Remote support system, remote support method

Abstract

To provide a remote support system for selecting and assigning an operator to handle a remote support request from a plurality of operators to an automatic vehicle that issues a remote support request, and to provide technology that can reduce the burden on each of the plurality of operators.SOLUTION: A remote support system according to the present disclosure includes: a storage device for storing a database managing feature values for each of a plurality of feature elements for a previous support scene for which a previously processed remote support request was made; and a processor for identifying one or more indicator elements for a current support scene for which a new remote support request is to be made, identifying one or more candidate operators whose previous support scene is similar to the current support scene based on the feature values of the one or more indicator elements, and performing a process of selecting an operator to process the new remote support request from the one or more candidate operators.SELECTED DRAWING: Figure 4

Description

Technical Field

[0001] The present disclosure relates to vehicle remote support technology.

Background Art

[0002] In recent years, a remote support system has been considered that selects and assigns an operator who processes remote support requests from a plurality of operators to an autonomous vehicle that issues a remote support request.

[0003] For example, Patent Document 1 discloses a remote support instruction system including a remote instruction point status recognition unit that recognizes the status of remote instruction points on a target route, a time prediction unit that predicts the start time and end time of monitoring of a remote commander (operator) for the status of remote instruction points on the target route, and a monitoring time allocation unit that allocates, to a plurality of remote commanders (operators), a monitoring time that is the time between the start time and the end time of monitoring based on the start time and end time of monitoring of the status of remote instruction points in a plurality of autonomous vehicles.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] It is assumed that remote support requests are issued from various autonomous vehicles in various scenes. Therefore, if an operator is randomly selected and assigned each time in response to a remote support request, the operator will be required to continuously recognize various traffic environments and respond to various tasks. Consequently, there is a risk that the operator will feel a very heavy burden compared to the case of remotely supporting a single autonomous vehicle on a dedicated basis.

[0006] One object of the present disclosure is to provide a technology capable of reducing the burden on an operator when selecting and assigning an operator who processes remote support requests from a plurality of operators to an autonomous vehicle that issues a remote support request, in view of the above problems.

Means for Solving the Problems

[0007] A first aspect of the present disclosure relates to a remote support system that provides a remote support function for a vehicle by a plurality of operators.

[0008] The remote support system according to the first aspect includes a storage device that stores a database for managing feature amounts for each of a plurality of feature elements regarding a previous support scene that was the target of a remote support request processed last time, for each of a plurality of operators, and one or more processors. The one or more processors receive a new remote support request from a vehicle, perform a process of specifying one or more index elements to be used as an index for similarity determination for the current support scene that is the target of the new remote support request, perform a process of specifying one or more candidate operators among the plurality of operators whose previous support scene is similar to the current support scene based on the feature amounts of the specified one or more index elements, and perform a process of selecting an operator who processes the new remote support request from the specified one or more candidate operators, and is characterized in being configured to execute the processes.

[0009] A second aspect of the present disclosure relates to a remote support method for providing, by a computer, a remote support function for a vehicle by a plurality of operators.

[0010] The remote support method according to the second aspect includes, for each of a plurality of operators, managing feature amounts for each of a plurality of feature elements regarding the previous support scene that was the target of the previously processed remote support request, receiving a new remote support request from a vehicle, identifying one or more index elements to be used as an index for similarity determination for the current support scene that is the target of the new remote support request, identifying one or more candidate operators among the plurality of operators whose previous support scene is similar to the current support scene based on the feature amounts of the identified one or more index elements, and selecting an operator from the identified one or more candidate operators to process the new remote support request.

Advantages of the Invention

[0011] According to the present disclosure, based on the feature amounts of the identified one or more index elements, one or more candidate operators among the plurality of operators whose previous support scene is similar to the current support scene that is the target of the new remote support request are identified. The one or more candidate operators thus identified are expected to have a smaller burden than other operators when processing a new remote support request. Then, an operator is selected from the one or more candidate operators to process the new remote support request. Thereby, the burden on each of the plurality of operators can be reduced.

Brief Description of the Drawings

[0012]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Mode for Carrying Out the Invention

[0013] Hereinafter, the present embodiment will be described with reference to the drawings.

[0014] 1. Configuration FIG. 1 is a block diagram showing the configuration of a remote support system 10 according to the present embodiment. The remote support system 10 according to the present embodiment provides a remote support function for a vehicle 100 by a plurality of operators 320. More specifically, the remote support system 10 according to the present embodiment selects and assigns an operator 320 who processes a remote support request from a plurality of operators 320 to a vehicle 100 that issues a remote support request. Then, the assigned operator 320 (hereinafter, also referred to as "assigned operator") performs remote support for the vehicle 100 that issues a remote support request by operating the remote support terminal 310.

[0015] In the present embodiment, the vehicle 100 is an autonomous vehicle. That is, the vehicle 100 performs recognition of the surrounding environment, driving judgment according to the recognition result, and autonomous driving according to the driving judgment by the autonomous driving function. And the vehicle 100 issues a remote support request when it cannot or has difficulty making a driving judgment in the autonomous driving function. Hereinafter, the vehicle 100 is referred to as an autonomous vehicle 100.

[0016] Figure 2(A) shows an example of the types and details of remote support requests issued by the autonomous vehicle 100. For example, when the autonomous vehicle 100 recognizes a pedestrian near a crosswalk in the vicinity of a crosswalk, and there is not enough certainty as to whether the pedestrian will cross the crosswalk, the autonomous vehicle 100 issues a remote support request for "permission to cross the crosswalk". Also, for example, when the autonomous vehicle 100 attempts to open the door to drop off a passenger, and it cannot determine with sufficient certainty the timing at which the opening of the door will not interfere with the surroundings, the autonomous vehicle 100 issues a remote support request for "permission to open / close the door". Note that the autonomous vehicle 100 may issue multiple types of remote support requests simultaneously. For example, the autonomous vehicle 100 may simultaneously issue remote support requests for "permission to open / close the door" and "attention to opening / closing the door".

[0017] Referring again to Figure 1. The autonomous vehicle 100 includes a sensor 101, a storage device 102, a control device 103, and a communication device 104. The control device 103 is connected so as to be able to transmit information to and from the sensor 101, the storage device 102, and the communication device 104 (thick lines in Figure 1). For example, the control device 103 is connected to these devices via a vehicle-mounted network configured by CAN (Control Area Network) or the like.

[0018] The sensor 101 is a sensor for recognizing the surrounding environment and road structure. Examples of the sensor 101 include LiDAR (Light Detection And Ranging), radar, cameras, etc. The recognition information of the sensor 101 is transmitted to the control device 103.

[0019] The storage device 102 stores data related to the autonomous driving function. The storage device 102 is composed of a recording medium such as an HDD or an SSD, for example. In particular, the storage device 102 stores map information 105. The map information 105 is typically information indicating the positions on the map of roads, structures, etc. The control device 103 can acquire the map information 105 by accessing the storage device 102.

[0020] The communication device 104 communicates with devices external to the autonomous vehicle 100 to transmit and receive information. In particular, the communication device 104 includes a device that communicates with the remote support server 200 and the remote support terminal 310. For example, the communication device 104 communicates with the remote support server 200 and the remote support terminal 310 via a mobile communication network and the Internet. Note that the communication device 104 may be configured to start communicating with the remote support terminal 310 corresponding to the assigned operator upon receiving that the assigned operator has been determined. The information received by the communication device 104 is transmitted to the control device 103.

[0021] Based on the recognition information of the sensor 101 and the map information 105, the control device 103 executes processing related to the autonomous driving function. That is, by the control device 103 executing the processing, the autonomous driving of the autonomous vehicle 100 is realized. The control device 103 is composed of, for example, one or more in-vehicle ECUs (Electronic Control Units). In particular, in the autonomous driving function, the control device 103 executes processing to generate a remote support request when it is unable or difficult to make a driving judgment. Then, the control device 103 transmits the generated remote support request to the remote support server 200 via the communication device 104. After that, the control device 103 receives the judgment of the assigned operator for the remote support request via the communication device 104, and autonomously drives the autonomous vehicle 100 according to the judgment of the assigned operator.

[0022] Here, when transmitting the remote support request to the remote support server 200, the control device 103 is configured to further transmit information (scene information) representing the scene targeted by the remote support request to the remote support server 200. Examples of the scene information include point cloud data detected by LiDAR, image data captured by a camera, the position of the autonomous vehicle 100 on the map, the map information 105 around the autonomous vehicle 100, the attributes of the autonomous vehicle 100 (vehicle specifications, body type, etc.), information regarding passengers (number, attributes, position, status, etc.). These scene information can be obtained as the processing results of the control device 103 and the recognition information of the sensor 101.

[0023] In the remote support system 10 according to the present embodiment, the remote support function is realized by the remote support server 200 executing processing. The remote support server 200 is typically a computer accessible via the Internet. In this case, the remote support server 200 may be a cloud server or a dedicated server. In particular, the remote support server 200 is configured to be able to communicate with the autonomous vehicle 100 and the remote support terminal 310. The configuration of the remote support server 200 will be described later.

[0024] The processing executed by the remote support server 200 is composed of an assignable operator identification processing unit P202, a scene feature calculation processing unit P203, an index element identification processing unit P204, a similarity determination processing unit P205, an assigned operator determination processing unit P206, and a remote support execution processing unit P207.

[0025] The assignable operator identification processing unit P202 receives a new remote support request from the autonomous vehicle 100 and identifies the assignable operator 320 among the plurality of operators 320. For example, the assignable operator identification processing unit P202 identifies one or more operators 320 who are not currently providing remote support or who are providing remote support but whose interruption by other remote support is permitted, based on the remote support status of the plurality of operators 320. If there are no assignable operators 320, the assignable operator identification processing unit P202 may be configured to wait for processing until there is one or more assignable operators. Alternatively, the assignable operator identification processing unit P202 may be configured to reject the remote support request from the autonomous vehicle 100.

[0026] Based on the scene information received from the autonomous vehicle 100, the scene feature calculation processing unit P203 calculates the features of the scene (scene features) that are the target of the remote support request received this time. Here, the scene features are represented by combinations of feature amounts for each of a plurality of feature elements. An example of the feature elements and the possible feature amounts for each feature element is shown in (B) of FIG. 2. For example, the scene feature calculation processing unit P203 calculates the feature amount of "operation area" as area A from the position of the autonomous vehicle 100 and the map information 105 acquired as scene information.

[0027] Table 1 shows an example of the scene features calculated by the scene feature calculation processing unit P203. The scene features shown in Table 1 are, for example, the case where the autonomous vehicle 100, which is a medium-sized bus, issues a remote support request for "permission to start from the bus stop" near bus stop B in area A.

[0028]

Table 1

[0029] Note that the remote support server 200 may be configured to acquire the scene features as scene information. In this case, the scene feature calculation processing unit P203 is realized in the autonomous vehicle 100. For example, it is realized by the control device 103 of the autonomous vehicle 100. Also, one or a plurality of feature elements may be determined in advance according to the environment to which the remote support system 10 according to the present embodiment is applied, or may be determined according to the type of the remote support request received this time.

[0030] Also, the mode of the possible feature amounts shown in (B) of FIG. 2 is an example, and other modes may be adopted. For example, the feature amounts of each feature element shown in (B) of FIG. 2 can also adopt other modes as follows.

[0031] The feature quantity of "confirmation target (gaze direction)" can also be expressed by a binary number representing the combination of cameras that present images to the operator 320. The feature quantity of "operation area" can also be expressed by the maximum and minimum values of latitude and longitude. The feature quantity of "location" can also be expressed by latitude and longitude (GPS position) or relative position on the map (localized position). The feature quantity of "road shape" can also be expressed by numerical values representing curvature, road width, number of lanes, number of branches, etc. The feature quantity of "vehicle shape" can also be expressed by numerical values representing overall length, overall width, vehicle type, etc. The feature quantity of "surrounding targets" can also be expressed by a list of vector information representing the type, relative position, relative speed, etc. of each target. The feature quantity of "passengers" can also be expressed by a list of vector information representing each piece of information for each passenger recognized in the vehicle interior.

[0032] Refer to FIG. 1 again. The index element specifying processing unit P204 specifies one or more index elements that serve as indices for similarity determination for the scene targeted by the remotely supported request received this time among a plurality of feature elements. For example, the index element specifying processing unit P204 specifies one or more index elements according to the type of remotely supported request received this time. In this case, the index element specifying processing unit P204 can specify one or more index elements by referring to a list that defines index elements for each type of remotely supported request.

[0033] FIG. 3(A) shows an example of a list that defines index elements for each type of remotely supported request. In the list shown in FIG. 3(A), for each type of remotely supported request, the feature elements marked with check marks are defined as index elements. For example, when the type of remotely supported request received this time is "permission to start at an intersection", the index element specifying processing unit P204 specifies "confirmation target", "operation area", "location", "road shape", and "surrounding targets" as index elements.

[0034] The index element specifying processing unit P204 may be configured to set priorities for the further specified one or more index elements. For example, the index element specifying processing unit P204 can specify one or more index elements and set priorities by referring to a list that defines index elements and priorities for each type of remote support request.

[0035] An example of a list defining index elements and priorities for each type of remote support request is shown in FIG. 3(B). In the list shown in FIG. 3(B), for each type of remote support request, a characteristic element with a numerical value is defined as an index element, and the priority of the index element is defined by the numerical value. For example, when the type of the remote support request received this time is "Passenger seating attention arousal", the index element specifying processing unit P204 specifies "Confirmation target", "Vehicle shape", and "Passenger" as index elements and sets the priorities to 1, 2, and 3 in order.

[0036] Note that the lists as shown in FIGS. 3(A) and 3(B) may be provided in advance according to the environment to which the remote support system 10 is applied. Here, the index elements defined for each type of remote support request may be characteristic elements among a plurality of characteristic elements that are related to the burden on the operator 320 when processing the remote support request of that type. Also, the priorities may be in the order of the magnitude of the influence on the burden on the operator 320 when processing the remote support request of that type.

[0037] Referring to FIG. 1 again. The similarity determination processing unit P205 performs a similarity determination between the scene to which the remote support request processed by the operator 320 last time (hereinafter referred to as the "previous support scene") and the scene to which the remote support request received this time (hereinafter referred to as the "current support scene") for each of the operators 320 to whom assignment is possible. Then, the similarity determination processing unit P205 outputs, as a processing result, the operator 320 (hereinafter referred to as the "candidate operator") among the operators 320 to whom assignment is possible whose previous support scene is similar to the current support scene.

[0038] The similarity determination processing unit P205 performs a similarity determination by comparing the scene features of the previous support scene and the scene features of the current support scene. Here, the scene features of the previous support scene for each of the plurality of operators 320 are managed as a scene feature database 215. That is, the similarity determination processing unit P205 can obtain the scene features of the previous support scene for each of the operators 320 to which assignment is possible by referring to the scene feature database 215.

[0039] In particular, the similarity determination processing unit P205 performs a similarity determination based on the feature amounts of the specified one or more index elements. Hereinafter, an example of the processing of the similarity determination processing unit P205 will be described with reference to FIGS. 4(A) and 4(B). FIGS. 4(A) and 4(B) show examples of the scene features of the current support scene and the scene features of the previous support scene of the assignable operator 320, respectively. In FIGS. 4(A) and 4(B), the assignable operators 320 are three people, 320a, 320b, and 320c. Also, in FIGS. 4(A) and 4(B), priorities are set for the specified one or more index elements.

[0040] First, refer to (A) of FIG. 4. The similarity determination processing unit P205 performs similarity determination of the feature amounts for the current support scene in the order of the index elements according to the priority. Therefore, the similarity determination processing unit P205 first performs similarity determination of the feature amounts of "confirmation target" with a priority of 1. Here, the similarity determination processing unit P205 can determine whether they are similar based on the coincidence of the feature amounts. Thus, based on the similarity determination of the feature amounts of "confirmation target", the operator 320b can determine that the previous support scene is not similar to the current support scene. Next, the similarity determination processing unit P205 performs similarity determination of the feature amounts of "operation area" with a priority of 2. At this time, the similarity determination processing unit P205 may exclude the operator 320b from the target of similarity determination. Based on the similarity determination of the feature amounts of "operation area", the operator 320c can determine that the previous support scene is not similar to the current support scene. At this point, among the assignable operators 320, the operator 320 (hereinafter referred to as "operator to be determined") for which it has not been determined that the previous support scene is not similar to the current support scene is only the operator 320a. Therefore, the similarity determination processing unit P205 sets the operator 320a as the candidate operator.

[0041] Next, refer to (B) of FIG. 4. Assuming that similarity determination is performed in the same manner as in the case described in (A) of FIG. 4, in the example shown in (B) of FIG. 4, it can be seen that the similarity determination processing unit P205 sets the operator 320c as the candidate operator.

[0042] In addition, when similarity determination of the feature amounts is performed for all of the specified one or more index elements, if a plurality of operators to be determined remain, the similarity determination processing unit P205 may set the remaining plurality of operators to be determined as candidate operators. Also, when there are no operators to be determined due to similarity determination of a certain feature amount, the similarity determination processing unit P205 may be configured to skip the similarity determination for that feature amount. Thereby, a situation where there are no candidate operators can be avoided.

[0043] In this way, the similarity determination processing unit P205 performs a similarity determination based on the feature amounts of the specified one or more index elements. On the other hand, the similarity determination processing unit P205 does not perform a similarity determination for feature elements that are not index elements. As a result, it can be expected that the candidate operator specified by the similarity determination processing unit P205 has a smaller burden than other operators 320 when processing the remotely supported request received this time.

[0044] For example, the example shown in Fig. 4(A) can be considered as a case where the autonomous vehicle 100 issues a remotely supported request for "permission to start at an intersection". In this case, the "confirmation target" and "operation area" which are index elements are considered to have a great influence on the burden when the operator 320 processes the remotely supported request. On the other hand, the "vehicle shape" and "passengers" are considered to have little influence on the burden when the operator 320 processes the remotely supported request. And in the example shown in Fig. 4(A), the operator 320a who was checking and judging the traffic outside the vehicle in the same area as the current supported scene is the candidate operator. When processing the remotely supported request for "permission to start at an intersection", the operator 320a can start the remote support in a state familiar with the traffic environment peculiar to the area. Consequently, it is expected that the operator 320a has a smaller burden when processing the remotely supported request received this time.

[0045] For example, the example shown in (B) of FIG. 4 can be considered as a case where the autonomous vehicle 100 issues a remote support request for "passenger seating attention reminder". In this case, the "confirmation target" and "vehicle shape" as index elements are considered to have a great impact on the burden when the operator 320 processes the remote support request. On the other hand, the "operation area" and "location" are considered to have little impact on the burden when the operator 320 processes the remote support request. And in the example shown in (B) of FIG. 4, the operator 320c who confirmed and judged the interior of the vehicle for a vehicle with the same vehicle shape as the current support scene has become the candidate operator. The operator 320c can start remote support in a state familiar with the vehicle structure when processing the remote support request for "passenger seating attention reminder". Consequently, it is expected that the operator 320c has less burden when processing the remote support request received this time.

[0046] Furthermore, in the examples shown in (A) and (B) of FIG. 4, similarity judgment of the feature amounts for the current support scene is performed in the order of the index elements according to the priority. By performing such similarity judgment, it is possible to preferentially identify the operator 320 who is particularly similar in terms of the index elements that have a great impact on the burden when processing the remote support request between the previous support scene and the current support scene as the candidate operator. Consequently, it is possible to preferentially identify the operator 320 who has less burden when processing the remote support request received this time as the candidate operator.

[0047] For example, in the example shown in (A) of FIG. 4, the number of similar index elements of the operator 320b is larger than that of the operator 320a. However, the operator 320b is not similar to the current support scene in terms of the "confirmation target", which is the index element with the highest priority and the greatest impact on the burden when processing the remote support request. Therefore, it is assumed that the operator 320b has a greater burden when processing the remote support request than the operator 320a. Thus, it can be seen that in the example shown in (A) of FIG. 4, the operator 320a, who is assumed to have less burden when processing the remote support request received this time, can be appropriately identified as the candidate operator.

[0048] The similarity determination processing unit P205 can also adopt processing from other viewpoints for the similarity determination based on the feature amounts of the specified one or more index elements.

[0049] One of the other viewpoints is to calculate the similarity between the previous support scene and the current support scene for each of the assignable operators 320 based on the feature amounts of the specified one or more index elements.

[0050] For example, for each of the one or more index elements, the similarity determination processing unit P205 calculates a similarity point indicating the degree of similarity between the feature amount of the previous support scene and the feature amount of the current support scene according to a predetermined criterion, and can calculate the sum of the similarity points for each of the one or more index elements as the similarity. In this case, the predetermined criterion gives, for example, 1 point for the similarity point between "outdoor traffic" and "indoor of the vehicle" for the index element of "object to be confirmed", and 3 points for the similarity point when they match in "outdoor traffic". Also, for example, when the feature amount is expressed numerically, the similarity determination processing unit P205 calculates the difference or a predetermined spatial distance between the feature amount of the previous support scene and the feature amount of the current support scene for each of the one or more index elements, and can calculate the reciprocal of the sum of the differences or distances for each of the one or more index elements as the similarity. Furthermore, the similarity determination processing unit P205 may be configured to calculate the sum by weighting the similarity points, differences, or distances of each of the one or more index elements. In this case, the index element specifying processing unit P204 may be configured to perform weighting for the specified one or more index elements. For example, the index element specifying processing unit P204 can perform the specification and weighting of the one or more index elements by referring to a list in which numerical values indicating weights are given for the types of remote support requests as in (B) of FIG. 3. By calculating the sum in this way with weighting, the contribution to the similarity of the index element that has a great influence on the burden when processing the remote support request can be made larger.

[0051] Then, the similarity determination processing unit P205 designates an operator 320 with a similarity equal to or higher than a predetermined threshold value as a candidate operator. Here, the predetermined threshold value may be appropriately determined according to the environment in which the remote support system 10 is applied. Also at this time, the similarity determination processing unit P205 may be configured to output the similarity of each of the identified one or more candidate operators as a processing result. By performing similarity determination in this way, it is possible to suppress the situation where an operator 320 that imposes a burden equal to or more than a certain level when processing the remote support request received this time is identified as a candidate operator.

[0052] From this perspective, the similarity determination processing unit P205 may also be configured to designate an operator 320 with the highest similarity as a candidate operator.

[0053] Also, these perspectives can be adopted in combination. For example, the similarity determination processing unit P205 can be configured to perform similarity determination of the feature amounts for the current support scene in the order of the index elements according to the priority for the operator 320 among the assignable operators 320 whose similarity is equal to or higher than a predetermined threshold value.

[0054] Furthermore, the similarity determination processing unit P205 may be configured to designate an operator 320 whose previous support scene is not managed in the scene feature database 215 as one of the candidate operators. By configuring in this way, it is possible to avoid a situation where an operator 320 who performs remote support for the first time is never identified as a candidate operator.

[0055] Also, in such a configuration, the similarity determination processing unit P205 may be configured to calculate the highest similarity for an operator 320 among the identified one or more candidate operators whose previous support scene is not managed in the scene feature database 215. This is because such an operator 320 can start remote support without preconceived notions and is expected to have a low burden for any remote support request.

[0056] Alternatively, the similarity determination processing unit P205 may be configured to calculate the similarity of such an operator 320 with a value of a certain height. This is because it can be expected that a candidate operator with an extremely high similarity has a lower burden than such an operator 320.

[0057] As described above, the similarity determination processing unit P205 outputs one or more candidate operators specified as a processing result.

[0058] Referring to FIG. 1 again, the assigned operator determination processing unit P206 determines the operator 320 selected from one or more candidate operators as the assigned operator. When there is one specified candidate operator, the assigned operator determination processing unit P206 may be configured to select that one candidate operator. When a plurality of candidate operators are specified, the assigned operator determination processing unit P206 randomly selects the operator 320 from the plurality of candidate operators, for example. Alternatively, when the similarity determination processing unit P205 outputs the similarity of each of the plurality of candidate operators as a processing result, the assigned operator determination processing unit P206 may be configured to select the candidate operator with the highest similarity. Or alternatively, the assigned operator determination processing unit P206 may be configured to select the operator 320 based on other indicators such as the number of times of remote support assignment so far. For example, the assigned operator determination processing unit P206 selects the operator 320 so as to reduce the bias in the number of times of remote support assignment.

[0059] The remote support execution processing unit P207 notifies the remote support terminal 310 corresponding to the determined assigned operator that the remote support of the autonomous driving vehicle 100 that has received the remote support request this time has been assigned. Upon receiving the notification, the remote support terminal 310 starts communication with the autonomous driving vehicle 100. Thereby, remote support between the remote support terminal 310 and the autonomous driving vehicle 100 is implemented. The remote support terminal 310 presents information necessary for remote support to the operator 320 and accepts an operation related to the remote support of the operator 320 (for example, input of a judgment on the remote support request). The operation related to the remote support received by the remote support terminal 310 is transmitted to the autonomous driving vehicle 100. Thereby, remote support of the autonomous driving vehicle 100 by the assigned operator is realized.

[0060] In addition, the remote support execution processing unit P207 updates the scene feature database 215 with the scene features of the current support scene as the scene features of the previous support scene for the assigned operator. Thereby, the scene feature database 215 can be constructed.

[0061] As described above, the remote support system 10 according to the present embodiment is configured. In the remote support system 10 according to the present embodiment, generally, there may be a plurality of autonomous driving vehicles 100 that are the targets of remote support.

[0062] Next, with reference to FIG. 5, the configuration of the remote support server 200 will be described. FIG. 5 is a block diagram showing a preferred example of the configuration of the remote support server 200. The remote support server 200 includes a processing unit 210 and a communication unit 220.

[0063] The communication unit 220 communicates with devices external to the remote support server 200 to transmit and receive information. In particular, the communication unit 220 communicates with the autonomous driving vehicle 100 and the remote support terminal 310. Typically, the communication unit 220 communicates with the autonomous driving vehicle 100 and the remote support terminal 310 via the Internet. The communication unit 220 realizes reception of remote support requests and scene information, and transmission of notifications of assignment of remote support to the remote support terminal 310.

[0064] The processing unit 210 executes processing related to the remote support function. The processing unit 210 is a computer including a storage device 211 and a processor 216.

[0065] The storage device 211 is coupled to the processor 216 and stores a plurality of instructions 213 executable by the processor 216 and various data 214 necessary for the execution of processing. The storage device 211 can be composed of recording media such as ROM, RAM, HDD, and SSD.

[0066] The plurality of instructions 213 are provided by a computer program 212. The plurality of instructions 213 are also configured to cause the processor 216 to execute processing related to the remote support function. That is, when the processor 216 operates according to the plurality of instructions 213, the processor 216 functions as an assignable operator identification processing unit P202, a scene feature calculation processing unit P203, an index element identification processing unit P204, a similarity determination processing unit P205, an assigned operator determination processing unit P206, and a remote support execution processing unit P207. The processor 216 can be composed of a CPU or the like including an arithmetic unit, a register, and the like.

[0067] The data 214 includes information acquired by the remote support server 200, parameter information of the computer program 212, and the like. In particular, the data 214 includes a scene feature database 215. The scene feature database 215 is updated by processing executed by the processor 216.

[0068] Here, the scene feature database 215 may be configured to continuously hold the scene features of the previous support scene for each of the plurality of operators 320, or may be configured such that the scene features of the previous support scene are deleted for some of the operators 320 by processing executed by the processor 216. For example, the processor 216 may be configured to execute processing for deleting the scene features managed in the scene feature database 215 as follows.

[0069] One example is to delete the scene features for the operator 320 for whom a certain period of time has elapsed since the last update.

[0070] Another example is to delete the scene features for the operator 320 who has left the seat or is taking a break. In this case, the determination of leaving the seat or taking a break can be made by detecting that the operation of the remote support terminal 310 has not been performed for a certain period of time or that an explicit operation (for example, the departure button has been pressed, etc.) by the operator 320 has been performed.

[0071] Another example is to delete the scene features for the operator 320 who has completed the day's work or the operator 320 who starts the day's work.

[0072] By configuring the processor 216 to execute the process of deleting the scene features in this way, it is possible to represent the operator 320 for whom it is assumed that the feeling of the previous support scene does not remain. Note that since such an operator 320 can be expected to be less burdened with any remote support request, instead of deleting the scene features in the same case as above, it is also conceivable to configure the processor 216 to execute a process of updating the feature amount of each of the plurality of feature elements to a special value similar to any feature amount.

[0073] As described above, the remote support server 200 is configured.

[0074] 2. Process Hereinafter, the process executed by the remote support server 200 will be described.

[0075] FIG. 6 is a flowchart showing the process executed by the remote support server 200, more specifically, the process executed by the processor 216. The flowchart shown in FIG. 6 typically starts when a new remote support request is received from the autonomous vehicle 100.

[0076] In step S100, the remote support server 200 identifies an operator 320 to whom assignment is possible among a plurality of operators 320 by the assignable operator identification processing unit P202. In step S200, the remote support server 200 calculates the scene features of the current support scene by the scene feature calculation processing unit P203. In step S300, the remote support server 200 identifies one or more index elements to be used as an index for similarity determination for the current support scene by the index element identification processing unit P204. Here, the processes according to steps S100, S200, and S300 may be executed in any order. Alternatively, the processes according to steps S100, S200, and S300 may be executed in parallel.

[0077] Next, in step S400, the remote support server 200 performs a similarity determination based on the feature amounts of the identified one or more index elements by the similarity determination processing unit P205, and identifies one or more candidate operators from the operators 320 to whom assignment is possible.

[0078] Next, in step S500, the remote support server 200 selects an operator 320 who processes the remote support request received this time from the identified one or more candidate operators by the assigned operator determination processing unit P206.

[0079] Next, in step S600, the remote support server 200 assigns remote support for the autonomous vehicle 100 to the determined assigned operator by the remote support execution processing unit P207.

[0080] Next, in step S700, the remote support server 200 updates the scene feature database 215 with the features of the current support scene as the scene features of the previous support scene for the assigned operator by the remote support execution processing unit P207. After step S700, the current process ends.

[0081] As described above, the processing is executed by the remote support server 200 (processor 216). Further, by the remote support server 200 executing the processing in this manner, a remote support method for providing the remote support function according to the present embodiment by a computer is realized.

Explanation of Signs

[0082] 100 Vehicle 200 Remote support server 310 Remote support terminal 320 Operator 210 Processing unit 211 Storage device 216 Processor 220 Communication unit

Claims

1. A remote support system that provides a remote support function for a vehicle by a plurality of operators, a storage device that stores a database for managing feature amounts for each of a plurality of feature elements regarding a previous support scene that was the target of a previously processed remote support request, for each of the plurality of operators; one or more processors; comprising: the one or more processors are configured to: receive a new remote support request from a vehicle, and perform a process of identifying one or more index elements to be used as an index for similarity determination with respect to the current support scene that is the target of the new remote support request; perform a process of identifying one or more candidate operators among the plurality of operators for whom the previous support scene is similar to the current support scene, based on the feature amounts of the one or more index elements; perform a process of selecting an operator to process the new remote support request from the one or more candidate operators; and are configured to execute the above. A remote support system characterized by the above.

2. The remote support system according to claim 1, wherein the process of identifying the one or more index elements further includes setting priorities for the identified one or more index elements, and the process of identifying the one or more candidate operators includes: performing similarity determination of the feature amounts with respect to the current support scene in the order of the index elements according to the priorities, and sequentially excluding from the determination target operators, operators for whom it is determined that the previous support scene is not similar to the current support scene; upon receiving that the number of determination target operators has become one, or that similarity determination of the feature amounts has been performed for all of the one or more index elements, setting the determination target operators as the one or more candidate operators; and includes the above. A remote support system characterized by the above.

3. The remote support system according to claim 1, wherein the process of identifying the one or more candidate operators includes: for each of the plurality of operators, calculating a similarity degree between the previous support scene and the current support scene based on the feature amounts of the one or more index elements; setting as the one or more candidate operators, operators for whom the similarity degree is equal to or greater than a predetermined threshold; and includes the above. A remote support system characterized by the above.

4. The remote support system according to any one of claims 1 to 3, The process of identifying the one or more candidate operators further includes using, as the one or more candidate operators, the operator(s) for which the previous support scenario is not managed in the database. A remote support system characterized by the above. **Claim 5** A remote support method for providing, by a computer, a remote support function for a vehicle by a plurality of operators, comprising: For each of the plurality of operators, managing feature amounts for each of a plurality of feature elements regarding the previous support scenario that was the target of the previously processed remote support request; Receiving a new remote support request from the vehicle and identifying one or more index elements to be used as an index for similarity determination with respect to the current support scenario that is the target of the new remote support request; Based on the feature amounts of the one or more index elements, identifying one or more candidate operators among the plurality of operators for which the previous support scenario is similar to the current support scenario; Selecting, from the one or more candidate operators, an operator to process the new remote support request; A remote support method characterized by including the above.

Images