JP2012203867A - Information providing system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide technology for, when there is a plurality of information to be provided, for example, when a user is driving a vehicle, providing information without impairing safety according to the driving conditions.SOLUTION: Evaluation values relating to a plurality of evaluation factors are given to each information. For example, when a user is a driver, the priority of the evaluation factors changing according to driving conditions is determined. Then, the significance of each information as a whole is determined from the evaluation values of the evaluation factors of each information and the priority of the evaluation factors. Furthermore, each information has a temporal and spatial effective range, and the scheduling of an information providing timing is performed either by providing the maximally large number of information within the effective range or by maximizing the sum of the priorities of the information to be provided. Under the driving conditions that the provision of the information is permitted, the information is controlled so as to be provided in descending order of priority. Also, the priority of the evaluation factors may be adaptively adjusted from the past history.

Description

  The present invention relates to an information providing system, and more particularly, to a technique for providing information selected based on driving conditions to a driver who is driving a vehicle at a timing that does not impair safety.

  In recent years, an apparatus for providing various information to a user who is driving a car has been proposed. For example, Patent Document 1 (Japanese Patent Laid-Open No. 2008-213609) discloses an image display control capable of improving display processing speed of a display screen and promptly transmitting information to a vehicle occupant at an appropriate timing. An apparatus is disclosed. In this image display control device, after image data input from a plurality of in-vehicle devices is developed in an image memory, an image to be displayed on the display device is determined based on the priority order of the images. In particular, in this image display control device, the image of the danger notification image for notifying the danger of driving the vehicle is preliminarily developed in the image memory, so that the display of the danger notification image is quickly and preferentially displayed. .

  Patent Document 2 (Japanese Patent Application Laid-Open No. 2010-179783) does not confuse the driver when a plurality of pieces of information are presented to the driver at the same time, and more to the driver at that time. A method for providing useful information is disclosed. In this method, information is notified to the driver in various notification modes according to a predetermined information notification priority.

  Further, in Patent Document 3 (Japanese Patent Application Laid-Open No. 2009-186407), when information is displayed based only on the priority order, the driver frequently feels annoyed because the display is frequently switched. In order to solve such a problem, a method has been proposed in which information to be provided on a predicted route is determined in advance, and scheduling is performed so that the number of display switching is reduced as much as possible based on a predetermined priority order.

  However, Patent Document 1 mentions that audio information is provided when the image data of the image to be displayed is accompanied with the audio information, but when the audio information is not attached to the image, There is no mention of how to transmit voice information to passengers promptly and at appropriate timing, and specifically, what evaluation factors determine the priority of voice information to be provided. Not.

  In addition, since the method disclosed in Patent Document 1 needs to prepare a rule table that defines the urgency level for each information in advance, the urgency level is adapted to the situation according to the driving situation and the driving situation. There is a problem that cannot be made. In addition, when there are a plurality of evaluation criteria for determining the priority of information, there is a problem that the method is difficult to apply. Furthermore, because the rule table that defines the urgency level is fixed, it is necessary to update the rule table that defines the urgency level of information when a new in-vehicle information service is added. However, since the rule table must be huge, there is a problem that the priority order of information cannot be determined when an in-vehicle information service is added later.

  In the methods disclosed in Patent Document 2 and Patent Document 3, since the priority order of information to be provided is determined in advance, there is a problem that it is impossible to cope with a change in the situation.

JP 2008-213609 A JP 2010-179783 A JP 2009-186407 A

  Nowadays, devices with communication functions are installed in the vehicle itself, and it is now possible to connect to external networks via mobile terminals. From outside, information service updates and new information services Can now be installed.

  Under such circumstances, it is practically impossible to thoroughly implement evaluation criteria for determining importance for each information in advance for all information services, and it is necessary to assume that unknown elements are added. And in order to realize appropriate information provision to the user in such an environment where information that can be provided to the user is constantly updated, a method that can evaluate the importance of the information as a whole Is needed.

  In order to evaluate the importance of information as a whole, it is practically impossible to deal with only one evaluation standard, and an evaluation using a plurality of evaluation standards is required. Moreover, because the environment in which users are placed changes frequently in time and space, it is not sufficient to simply perform evaluation using multiple evaluation criteria. The importance of information as a whole It is required to make a dynamic decision according to a change in the user's environment.

  The present invention has been made in view of such problems, and the purpose of the present invention is when there are a plurality of information that can be provided to the user, according to changes in the environment in which the user is placed, A technology that dynamically determines the priority of such information, and controls the selection and provision timing of information so as to provide information without compromising driving safety even when the user is driving the vehicle. Is to provide.

  In order to solve such a problem, the information providing system of the present invention includes an environment information sensing unit that acquires environment information related to a situation surrounding the user, and information that generates information to be provided to the user based on the environment information. A provision service unit; and a provision information integration unit that determines a priority order of the plurality of provision information according to a situation surrounding the user when there are a plurality of provision information, and the information provision service unit includes the provision information For each piece of information, a plurality of evaluation factors and evaluation values related to the evaluation factors are attached and transmitted to the provided information integration unit. The provided information integration unit includes an evaluation factor importance level management unit, an information importance level evaluation unit, And the evaluation factor importance management unit changes the importance of each of the plurality of evaluation factors when the situation surrounding the user changes, and the information importance evaluation The section changes the importance of the plurality of provision information based on the evaluation factor importance after the change, and determines the priority of the provision information in order of the provision information importance after the change .

  In addition, the provision information integration unit schedules the information provision timing so that the number of pieces of information satisfying the predetermined condition or the sum of their importance levels is maximized within a range that satisfies the provision condition of the predetermined information. It is good also as an aspect provided with the provision information scheduling part. The range that satisfies the predetermined information provision possibility condition is, for example, a predetermined temporal range and a spatial range. The number of pieces of information satisfying the predetermined condition or the sum of the importance levels thereof is, for example, the number of pieces of information that can be provided to the user or the sum of importance levels of a plurality of pieces of information to be provided.

  Further, the evaluation factor importance degree management unit includes an evaluation factor importance degree learning unit that adaptively adjusts the evaluation factor importance degree to an environment surrounding the user, and the evaluation factor importance degree learning unit includes the evaluation factor importance degree learning unit, The remuneration value for the future of the reward with the evaluation factor importance as the reward, the number of information provided within the predetermined time range and the spatial range or the sum of the importance of the plurality of information provided. It is good also as an aspect learned by changing the importance of an evaluation factor so that the expected value of may be maximized.

  The evaluation factor importance degree management unit includes an evaluation factor importance degree learning unit that adaptively adjusts the evaluation factor importance degree to an environment surrounding the user, and the evaluation factor importance degree learning unit includes a plurality of evaluation factor importance degree learning units. The evaluation factor set (policy) may be managed, and policy learning may be performed to select the most appropriate evaluation factor set in the environment surrounding the user.

  According to the present invention, there are a plurality of information providing services related to the movement of the user, and the information presentation request from the in-vehicle information service to the driver even when the in-vehicle information service is changed or added later, for example. In consideration of the risk level of the driving situation of the vehicle and the degree of load related to the driving of the driver, it is possible to provide as much safety and information as possible at an appropriate timing.

  The in-vehicle information service side does not need to define the importance of the information to be provided as a whole, and only needs to determine the importance value (importance score) of the service alone. Therefore, the importance score determination logic in each in-vehicle information service can be independently developed without depending on each other. Therefore, it becomes easy to add or change services later.

  In addition to providing information, the present invention can be used to determine the importance of resources to be used (memory used, network bandwidth, CPU time, etc.).

It is a block diagram for demonstrating the structural example of the information provision system of this invention. It is a block diagram for demonstrating the structural example of the provision information integration part with which the information provision system of this invention is provided. It is a flowchart for demonstrating the processing content performed with the information provision system of this invention. It is a table in which information is classified into a predetermined hierarchy according to its level. The contents of data set in the information I _i is a diagram for explaining by way of example. It is a figure which shows a mode that the node was allocated with respect to information set [ _Ii ]. Is a diagram showing a state of setting the node I ₀ of the dummy information to determine the overall significance. It is a figure for demonstrating a graph cut vector.

  Embodiments of an information providing system and an information providing method of the present invention will be described below with reference to the drawings. In the following description, it is assumed that the user of the present invention is a driver who drives an automobile.

  FIG. 1 is a block diagram for explaining a configuration example of an information providing system according to the present invention. The information providing system 100 includes an environmental information sensing unit 101, an environmental state management unit 102, an operation state management unit 103, an information provision service unit 104, a provision information integration unit 105, and an information output unit 106.

  The environment information sensing unit 101 is connected to the vehicle movement information such as time, position, vehicle speed, acceleration, and yaw rate, vehicle operation information such as the steering angle and accelerator opening, driver images, sounds, and pulses via the in-vehicle LAN. Get driver information and other information as environmental information.

  The environmental state management unit 102 has a function of estimating the state around the current vehicle position based on the environmental information acquired by the environmental information sensing unit 101. The environmental state management unit 102 has a function of estimating, for example, the amount of sunlight, the outside air temperature, the rainfall state, and the road surface state.

  The driving state management unit 103 has a function of estimating the current driving state based on the information acquired by the environment information sensing unit 101. Here, the driving state includes states such as stop, progress, right turn, and left turn.

  The information provision service unit 104 is means for generating information to be provided to the user and outputting the information to the provision information management unit 201 provided in the provision information integration unit 104. The information providing service unit 104 shown in FIG. 1 includes M information providing service units 104-1 to 104-M each handling information of different categories.

  Each of the information providing service units 104-1 to 104-M can operate with independence, but can also be operated with mutual dependency.

  For example, the information providing service unit 104-1 performs an information providing service that gives a warning such as the approach of a preceding vehicle or detection of a pedestrian. The information providing service unit 104-2 performs, for example, an information providing service (navigation service) that provides route guidance. The information providing service unit 104-3 performs, for example, an information providing service that recommends a stop-in place. The information providing service unit 104-M performs an information providing service such as eco driving advice.

  These information provision service units 104-1 to 104-M acquire necessary environment information from the environment information sensing unit 101, generate information to be provided to the user, and provide information provided in the provision information integration unit 104 The data is output to the management unit 201. Note that the output status of the provision information integration unit 104 may be acquired as feedback, and this may be used when generating provision information for the user next time.

  The provided information integration unit 105 prioritizes each piece of information received from the information providing service units 104-1 to 104-M according to the driving situation. Then, this prioritized information is provided to the user at an appropriate timing so that more information can be obtained within a given time and space, or the sum of the importance of the information to be provided becomes larger. It is a means for controlling to be able to.

  The information output unit 106 is a unit that outputs information to the user in accordance with an information provision command output from the information output control unit 207 provided in the provision information integration unit 105. In the example shown in this figure, the information output unit 106 is provided with N information output devices. Such information output devices include, for example, a video display device, an audio output device, an indicator, an actuator device such as a vibration, and the like according to the type of provided information, and may be a device in which these are combined.

  FIG. 2 is a block diagram for explaining a configuration example of the provision information integration unit 105 provided in the information provision system 100 of the present invention. The provision information integration unit 105 includes a provision information management unit 201, an evaluation factor importance degree management unit 202, an information importance degree evaluation unit 203, an information provision scheduling unit 204, a route prediction unit 205, a workload prediction unit 206, and an information output. A control unit 207 is provided.

  The provided information management unit 201 stores and accumulates the information generated by the information providing service units 104-1 to 104-M in the input buffer. The provided information management unit 201 also calls information to be provided to the user in accordance with an information providing schedule to be described later from the information stored in the input buffer into the output buffer, and temporarily and temporarily stores these information. Memorize and store. Further, the provision information management unit 201 instructs the reporting provision scheduling unit 204 to create an information provision schedule or performs processing necessary for changing the information provision schedule when the state surrounding the user changes. To delete invalid information from the input buffer and the output buffer.

  The evaluation factor importance level management unit 202 includes an evaluation factor management unit 202-1 and an evaluation factor importance level learning unit 202-2.

  Each of the information providing service units 104-1 to 104-M is provided with a plurality of evaluation factors and evaluation values for each evaluation factor. However, since the information providing service units 104-1 to 104-M operate independently, the configuration of the evaluation factors does not necessarily match.

  The evaluation factor importance degree management unit 202 records the evaluation factor assigned to the information input from the information provision service units 104-1 to 104-M in the provision information management unit 201 and the evaluation value for each evaluation factor, and , Evaluate and manage the importance of evaluation factors. This evaluation factor importance is used when evaluating the importance of information generated by the information providing service units 104-1 to 104-M as information to be provided to the user.

  A set of importance values of evaluation factors is determined for each operation state. When the current driving state estimated by the driving state management unit 103 changes, the evaluation factor management unit 202-1 switches the evaluation factor importance determined according to the driving state.

  The evaluation factor importance determined according to the driving state is adaptively adjusted to the environment by the evaluation factor importance learning unit 202-2.

  The information importance level evaluation unit 203 includes a graph cut vector generation unit 203-1 and a mathematical programming problem processing unit 203-2, and stores information stored in the provided information management unit 201 according to the current operating state. Assess importance. This information importance evaluation is comprehensively performed in consideration of all evaluation factors. In particular, in the present embodiment, a technique that is applicable even if there is a mismatch in evaluation factors between information is adopted by an extended technique of the hierarchical analysis method.

  The information provision scheduling unit 204 is a means for scheduling information provision timing so that the number of pieces of information satisfying the predetermined condition or the sum of the importance thereof is maximized within a range satisfying the provisionable condition of the predetermined information. It is.

  Here, the range that satisfies the above-described predetermined information provision possibility condition is, for example, within a predetermined temporal range and a spatial range. Further, the number of pieces of information satisfying the above-described predetermined condition or the sum of the importance levels thereof is, for example, the number of pieces of information that can be provided to the user or the sum of importance levels of a plurality of pieces of information to be provided.

  In the configuration example shown in the figure, the information provision scheduling unit 204 includes a provision information number optimization unit 204-A, and the information importance degree that has been comprehensively taken into consideration in consideration of all evaluation factors is assigned to each. Information allocation processing is performed in order to provide a plurality of information to the user as much as possible within a predetermined time and space. This allocation process is preferentially performed in descending order of information importance score. Each information is provided with data instructing an operating state in which the information can be provided and a providing time. Based on these data, it is possible to provide more information by optimizing the timing (order) for providing the information.

  Information to be provided to the user is scheduled by such information allocation processing based on the overall information importance taking into account all evaluation factors and information provision timing optimization processing. Hereinafter, the result is referred to as an information provision schedule.

  The route prediction unit 205 manages the route from the current location to the destination when there is a destination set in the navigation system. In addition, the route prediction unit 205 compares the past route prediction with the current environment information, evaluates the necessity of route prediction based on the result, and determines that the route prediction is necessary. It has a function to predict the route.

  The route prediction unit 205 further has a function of determining whether or not the environment information acquired by the environment information sensing unit 101 is inconsistent with the predicted route. If the result of this determination is inconsistent, the current location may be more than a predetermined distance away from the original predicted route. Therefore, the route prediction unit 205 updates (reroutes) the predicted route and notifies the provision information management unit 201 of the update of the predicted route when the result of the determination is inconsistent.

  The route predicting unit 205 has a function of automatically predicting a route from various conditions such as the current position, day of the week, and time based on past learning results when a destination is not set. Also good.

  The workload prediction unit 206 is mentally or cognitively related to the driving of the driver based on the state around the current vehicle position estimated by the environmental state management unit 102 (the amount of sunlight, the outside temperature, the rainfall state, the road surface state, etc.). Appreciable load. The workload prediction unit 206 manages information related to areas with high driving loads such as intersections and sharp curves, and has a function of predicting future driving loads from current traffic conditions.

  The information output control unit 207 controls output to the information output unit 106. As described above, the information output unit 106 is provided with an information output device. The information output control unit 207 controls an information providing command issued to the information output unit 106 for each piece of information provided to the user. The information provided to the user also includes special information such as temporarily stopping or canceling output from a specific information providing device with respect to information currently being output. For example, when providing AV information, video information and audio information are usually provided to the user at the same time. However, when it is necessary to stop providing only video information, information output from the audio output device is performed. Is continued, but information output from the video display device is canceled.

  Next, processing contents performed by the above-described components will be described.

  FIG. 3 is a flowchart for explaining the processing contents performed in the information providing system of the present invention. Although each component operates in parallel, the description will be made along the flow of processing information output from the information providing service unit 104.

  First, by the environment information sensing unit 101, various environment information (time, position, vehicle speed, acceleration, yaw rate, etc.), vehicle operation information such as steering angle, accelerator opening, etc. Driver information such as an image, sound, or pulse) is acquired (S101). The environment information sensing unit 101 includes a plurality of various sensors. Each sensor measures the environmental state at a constant cycle or when a change occurs in the sensing environmental state. The environment information sensing unit 101 transmits the environment information acquired by these sensors to the information providing service unit 104 and the provided information management unit 201 at a predetermined timing.

  The information providing service unit 104 generates information to be provided to the user based on the environment information transmitted from the environment information sensing unit 101 (S102). In the above-described example, when the information to be provided is related to a warning such as approach of a preceding vehicle or detection of a pedestrian, the information providing service unit 104-1 generates the information. When the information to be provided is related to route guidance or the like, the information providing service unit 104-2 generates the information. Further, when the information to be provided is related to the recommendation of a stop-in place, the information providing service unit 104-3 generates the information. Furthermore, when the information to be provided is related to eco-driving advice or the like, the information generation service unit 104-M generates the information.

  The information provision service unit 104 transmits the information generated in this way to the provision information management unit 201 at a predetermined timing (S103).

  The provision information management unit 201 stores this information in an input buffer that is a means for temporarily and temporarily storing and accumulating information generated by the information provision service unit 104 and transmitted to the provision information management unit 201 ( S104).

  The provision information management unit 201 instructs the information provision scheduling unit 204 to create an information provision schedule, and the information provision scheduling unit 204 creates an information provision schedule (S105). The provision information management unit 201 also calls information to be provided to the user according to the information provision schedule from the information accumulated in the input buffer, and temporarily and temporarily stores and accumulates the information. The data is stored in an output buffer, which is a means to keep (S106).

[Procedure for changing the information schedule]
The information provision schedule is changed according to the following procedure as the environment surrounding the user changes. When the provided information management unit 201 receives a timer interruption at a predetermined period or a state change notification from the environmental state management unit 102, the driving state management unit 103, the route prediction unit 205, or the like, Since it is necessary to change the information provision schedule (S107: Yes), processing necessary for the change is instructed to each component (S111). If it is not necessary to change the information provision schedule (S107: No), the process proceeds to step S108 described later.

  The information provision schedule changing process (S112) includes (i) a process of deleting “invalid information” from the input buffer and the output buffer by the provision information management unit 201, and (ii) an input buffer and a process by the importance level evaluation unit 203. The process is executed by three sub-processes: a process for evaluating the importance of information stored in the output buffer; and (iii) a process for scheduling information provision on the predicted route by the provision information scheduling unit 204. The processing contents of the information provision schedule by the information provision scheduling unit 204 in step S105 are the same.

  Here, the above-mentioned “invalid information” means information that is no longer meaningful to be provided to the user due to changes in the environment surrounding the user, information that should not be provided, etc. This information should be invalid or invalid.

  Further, in the process (ii) for evaluating the importance of the information stored in the input buffer and the output buffer, (a) the priority order of the evaluation factors is determined based on an “evaluation factor hyperscore” described later. A process, (b) a process for determining the priority of information relating to each evaluation factor, and (c) a process for determining an overall priority from the priority of the evaluation factor and the priority of each information in each evaluation factor. It consists of three sub-processes.

  Specifically, the process of deleting the invalid information from the input buffer and the output buffer by the provided information management unit 201 is performed as follows. When the provision information management unit 201 receives a timer interruption at a predetermined period or a notification of a state change from the environmental state management unit 102, the driving state management unit 103, the route prediction unit 205, or the like, the invalid information is displayed. The stored information in the buffer is updated by deleting from the above-described input buffer or output buffer or both of these buffers. For example, if the information stored in the buffer has already expired at the current time (when the valid time has passed), or the information stored outside the valid area in light of the current position and the predicted route In this case, the provision information management unit 201 discards the invalid information from the input buffer or the output buffer.

  In addition, the above-described processing for evaluating the importance of information stored in the input buffer and the output buffer by the importance evaluation unit 203 is performed when information remains in either the input buffer or the output buffer. It is executed by evaluating the magnitude relation of the importance of the value of the evaluation factor (evaluation factor score) given to.

Further, the information provision scheduling process on the predicted route by the provision information scheduling unit 204 described above is performed based on the total importance of the information I _i obtained by the procedure described later.

  The details of the process of evaluating the importance of the information stored in the input buffer and the output buffer by the importance evaluation unit 203 will be described below.

[Evaluation factor information]
Information that is generated and output by each of the information providing service units 104-1 to 104-M provided in the information providing service unit 104 is given a value of an evaluation factor for evaluating the importance of the information. Note that the value of this evaluation factor is a non-negative real value. The value of this evaluation factor is referred to as “evaluation factor score” for convenience.

  FIG. 4 is a table in which information is classified into a predetermined hierarchy according to its level. In the example shown in this figure, there are four layers, namely “alarm information”, “safety information”, “navigation information”, and “media information”. “Alarm information” is considered to be more important than “safety information”, “safety information” is more important than “navigation information”, and “navigation information” is more important than “media information”. For this reason, a high evaluation factor score (information hierarchy score) is given in this order.

  This information hierarchy is one of special evaluation factors, and is defined indispensable and commonly in each of the information providing service units 104-1 to 104-M. In addition, each information provision service part 104-1 to 104-M shall be permitted to define a unique evaluation factor in addition to such an essential and commonly defined evaluation factor.

  Information about the evaluation factor itself is accumulated and managed in the evaluation factor management unit 202-1. Further, the evaluation factor management unit 202-1 also sets an importance score (evaluation factor hyperscore) for an evaluation factor that determines which evaluation factor is more important.

  It is assumed that a predetermined evaluation factor hyperscore is determined in advance for each environmental state and driving state as the essential common evaluation factors such as the information hierarchy described above. In other words, the hyper score of the essential common evaluation factor is not changed depending on the environmental state and the driving state.

  Other evaluation factors have predetermined evaluation factor hyper-scores determined in advance for each environmental state and driving state in the same manner as the essential common evaluation factors described above. Unlike the required common evaluation factors, the initial values are set and then dynamically changed. In other words, the hyperscores of the evaluation factors other than the essential common evaluation factors change dynamically with changes in the environmental state and the driving state. The procedure for dynamically changing the hyperscore of the evaluation factor will be described later.

  The essential common evaluation factors are known in advance determined in all the information provision service units 104-1 to 104-M and the provision information integration unit 105. On the other hand, the evaluation factors defined and determined for each of the information providing service units 104-1 to 104-M may exist independently of other information providing service units or the provided information integration unit 105.

  The provided information management unit 201 transmits the evaluation factor assigned to each information input from the information providing service unit 104 to the evaluation factor management unit 202-1. The evaluation factor management unit 202-1 determines whether or not these evaluation factors are already registered. If the evaluation factor management unit 202-1 is an unregistered evaluation factor, it is registered as a new evaluation factor and the evaluation factor hyperscore is updated. Set the initial value.

  The magnitude relationship of importance between information with different information hierarchies is determined from the magnitude relationship of a predetermined score given in advance to an information hierarchy as one of evaluation factors. On the other hand, for information belonging to the same information hierarchy, the magnitude relationship of importance cannot be determined only from the information hierarchy. Therefore, with regard to these pieces of information, the magnitude relation of importance is determined from the magnitude relation of scores of evaluation factors other than the information hierarchy.

  However, these evaluation factors do not always share all information. However, even if there is a deficiency in such an evaluation factor, in the present invention, as described later, all evaluation factors are integrated to determine the magnitude relationship of importance.

[Information importance calculation]
In the following, it is assumed that there are N pieces of information remaining in the input buffer and the output buffer that belong to a certain information hierarchy, and the procedure for determining the magnitude relationship of the importance of these pieces of information will be described. Here, each of the N pieces of information is I _i (I = 1, 2,..., N). The processing described below is performed for each information hierarchy.

FIG. 5 is a diagram for explaining the contents of data set in the information I _i by way of example. As for the value of the evaluation factor, it is assumed that there are A _i evaluation factors given to the information I _i constituting a part of such data, and a specific evaluation factor a (a = 1, 2,. .., the evaluation factor data a _i) is denoted as f [ia], the set of a _i number of evaluation factors data provided to the information I _i is [f [ia]] (a = 1, 2,. .., A _i ).

Here, the data f [ia] of the specific evaluation factor a in the information I _i is fid [ia] which is identification ID data of the evaluation factor a and fv [ia] which is score data of the evaluation factor a It consists of a set of. That is, the data f [ia] of the specific evaluation factor a of the information I _i is expressed as f [ia] = (fid [ia], fv [ia]).

Now, consider an evaluation factor fid specified by a certain ID data. For convenience, this will be referred to as evaluation factor a. This evaluation factor a is given to any information included in the set {I _i } of information I _i .

First, as shown in FIG. 6A, a node is assigned to the information set [I _i ]. The information I _i having the evaluation factor a and the information I _j are connected by an arrow from I _i to I _j . Such an arrow is simply called an edge. Here, i <j. If i> j, the edge direction _changes from I _j to I _i . In this way, a graph (evaluation network) expressing the relationship between the evaluation factors is obtained.

  In such an evaluation network, there is no edge between information that does not have a common evaluation factor. The evaluation network obtained in this way is generally divided (grouped) into a plurality of connected components. In the example shown in FIG. 6A, it is divided into four connected components.

When the evaluation network includes a plurality of connected components, it is not possible to compare the importance of information belonging to different connected components. Therefore, in order to obtain the overall importance, a node I _{0 of} dummy information is set as shown in FIG. 6B. Here, it is assumed that the evaluation factor score given to the node I ₀ of the dummy information is not larger than the minimum value of the evaluation factor score set to [I _i | fid∈ [f [ia]]]. Further, for the case of [I _i | [fid] ∩ [f [ia]] = φ], the pair comparison weight values described later are equivalent.

  In the above example, a dummy information node is added, but even if such a dummy is not added, after the processing described below is first performed for each connected component, the importance between the connected components is evaluated. Thus, it is possible to determine the magnitude relationship of the importance of the provided information by using the product of the importance score of each connected component and the importance score of the connected component as the importance score of the information as a whole. The importance between the connected components can be, for example, a number obtained by dividing the number of nodes in the connected components by the total number of nodes. In this case, a large connected component is more important.

In the information I _i, it is considered that the importance w _i determined from the whole relationship with other provided information is defined for a certain evaluation factor. This importance is determined as follows.

It is assumed that there is a functional relationship such as the following equation (1) between the importance w _i and the score of the evaluation factor.

  Here, it is assumed that the function F (x) has a property as shown in the following equation (2).

  As the function F (x), for example, a function such as the following expression (3) can be selected.

  here,

Then,

It becomes.

  Take the logarithm of equation (1) above,

Put it. Where

It is. However, regarding the edge with the dummy information node,

And

  Here, the adjacency matrix in the evaluation network illustrated in FIG. 6B is defined by the following equation (9).

  As a result, the relationship of the above equation (6) can be summarized as follows.

  here,

  Also,

It is.

  The vector of the above equation (12) is called a graph cut vector.

  FIG. 7 is a diagram for explaining the graph cut vector. This graph cut vector is an amount determined from the evaluation factor score set in the information.

  In general, the simultaneous linear equations given by the above equation (10) have more equations than variables, and there is no algebraic solution. Therefore,

Then, the vector of the above equation (11) is determined so that the sum of squares of each component of the error vector on the left side of the above equation (13) is minimized. According to such a procedure, the vector of the above equation (11) can be determined regardless of the number of equations, and the following equation (14) is obtained.

  From the vector determined by equation (14) above,

Is obtained.

The left side (w _i ) of the above equation (15) is the importance of the information I _i .

  In the present invention, the above processing is executed for all the evaluation factors registered in the evaluation factor management unit 202-1, and the importance for each evaluation factor is determined for each piece of information.

  In addition, the same processing as described above is performed in other information layers to determine the importance of information.

Further, for the evaluation factor f, the importance (v _f ) of the evaluation factor f is determined by the same procedure as that for determining the importance of the information described above. In determining the importance of the evaluation factor, the hyperscore of the evaluation factor is used.

Based on the importance of information obtained in this way and the importance of evaluation factors, the overall importance (S _i ) of information using all evaluation factors is defined. Here, the importance (S _i ) of comprehensive information is defined by the following equation (16).

Here, w _i ^f is the total importance of the information I _i regarding the evaluation factor f.

[Scheduling]
Details of scheduling of information provision on the predicted route by the provision information scheduling unit 204 will be described below. This information provision scheduling is performed based on the total importance of the information I _i obtained by the above-described procedure.

  The information provision timing is determined in order from the information with the highest overall importance. At this time, the information provision timing is determined so that the sum of the importance of the information that can be provided is maximized. However, the cost is very high if the optimum solution for the timing is calculated. Therefore, in this embodiment, a solution (approximate solution) that is as close to the optimum solution as possible within an allowable cost range is obtained. This approximate solution is obtained by the following method.

  First, the most initial time within the provision time given to the information is provisionally set. Confirm that the predicted position at that time is within the provision area. Here, the predicted position is predicted based on the predicted route information and the average hourly speed within the latest fixed time. If the provision area is not designated, the process proceeds to the next process as it is. In addition, when it is outside the provision area, the provision time is shifted to a time after a predetermined time, and the time to enter the provision area is obtained. If the information does not enter the provision area within the time that is valid as provision information to the user (valid time), the information is returned to the input buffer, and the provision timing of the next most important information is determined. Transition to processing.

  When the provision time can be determined within the provision area, the workload prediction at the predicted provision position is further used to confirm that the workload prediction value is less than or equal to the upper limit of the workload permitted in the information hierarchy to which the information belongs. . If other highly important information is provided within the period from the provision time to the provision time at that time, the provision time is shifted to a time after a predetermined time, and the above provision is made. Repeat the time determination.

  If the provision time that can be set is not found within the valid time, the information is returned to the input buffer, and the process proceeds to the process for determining the provision timing of the next most important information.

  In the information, the provision position is recorded based on the provision time determined by the above processing. This position is used as a trigger for actually outputting information. Further, the information is stored in the output buffer in the order of providing time.

  The above processing is performed for all information.

  In the above example, the importance of the information itself was used, but dividing by the provision time and the size of the provision area, in order from the importance (importance density) per unit time and unit area, You may schedule by performing the said process.

[Evaluation factor importance learning]
As described above, the evaluation factor importance determined according to the driving state is adaptively adjusted to the environment by the evaluation factor importance learning unit 202-2.

  For example, the evaluation factor importance degree learning unit 202-2 calculates the evaluation factor importance degree by calculating the number of pieces of information provided within a predetermined temporal range and a spatial range or the sum of importance levels of a plurality of pieces of provided information. As a reward, learning is performed by changing the importance of the evaluation factor so that the expected value of all reward values in the future of the reward is maximized.

  In addition, the evaluation factor importance degree learning unit 202-2 manages, for example, a plurality of evaluation factor sets (policies), and performs policy learning that selects the most suitable evaluation factor set in the environment surrounding the user.

  In the present embodiment, the evaluation factor importance is the expected value of the sum of reward values in the future (abbreviated as expected reward) with the sum of the importance of information provided within a predetermined range or the number of provided information as a reward. Is made to learn by changing the hyperscore of the evaluation factor so as to maximize. The expected reward is a function of the running state and a function of the evaluation factor set. The expected reward is also a function of the discount rate for future rewards, but here the discount rate is a constant value.

  In order to execute this learning, a pseudo expected reward value depending on the importance of each evaluation factor is calculated based on Bellmann's principle instead of the expected reward itself. By using this pseudo expected reward, the evaluation factor hyperscore is learned in the direction in which the reward is maximized by setting the evaluation factor hyperscore in a direction in which the gradient is positive.

  To correct the evaluation factor hyperscore in the positive direction of the gradient, first select one of the evaluation factors, calculate the reward when the value is changed by a small positive amount, and the reward is selected Compared to the reward calculated before changing the hyperscore of the evaluated factor. When the reward value increases, the hyperscore of the selected evaluation factor is corrected to the value after the change. The pseudo-expected reward is also corrected by adding a reward increment. Such correction may be performed for only one evaluation factor at a time, or may be repeated for all evaluation factors.

  In addition, this method also enables policy selection as to which evaluation factor set should be used in each driving situation. That is, by selecting an evaluation factor set that maximizes the pseudo expected reward value in each driving situation, an evaluation factor set suitable for the driving situation can be selected.

[Output buffer information addition and output timing processing]
In a predetermined cycle, the information providing position at the head of the output buffer is confirmed, and it is searched whether it is within the processing distance from the current position of the user (S108). In other words, this process is an evaluation of whether or not the provision information at the head of the output buffer is at a timing to provide information in the environment where the current user is placed. It should be noted that the process of confirming the information provision position at the head of the output buffer operates asynchronously with the scheduling described above.

  When valid head provision information in the output buffer is found (S108: Yes), information to be output is transmitted to the information output unit 106 (S109), and the information is output from the information output unit 106 and provided to the user. (S110).

  On the other hand, when the information providing position at the head of the output buffer is confirmed, if the information is not inserted in the output buffer (S108: No), the process is terminated.

  As described above, according to the present invention, when there are a plurality of pieces of information to be provided, for example, when the user is a driver, the information can be provided without impairing safety according to the driving situation.

  As described above, according to the present invention, when there are a plurality of pieces of information that can be provided to the user, the priority of the information is dynamically determined according to a change in the environment where the user is placed, and In order to provide information without impairing driving safety even when a user is driving a vehicle, a technique for controlling the selection and provision timing of information is provided.

DESCRIPTION OF SYMBOLS 100 Information provision system 101 Environmental information sensing part 102 Environmental state management part 103 Operation state management part 104 Information provision service part 105 Provision information integration part 106 Information output part 201 Provision information management part 202 Evaluation factor importance degree management part 203 Information importance degree evaluation Unit 204 information provision scheduling unit 205 route prediction unit 206 workload prediction unit 207 information output control unit

Claims (6)

An environmental information sensing unit that obtains environmental information about the situation surrounding the user;
An information provision service unit that generates provision information to the user based on the environment information;
A provision information integration unit that determines a priority order of the plurality of provision information according to a situation surrounding the user when there is a plurality of the provision information;
The information provision service unit attaches a plurality of evaluation factors and evaluation values related to the evaluation factors to each of the provision information and transmits the evaluation information to the provision information integration unit,
The provided information integration unit includes an evaluation factor importance management unit and an information importance evaluation unit,
The evaluation factor importance management unit changes the importance of each of the plurality of evaluation factors when the situation surrounding the user changes,
The information importance level evaluation unit changes the importance level of the plurality of provided information based on the changed evaluation factor importance level, and determines the priority order of the provided information in the order of the changed provided information importance level. An information providing system characterized by that.   The provision information integration unit provides a schedule for providing information so that the number of pieces of information satisfying a predetermined condition or the sum of their importance levels is maximized within a range satisfying a provisionable condition of the predetermined information. The information providing system according to claim 1, further comprising an information scheduling unit.   The information providing system according to claim 2, wherein the range satisfying the provision condition of the predetermined information is within a predetermined temporal range and a spatial range.   The information provision according to claim 2 or 3, wherein the number of pieces of information satisfying the predetermined condition or the sum of the degrees of importance thereof is the number of pieces of information that can be provided to the user or the sum of the degrees of importance of a plurality of pieces of information to be provided. system. The evaluation factor importance management unit includes an evaluation factor importance learning unit that adaptively adjusts the evaluation factor importance to an environment surrounding the user,
The evaluation factor importance degree learning unit uses the evaluation factor importance degree as a reward for the number of pieces of information provided within the predetermined time range and the spatial range or a sum of importance levels of a plurality of pieces of information provided. The information providing system according to any one of claims 2 to 4, wherein learning is performed by changing an importance of an evaluation factor so that an expected value of all reward values over the future of the reward is maximized. The evaluation factor importance management unit includes an evaluation factor importance learning unit that adaptively adjusts the evaluation factor importance to an environment surrounding the user,
The evaluation factor importance degree learning unit manages a plurality of evaluation factor sets (policies) and performs policy learning for selecting an evaluation factor set most suitable in an environment surrounding a user. Information providing system described in 1.

Images