Vehicle behavior prediction device and vehicle behavior prediction method

Abstract

To provide a vehicle action prediction device and a vehicle action prediction method that can predict a lane change position of an adjacent vehicle into a target lane, taking into consideration an overlap degree between a target vehicle and the adjacent vehicle.SOLUTION: A vehicle action prediction device calculates an overlap degree between a positional range of a predicted target vehicle and a positional range of a target vehicle in a forward / backward direction based on position information and shape information of the predicted target vehicle set from an adjacent vehicle, and shape information of the target vehicle; and the vehicle action prediction device predicts a lane change position of the predicted target vehicle into the target lane using a prediction model to which the position information and speed information of the predicted target vehicle and the overlap degree are input.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present application relates to a vehicle behavior prediction device and a vehicle behavior prediction method. [Background technology]

[0002] A technology has been proposed for estimating the position of a lane change destination when an adjacent vehicle traveling in an adjacent lane adjacent to the lane in which the host vehicle is traveling changes lanes into the host vehicle's lane.

[0003] For example, the technology in Patent Document 1 assumes a first vehicle traveling ahead of the vehicle, and a second vehicle and a third vehicle traveling in lanes adjacent to the vehicle's own lane, and estimates the possibility that the third vehicle will change lanes into the vehicle's own lane, taking into account the time to collision between the third vehicle and each of the other vehicles.

[0004] The technology in Non-Patent Document 1 analyzes risk factors during merging using canonical discrimination. It suggests that the potential risk of collision during merging increases when a vehicle merging into the own lane is large.

[0005] The technology of Non-Patent Document 2 predicts the cut-in position of a merging vehicle using quadratic discriminant analysis or random forest based on the headway distance and relative speed between the merging vehicle and a vehicle traveling on the main line. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Patent No. 6494121 [Non-patent literature]

[0007] [Non-Patent Document 1] Hiroshi Suzuki and Yuki Matsumura, "Study on Collision Risk Assessment in Complex Events at Urban Expressway Merging Sections," Journal of the Japan Society of Civil Engineers, D3 (Civil Engineering Planning), Vol. 71, No. 5, 2015 [Non-patent document 2] Koji Tanida, Masahiro Kimura, and Yuichi Yoshida, "A Model of Merging Behavior on Expressways for Autonomous Vehicle Control," Transactions of the Society of Automotive Engineers of Japan 48.4 (2017): 885-890 Summary of the Invention [Problem to be solved by the invention]

[0008] However, the inventors have conducted research and found that even if the positional relationship between the representative position of the host vehicle and the representative position of the adjacent vehicle is the same, the lane change destination of the adjacent vehicle varies depending on the degree of overlap between the position range of the host vehicle and the position range of the adjacent vehicle in the longitudinal direction, and that using the degree of overlap in prediction is important. This degree of overlap varies depending on the representative position and length of each vehicle. Predicting the lane change destination of the adjacent vehicle using the representative position and length of each vehicle without using the degree of overlap increases the number of input parameters required for prediction, complicates the prediction model, and increases the processing load using the prediction model. Furthermore, indirect parameters may not adequately represent the degree of overlap, resulting in reduced prediction accuracy. The above-mentioned documents do not disclose estimating the lane change destination using the degree of overlap.

[0009] Therefore, the present application aims to provide a vehicle behavior prediction device and a vehicle behavior prediction method that can predict the location where an adjacent vehicle will change lanes into a target lane, taking into account the degree of overlap between the target vehicle and adjacent vehicles. [Means for solving the problem]

[0010] The vehicle behavior prediction device according to the present application comprises: an information acquisition unit that acquires a surrounding state of a target vehicle and a vehicle state of the target vehicle, and acquires position information, speed information, and shape information of an adjacent vehicle traveling in an adjacent lane adjacent to a target lane in which the target vehicle is traveling, based on the surrounding state and the vehicle state of the target vehicle; a prediction target setting unit that sets a prediction target vehicle from the adjacent vehicles; a feature amount calculation unit that calculates an overlapping degree between a position range of the predicted target vehicle and a position range of the target vehicle in a front-rear direction based on the position information and the shape information of the predicted target vehicle and the shape information included in the vehicle state of the target vehicle; and a lane change prediction unit that predicts the position of the predicted target vehicle at which it will change lanes into the target lane using a prediction model to which at least one of the position information and the speed information of the predicted target vehicle and the degree of overlap between the target vehicle and the predicted target vehicle are input.

[0011] The vehicle behavior prediction method according to the present application includes: an information acquisition step of acquiring a surrounding state of a target vehicle and a vehicle state of the target vehicle, and acquiring position information, speed information, and shape information of an adjacent vehicle traveling in an adjacent lane adjacent to a target lane in which the target vehicle is traveling based on the surrounding state and the vehicle state of the target vehicle; a prediction target setting step of setting a prediction target vehicle from the adjacent vehicles; a feature amount calculation step of calculating an overlap degree between a position range of the predicted target vehicle and a position range of the target vehicle in a front-rear direction based on the position information and the shape information of the predicted target vehicle and the shape information included in the vehicle state of the target vehicle; and a lane change prediction step for predicting the position of the predicted target vehicle to change lanes into the target lane using a prediction model to which at least one of the position information and the speed information of the predicted target vehicle and the degree of overlap between the target vehicle and the predicted target vehicle are input. [Effects of the Invention]

[0012] Even if the relative position and relative speed of the prediction target vehicle with respect to the target vehicle are the same, the lane change destination position of the prediction target vehicle changes depending on the degree of overlap between the target vehicle and the prediction target vehicle. Therefore, the characteristics of the change in the lane change destination position can be captured based on the degree of overlap between the target vehicle and the prediction target vehicle. The lane change destination position of the prediction target vehicle is predicted using a prediction model that inputs the degree of overlap between the target vehicle and the prediction target vehicle in addition to at least one of the position information and speed information of the prediction target vehicle, thereby improving prediction accuracy. Since the degree of overlap is input directly, the number of input parameters required for prediction can be reduced compared to when multiple other parameters related to the degree of overlap are input, the prediction model can be simplified, and the processing load using the prediction model can be reduced. Furthermore, prediction accuracy is improved compared to when an indirect parameter related to the degree of overlap is input. [Brief explanation of the drawings]

[0013] [Figure 1] 1 is a schematic block diagram of a vehicle behavior prediction device and a driving assistance device according to a first embodiment. [Figure 2] 1 is a schematic hardware configuration diagram of a vehicle behavior prediction device according to a first embodiment. [Figure 3] 1 is a schematic hardware configuration diagram of a vehicle behavior prediction device according to a first embodiment. [Figure 4] FIG. 2 is a diagram for explaining a host vehicle coordinate system according to the first embodiment. [Figure 5] FIG. 4 is a diagram for explaining calculation of the degree of overlap between a prediction target vehicle and a host vehicle according to the first embodiment. [Figure 6] FIG. 4 is a diagram for explaining a change in a lane change destination due to a change in the degree of overlap according to the first embodiment. [Figure 7] FIG. 4 is a diagram for explaining a change in a lane change destination due to a change in the degree of overlap according to the first embodiment. [Figure 8] FIG. 4 is a diagram for explaining a change in a lane change destination due to a change in the degree of overlap according to the first embodiment. [Figure 9]FIG. 4 is a diagram for explaining a change in a lane change destination due to a change in the degree of overlap according to the first embodiment. [Figure 10] FIG. 10 is a diagram for explaining the setting of the degree of overlap in the case of complete inclusion according to the first embodiment. [Figure 11] FIG. 10 is a diagram for explaining the setting of the degree of overlap in the case of complete inclusion according to the first embodiment. [Figure 12] FIG. 4 is a diagram for explaining calculation of the degree of overlap taking into consideration the safe distance according to the first embodiment. [Figure 13] 5 is a diagram for explaining calculation of the degree of overlap between a preceding vehicle and a following vehicle of the host vehicle and a prediction target vehicle according to the first embodiment. FIG. [Figure 14] FIG. 10 is a diagram for explaining calculation of the degree of overlap using a probability distribution according to the first embodiment. [Figure 15] 5 is a diagram for explaining calculation of a feature amount related to an intimidating feeling according to the first embodiment. FIG. [Figure 16] FIG. 2 is a diagram for explaining a rule-based model as a prediction model according to the first embodiment. [Figure 17] 10 is a diagram for explaining calculation of the degree of likelihood of collision between the target vehicle and a preceding vehicle and a following vehicle of the target vehicle according to the second embodiment. FIG. [Figure 18] FIG. 10 is a diagram for explaining calculation of the remaining time until the start of forced merging according to the second embodiment. [Figure 19] 4 is a flowchart for explaining the processing of the vehicle behavior prediction device according to the first embodiment. [Figure 20] 1 is a schematic block diagram of a vehicle behavior prediction device according to a first embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0014] 1. First Embodiment A vehicle behavior prediction device 1 according to a first embodiment will be described with reference to the drawings. In this embodiment, the target vehicle is set to a host vehicle, which is a vehicle equipped with the vehicle behavior prediction device 1, and the target lane in which the target vehicle is traveling is referred to as the host lane. The vehicle behavior prediction device 1 is incorporated into a driving assistance device 50. The driving assistance device 50 is provided in the host vehicle.

[0015] As shown in FIG. 1, the vehicle is equipped with a surroundings monitoring device 31, a position detection device 32, a vehicle state detection device 33, a map information database 34, a wireless communication device 35, a driving assistance device 50, a drive control device 36, a power plant 8, an electric steering device 7, an electric braking device 9, and a human interface device 37.

[0016] The periphery monitoring device 31 is a device such as a camera or radar that monitors the periphery of the vehicle. The radar may be a millimeter wave radar, a laser radar, an ultrasonic radar, etc. The wireless communication device 35 performs wireless communication with a base station using a cellular wireless communication standard such as 4G or 5G.

[0017] The position detection device 32 is a device that detects the current position (latitude, longitude, altitude) of the vehicle, and uses a GPS antenna or the like that receives signals output from artificial satellites such as the Global Navigation Satellite System (GNSS). Note that various methods may be used to detect the current position of the vehicle, such as a method using the lane number of the vehicle, a map matching method, a dead reckoning method, or a method using detected information around the vehicle.

[0018] The map information database 34 stores road information such as road shapes (for example, the number of lanes, the position of each lane, the shape of each lane, the type of each lane, the road type, the speed limit, etc.), signs, traffic lights, etc. The map information database 34 is mainly composed of a storage device. The map information database 34 may be provided in a server outside the vehicle connected to a network, and the driving assistance device 50 may obtain necessary road information from the server outside the vehicle via the wireless communication device 35.

[0019] The drive control device 36 includes a power control device, a brake control device, an automatic steering control device, a light control device, etc. The power control device controls the output of a power machine 8 such as an internal combustion engine or a motor. The brake control device controls the braking operation of an electric brake device 9. The automatic steering control device controls the electric steering device 7. The light control device controls turn signals, hazard lights, etc.

[0020] The vehicle state detection device 33 is a detection device that detects the state of the host vehicle, which is the driving state and running state of the host vehicle. In this embodiment, the vehicle state detection device 33 detects the speed, acceleration, yaw rate, steering angle, lateral acceleration, etc. of the host vehicle as the running state of the host vehicle. For example, the vehicle state detection device 33 may be provided with a speed sensor, an acceleration sensor, an angular velocity sensor, a steering angle sensor, etc. that detect the rotational speed of the wheels.

[0021] The driving state of the vehicle is detected by detecting acceleration / deceleration operations, steering angle operations, and lane change operations by the driver. For example, the vehicle state detection device 33 is provided with an accelerator position sensor, a brake position sensor, a steering angle sensor (handle angle sensor), a steering torque sensor, a turn signal position switch, and the like.

[0022] The human interface device 37 is a device that receives input from the driver through a speaker, a display screen, an input device, etc., and transmits information to the driver.

[0023] 1-1. Driving assistance device 50 (vehicle behavior prediction device 1) The driving assistance device 50 (vehicle behavior prediction device 1) includes functional units such as an information acquisition unit 51, a prediction target setting unit 52, a feature calculation unit 53, a lane change prediction unit 54, and a driving assistance unit 55. Each function of the driving assistance device 50 is realized by a processing circuit included in the driving assistance device 50. Specifically, as shown in FIG. 2 , the driving assistance device 50 includes an arithmetic processing device 90 such as a CPU (Central Processing Unit), a storage device 91, an input / output device 92 that inputs and outputs external signals to and from the arithmetic processing device 90, and the like.

[0024] The arithmetic processing device 90 may be an ASIC (Application Specific Integrated Circuit), an IC (Integrated Circuit), a DSP (Digital Signal Processor), an FPGA (Field Programmable Gate Array), a GPU (Graphics Processing Unit), an AI (Artificial Intelligence) chip, various logic circuits, various signal processing circuits, etc. Furthermore, the arithmetic processing device 90 may be a plurality of the same or different types, and each process may be shared and executed. The storage device 91 may be a variety of storage devices, such as a RAM (Random Access Memory), a ROM (Read Only Memory), a flash memory, an EEPROM (Electrically Erasable Programmable Read Only Memory), a hard disk, etc.

[0025] The input / output device 92 is equipped with a communication device, an A / D converter, an input / output port, a drive circuit, etc. The input / output device 92 is connected to the surroundings monitoring device 31, the position detection device 32, the vehicle state detection device 33, the map information database 34, the wireless communication device 35, the drive control device 36, the human interface device 37, etc., and communicates with these devices.

[0026] The functions of the functional units 51 to 55 of the driving support device 50 are realized by the arithmetic processing device 90 executing software (programs) stored in the storage device 91 and cooperating with other hardware of the driving support device 50, such as the storage device 91 and the input / output device 92. Setting data such as the speed increase amount used by the functional units 51 to 55 is stored in the storage device 91, such as an EEPROM.

[0027] Alternatively, the driving assistance device 50 may be provided with dedicated hardware 93 as a processing circuit, such as a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC, an FPGA, a GPU, an AI chip, or a circuit that combines these, as shown in Fig. 3. Each function of the driving assistance device 50 will be described in detail below.

[0028] 1-1-1. Information acquisition section 51 <Acquisition of surrounding conditions and driving conditions of the vehicle> The information acquisition unit 51 acquires the surrounding conditions of the host vehicle and the vehicle conditions of the host vehicle. The vehicle conditions of the host vehicle include the running conditions and shape information of the host vehicle. In this embodiment, the information acquisition unit 51 acquires, as the running conditions of the host vehicle, the position, moving direction, speed, acceleration, and whether or not a lane change has occurred of the host vehicle, based on the position information of the host vehicle acquired from the position detection device 32 and the vehicle conditions acquired from the vehicle condition detection device 33. The information acquisition unit 51 also acquires preset vehicle information of the host vehicle (vehicle type, shape information, etc.) from a storage device or the like.

[0029] As the surrounding conditions of the host vehicle, the information acquisition unit 51 acquires the driving conditions and vehicle information of surrounding vehicles present around the host vehicle. In this embodiment, the information acquisition unit 51 acquires the positions, moving directions, speeds, accelerations, and turn signal operation states of the surrounding vehicles based on the detection information acquired from the periphery monitoring device 31 and the position information of the host vehicle acquired from the position detection device 32. In addition to the surrounding vehicles, the information acquisition unit 51 also acquires information on obstacles, pedestrians, signs, traffic regulations such as lane restrictions, and the like. Based on the detection information acquired from the periphery monitoring device 31, the information acquisition unit 51 acquires the vehicle types (e.g., light vehicle, compact vehicle, medium-sized vehicle, medium-sized truck, large truck, medium-sized trailer, large trailer, ambulance, police car, motorcycle, various special-purpose vehicle, etc.) and shape information (e.g., vehicle length, vehicle width, vehicle height, etc.) of the surrounding vehicles as vehicle information of the surrounding vehicles. The information acquisition unit 51 also acquires information on the reliability of each piece of acquired information on the surrounding vehicles.

[0030] The information acquisition unit 51 can acquire the driving conditions of surrounding vehicles, lane information of surrounding vehicles, and vehicle information of surrounding vehicles (vehicle type, shape information, etc.) via communication from outside the vehicle. For example, the information acquisition unit 51 may acquire the driving conditions of surrounding vehicles (position, moving direction, speed, operation state of turn signal, target driving trajectory, etc. of surrounding vehicles) via wireless communication or the like from surrounding vehicles. Furthermore, the information acquisition unit 51 may acquire the driving conditions of surrounding vehicles present in a monitoring area (position, moving direction, speed, acceleration, operation state of turn signal, etc. of surrounding vehicles), vehicle information (vehicle type, shape information, etc.), information on obstacles and pedestrians, road shape, traffic regulations, traffic conditions, etc., via wireless communication or the like from roadside devices such as cameras that monitor road conditions, etc.

[0031] In this embodiment, the information acquisition unit 51 acquires the relative positions and relative speeds of surrounding vehicles and the like with respect to the host vehicle in a coordinate system of the host vehicle based on the current position of the host vehicle. As shown in FIG. 4, the coordinate system of the host vehicle is a coordinate system having axes in the longitudinal direction X of the current host vehicle and in the lateral direction Y of the host vehicle. Note that the information acquisition unit 51 may also acquire the relative positions and relative speeds of surrounding vehicles in a coordinate system in the longitudinal and lateral directions of the host vehicle's lane in which the host vehicle is traveling. The information acquisition unit 51 may also acquire the absolute position (latitude, longitude), absolute movement direction (orientation), absolute speed, absolute acceleration, etc. of each vehicle.

[0032] As the surrounding conditions of the host vehicle, the information acquisition unit 51 acquires road information about the surroundings of the host vehicle from the map information database 34 based on the location information of the host vehicle acquired from the position detection device 32. The acquired road information includes information such as the number of lanes, the position of each lane, the shape of each lane, the type of each lane, the road type, and the speed limit. The shape of each lane includes the position of the lane, the curvature of the lane, the longitudinal gradient of the lane, the cross gradient of the lane, and the width of the lane. The shape of the lane is set at each point along the longitudinal direction of the lane. The type of each lane includes a merging lane, a main lane into which the merging lane merges, and the like. The shape of the lane also includes a merging start position where the merging lane can start merging into the main lane, and the end position of the merging lane. The information acquisition unit 51 also acquires information about traffic regulations, such as lane restrictions due to construction, from an external server or the like.

[0033] In addition, the information acquisition unit 51 detects the shape and type of road dividing lines, etc. based on the detection information of white lines, road shoulders, etc. acquired from the perimeter monitoring device 31, and determines the shape and position of each lane, the number of lanes, the type of each lane, etc. based on the detected shape and type of road dividing lines, etc.

[0034] The information acquisition unit 51 acquires lane information corresponding to the lane in which the vehicle is traveling, based on the position of the vehicle. The information acquisition unit 51 also acquires lane information corresponding to the lane in which each of the surrounding vehicles is traveling, based on the positions of each of the surrounding vehicles. The acquired lane information includes the shape, position, and type of the lane, as well as lane information for surrounding lanes.

[0035] <Acquisition of information on adjacent vehicles> Based on the acquired surrounding conditions of the host vehicle and the traveling conditions of the host vehicle, the information acquisition unit 51 acquires position information, speed information, and shape information (vehicle length, vehicle width, vehicle height, etc.) of adjacent vehicles traveling in adjacent lanes adjacent to the host vehicle's traveling lane. If there are multiple adjacent vehicles, information on each of the adjacent vehicles is acquired.

[0036] The information acquisition unit 51 determines the lane in which the vehicle is traveling based on the position of the vehicle and road information (such as the position and shape of each lane), and determines adjacent lanes to the vehicle based on the determined information on the vehicle's lane and the road information. The adjacent lanes include one or both of the lanes to the left and right of the vehicle traveling in the same direction as the vehicle. The information acquisition unit 51 determines adjacent vehicles that are nearby vehicles traveling in the adjacent lanes based on the determined information on the adjacent lanes and the traveling states of the nearby vehicles.

[0037] The position information of the adjacent vehicle includes the relative position of the adjacent vehicle with respect to the host vehicle. Furthermore, the speed information of the adjacent vehicle includes the relative speed of the adjacent vehicle with respect to the host vehicle. As described above, the relative position and relative speed of the adjacent vehicle may be acquired in the coordinate system of the host vehicle (forward / backward and lateral directions of the host vehicle) or in the coordinate system of the host lane (forward / backward and lateral directions of the host lane).

[0038] The relative positions may be set to the center positions of the respective vehicles, or the positions of the respective vehicles may be set to the front end positions or rear end positions of the respective vehicles. The position of the own vehicle may be set to the front end position and the position of the adjacent vehicle may be set to the rear end position. Conversely, the position of the own vehicle may be set to the rear end position and the position of the adjacent vehicle may be set to the front end position. The inter-vehicle distance in the longitudinal direction may be acquired as the relative position.

[0039] The information acquisition unit 51 acquires position information (relative position, etc.), speed information (relative speed, etc.), and shape information (vehicle length, vehicle width, vehicle height, etc.) of each of the preceding and following vehicles traveling in the same lane in front of and behind the host vehicle, based on the surrounding conditions of the host vehicle and the traveling conditions of the host vehicle. Note that if there is no vehicle preceding the host vehicle, information about the preceding vehicle is not acquired, and if there is no vehicle following the host vehicle, information about the following vehicle is not acquired.

[0040] The information acquisition unit 51 acquires position information (relative position, etc.), speed information (relative speed, etc.), and shape information (vehicle length, vehicle width, vehicle height, etc.) of each of the preceding vehicle and the following vehicle traveling in the adjacent lanes in front of and behind the prediction target vehicle, based on the surrounding conditions of the subject vehicle and the vehicle condition (for example, the traveling condition) of the subject vehicle. Note that if there is no vehicle preceding the prediction target vehicle, information of the preceding vehicle is not acquired, and if there is no vehicle following the prediction target vehicle, information of the following vehicle is not acquired.

[0041] The information acquisition unit 51 calculates the relative positions and relative speeds of the preceding vehicle and the following vehicle relative to the prediction target vehicle based on the position information and speed information of the prediction target vehicle and the position information and speed information of the preceding vehicle and the following vehicle of the prediction target vehicle.

[0042] <Setting the length of the vehicle to be predicted> The information acquisition unit 51 acquires the vehicle length Lm of the prediction target vehicle and the vehicle type of the prediction target vehicle from the periphery monitoring device 31 or the like, and if the difference between the acquired vehicle length Lm of the prediction target vehicle and the vehicle length Lmest estimated from the vehicle type of the prediction target vehicle is equal to or greater than a judgment value, sets the vehicle length Lmest estimated from the vehicle type of the prediction target vehicle as the official vehicle length Lm of the prediction target vehicle to be used in calculating the degree of overlap.On the other hand, if the difference is less than the judgment value, the information acquisition unit 51 sets the vehicle length Lm of the prediction target vehicle acquired from the periphery monitoring device 31 or the like as the official vehicle length Lm of the prediction target vehicle to be used in calculating the degree of overlap.

[0043] There may be a large error in the recognition of the vehicle length Lm of the prediction target vehicle obtained from the periphery monitoring device 31, etc. According to the above configuration, by comparing the acquired vehicle length Lm with the vehicle length Lmest estimated from the vehicle type, if the recognition error of the acquired vehicle length Lm is large, the error in the vehicle length Lm of the prediction target vehicle used to calculate the degree of overlap can be reduced by using the vehicle length Lmest estimated from the vehicle type.

[0044] 1-1-2. Prediction target setting unit 52 The prediction target setting unit 52 sets a prediction target vehicle from adjacent vehicles. The prediction target setting unit 52 determines whether or not there is a possibility that the adjacent vehicle will change lanes from an adjacent lane to the own lane, and sets the adjacent vehicle determined to have a possibility of changing lanes as the prediction target vehicle. When multiple prediction target vehicles are set, prediction processing is performed for each prediction target vehicle.

[0045] The prediction target setting unit 52 determines whether or not there is a possibility that the adjacent vehicle will change lanes into the own lane, based on the traveling state of the adjacent vehicle, the surrounding state of the adjacent vehicle, etc. For example, the prediction target setting unit 52 determines that there is a possibility that the adjacent vehicle will change lanes if the adjacent lane in which the adjacent vehicle is traveling is a merging lane that merges into the own lane. The prediction target setting unit 52 determines that there is a possibility that the adjacent vehicle will change lanes if the adjacent vehicle is operating a turn signal toward the own lane. The prediction target setting unit 52 determines that there is a possibility that the adjacent vehicle will change lanes if information indicating that the adjacent vehicle will change lanes into the own lane is transmitted from the adjacent vehicle via wireless communication. The prediction target setting unit 52 determines whether or not there is a possibility that the adjacent vehicle will change lanes, based on the lateral relative speed and relative position of the adjacent vehicle.

[0046] 1-1-3. Feature calculation unit 53 The feature amount calculation unit 53 calculates the degree of overlap IoU between the position range of the prediction target vehicle and the position range of the host vehicle in the forward / backward direction X, based on the position information and shape information of the prediction target vehicle and the shape information of the host vehicle. The forward / backward direction X may be the forward / backward direction of the host vehicle or the forward / backward direction of the host lane.

[0047] 5, the feature calculation unit 53 calculates an overlap length Lov of the position range in the longitudinal direction X where the position range of the prediction target vehicle overlaps with the position range of the host vehicle, based on the relative position of the prediction target vehicle in the longitudinal direction X with respect to the host vehicle, the vehicle length Lm of the prediction target vehicle, and the vehicle length Lego of the host vehicle. Note that if there is no overlap, the overlap length Lov is 0. The position range of the prediction target vehicle is set based on the position information and shape information of the prediction target vehicle acquired by the information acquisition unit 51. For example, the position range of the prediction target vehicle may be set to the position range in the longitudinal direction X where the prediction target vehicle is recognized, or may be set based on the representative position and vehicle length of the prediction target vehicle.

[0048] In this embodiment, the feature amount calculation unit 53 calculates the degree of overlap IoU based on the overlap length Lov and the length of the union of the position range of the prediction target vehicle and the position range of the host vehicle in the forward / backward direction X. For example, as shown in the following equation, the feature amount calculation unit 53 calculates the degree of overlap IoU by dividing the overlap length Lov by the length of the union. This degree of overlap IoU is IoU (Intersection over Union), and is obtained by dividing the common part of the two regions by the union. Note that if there is no overlap, the degree of overlap IoU is 0. IoU=Lov / (Lm+Lego-Lov) ···(1)

[0049] By using this overlap degree IoU, as shown in Figures 6 and 7, even if the overlap length Lov is the same, the overlap degree IoU decreases as the vehicle length Lm of the prediction target vehicle increases relative to the vehicle length Lego of the host vehicle. As shown in Figure 6, when the prediction target vehicle is located behind the host vehicle, the likelihood that the prediction target vehicle will change lanes behind the host vehicle increases as the vehicle length Lm of the prediction target vehicle increases relative to the vehicle length Lego of the host vehicle. On the other hand, as shown in Figure 7, when the prediction target vehicle is located ahead of the host vehicle, the likelihood that the prediction target vehicle will change lanes ahead of the host vehicle increases as the vehicle length Lm of the prediction target vehicle increases relative to the vehicle length Lego of the host vehicle. Therefore, even if the overlap length Lov is the same, the overlap degree IoU can capture the characteristics of changes in the position of the lane change destination according to changes in the vehicle length Lm of the prediction target vehicle relative to the vehicle length Lego of the host vehicle.

[0050] Furthermore, as shown in FIG. 8, when the prediction target vehicle is located behind the host vehicle, even if the distance between the front end of the prediction target vehicle and the front end of the host vehicle in the longitudinal direction X is the same, the degree of overlap IoU increases as the vehicle length Lm of the prediction target vehicle becomes shorter relative to the vehicle length Lego of the host vehicle. As the vehicle length Lm of the prediction target vehicle becomes shorter relative to the vehicle length Lego of the host vehicle, the possibility that the prediction target vehicle will change lanes behind the host vehicle decreases. On the other hand, as shown in FIG. 9, when the prediction target vehicle is located ahead of the host vehicle, even if the distance between the front end of the prediction target vehicle and the front end of the host vehicle in the longitudinal direction X is the same, the degree of overlap IoU increases as the vehicle length Lm of the prediction target vehicle becomes longer relative to the vehicle length Lego of the host vehicle. As the vehicle length Lm of the prediction target vehicle becomes longer relative to the vehicle length Lego of the host vehicle, the possibility that the prediction target vehicle will change lanes ahead of the host vehicle decreases. Therefore, even if the distance between the front end of the vehicle to be predicted and the front end of the own vehicle is the same, the overlapping degree IoU can capture the characteristics of the change in the position of the lane change destination according to the change in the vehicle length Lm of the vehicle to be predicted relative to the vehicle length Lego of the own vehicle.

[0051] Alternatively, the feature amount calculation unit 53 may calculate the degree of overlap IoU based on the overlap length Lov and the length Lm of the prediction target vehicle. For example, the feature amount calculation unit 53 may calculate the degree of overlap IoU by dividing the overlap length Lov by the length Lm of the prediction target vehicle. Even in this calculation, the length Lego of the host vehicle does not change, so it is possible to capture the feature of a change in the position of the lane change destination in accordance with a change in the length Lm of the prediction target vehicle relative to the length Lego of the host vehicle.

[0052] Alternatively, the feature amount calculation unit 53 may calculate the overlap degree IoU based on the overlap length Lov and the vehicle length Lego of the host vehicle. For example, the feature amount calculation unit 53 may calculate the overlap degree IoU by dividing the overlap length Lov by the vehicle length Lego of the host vehicle. This calculation also makes it possible to capture the features of a change in the position of the lane change destination when the vehicle length Lego of the host vehicle is long.

[0053] <In case of complete inclusion> The feature calculation unit 53 may set the overlapping degree IoU to a preset value (for example, 1) when the position range of the prediction target vehicle in the forward / backward direction X completely encompasses the position range of the host vehicle, or when the position range of the host vehicle in the forward / backward direction X completely encompasses the position range of the prediction target vehicle. 1) When Lov=Lego or Lov=Lm, IoU=1 2) Otherwise, IoU=Lov / (Lm+Lego-Lov) ···(2)

[0054] 10 and 11 have the same overlapping degree IoU calculated by equation (1), but in Fig. 11, which does not completely encompass, the prediction target vehicle is more likely to change lanes behind the own vehicle. Therefore, when one of the own vehicle and the prediction target vehicle is extremely long, setting the overlapping degree IoU to a predetermined value in the case of complete encompassment makes it possible to capture the characteristics of the change in the position of the lane change destination.

[0055] <Increase in vehicle length due to safety distance> The feature calculation unit 53 may virtually increase the position range and vehicle length of one or both of the host vehicle and the prediction target vehicle used to calculate the overlap degree IoU, based on the minimum safe distance Lsf that must be secured in front of and behind the vehicle.

[0056] FIG. 12 shows a case where the position range and vehicle length Lego of the host vehicle are increased by the safety distance Lsf. The safety distance Lsf is added to the front and rear of the actual position range of the host vehicle, and the vehicle length Lego is longer by the front and rear safety distance Lsf. The overlap length Lov is calculated based on the position range of the host vehicle after the safety distance Lsf is added. The overlap degree IoU is calculated based on the host vehicle's length Lego after the safety distance Lsf is added.

[0057] The position range and vehicle length Lm of the target vehicle may be increased by the safety distance Lsf. The safety distance Lsf in front of the vehicle and the safety distance Lsf in rear of the vehicle may be different. For example, the safety distance Lsf in front of the vehicle may be longer than the safety distance Lsf in rear of the vehicle. The safety distance Lsf set in front of and behind each vehicle may be set to a predetermined value, or may be changed according to the vehicle type of each vehicle or the speed of each vehicle.

[0058] Generally, when a driver or an automatic driving device determines a lane change destination, the safe distance Lsf in front of and behind the vehicle is also taken into consideration. According to the above configuration, by taking the safe distance Lsf into consideration when calculating the overlap degree IoU, the characteristics of the lane change destination position can be captured more accurately.

[0059] <Calculation of the degree of overlap between the predicted vehicle and the preceding and following vehicles> The feature calculation unit 53 may calculate the degree of overlap between the position range of the prediction target vehicle in the forward / backward direction X and the position ranges of the preceding vehicle and the following vehicle of the subject vehicle, based on the position information and shape information of the prediction target vehicle and the position information and shape information of each of the preceding vehicle and the following vehicle of the subject vehicle.

[0060] FIG. 13 shows an example of calculating the overlapping degree IoUbk between the vehicle to be predicted and the following vehicle. If the position range of the vehicle to be predicted overlaps not only the positioning range of the subject vehicle but also the positioning range of the following vehicle, there is not enough space between the subject vehicle and the following vehicle for the vehicle to change lanes. Therefore, even if the overlapping degree IoU between the subject vehicle and the subject vehicle is the same, the greater the overlapping degree IoUbk between the subject vehicle and the following vehicle, the lower the possibility that the vehicle to be predicted will change lanes behind the subject vehicle. Therefore, the overlapping degree IoUbk between the subject vehicle and the following vehicle can capture the characteristics of the lane change destination of the vehicle to be predicted. The same applies to the overlapping degree IoUfr between the subject vehicle and the preceding vehicle.

[0061] The feature calculation unit 53 calculates the overlap degree IoUbk between the prediction target vehicle and the following vehicle and the overlap degree IoUfr between the prediction target vehicle and the preceding vehicle using a method similar to that used to calculate the overlap degree IoU between the prediction target vehicle and the host vehicle. Here, Lovbk is the overlap length of the position ranges where the position ranges of the prediction target vehicle and the following vehicle overlap in the longitudinal direction X, Lbk is the vehicle length of the following vehicle, Lovfr is the overlap length of the position ranges where the position ranges of the prediction target vehicle and the preceding vehicle overlap in the longitudinal direction X, and Lfr is the vehicle length of the preceding vehicle. IoUbk=Lovbk / (Lm+Lbk-Lovbk) IoUfr=Lovfr / (Lm+Lfr-Lovfr) ···(3)

[0062] If the prediction target vehicle and the following vehicle do not overlap, or if there is no following vehicle, Lovbk is 0. Also, if the prediction target vehicle and the preceding vehicle do not overlap, or if there is no preceding vehicle, Lovfr is 0. The above-mentioned safety distance Lsf may be taken into consideration when calculating Lovbk and Lovfr.

[0063] <Calculating the degree of overlap using probability distribution> As shown in FIG. 14 , the feature calculation unit 53 may calculate a probability distribution in which the prediction target vehicle exists in the forward / backward direction X as the position range of the prediction target vehicle, calculate a probability distribution in which the host vehicle exists in the forward / backward direction X as the position range of the host vehicle, and calculate the degree of overlap based on the degree of overlap between the probability distribution of the prediction target vehicle and the probability distribution of the host vehicle.

[0064] The position information and shape information of the target vehicle contain perception errors. By calculating the degree of overlap of the probability distributions, it is possible to calculate the degree of overlap taking perception errors into account.

[0065] The probability distribution is set based on the position range of each vehicle acquired as described above and a variance or standard deviation that takes into account recognition errors. For example, the feature calculation unit 53 generates a probability distribution that represents a predetermined variance or standard deviation in which the longitudinal direction X is non-dimensionalized, and multiplies the dimensionless longitudinal direction X by the length of the position range of each vehicle to generate the probability distribution of each vehicle. Note that the center position of the probability distribution in the longitudinal direction X is set to the center position of each vehicle in the longitudinal direction X. The variance or standard deviation of each vehicle may be changed depending on the reliability of the position range of each vehicle acquired by the information acquisition unit 51.

[0066] For example, the feature amount calculation unit 53 calculates an overlap length Lov of a position range in the longitudinal direction X where the probability distribution of the prediction target vehicle is equal to or greater than a predetermined value and a position range in which the probability distribution of the host vehicle is equal to or greater than a predetermined value overlap. Then, the feature amount calculation unit 53 calculates the degree of overlap IoU using equation (1). Note that the length of the union may be the length of the union of the position range in the longitudinal direction X where the probability distribution of the prediction target vehicle is equal to or greater than a predetermined value and the position range in which the probability distribution of the host vehicle is equal to or greater than a predetermined value.

[0067] Alternatively, the feature calculation unit 53 may calculate the overlap area where the probability distribution of the prediction target vehicle and the probability distribution of the own vehicle overlap, calculate the area of ​​the union of the probability distribution of the prediction target vehicle and the probability distribution of the own vehicle, and calculate the overlap area divided by the area of ​​the union as the degree of overlap IoU.

[0068] The feature amount calculation unit 53 may also calculate the degree of overlap between the prediction target vehicle and the preceding and following vehicles of the own vehicle based on the degree of overlap between the probability distributions of the locations of the respective vehicles.

[0069] <Calculating the degree of overlap taking into account the sense of intimidation> The feature amount calculation unit 53 may further calculate the degree of overlap IoU based on a feature amount relating to the sense of intimidation given to the driver of the other vehicle between the prediction target vehicle and the own vehicle.

[0070] When a driver decides where to change lanes, they are influenced by the sense of intimidation of other vehicles. With the above configuration, by taking into account the feature amount related to the sense of intimidation when calculating the overlap degree IoU, it is possible to more accurately capture the characteristics of the location of the lane change destination.

[0071] 15, the feature amount calculation unit 53 calculates the feature amount related to the intimidating feeling by dividing the vehicle length Lm of the prediction target vehicle by the vehicle length Lego of the subject vehicle. In addition to the vehicle length, vehicle features that give an impression to the driver of another vehicle, such as vehicle height, vehicle model, and vehicle color, may also be taken into consideration.

[0072] For example, the feature calculation unit 53 corrects the overlap degree to decrease as the intimidation of the prediction target vehicle relative to the subject vehicle increases, and corrects the overlap degree to increase as the intimidation of the subject vehicle relative to the prediction target vehicle increases. As the intimidation of the prediction target vehicle relative to the subject vehicle increases, the prediction target vehicle tends to decide on a lane change destination without considering the degree of overlap with the subject vehicle. Furthermore, as the intimidation of the subject vehicle relative to the prediction target vehicle increases, the prediction target vehicle tends to decide on a lane change destination with consideration of the degree of overlap with the subject vehicle. Therefore, by increasing or decreasing the overlap degree IoU using the feature related to the intimidation, it is possible to more accurately capture the characteristics of changes in the position of the lane change destination.

[0073] 1-1-4. Lane change prediction unit 54 The lane change prediction unit 54 predicts the position of the lane change destination of the prediction target vehicle into the own lane, using a prediction model to which at least one of position information and speed information of the prediction target vehicle and the overlapping degree IoU between the own vehicle and the prediction target vehicle are input. In this embodiment, both the position information and speed information of the prediction target vehicle are input to the prediction model.

[0074] Even if the relative position and relative speed of the prediction target vehicle with respect to the host vehicle are the same, the lane change destination position of the prediction target vehicle changes depending on the change in the overlapping degree IoU between the host vehicle and the prediction target vehicle. Therefore, the characteristics of the change in the lane change destination position can be captured based on the overlapping degree IoU between the vehicle and the prediction target vehicle. According to the above configuration, the lane change destination position of the prediction target vehicle is predicted using a prediction model that inputs the overlapping degree IoU between the host vehicle and the prediction target vehicle in addition to the position information and speed information of the prediction target vehicle, thereby improving prediction accuracy.

[0075] As the position information and speed information of the prediction target vehicle, the relative position and relative speed of the prediction target vehicle with respect to the host vehicle in the forward / backward direction X are input.

[0076] The position of the lane change destination of the prediction target vehicle is the output of the prediction model. The lane change prediction unit 54 may predict whether the lane change destination position of the prediction target vehicle is in front of or behind the host vehicle as the lane change destination position of the prediction target vehicle. Furthermore, the lane change prediction unit 54 may predict the lane change destination position relative to the host vehicle as the lane change destination position of the prediction target vehicle.

[0077] In this embodiment, a statistical model or machine learning model is used as the prediction model, which represents the relationship between the position information and speed information of the vehicle to be predicted, the degree of overlap IoU between the host vehicle and the vehicle to be predicted, and the position of the lane change destination. The statistical model or machine learning model is a mathematical model that represents the statistical relationship between input and output. The statistical model or machine learning model is trained or designed in advance using a large number of input and output data sets. The structure of the prediction model and each constant are stored in a storage device such as an EEPROM. The prediction model may be updated using a newly acquired data set.

[0078] For example, various known statistical models or machine learning models such as a support vector machine (SVM), a decision tree, a neural network, etc. Since the various statistical models or machine learning models are known, their description will be omitted.

[0079] According to this configuration, the position information and speed information of the vehicle to be predicted, as well as the overlapping degree IoU between the host vehicle and the vehicle to be predicted, are input into a prediction model using a statistical model or a machine learning model, and the output of the position of the lane change destination of the vehicle to be predicted can be calculated, capturing the characteristics of the change in the position of the lane change destination due to changes in the overlapping degree IoU.

[0080] Alternatively, a rule-based model may be used as the prediction model, which represents the relationship between the position information and speed information of the vehicle to be predicted, the degree of overlap IoU between the vehicle itself and the vehicle to be predicted, and the position of the lane change destination.

[0081] For example, in the rule-based model, as shown in FIG. 16, a judgment is made on a preset rule (condition) based on the position information and speed information of the vehicle to be predicted, and the degree of overlap IoU between the vehicle itself and the vehicle to be predicted, and a preset lane change destination position is output corresponding to the judgment result of each condition.

[0082] The lane change prediction unit 54 may predict the lane change destination position of the prediction target vehicle into the own lane using a prediction model that inputs one of the position information and speed information of the prediction target vehicle and the overlap degree IoU between the own vehicle and the prediction target vehicle. For example, a prediction model that inputs the speed information and overlap degree IoU of the prediction target vehicle may be used. As the relative speed of the prediction target vehicle with respect to the own vehicle increases and as the overlap degree IoU decreases, the likelihood that the lane change destination position of the prediction target vehicle will be ahead of the own vehicle increases. Alternatively, a prediction model that inputs the position information and overlap degree IoU of the prediction target vehicle may be used. As the position information of the prediction target vehicle becomes further ahead with respect to the own vehicle and as the overlap degree IoU decreases, the likelihood that the lane change destination position of the prediction target vehicle will be ahead of the own vehicle increases.

[0083] <Input the degree of overlap between the preceding and following vehicles> The lane change prediction unit 54 may predict the lane change destination of the target vehicle using a prediction model that further inputs position information and speed information of the preceding and following vehicles of the target vehicle, as well as overlapping degrees IoUfr and IoUbk between the preceding and following vehicles of the target vehicle and the target vehicle. That is, the number of parameters input to the prediction model is increased. In this case, a statistical model, a machine learning model, or a rule-based model is used as the prediction model.

[0084] According to this configuration, as described above, the overlapping degrees IoUfr and IoUbk of the preceding and following vehicles can represent the characteristics of whether there is sufficient space between the host vehicle and the preceding and following vehicles for the vehicle to change lanes. Therefore, even if the overlapping degree IoU of the host vehicle is the same, it is possible to calculate the lane change destination position of the vehicle to be predicted, capturing the characteristics of the change in the lane change destination position due to changes in the overlapping degrees IoUfr and IoUbk of the preceding and following vehicles.

[0085] As the position information and speed information of the preceding vehicle and the following vehicle, the relative positions and relative speeds of the preceding vehicle and the following vehicle relative to the host vehicle in the forward / backward direction X are input.

[0086] 1-1-5. Driving Support Department 55 The driving assistance unit 55 provides driving assistance for the host vehicle based on the prediction result of the lane change destination position of the prediction target vehicle. For example, the driving assistance unit 55 notifies the driver of the prediction result of the lane change destination position via the human interface device 37, such as a speaker and a display screen. The driving assistance unit 55 increases or decreases the speed of the host vehicle based on the prediction result of the lane change destination position. For example, if the predicted lane change destination position is ahead of the host vehicle, the driving assistance unit 55 decreases the speed of the host vehicle, and if the predicted position is behind the host vehicle, the driving assistance unit 55 increases the speed of the host vehicle.

[0087] The driving assistance unit 55 determines a target output torque and a target braking force for increasing or decreasing the speed of the vehicle, and transmits these to a power control device and a brake control device serving as the drive control device 36. For example, the driving assistance unit 55 changes the target output torque and the target braking force by feedback control or the like so that actual values ​​approach target values ​​such as a target inter-vehicle distance, a target speed, and a target acceleration. The power control device then controls the output torque of a power machine 8 such as an internal combustion engine or a motor in accordance with the target output torque. Furthermore, the brake control device controls the braking operation of an electric brake device 9 in accordance with the target braking force.

[0088] The driving assistance unit 55 may perform steering control to change the target steering angle so that the host vehicle travels within the host vehicle's lane. Alternatively, the driving assistance unit 55 may generate a target driving trajectory and change the target steering angle so that the host vehicle travels along the target driving trajectory. The driving assistance unit 55 transmits the target steering angle to the automatic steering control device, and the automatic steering control device controls the electric steering device 7 so that the steering angle follows the target steering angle.

[0089] Alternatively, to perform more advanced automated driving, the driving assistance unit 55 may generate a time-series target driving trajectory. The time-series target driving trajectory is a time-series driving plan including the target position, target traveling direction, target speed, target acceleration, etc. of the host vehicle at each future time. The driving assistance unit 55 changes the target driving trajectory based on the prediction result of the position of the lane change destination. When increasing the speed of the host vehicle, the driving assistance unit 55 increases the target speed and target acceleration at each future time according to the increase in the target speed. The driving assistance unit 55 controls the host vehicle so that the host vehicle travels along the target driving trajectory. For example, the driving assistance unit 55 determines a target output torque, a target braking force, a target steering angle, a turn signal operation command, etc., and transmits each determined command value to the power control device, brake control device, automatic steering control device, light control device, etc., which are the drive control device 36.

[0090] <Flowchart> 19 is a schematic flowchart illustrating the processing (vehicle behavior prediction method) of the vehicle behavior prediction device 1 according to this embodiment. The processing in FIG. 19 is executed, for example, at every predetermined calculation cycle.

[0091] In step S01, as described above, the information acquisition unit 51 acquires the surrounding conditions and the running conditions of the host vehicle, and acquires the position information, speed information, and shape information of adjacent vehicles running in the adjacent lane adjacent to the host vehicle's running lane based on the surrounding conditions and the running conditions of the host vehicle. The information acquisition unit 51 also acquires the various types of information described above.

[0092] In step S02, as described above, the prediction target setting unit 52 sets a prediction target vehicle from among the adjacent vehicles.

[0093] In step S03, as described above, the feature amount calculation unit 53 calculates the degree of overlap between the position range of the prediction target vehicle and the position range of the host vehicle in the longitudinal direction, based on the position information and shape information of the prediction target vehicle and the shape information of the host vehicle. As described above, the feature amount calculation unit 53 calculates the degree of overlap between the prediction target vehicle and the preceding vehicle and the following vehicle.

[0094] In step S04, as described above, the lane change prediction unit 54 predicts the position of the lane change destination of the prediction target vehicle into the own lane using a prediction model that inputs the position information and speed information of the prediction target vehicle and the degree of overlap between the own vehicle and the prediction target vehicle.

[0095] In step S05, as described above, the driving assistance unit 55 performs driving assistance for the vehicle itself based on the prediction result of the lane change destination position of the prediction target vehicle.

[0096] 2. Second Embodiment Next, a vehicle behavior prediction device 1 according to a second embodiment will be described. A description of the same components as those in the first embodiment will be omitted. The basic configuration of the vehicle behavior prediction device 1 according to this embodiment is the same as that of the first embodiment, but differs from the first embodiment in that a feature calculated in the feature calculation unit 53 and input to the prediction model is added.

[0097] As in the first embodiment, the feature amount calculation unit 53 calculates the overlapping degree IoU between the prediction target vehicle and the host vehicle, and the overlapping degree IoU is input to the prediction model. Also, as in the first embodiment, the feature amount calculation unit 53 may calculate the overlapping degrees IoUfr, IoUbk between the prediction target vehicle and the preceding vehicle and the following vehicle, and the overlapping degrees IoUfr, IoUbk may be input to the prediction model.

[0098] <Calculation and input of collision probability> In this embodiment, the feature calculation unit 53 calculates the degree of collision probability of the vehicle to be predicted with each of the preceding and following vehicles of the vehicle to be predicted, based on the position information and speed information of the vehicle to be predicted and the position information and speed information of each of the preceding and following vehicles of the vehicle to be predicted.

[0099] 17, the feature amount calculation unit 53 calculates the time to collision TTCfr and TTCbk of the target vehicle relative to the preceding vehicle and the following vehicle, respectively, as the degree of collision possibility. The feature amount calculation unit 53 calculates the time to collision TTCfr of the target vehicle relative to the preceding vehicle by dividing the relative position of the preceding vehicle relative to the target vehicle by the relative speed of the preceding vehicle relative to the target vehicle. The feature amount calculation unit 53 also calculates the time to collision TTCbk of the target vehicle relative to the following vehicle by dividing the relative position of the following vehicle relative to the target vehicle by the relative speed of the following vehicle relative to the target vehicle.

[0100] The lane change prediction unit 54 predicts the lane change destination position of the vehicle to be predicted, using a prediction model to which the collision probability levels TTCfr, TTCbk of the vehicle to be predicted with respect to the preceding vehicle and the following vehicle of the vehicle to be predicted are further input. That is, the number of parameters input to the prediction model is increased compared to that of the first embodiment. In this case as well, a statistical model or a rule-based model is used as the prediction model.

[0101] Even if the overlap degree IoU is the same, as the time to collision with the preceding vehicle TTCfr becomes shorter, the likelihood that the prediction target vehicle will change lanes to behind the host vehicle to avoid a collision with the preceding vehicle decreases. Also, even if the overlap degree IoU is the same, as the time to collision with the following vehicle TTCbk becomes shorter, the likelihood that the prediction target vehicle will change lanes to behind the host vehicle to avoid a collision with the following vehicle decreases. Therefore, even if the overlap degree IoU is the same, it is possible to calculate the lane change destination position of the prediction target vehicle by capturing the characteristics of changes in the lane change destination position due to changes in the collision probability degrees TTCfr and TTCbk with the preceding and following vehicles.

[0102] <Calculation and input of acceleration / deceleration tendency> In this embodiment, the feature amount calculation unit 53 calculates the tendency of acceleration / deceleration of the vehicle to be predicted based on time-series speed information of the vehicle to be predicted.

[0103] For example, the feature amount calculation unit 53 determines the maximum value Vmax and minimum value Vmin of the speed of the prediction target vehicle within a past determination period based on the time-series speed information of the prediction target vehicle. Then, as shown in the following equation, the feature amount calculation unit 53 calculates the value obtained by subtracting the minimum value Vmin from the current vehicle speed Vnow of the prediction target vehicle as an acceleration tendency, and calculates the value obtained by subtracting the maximum value Vmax from the current vehicle speed Vnow of the prediction target vehicle as a deceleration tendency. Acceleration tendency = Vnow - Vmin Deceleration tendency = Vnow - Vmax (4)

[0104] Then, the feature amount calculation unit 53 calculates the acceleration / deceleration tendency as the larger absolute value or the average value of the acceleration tendency or the deceleration tendency. Alternatively, both the acceleration tendency and the deceleration tendency may be calculated as the acceleration / deceleration tendency. Note that the time-series speed information of the prediction target vehicle may be the time-series relative speed of the prediction target vehicle with respect to the host vehicle.

[0105] The lane change prediction unit 54 predicts the lane change destination of the target vehicle using a prediction model to which the acceleration / deceleration tendency of the target vehicle is further input. That is, the parameters input to the prediction model are increased. In this case, a statistical model or a rule-based model is also used as the prediction model.

[0106] Even if the overlapping degree IoU is the same, the greater the tendency for acceleration, the greater the likelihood that the prediction target vehicle will change lanes ahead of the host vehicle. Also, even if the overlapping degree IoU is the same, the greater the tendency for deceleration, the greater the likelihood that the prediction target vehicle will change lanes behind the host vehicle. Therefore, even if the overlapping degree IoU is the same, it is possible to calculate the lane change destination position of the measurement target vehicle by capturing the characteristics of changes in the lane change destination position due to changes in the tendency for acceleration and deceleration.

[0107] <Calculating and inputting the remaining time to the end of the merging lane> When the adjacent lane is a merging lane, the information acquisition unit 51 acquires the distance from the adjacent vehicle to the end of the merging lane. The prediction target setting unit 52 sets the adjacent vehicle traveling in the merging lane as a prediction target vehicle.

[0108] The feature amount calculation unit 53 calculates the remaining time until the prediction target vehicle reaches the end of the merging lane based on the distance from the prediction target vehicle to the end of the merging lane and the speed information of the prediction target vehicle.

[0109] For example, the feature amount calculation unit 53 calculates the remaining time by dividing the distance from the prediction target vehicle to the end of the merging lane by the speed of the prediction target vehicle.

[0110] Then, the lane change prediction unit 54 predicts the lane change destination position of the prediction target vehicle using a prediction model to which the remaining time until the vehicle reaches the end of the merging lane is further input. In other words, the parameters input to the prediction model are increased.

[0111] The position of the lane change destination where the remaining time has elapsed and the lane change will be made is no longer predicted, thereby improving prediction accuracy.

[0112] <Calculating and entering the remaining time until the start of the forced merge> When the adjacent lane is a merging lane, the information acquisition unit 51 acquires the distance from the adjacent vehicle to the end of the merging lane and road width information of the merging lane. The prediction target setting unit 52 sets the adjacent vehicle traveling in the merging lane as a prediction target vehicle.

[0113] The feature calculation unit 53 calculates the remaining time until the predicted vehicle will need to forcibly merge into its own lane based on the distance from the predicted vehicle to the end of the merging lane, road width information of the merging lane, and speed information of the predicted vehicle.

[0114] 18 and the following equation, the feature amount calculation unit 53 determines the point where the road width of the merging lane starts to decrease from the road width information at each point on the merging lane, calculates the distance Ltpr from the vehicle to be predicted to the point where the road width starts to decrease, and calculates a first remaining time Tmgst1 until the forced merging starts by dividing the distance Ltpr to the point where the decrease starts by the longitudinal speed Vx of the vehicle to be predicted. The feature amount calculation unit 53 also calculates the longitudinal distance Ly required for the vehicle to move laterally from the merging lane to its own lane, and calculates a second remaining time Tmgst2 until the forced merging starts by subtracting the distance Ly required for the lateral movement from the distance Lend to the end of the merging lane by the longitudinal speed Vx of the vehicle to be predicted. Then, the feature calculation unit 53 calculates the smaller of the first remaining time Tmgst1 and the second remaining time Tmgst2 as the remaining time Tmgst until the final forced merging starts. Here, Wmg is the lane width of the merging lane, and Vymg is the lateral speed of the prediction target vehicle at the time of lane change, which is set according to the vehicle type and current speed of the prediction target vehicle. Tmgst1=Ltpr / Vx Ly=Wmg / Vymg×Vx Tmgst2=(Lend-Ly) / Vx Tmgst=MIN(Tmgst1, Tmgst2) ···(5)

[0115] Then, the lane change prediction unit 54 predicts the lane change destination position of the prediction target vehicle using a prediction model to which the remaining time Tmgst until the start of the forced merging is further input. In other words, the parameters input to the prediction model are increased.

[0116] The position of the lane change destination where the lane change occurs after the remaining time Tmgst until the start of the forced merging has elapsed is no longer predicted, thereby improving prediction accuracy.

[0117] In addition, it is not necessary for all of the following to be input into the prediction model: the degree of collision probability, the tendency to accelerate or decelerate, the remaining time until the end of the merging lane is reached, and the remaining time Tmgst until the forced merging begins; one or more of these may be input into the prediction model.

[0118] 3. Embodiment 3 Next, a vehicle behavior prediction device 1 according to a second embodiment will be described. A description of the same components as those in the first embodiment will be omitted. The basic configuration of the vehicle behavior prediction device 1 according to this embodiment is the same as that of the first embodiment.

[0119] However, in this embodiment, the target vehicle is set to a specific vehicle present in the control area. For example, the vehicle behavior prediction device 1 sets multiple control target vehicles present in the control area as target vehicles in order, and performs vehicle behavior prediction processing for the set target vehicles. The vehicle behavior prediction device 1 is provided in a server connected to a network. Then, the lane change prediction unit 54 transmits a prediction result of the destination position of the predicted target vehicle to the target lane to the target vehicle. The target vehicle (driving assistance unit) performs driving assistance for the host vehicle based on the transmitted prediction result of the destination position of the lane change, as in the first embodiment. The host vehicle in the first embodiment is replaced with the target vehicle, and the host lane in the first embodiment is replaced with the target lane. FIG. 20 shows a schematic block diagram of the vehicle behavior prediction device 1.

[0120] The controlled area may be set in an area of ​​a public road, or in an area within various facilities such as a factory, a logistics base, or a resort facility.

[0121] The information acquisition unit 51 acquires, via wireless and wired communication, the surrounding conditions and vehicle conditions recognized by each of the multiple vehicles 100 present in the control area and the monitoring devices 101, such as roadside units or monitoring cameras. The acquired information includes the surrounding conditions and vehicle conditions of the target vehicle. Based on the surrounding conditions and vehicle conditions of the target vehicle, the information acquisition unit 51 acquires position information, speed information, and shape information of adjacent vehicles traveling in an adjacent lane adjacent to the target lane in which the target vehicle is traveling.

[0122] The processes performed by the prediction target setting unit 52, the feature amount calculation unit 53, the lane change prediction unit 54, and the like are the same as those in the first or second embodiment, and therefore will not be described here.

[0123] The driving assistance unit 55 may be provided in the vehicle behavior prediction device 1 and may perform driving assistance for the target vehicle via wireless communication. In this case, the processing of the driving assistance unit 55 itself is the same as in the first embodiment, and therefore a description thereof will be omitted.

[0124] <Summary of various aspects of the present application> Various aspects of the present application will be summarized below as appendices.

[0125] (Appendix 1) an information acquisition unit that acquires a surrounding state of a target vehicle and a vehicle state of the target vehicle, and acquires position information, speed information, and shape information of an adjacent vehicle traveling in an adjacent lane adjacent to a target lane in which the target vehicle is traveling, based on the surrounding state and the vehicle state of the target vehicle; a prediction target setting unit that sets a prediction target vehicle from the adjacent vehicles; a feature amount calculation unit that calculates an overlapping degree between a position range of the predicted target vehicle and a position range of the target vehicle in a front-rear direction based on the position information and the shape information of the predicted target vehicle and the shape information included in the vehicle state of the target vehicle; and a lane change prediction unit that predicts the position of the vehicle to be predicted into the target lane using a prediction model to which at least one of the position information and the speed information of the vehicle to be predicted and the degree of overlap between the target vehicle and the vehicle to be predicted is input.

[0126] (Appendix 2) The vehicle behavior prediction device described in Appendix 1, wherein the prediction model is a statistical model or a machine learning model that represents at least one of the position information and the speed information of the prediction target vehicle, as well as the degree of overlap between the target vehicle and the prediction target vehicle, and the position of the lane change destination.

[0127] (Appendix 3) the information acquisition unit determines, based on the surrounding state, whether the adjacent lane in which the adjacent vehicle is traveling is a merging lane that merges into the target lane; 3. The vehicle behavior prediction device according to claim 1, wherein the prediction target setting unit sets the adjacent vehicle traveling in the merging lane as the prediction target vehicle when the adjacent lane is the merging lane.

[0128] (Appendix 4) 4. The vehicle behavior prediction device according to claim 1, wherein the feature calculation unit calculates the degree of overlap based on an overlap length of the position ranges of the prediction target vehicle and the target vehicle in the longitudinal direction, and a length of a union of the position range of the prediction target vehicle and the position range of the target vehicle.

[0129] (Appendix 5) the information acquisition unit acquires position information, speed information, and shape information of a preceding vehicle and a following vehicle traveling in the target lane in front of and behind the target vehicle based on the surrounding state and the vehicle state of the target vehicle; the feature amount calculation unit calculates a degree of overlap between a position range of the prediction target vehicle in the longitudinal direction and each of the position ranges of the preceding vehicle and the following vehicle, based on the position information and the shape information of the prediction target vehicle and the position information and the shape information of each of the preceding vehicle and the following vehicle of the target vehicle; The vehicle behavior prediction device according to any one of appendices 1 to 4, wherein the lane change prediction unit predicts the lane change destination position of the prediction target vehicle using the prediction model to which at least one of the position information and the speed information of each of the preceding vehicle and the following vehicle of the target vehicle, and the degree of overlap between each of the preceding vehicle and the following vehicle of the target vehicle and the prediction target vehicle are further input.

[0130] (Appendix 6) the information acquisition unit acquires position information and speed information of a preceding vehicle and a following vehicle traveling in the adjacent lane in front of and behind the prediction target vehicle based on the surrounding state and the vehicle state of the target vehicle; the feature calculation unit calculates a degree of collision likelihood of the prediction target vehicle with respect to each of the preceding vehicle and the following vehicle of the prediction target vehicle based on the position information and the speed information of the prediction target vehicle and the position information and the speed information of each of the preceding vehicle and the following vehicle of the prediction target vehicle; The vehicle behavior prediction device according to any one of appendices 1 to 5, wherein the lane change prediction unit predicts the lane change destination position of the vehicle to be predicted using the prediction model to which the degree of collision likelihood of the vehicle to be predicted with respect to each of the preceding vehicle and the following vehicle of the vehicle to be predicted is further input.

[0131] (Appendix 7) the information acquisition unit determines, based on the surrounding conditions, whether the adjacent lane in which the adjacent vehicle is traveling is a merging lane that merges into the target lane, and if the adjacent lane is the merging lane, acquires a distance from the adjacent vehicle to an end of the merging lane; the prediction target setting unit sets the adjacent vehicle traveling in the merging lane as the prediction target vehicle when the adjacent lane is the merging lane; the feature amount calculation unit calculates a remaining time until the prediction target vehicle reaches an end of the merging lane based on a distance to the end of the merging lane and the speed information of the prediction target vehicle; and The vehicle behavior prediction device according to any one of appendix 1 to 6, wherein the lane change prediction unit predicts the lane change destination position of the prediction target vehicle using the prediction model to which the remaining time until the vehicle reaches an end of a merging lane is further input.

[0132] (Appendix 8) the information acquisition unit determines, based on the surrounding conditions, whether the adjacent lane in which the adjacent vehicle is traveling is a merging lane that merges into the target lane, and if the adjacent lane is the merging lane, acquires a distance from the adjacent vehicle to an end of the merging lane and road width information of the merging lane; the prediction target setting unit sets the adjacent vehicle traveling in the merging lane as the prediction target vehicle when the adjacent lane is the merging lane; the feature amount calculation unit calculates a remaining time until the prediction target vehicle needs to forcibly merge into the target lane based on a distance to an end of the merging lane, road width information of the merging lane, and the speed information of the prediction target vehicle; The vehicle behavior prediction device according to any one of appendices 1 to 7, wherein the lane change prediction unit predicts the lane change destination position of the prediction target vehicle using the prediction model to which the remaining time until the start of a forced merging is further input.

[0133] (Appendix 9) the feature calculation unit calculates an acceleration / deceleration tendency of the prediction target vehicle based on the time-series speed information of the prediction target vehicle; The vehicle behavior prediction device according to any one of appendices 1 to 8, wherein the lane change prediction unit predicts the lane change destination position of the vehicle to be predicted using the prediction model to which the acceleration / deceleration tendency of the vehicle to be predicted is further input.

[0134] (Appendix 10) 10. The vehicle behavior prediction device according to claim 1, wherein the feature calculation unit virtually increases the position range and the length of each vehicle used in calculating the degree of overlap based on a minimum safe distance that must be secured in front of and behind the vehicle.

[0135] (Appendix 11) 11. The vehicle behavior prediction device according to any one of claims 1 to 10, wherein the feature calculation unit sets the degree of overlap to a preset value when the position range of the predicted target vehicle completely encompasses the position range of the target vehicle, or when the position range of the target vehicle completely encompasses the position range of the predicted target vehicle.

[0136] (Appendix 12) the feature calculation unit calculates a probability distribution in which the prediction target vehicle exists in the longitudinal direction as a position range of the prediction target vehicle, calculates a probability distribution in which the target vehicle exists in the longitudinal direction as a position range of the target vehicle, and calculates a degree of overlap based on a degree of overlap between the probability distribution of the prediction target vehicle and the probability distribution of the target vehicle.

[0137] (Appendix 13) 13. The vehicle behavior prediction device according to claim 1, wherein the feature calculation unit further calculates the degree of overlap based on a feature related to a sense of intimidation given to a driver of another vehicle between the prediction target vehicle and the target vehicle.

[0138] (Appendix 14) The vehicle behavior prediction device according to any one of appendices 1 to 13, wherein the information acquisition unit acquires the length of the vehicle to be predicted and the vehicle type of the vehicle to be predicted, and if a difference between the acquired length of the vehicle to be predicted and the length estimated from the vehicle type of the vehicle to be predicted is equal to or greater than a determination value, sets the length estimated from the vehicle type of the vehicle to be predicted as the official length of the vehicle to be used in calculating the degree of overlap.

[0139] (Appendix 15) an information acquisition step of acquiring a surrounding state of a target vehicle and a vehicle state of the target vehicle, and acquiring position information, speed information, and shape information of an adjacent vehicle traveling in an adjacent lane adjacent to a target lane in which the target vehicle is traveling based on the surrounding state and the vehicle state of the target vehicle; a prediction target setting step of setting a prediction target vehicle from the adjacent vehicles; a feature amount calculation step of calculating an overlap degree between a position range of the predicted target vehicle and a position range of the target vehicle in a front-rear direction based on the position information and the shape information of the predicted target vehicle and the shape information included in the vehicle state of the target vehicle; a lane change prediction step of predicting a position where the prediction target vehicle will change lanes into the target lane using a prediction model to which at least one of the position information and the speed information of the prediction target vehicle and the degree of overlap between the target vehicle and the prediction target vehicle are input.

[0140] Although various exemplary embodiments and examples are described in this application, the various features, aspects, and functions described in one or more embodiments are not limited to the application of a particular embodiment, but may be applied to the embodiments alone or in various combinations. Therefore, countless variations not illustrated are contemplated within the scope of the technology disclosed in this specification. For example, this includes cases where at least one component is modified, added, or omitted, or where at least one component is extracted and combined with components of another embodiment. [Explanation of symbols]

[0141] 1: Vehicle behavior prediction device, 51: Information acquisition unit, 52: Prediction target setting unit, 53: Feature calculation unit, 54: Lane change prediction unit, 55: Driving assistance unit, Lego: Vehicle length, Lend: Distance to the end of the merging lane, Lm: Vehicle length of the vehicle to be predicted, Lsf: Safety distance, X: Forward / backward direction

Claims

1. an information acquisition unit that acquires a surrounding state of a target vehicle and a vehicle state of the target vehicle, and acquires position information, speed information, and shape information of an adjacent vehicle traveling in an adjacent lane adjacent to a target lane in which the target vehicle is traveling, based on the surrounding state and the vehicle state of the target vehicle; a prediction target setting unit that sets a prediction target vehicle from the adjacent vehicles; a feature amount calculation unit that calculates an overlapping degree between a position range of the predicted target vehicle and a position range of the target vehicle in a front-rear direction based on the position information and the shape information of the predicted target vehicle and the shape information included in the vehicle state of the target vehicle; and a lane change prediction unit that predicts the position of the vehicle to be predicted into the target lane using a prediction model to which at least one of the position information and the speed information of the vehicle to be predicted and the degree of overlap between the target vehicle and the vehicle to be predicted is input.

2. 2. The vehicle behavior prediction device according to claim 1, wherein the prediction model is a statistical model or a machine learning model that represents a relationship between at least one of the position information and the speed information of the vehicle to be predicted, the degree of overlap between the target vehicle and the vehicle to be predicted, and the position of the lane change destination.

3. the information acquisition unit determines, based on the surrounding state, whether the adjacent lane in which the adjacent vehicle is traveling is a merging lane that merges into the target lane; The vehicle behavior prediction device according to claim 1 , wherein the prediction target setting unit sets the adjacent vehicle traveling in the merging lane as the prediction target vehicle when the adjacent lane is the merging lane.

4. 2. The vehicle behavior prediction device according to claim 1, wherein the feature calculation unit calculates the degree of overlap based on an overlap length of the position ranges of the prediction target vehicle and the target vehicle in the longitudinal direction, and a length of a union of the position ranges of the prediction target vehicle and the target vehicle.

5. the information acquisition unit acquires position information, speed information, and shape information of a preceding vehicle and a following vehicle traveling in the target lane in front of and behind the target vehicle based on the surrounding state and the vehicle state of the target vehicle; the feature amount calculation unit calculates a degree of overlap between a position range of the prediction target vehicle in the longitudinal direction and each of the position ranges of the preceding vehicle and the following vehicle, based on the position information and the shape information of the prediction target vehicle and the position information and the shape information of each of the preceding vehicle and the following vehicle of the target vehicle; 2. The vehicle behavior prediction device according to claim 1, wherein the lane change prediction unit predicts the lane change destination position of the target vehicle using the prediction model to which at least one of the position information and the speed information of each of the preceding vehicle and the following vehicle of the target vehicle, and the degree of overlap between each of the preceding vehicle and the following vehicle of the target vehicle and the target vehicle are further input.

6. the information acquisition unit acquires position information and speed information of a preceding vehicle and a following vehicle traveling in the adjacent lane in front of and behind the prediction target vehicle based on the surrounding state and the vehicle state of the target vehicle; the feature calculation unit calculates a degree of collision likelihood of the prediction target vehicle with respect to each of the preceding vehicle and the following vehicle of the prediction target vehicle based on the position information and the speed information of the prediction target vehicle and the position information and the speed information of each of the preceding vehicle and the following vehicle of the prediction target vehicle; 2. The vehicle behavior prediction device according to claim 1, wherein the lane change prediction unit predicts the lane change destination position of the vehicle to be predicted using the prediction model to which the degree of collision likelihood of the vehicle to be predicted with respect to each of the preceding vehicle and the following vehicle of the vehicle to be predicted is further input.

7. the information acquisition unit determines, based on the surrounding conditions, whether the adjacent lane in which the adjacent vehicle is traveling is a merging lane that merges into the target lane, and if the adjacent lane is the merging lane, acquires a distance from the adjacent vehicle to an end of the merging lane; the prediction target setting unit sets the adjacent vehicle traveling in the merging lane as the prediction target vehicle when the adjacent lane is the merging lane; the feature amount calculation unit calculates a remaining time until the prediction target vehicle reaches an end of the merging lane based on a distance to the end of the merging lane and the speed information of the prediction target vehicle; and 2. The vehicle behavior prediction device according to claim 1, wherein the lane change prediction unit predicts the lane change destination position of the prediction target vehicle using the prediction model to which the remaining time until the vehicle reaches an end of a merging lane is further input.

8. the information acquisition unit determines, based on the surrounding conditions, whether the adjacent lane in which the adjacent vehicle is traveling is a merging lane that merges into the target lane, and if the adjacent lane is the merging lane, acquires a distance from the adjacent vehicle to an end of the merging lane and road width information of the merging lane; the prediction target setting unit sets the adjacent vehicle traveling in the merging lane as the prediction target vehicle when the adjacent lane is the merging lane; the feature calculation unit calculates a first remaining time by dividing the distance from the prediction target vehicle to a point where the road width of the merging lane starts to decrease by the speed of the prediction target vehicle based on the distance to the end of the merging lane, road width information of the merging lane, and the speed information of the prediction target vehicle; calculates a second remaining time by dividing the distance obtained by subtracting a longitudinal distance required for the prediction target vehicle to move laterally from the merging lane to the target lane from the distance to the end of the merging lane by the speed of the prediction target vehicle; and calculates the smaller of the first remaining time and the second remaining time as the remaining time; The vehicle behavior prediction device according to claim 1 , wherein the lane change prediction unit predicts the lane change destination position of the prediction target vehicle using the prediction model to which the remaining time is further input.

9. the feature calculation unit calculates an acceleration / deceleration tendency of the prediction target vehicle based on the time-series speed information of the prediction target vehicle; 2. The vehicle behavior prediction device according to claim 1, wherein the lane change prediction unit predicts the lane change destination position of the vehicle to be predicted using the prediction model to which the acceleration / deceleration tendency of the vehicle to be predicted is further input.

10. The vehicle behavior prediction device according to claim 1 , wherein the feature calculation unit virtually increases the position range and vehicle length of each vehicle used to calculate the degree of overlap based on the minimum safe distance that must be maintained in front of and behind the vehicle.

11. 2. The vehicle behavior prediction device according to claim 1, wherein the feature calculation unit sets the degree of overlap to a predetermined value when the position range of the predicted target vehicle completely encompasses the position range of the target vehicle, or when the position range of the target vehicle completely encompasses the position range of the predicted target vehicle.

12. 2. The vehicle behavior prediction device according to claim 1, wherein the feature calculation unit calculates, as the position range of the prediction target vehicle, a probability distribution in the longitudinal direction in which the target vehicle exists; calculates, as the position range of the target vehicle, a probability distribution in the longitudinal direction in which the target vehicle exists; and calculates, as the degree of overlap, a value obtained by dividing an overlap length between a first position range in which the probability distribution of the prediction target vehicle is equal to or greater than a predetermined value and a second position range in which the probability distribution of the target vehicle is equal to or greater than the predetermined value by a length of a union of the first position range and the second position range; or calculates, as the degree of overlap, a value obtained by dividing an overlap area in which the probability distribution of the prediction target vehicle and the probability distribution of the target vehicle overlap by an area of ​​a union of the probability distribution of the prediction target vehicle and the probability distribution of the target vehicle.

13. 2. The vehicle behavior prediction device according to claim 1, wherein the feature calculation unit further calculates the feature by dividing the vehicle length of the prediction target vehicle by the vehicle length of the target vehicle, and calculates the degree of overlap based on the feature.

14. 2. The vehicle behavior prediction device according to claim 1, wherein the information acquisition unit acquires the vehicle length of the prediction target vehicle and the vehicle type of the prediction target vehicle, and if a difference between the acquired vehicle length of the prediction target vehicle and the vehicle length estimated from the vehicle type of the prediction target vehicle is equal to or greater than a determination value, sets the vehicle length estimated from the vehicle type of the prediction target vehicle as the official vehicle length of the prediction target vehicle to be used in calculating the degree of overlap.

15. an information acquisition step of acquiring a surrounding state of a target vehicle and a vehicle state of the target vehicle, and acquiring position information, speed information, and shape information of an adjacent vehicle traveling in an adjacent lane adjacent to a target lane in which the target vehicle is traveling based on the surrounding state and the vehicle state of the target vehicle; a prediction target setting step of setting a prediction target vehicle from the adjacent vehicles; a feature amount calculation step of calculating an overlap degree between a position range of the predicted target vehicle and a position range of the target vehicle in a front-rear direction based on the position information and the shape information of the predicted target vehicle and the shape information included in the vehicle state of the target vehicle; a lane change prediction step of predicting a position where the prediction target vehicle will change lanes into the target lane using a prediction model to which at least one of the position information and the speed information of the prediction target vehicle and the degree of overlap between the target vehicle and the prediction target vehicle are input.

Images