condition sensor 11.

[0069] <Third Implementation Method> The vehicle 1 that applies the other vehicle movement prediction device 15 of the third embodiment is configured in the same way as the vehicle 1 that applies the other vehicle movement prediction device 15 of the first or second embodiment, except for the points described later.

[0070] As described above, in the vehicle 1 using the other vehicle action prediction device 15 of the first embodiment, the vehicle control device 14 has an autonomous driving function that allows the vehicle 1 to drive autonomously without the driver needing to operate the steering actuator 14A, brake actuator 14B, and drive actuator 14C. Specifically, the vehicle control device 14 generates a driving plan for the vehicle 1 to reach its destination, for example, based on map information, the location information of the vehicle 1, and information indicating the destination of the vehicle 1. Furthermore, the vehicle control device 14 drives the vehicle 1 autonomously according to this driving plan. More specifically, the vehicle control device 14 uses the detection results of the surrounding condition sensor 11 and other vehicle OV (OV) obtained by the other vehicle action prediction device 15 (see reference 15) to determine the location of the vehicle 1. Figure 2 A and Figure 2 Based on the predicted results of B), the driving plan is modified to avoid collisions between the vehicle 1 and other vehicles, such as OV, so that the vehicle 1 can drive autonomously.

[0071] On the other hand, in the vehicle 1 that applies the other vehicle movement prediction device 15 of the third embodiment, the vehicle control device 14 has a driving assistance function. Specifically, the vehicle control device 14, based on the detection results of the surrounding situation sensor 11 and the other vehicle movement prediction device 15, obtains the other vehicle movement (OV) (refer to...) Figure 2 A and Figure 2 If the predicted result of the action of B) indicates that it is necessary to avoid collisions between this vehicle 1 and other vehicles OV, an alarm indicating this purpose will be output to HMI13.

[0072] As described above, embodiments of other vehicle motion prediction devices, other vehicle motion prediction methods, and program products of this disclosure have been described with reference to the accompanying drawings. However, other vehicle motion prediction devices, other vehicle motion prediction methods, and program products of this disclosure are not limited to the above embodiments, and appropriate modifications can be made without departing from the spirit of this disclosure. The configurations of the various examples of the above embodiments can also be appropriately combined. In the examples of the above embodiments, the processing performed in the other vehicle motion prediction device 15 has been described as software processing performed by executing a program, but the processing performed in the other vehicle motion prediction device 15 can also be hardware processing. Alternatively, the processing performed in the other vehicle motion prediction device 15 can be a combination of software and hardware processing. Furthermore, the program stored in the memory 152 of the other vehicle motion prediction device 15 (the program that implements the functions of the processor 153 of the other vehicle motion prediction device 15) can be provided, circulated, etc., for example, recorded on a computer-readable storage medium (program product) such as a semiconductor memory, magnetic recording medium, or optical recording medium.

Claims

1. An additional vehicle movement prediction device, comprising a processor, wherein, The processor is configured to: Based on the detection results of the surrounding condition sensors that detect the surrounding conditions of the vehicle, it is predicted whether there is a driver among other vehicles located around the vehicle. as well as The actions of the other vehicles are predicted based on the prediction of whether the driver is present in the other vehicles and the detection results of the time series of the other vehicles obtained by the surrounding condition sensors.

2. The other vehicle movement prediction device according to claim 1, wherein, The predicted actions of the other vehicles when the driver is predicted to be present in the other vehicles are different from the predicted actions when the driver is predicted to be absent in the other vehicles.

3. The other vehicle movement prediction device according to claim 1, wherein, The processor uses a prediction model to predict the likelihood of other vehicles changing lanes, based on a prediction of whether the driver is present in the other vehicles and the detection results of the other vehicles' time series obtained by the surrounding condition sensor. The prediction model is obtained by learning using training data, which is a dataset of detection results and labels of other learning vehicles' time series from a first time point to a second time point obtained by the surrounding condition sensor mounted on the learning vehicle. The labels are labels indicating whether a driver is present in the other learning vehicles and whether the other learning vehicles changed lanes at a third time point after the second time point.

4. The other vehicle movement prediction device according to claim 1, wherein, The processor predicts whether the driver is present in any of the other vehicles based on the behavior of the other vehicles detected by the surrounding environment sensors.

5. The other vehicle movement prediction device according to claim 1, wherein, The processor predicts whether the driver is present in any of the other vehicles based on images of those vehicles captured by a camera that acts as a surrounding condition sensor.

6. The other vehicle movement prediction device according to claim 1, wherein, The processor predicts whether the driver is in any of the other vehicles based on information obtained via wireless communication with the outside of the vehicle through a wireless communication device that acts as a surrounding condition sensor, indicating whether the driver is in any of the other vehicles.

7. Another method for predicting vehicle movement, comprising: Based on the detection results of the surrounding condition sensors that detect the surrounding conditions of the vehicle, it is predicted whether there is a driver among other vehicles located around the vehicle. as well as The actions of the other vehicles are predicted based on the prediction of whether the driver is present in the other vehicles and the detection results of the time series of the other vehicles obtained by the surrounding condition sensors.

8. A program product having a computer program recorded thereon, the computer program being configured to cause a processor to perform the following actions: Based on the detection results of the surrounding environment sensors that detect the surrounding environment of the vehicle, the system predicts whether there is a driver among other vehicles located in the vicinity of the vehicle; and The actions of the other vehicles are predicted based on the prediction of whether the driver is present in the other vehicles and the detection results of the time series of the other vehicles obtained by the surrounding condition sensors.

Images