Vehicle control system

Abstract

To improve the technology for determining a person's posture based on the likelihood of their posture. [Solution] The vehicle control device 20 includes a control unit 200, which derives a reference posture likelihood, indicating the likelihood that the posture of the object at the first time is the reference posture, and a nearest posture likelihood, indicating the likelihood that the posture of the object is the nearest posture, from an image of the object at the first time. It also calculates a likelihood ratio, which is the ratio of the reference posture likelihood to the nearest posture likelihood at the first time. The nearest posture is a posture that the object may take before or after transitioning to the reference posture. If the reference posture likelihood at the first time is greater than or equal to a threshold, and the likelihood ratio at the first time is greater than or equal to a first predetermined value, a flag indicating that the posture of the object is the reference posture is turned ON. If the reference posture likelihood at the first time is less than a threshold, or if the likelihood ratio at the first time is less than a first predetermined value, the flag is turned OFF.

Description

Technical Field

[0001] This disclosure relates to a vehicle control device.

Background Art

[0002] Conventionally, a technique for determining a person's posture based on a posture likelihood indicating the likelihood that the person's posture is a specific posture has been known. For example, Patent Document 1 discloses a technique for outputting the likelihood of a person's posture and determining that the person's posture is a specific posture (for example, a lying position) when the likelihood is equal to or greater than a threshold value.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] When using posture determination to determine permission to start a vehicle, it is necessary to determine that the posture of a person in the vehicle is a safe posture, for example, a sitting posture. When performing sitting determination by image recognition, the sitting posture can be determined based on whether a posture likelihood indicating the likelihood that the target's posture is a specific posture, for example, a sitting likelihood, exceeds a threshold value. However, due to factors such as the way a person sits, the position of the seat, the internal structure of the vehicle, variations in the input image signal due to the brightness at the time of imaging, and the performance of the camera, the sitting likelihood may increase or decrease. Therefore, in order to accurately determine the sitting posture, it is necessary to set the threshold value high. If the threshold value is high, the time required for the sitting likelihood to exceed the threshold value increases, which may cause stress to the passengers. In addition, the sitting determination is performed for all people in the vehicle. Therefore, waiting until the sitting likelihood of all people exceeds the threshold value may cause further stress to the passengers. On the other hand, if the threshold value is low, it becomes difficult to accurately determine the sitting posture.

[0005] In light of these circumstances, the purpose of this disclosure is to improve the technology for determining a person's posture based on the likelihood of that posture. [Means for solving the problem]

[0006] A vehicle control device according to one embodiment of the present disclosure comprises a control unit, The control unit, From the image of the object at the first time step, the likelihood of the object's posture being the reference posture at the first time step is derived, and the likelihood of the object's posture being the most recent posture is derived. Furthermore, the likelihood ratio, which is the ratio of the likelihood of the reference posture to the likelihood of the most recent posture at the first time step, is calculated, and the most recent posture is a posture that the object may take before or after transitioning to the reference posture. If the likelihood of the reference posture at the first time is greater than or equal to a threshold, and the likelihood ratio at the first time is greater than or equal to a first predetermined value, the flag indicating that the posture of the object is the reference posture is turned ON. If the likelihood of the reference posture at the first time is less than the threshold, or if the likelihood ratio at the first time is less than the first predetermined value, the flag is turned OFF. [Effects of the Invention]

[0007] According to one embodiment of the present disclosure, the technology for determining a person's posture based on the likelihood of that posture is improved. [Brief explanation of the drawing]

[0008] [Figure 1] This block diagram shows the schematic configuration of the vehicle control system according to this embodiment. [Figure 2] This is a schematic diagram illustrating the typical temporal changes in posture and posture likelihood before and after a subject transitions to a seated position. [Figure 3] This flowchart shows the operation of the vehicle control device according to this embodiment. [Modes for carrying out the invention]

[0009] The embodiments of this disclosure will be described below with reference to the drawings.

[0010] (Summary of this embodiment) Referring to Figure 1, an overview of the vehicle control system 1 according to this embodiment will be described. The vehicle control system 1 comprises an imaging device 10 and a vehicle control device 20. The imaging device 10 and the control device 20 are connected to each other so as to be able to communicate with each other via a network 30, which includes, for example, the Internet and a mobile communication network.

[0011] The imaging device 10 is at least one in-vehicle camera. The imaging device 10 captures images of the seats and objects near the seats inside the vehicle while the vehicle is temporarily stopped.

[0012] The vehicle is any vehicle capable of carrying one or more passengers, such as a car, bus, or shuttle bus. The entrance and seating may be located in different positions. In this embodiment, the vehicle is an autonomous vehicle capable of autonomous driving at a level 1 to 5 as defined by the Society of Automotive Engineers (SAE). The vehicle may also be a manually driven vehicle at level 0. The vehicle may be remotely monitored by an observer outside the vehicle. The vehicle may also be a vehicle dedicated to MaaS. "MaaS" is an abbreviation for Mobility as a Service.

[0013] The vehicle control device 20 is an electronic device mounted in the vehicle, such as a computer. The control device 20 detects an object from an image captured by the imaging device 10 using any object detection technology. The image may be a still image or a moving image. The object is a person inside the vehicle. The control device 20 uses any posture estimation technology, such as machine learning, to derive a posture likelihood from the image that indicates the likelihood that the object's posture is a specific posture. The control device 20 uses multiple posture likelihoods to estimate the object's posture and, based on the estimation result, decides whether to allow the vehicle to start.

[0014] First, an overview of this embodiment will be described, and details will be described later. The vehicle control device according to this embodiment includes a control unit. The control unit derives a reference posture likelihood, which indicates the likelihood that the posture of the object at the first time is the reference posture, and a nearest posture likelihood, which indicates the likelihood that the posture of the object is the nearest posture, from an image of the object at the first time. It also calculates a likelihood ratio, which is the ratio of the reference posture likelihood to the nearest posture likelihood at the first time. The nearest posture is the posture that the object may take before or after transitioning to the reference posture. If the reference posture likelihood at the first time is greater than or equal to a threshold, and the likelihood ratio at the first time is greater than or equal to a first predetermined value, the control unit turns on a flag indicating that the posture of the object is the reference posture. If the reference posture likelihood at the first time is less than a threshold, or if the likelihood ratio at the first time is less than a first predetermined value, the control unit turns off the flag.

[0015] Due to differences in sitting posture and other factors, it may be easier for some people to determine their most recent posture than their standard sitting posture. In this embodiment, even when it is difficult to determine seating using only the likelihood of sitting, i.e., when the likelihood of sitting is below a predetermined threshold, the seating posture can be accurately determined by comparing the likelihood of sitting with the likelihood of the most recent posture. This reduces the time required for seating determination and lowers passenger stress.

[0016] Reducing the time required for seating determination is beneficial for driverless autonomous vehicles. Such vehicles are remotely monitored, for example, by a monitor at a base station. If the vehicle does not immediately depart, the passengers may assume there is a problem and contact the monitor to ask for assistance. An increase in the frequency of communication between passengers and monitors can lead to an increased burden on the monitor. This burden on the monitor can increase even further if one monitor is responsible for multiple autonomous vehicles. The control device according to this embodiment can reduce the time required for seating determination, thereby reducing the burden on such monitors.

[0017] When using multiple cameras or other sensors to improve the accuracy of seating determination, the cost of seating determination increases. The control device according to the present embodiment can reduce the number of necessary cameras or other sensors and reduce the cost of seating determination by utilizing the likelihoods of the reference posture and the most recent posture derived from one image.

[0018] Therefore, according to the present embodiment, the technology related to posture determination based on the likelihood of a person's posture is improved.

[0019] Next, each component of the vehicle control system 1 will be described.

[0020] (Configuration of the imaging device 10) The imaging device 10 is an arbitrary imaging module installed inside the vehicle and capable of imaging all seats and targets inside the vehicle. The imaging module includes one or more cameras. In the present embodiment, the imaging device 10 is one camera installed near the ceiling at the entrance of the vehicle. The imaging device 10 may be two cameras installed at other positions, for example, at the center and rear of the ceiling of the vehicle. The imaging device 10 may be a 180-degree camera or a 360-degree camera. The imaging device 10 transmits the captured image to the control device 20 via the network 30.

[0021] (Vehicle control device 20) The vehicle control device 20 includes a control unit 200, a communication unit 201, and a storage unit 202.

[0022] The control unit 200 includes one or more processors, one or more programmable circuits, one or more dedicated circuits, or a combination thereof. The control unit 200 controls the operation of the control device 20.

[0023] The communication unit 201 includes at least one communication interface connected to the network 30. The communication interface supports, for example, a mobile communication standard such as 4G or 5G, a V2X communication standard such as DSRC or cellular V2X, or a wireless LAN communication standard such as IEEE 802.11. "4G" is an abbreviation for 4th generation. "5G" is an abbreviation for 5th generation. "DSRC" is an abbreviation for dedicated short range communications. "V2X" is an abbreviation for vehicle-to-everything. "IEEE" is an abbreviation for Institute of Electrical and Electronics Engineers.

[0024] The storage unit 202 includes one or more memories. Each memory included in the storage unit 202 may function as, for example, a main memory, an auxiliary memory, or a cache memory. The storage unit 202 stores any information used in the operation of the control device 20. For example, the storage unit 202 stores system programs, application programs, embedded software, and any data used for target detection and posture estimation. The storage unit 202 may pre-store information on the position and shape of each seat, or postures used for seating estimation, such as a reference posture and a model of the most recent posture. These models may be generated by machine learning. The information stored in the storage unit 202 may be updated by the control unit 200. In this embodiment, a flag indicating that the target posture is a seated posture is stored in the storage unit 202 and updated by the control unit 200.

[0025] (Posture Likelihood) Next, we will explain posture likelihood. Posture likelihood indicates the likelihood that the subject's posture is a specific posture. The control unit 200 of the control device 20 derives a reference posture likelihood, which indicates the likelihood that the subject's posture is a reference posture, and a recent posture likelihood, which indicates the likelihood that the subject's posture is a recent posture. The recent posture is a posture that the subject may take before or after transitioning to the reference posture. Examples of recent postures in seating determination include a crouched posture, a standing posture, a walking posture, and a squatting posture. In this embodiment, the reference posture is the seated posture, and the recent posture is the crouched posture. The seated posture is the posture when the subject is seated in a seat. The crouched posture is the bent posture immediately before the subject sits down in a seat. The standing posture is the posture while standing. The walking posture is the posture while walking. The squatting posture is the posture when the subject is squatting on the floor or on a seat.

[0026] The typical temporal changes in posture and posture likelihood before and after the subject transitions to a seated posture will be explained with reference to Figure 2. The upper part of Figure 2 shows the change in the subject's posture. In Figure 2, the subject's posture transitions from the immediate posture to a seated posture, and then returns to the immediate posture. The change in the subject's posture is classified into states A to E. In state A, the subject's posture is the immediate posture, transitioning from a walking posture to a standing posture and then to a crouching posture. In state B, the subject's posture gradually transitions from the immediate posture (crouching posture) to a seated posture. In state C, the subject's posture is a seated posture. In state D, the subject's posture gradually transitions from a seated posture to the immediate posture. In state E, the subject's posture is the immediate posture, transitioning from a crouching posture to a standing posture and then to a walking posture.

[0027] The lower part of Figure 2 shows the changes in the seated likelihood X and the crouched likelihood Y. In state A, the seated likelihood X is less than the second threshold X2, and the crouched likelihood Y is higher than the seated likelihood X. In state B, the seated likelihood X is greater than or equal to the second threshold X2 and less than the first threshold X1. As posture changes, the seated likelihood X increases while the crouched likelihood Y decreases, and the seated likelihood X becomes higher than the crouched likelihood Y. In state C, the seated likelihood X is greater than or equal to the first threshold X1. The seated likelihood X reaches its highest value, and the crouched likelihood Y reaches its minimum value. In state D, the seated likelihood X is greater than or equal to the third threshold X3 and less than the first threshold X1. As posture changes, the seated likelihood X decreases while the crouched likelihood Y increases, and the crouched likelihood Y becomes higher than the seated likelihood X. In state E, the seated likelihood X is less than the third threshold X3, and the crouched likelihood Y is higher than the seated likelihood X. The first threshold X1 is set higher than the second threshold X2 and the third threshold X3. In Figure 2, the second threshold X2 is set higher than the third threshold X3, but it may also be set to be the same as or less than the third threshold X3.

[0028] Although not shown in Figure 2, in state A, the walking posture, standing posture, and crouching posture reach their maximum values ​​in the order of walking posture, standing posture, and crouching posture, while in state E, they reach their maximum values ​​in the reverse order.

[0029] In states B and D, the boundary between a seated position and a crouched position is ambiguous, making it difficult to perform accurate seating determination using only the seating likelihood. The vehicle control device 20 according to this embodiment can accurately perform seating determination even in states B and D by utilizing both the seating likelihood and the crouching likelihood.

[0030] The posture of the subject may change in ways different from those shown in Figure 2. For example, if the subject is a child, their posture may transition from a squatting position to a half-crouching position, then to a sitting position, and then back to a squatting position in the reverse order. The likelihood of these postures reaches its maximum in the order of squatting, half-crouching, sitting, half-crouching, and squatting.

[0031] (Operation flow of the control unit 200) Referring to Figure 3, the operation of the control unit 200 in the vehicle control device 20 according to this embodiment will be described. The control unit 200 executes the following S101 to S106, for example, while the vehicle is temporarily stopped at a bus stop. In S101 to S106, the control unit 200 determines whether the target posture has transitioned to a seated posture. The flag indicating that the target posture is a seated posture is initially set to OFF.

[0032] S101: The control unit 200 detects the object from the image at the first time step.

[0033] The first time point is the time when the imaging device 10 captures images of the seats and objects near the seats while the vehicle is stopped. The control unit 200 receives the images captured at the first time point from the imaging device 10 via the communication unit 201. The control unit 200 detects objects from the images at the first time point using any object detection technique.

[0034] S102: The control unit 200 derives the seated likelihood and crouching likelihood at the first time from the image of the target at the first time, and also calculates the likelihood ratio at the first time.

[0035] The likelihood ratio is the ratio of the reference posture likelihood to the most recent posture likelihood, and is expressed in the form of the equation R = X / Y. Here, R is the likelihood ratio, X is the reference posture likelihood, and Y is the most recent posture likelihood. In this embodiment, the reference posture likelihood is the seated likelihood, and the most recent posture likelihood is the crouched likelihood.

[0036] S103: The control unit 200 determines whether the seating likelihood is greater than or equal to the first threshold X1. If it is determined that the seating likelihood is greater than or equal to the first threshold X1 (S103-YES), the process proceeds to S105. Otherwise (S103-NO), the process proceeds to S104.

[0037] S104: The control unit 200 determines whether the seating likelihood at the first time step is greater than or equal to the second threshold X2, and whether the likelihood ratio at the first time step is greater than or equal to the first predetermined value R1. If it is determined that the seating likelihood at the first time step is greater than or equal to the second threshold X2, and whether the likelihood ratio at the first time step is greater than or equal to the first predetermined value R1 (S104-YES), the process proceeds to S105. Otherwise (S104-NO), the process proceeds to S106.

[0038] Steps S103 to S104 correspond to the seating determination in states A to C in Figure 2. The control unit 200 does not have to perform the determination in S103. In S104, the control unit 200 does not have to perform the determination of whether the seating likelihood at the first time step is greater than or equal to the second threshold X2.

[0039] S105: The control unit 200 turns the flag ON. If the flag is already ON, the control unit 200 does nothing.

[0040] S106: The control unit 200 turns the flag OFF. The process then returns to S102.

[0041] The control unit 200 repeats steps S102 to S106 until the flag is turned ON.

[0042] The control unit 200 executes processes S101 to S106 for each object in the vehicle. If the flag is ON for all objects in the vehicle, the control unit 200 permits the vehicle to start. If the flag is OFF for one or more objects, the control unit 200 does not permit the vehicle to start.

[0043] After the flag is turned ON in S106 and the vehicle starts moving, the control unit 200 executes the following S107 to S109 while the vehicle is stopped again, for example, at a pedestrian crossing or traffic light. In S107 to S109, the control unit 200 determines whether the target's posture has changed from a seated posture.

[0044] S107: The control unit 200 derives the seated likelihood and crouched likelihood at the second time from the image at the second time, and also calculates the likelihood ratio at the second time.

[0045] The second time point is the time after the flag is turned ON in S105 and the vehicle starts moving. In this embodiment, the second time point is the time when the imaging device 10 captures images of the seats and objects near the seats while the vehicle is stopped again after the flag is turned ON in S105 and the vehicle starts moving. The control unit 200 receives the images captured at the second time point from the imaging device 10 via the communication unit 201.

[0046] S108: The control unit 200 determines whether the seating likelihood at the second time step is greater than or equal to the first threshold X1. If it is determined that the seating likelihood at the second time step is greater than or equal to the first threshold X1 (S108-YES), the process returns to S105. If it is determined that the seating likelihood is less than the first threshold X1 (S108-NO), the process proceeds to S109.

[0047] S109: The control unit 200 determines whether the seating likelihood at the second time step is greater than or equal to the third threshold X3, and whether the likelihood ratio at the second time step is greater than or equal to the second predetermined value R2. If it is determined that the seating likelihood at the second time step is greater than or equal to the third threshold X3, and whether the likelihood ratio at the second time step is greater than or equal to the second predetermined value R2 (S109-YES), the process returns to S105. Otherwise (S109-NO), the process returns to S106.

[0048] Steps S108 to S109 correspond to the seating determination in states C to E in Figure 2. The control unit 200 does not have to perform the determination in S108. In S109, the control unit 200 does not have to perform the determination of whether the seating likelihood at the second time step is greater than or equal to the third threshold X3. If the process returns from S108 or S109 to S106, the control unit 200 executes steps S102 to S106 again.

[0049] The control unit 200 executes S107 to S109 for each object in the vehicle. If the flag is ON for all objects, the control unit 200 permits the vehicle to start. If the flag is OFF for one or more objects, the control unit 200 does not permit the vehicle to start. The control unit 200 executes S107 to S109 each time the vehicle stops.

[0050] In another embodiment, in S102, S104, S107, and S109, the most recent posture may be a standing posture, a walking posture, or a squatting posture. In another embodiment, in S102 and S107, the control unit 200 may derive the likelihoods of a crouching posture, a standing posture, a walking posture, and a squatting posture, and calculate each likelihood ratio based on each posture likelihood. In S104, the control unit 200 may turn on a flag if the seating likelihood at the first time is greater than or equal to the second threshold X2, and at least one of these likelihood ratios at the first time is greater than or equal to the first predetermined value R1; otherwise, the flag may be turned off. In S108, the control unit 200 may turn on a flag if the seating likelihood at the second time is greater than or equal to the second threshold X2, and at least one of these likelihood ratios at the second time is greater than or equal to the second predetermined value R2; otherwise, the flag may be turned off. If the subject's behavior is rapid, the determination of a crouching posture may fail to derive the likelihood of crouching. In this alternative embodiment, even if the derivation of a particular immediate posture likelihood fails, other immediate likelihoods can be derived to perform the seating determination.

[0051] The second predetermined value R2 may be different from the first predetermined value R1. When the time for seating determination is limited, for example, when the vehicle is stopped near a pedestrian crossing or traffic light, it is necessary to reduce the time for seating determination. For this reason, the second predetermined value R2 may be lower than the first predetermined value R1.

[0052] The first predetermined value R1 and the second predetermined value R2 may be set differently for each posture in the most recent posture. As shown in the upper part of Figure 2, when the subject sits down, the subject's posture transitions from walking posture to standing posture and then to a seated posture in that order, and the subject's safety increases in this order. Therefore, in order to increase the subject's safety, the first predetermined value R1 may be set to increase in the order of likelihood of being in a seated posture, standing likelihood, and walking likelihood.

[0053] The first predetermined value R1 and the second predetermined value R2 may be set differently depending on the orientation of the subject (person) relative to the camera. It is more difficult to grasp the characteristic points of a person's posture from the front than from the side. For example, a person stands upright in a standing posture, while in a walking posture, the legs move forward relative to the torso. It is more difficult to grasp the difference between a standing posture and a walking posture when viewing a person from the front than when viewing a person from the side. Therefore, the first predetermined value R1 and the second predetermined value R2 when the subject is facing forward relative to the camera may be set to be less than or equal to the first predetermined value R1 and the second predetermined value R2 when the subject is facing sideways relative to the camera. "Front direction" refers to the direction from the subject toward the camera and the direction tilted from that direction in a range of greater than 0 degrees and less than 30 degrees. "Side direction" refers to the direction tilted from the subject toward the camera in a range of 30 degrees or more and 150 degrees or less.

[0054] Table 1 shows examples of the first predetermined value R1 and the second predetermined value R2 when the object is facing forward to the camera. Table 2 shows examples of the first predetermined value R1 and the second predetermined value R2 when the object is facing sideways to the camera. [Table 1] [Table 2]

[0055] To perform seating determination in S104 and S109 more quickly, the likelihood ratio may be defined as follows: R = (X + α) / Y. Here, R is the likelihood ratio, X is the reference posture likelihood, Y is the most recent posture likelihood, and α is any constant greater than or equal to 0. As α increases, the time it takes for the likelihood ratio to exceed the first predetermined value R1 or the second predetermined value R2 decreases. Table 3 shows examples of α when the object is facing forward or sideways relative to the camera. [Table 3] To enhance the safety of the subject, α may be set to increase in the order of crouching likelihood, standing likelihood, and walking likelihood, as shown in Table 3.

[0056] As described above, the vehicle control device according to this embodiment includes a control unit. The control unit derives a reference posture likelihood, which indicates the likelihood that the posture of the object at the first time is the reference posture, and a nearest posture likelihood, which indicates the likelihood that the posture of the object is the nearest posture, from an image of the object at the first time. It also calculates a likelihood ratio, which is the ratio of the reference posture likelihood to the nearest posture likelihood at the first time. The nearest posture is a posture that the object may take before or after transitioning to the reference posture. If the reference posture likelihood at the first time is greater than or equal to a threshold, and the likelihood ratio at the first time is greater than or equal to a first predetermined value R1, the control unit turns on a flag indicating that the posture of the object is the reference posture. If the reference posture likelihood at the first time is less than a threshold, or if the likelihood ratio at the first time is less than a first predetermined value R1, the control unit turns off the flag.

[0057] Due to differences in sitting posture and other factors, it may be easier for some people to determine their current posture than their standard sitting posture. With this configuration, even in situations where it is difficult to determine seating using only the likelihood of sitting, i.e., when the likelihood of sitting is below a predetermined threshold, the seating posture can be accurately determined by comparing the likelihood of sitting with the likelihood of the current posture. This reduces the time required for seating determination and lowers passenger stress.

[0058] Reducing the time required for seating determination is beneficial for driverless autonomous vehicles. Such vehicles are remotely monitored, for example, by a monitor at a base station. If the vehicle does not start immediately, the passengers may assume that a problem has occurred and contact the monitor to ask for help resolving the issue. An increase in the frequency of communication between passengers and monitors can lead to an increased burden on the monitor. If one monitor is responsible for monitoring multiple autonomous vehicles, this burden on the monitor may increase even further. The control device according to this embodiment can reduce such a burden on the monitor by reducing the time required for seating determination.

[0059] Using multiple cameras or other sensors to improve the accuracy of seating detection increases the cost of seating detection. The control device according to this embodiment can reduce the number of cameras or other sensors required and thus reduce the cost of seating detection by utilizing the likelihood of a reference posture and a nearest posture derived from a single image.

[0060] While this disclosure has been described based on the drawings and embodiments, those skilled in the art will understand that various modifications and alterations may be made based on this disclosure. Therefore, it should be noted that these modifications and alterations are within the scope of this disclosure.

[0061] The multiple values ​​and their relative magnitudes described in the above-described embodiment may be changed as appropriate. In this embodiment, seating determination is performed in S104 and S109 based on the likelihood ratio. In another embodiment, seating determination may be performed based on the likelihood difference, which is the difference between the reference posture likelihood and the immediate prior posture likelihood.

[0062] The functions and other elements included in each component or step can be rearranged in a logically consistent manner, and multiple components or steps can be combined into one or separated. For example, in the above-described embodiment, it is also possible to distribute the configuration and operation of the control device 20 to multiple computers that can communicate with each other. Furthermore, in the above-described embodiment, it is also possible to provide part or all of the imaging device 10 and the control device 20 in the same device. [Explanation of Symbols]

[0063] 1 System 10 Imaging device 20 Vehicle control system 200 Control Unit 201 Communications Department 202 Storage section 30 Networks

Claims

1. A vehicle control device comprising a control unit, The control unit, From the image of the object at the first time step, the likelihood of the object's posture being the reference posture at the first time step is derived, and the likelihood of the object's posture being the most recent posture is derived. Furthermore, the likelihood ratio, which is the ratio of the likelihood of the reference posture to the likelihood of the most recent posture at the first time step, is calculated. The most recent posture is a posture that the object may take before or after transitioning to the reference posture. If the likelihood of the reference posture at the first time is greater than or equal to a threshold, and the likelihood ratio at the first time is greater than or equal to a first predetermined value, the flag indicating that the posture of the object is the reference posture is turned ON. A vehicle control device that turns off the flag if the likelihood of the reference posture at the first time is less than the threshold, or if the likelihood ratio at the first time is less than the first predetermined value.

2. A vehicle control device according to claim 1, wherein the control unit is From the image of the target at the second time point after the flag is turned ON, the reference pose likelihood and the most recent pose likelihood at the second time point are derived, and the likelihood ratio at the second time point is calculated. A vehicle control device that turns off the flag if the likelihood of the reference posture at the second time is less than a threshold, or if the likelihood ratio at the second time is less than a second predetermined value.

3. A vehicle control device according to claim 2, wherein the second predetermined value is lower than the first predetermined value.

4. A vehicle control device according to claim 1, wherein the reference posture is a seated posture, and the immediate posture is a crouching posture, a standing posture, a walking posture, or a squatting posture.

5. A vehicle control device according to claim 1, wherein the control unit is If the aforementioned flag is ON, the vehicle is permitted to start. A vehicle control device that, when the aforementioned flag is OFF, does not permit the vehicle to start.

6. A vehicle control device according to claims 1 to 5, wherein the vehicle is an autonomous vehicle.

7. A method for providing MaaS (Mobility as a Service) using the vehicle control device described in claim 1.

Images