JP2012216039A - Meta-information generation device and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a meta-information generation device and program that can generate meta-information about a position of a subject to be monitored from sensor information in accordance with the situation of the subject.SOLUTION: A meta-information generation device comprises: multiple meta-information generation modules for receiving input of sensor information from multiple sensor nodes, processing the sensor information in accordance with different algorithms, and generating meta-information of equivalent quality; a meta-information integration module for receiving input of meta-information from the meta-information generation modules, processing the meta-information in accordance with one algorithm, and generating the maximum likelihood meta-information; and a database for accumulating maximum likelihood meta-information. The sensor information includes information about a subject to be monitored, and the meta-information includes positional information for the subject to be monitored.

Description

  The present invention relates to a meta information generation device and a program.

  A meta information generation apparatus that generates context or meta information from sensor information detected by a plurality of sensors is known. The generated meta information can be used for building maintenance, hospital patient management, and the like. Building maintenance includes, for example, air conditioning management and control. On the other hand, patient management in a hospital includes, for example, management of the patient's condition or posture.

  However, as the types and number of sensors increase, it becomes necessary to process a huge amount of sensor information, and it is difficult to generate highly reliable meta information with a relatively simple system configuration.

JP-T-2003-527713

  In the conventional meta information generation device, there is a problem that it is difficult to generate meta information related to the user's position and the like from the sensor information according to the situation of the user who is the monitoring target.

  Therefore, an object of the present invention is to provide a meta information generation device and a program that can generate meta information related to the position of a monitoring target from sensor information according to the status of the monitoring target.

  According to one aspect of the present invention, sensor information from a plurality of sensor nodes is input, a plurality of meta information generation modules that generate meta information of the same quality by processing the sensor information according to different algorithms, and the plurality The meta information from the meta information generation module is input, and the meta information integration module that generates the most likely meta information by processing the meta information according to a certain algorithm and the database that accumulates the most likely meta information are formed There is provided a meta information generating apparatus, wherein the sensor information includes information of a monitoring target, and the meta information includes position information of the monitoring target.

  According to one aspect of the present invention, a computer program for generating meta information is obtained by inputting sensor information from a plurality of sensor nodes and processing the sensor information according to algorithms different from each other. A plurality of meta information generation modules to be generated, and meta information from the plurality of meta information generation modules are input, and the meta information integration module that generates the most likely meta information by processing the meta information according to a certain algorithm A computer is made to function, the most likely meta information is stored in a storage unit forming a database, the sensor information includes information on a monitoring target, and the meta information includes position information on the monitoring target, etc. A program is provided.

  According to the disclosed meta information generation apparatus and program, it is possible to generate meta information related to the position of the monitoring target from the sensor information according to the status of the monitoring target.

It is a figure which shows an example of the meta information production | generation apparatus in one Example of this invention. It is a figure which shows an example of the kind of information which each meta information generation module of a meta information generation module group subscribes. It is a figure explaining an example of information processing which generates meta information. It is a figure which shows an example of the system with a feedback loop. It is a figure which shows the reliability of the time stamp added and information, and meta-information. It is a figure explaining an example of composition of a sensor node in a office, a meta information generating device, and a layout of an office desk. It is a figure explaining the meta-information production | generation process regarding the office worker in an office. It is a figure which shows an example of the meta-information production | generation apparatus using the communication part of a Client-Server system. It is a figure explaining an example of composition of a meta information generating device in a hospital, and a layout, such as a bed. It is a figure which shows an example of the meta-information production | generation apparatus in the other Example of this invention. It is a flowchart explaining an example of operation | movement of a fall detection module.

  In the disclosed meta information generation apparatus and program, sensor information from a plurality of sensor nodes is input, and the sensor information is processed according to different algorithms to allow a plurality of meta information generation modules to generate the same quality meta information. To do. Meta information from a plurality of homogeneous or heterogeneous meta information generation modules is input, and the meta information integration module generates the most likely meta information by processing the meta information according to a certain algorithm. The most likely meta information is stored in a database. The sensor information includes monitoring target information, and the meta information includes monitoring target position information.

  Hereinafter, embodiments of the disclosed meta information generation apparatus and program will be described with reference to the drawings.

  FIG. 1 is a diagram illustrating an example of a meta information generating apparatus according to an embodiment of the present invention. The meta information generating apparatus 100 shown in FIG. 1 can be formed by a general-purpose computer system including a processor such as a CPU (Central Processing Unit) and a storage unit. When the CPU executes the program stored in the storage unit, the computer system functions as a meta information generation device, and meta information generation processing is executed. The storage unit can be formed by a computer-readable storage medium, and may include various databases described later. Needless to say, the storage unit may be formed of a plurality of storage devices.

  In this example, a publish-subscribe (Publish-Subscribe) message communication unit is used for message communication between modules constituting the meta information generating apparatus 100, and in the following description, it is described as Pub / Sub. The Pub / Sub (publishing subscription type) messaging model is a messaging model for sending and receiving messages between modules of a publisher (Publisher) that sends a message and a subscriber (Subscriber) that receives the message. As a specific implementation of Pub / Sub, there is JMS (Java (registered trademark) Message Service). In the Pub / Sub messaging model, a message sender (Publisher) sends a message without assuming a specific receiver (Subscriber). Also, the receiving side receives only the specified message and has no knowledge about the transmitting side. As described above, since the Pub / Sub model has a low degree of coupling between the message transmitting side and the receiving side, it is easy to add Publisher or Subscriber, and the meta information generating apparatus 100 can be configured dynamically. In the meta information generating apparatus 100, when adding meta information (or context) to be extracted later, there are cases where it is desired to increase the types of necessary sensors or increase the number of modules that receive the sensor information and generate meta information. is there. By using the Pub / Sub model, the meta information generating apparatus 100 can be configured in a scalable and dynamic manner.

  In FIG. 1, for example, sensor information is published from the sensor nodes 1 to N (N is a natural number of 2 or more) of the sensor node group 101. FIG. 2 is a diagram illustrating an example of types of information subscribed to by each of the meta information generation modules 2-1 to 2-M (M is a natural number of 2 or more) of the meta information generation module group 102. The types of information to be subscribed to include on / off information including a combination of multiple types of on / off events, continuous information such as time series signals, a priori information, It includes past data and / or information mined from past data and combinations of the above information. A priori information is stored in the a priori database 111.

  Specific examples of the on / off information include, for example, information that a PC (Personal Computer) keylogger has reacted / stopped, and that the proximity sensor is on / off. The continuous amount information includes, for example, sensor information obtained by sampling a time series signal such as an acceleration sensor, data obtained by sampling a time series audio signal collected by a microphone, a moving image taken by a camera, and the like. The a priori information includes, for example, building structure information, layout information such as desks and shelves in the office, information that the passage is a straight line or passage A and passage B intersect, and a schedule. As an example of past data and / or information mined from past data, there is a place where a user was present between time t and time t + α (α> 0), a travel route frequently used by a certain user, and the like.

  The meta information class C1 illustrated in FIG. 1 includes a meta information generation module group 102 and a meta information integration module 104. The meta information class C2 includes meta information generation modules 22-1 to 22-K (K is a natural number of 2 or more) and a meta information integration module 124.

  The meta information generation module group 102 has a function of generating meta information by processing sensor information according to a certain rule or algorithm, and publishing the generated meta information. The combination of sensor information input to each meta information generation module 2-1 to 2-M may be different. Further, the rules or algorithms used in the meta information generation modules 2-1 to 2-M are different from each other. An example of information processing for generating meta information will be described with reference to FIG. FIG. 3 is a diagram illustrating an example of information processing for generating meta information. FIG. 3 shows processing when the type of information to be subscribed is on / off information, continuous amount information, a combination of on / off information and continuous amount, a priori information and on / off information combination, and mined information. The method and its specific example are shown. For example, when the type of information to be subscribed is on / off information, the processing method publishes information corresponding to on / off. A specific example of this processing method is that when the PC keylogger reaction is on, the position of the corresponding PC is the current position of the user having authority to use the PC, and the likelihood (or accuracy) that the user is at that position is 1.0. To output the information. When the response of the PC key logger is off, information indicating that the likelihood that the user is at the position of the corresponding PC is 0.0 is output. In this example, the likelihood is represented by a real value in the range of 0.0 to 1.0, but another representation such as a percentage may be used.

  Meta information generation modules 2-1 to 2-M that output meta information of the same quality are handled as one meta information class C1. Outputs of the meta information generation modules 2-1 to 2-M in the meta information class C1 are input to the meta information integration module 104. The meta information integration module 104 generates maximum likelihood meta information using outputs of the meta information generation modules 2-1 to 2-M based on a certain rule or algorithm.

  Other meta information generation modules 133 and 132 or meta information generation modules 131 and 132 may be connected in cascade to generate higher-level meta information and store it in the context database 105.

  A combination of the meta information generation modules may constitute a feedback loop as shown in FIG. FIG. 4 is a diagram illustrating an example of a system having a feedback loop. When the feedback loop is configured, the meta information generation module to which feedback information is input may become unstable in estimating meta information. A specific example of such a case will be described below with reference to FIG.

  In FIG. 4, position information (for example, x, y coordinates) of a certain monitoring target user is input to a module 1004 from a wireless positioning system (sensor node) 1001 such as GPS (Global Positioning System). This positioning system 1001 includes an observation error, for example, an observation error of 10 m. On the other hand, the module 1005 is a pedestrian (user) dead reckoning calculated on a passage detection sensor 1002 that detects that a user has passed and a smartphone with a motion sensor that detects motion (motion) of the user. Pedestrian (User) Dead Reckoning) information is input, and the positioning result of the monitoring target user estimated by the module 1005 is obtained with an error accuracy of 1 m. The module 1004 performs global position determination (ie, tracking processing) based on the wireless positioning system, and inputs an output to the module 1006. The module 1005 determines a global position based on the passing sensor 1002 and the dead reckoning 1003, and inputs an output to the module 1006. In this example, the module 1004 performs processing at a cycle of 1 second, and the module 1005 performs processing at a cycle of 10 seconds, which is longer than the module 1004.

  When the module 1006 outputs the output of the module 1005 as the confirmed position of the user, the output of the confirmed position of the user is fed back to the module 1004. When the information fed back from the module 1006 and the information from the sensor node 1001 are input to the module 1004, both of the input information are user position information (for example, x and y coordinates). At this time, the output of the sensor node 1001 is information of the latest time but has an error of 10 m, while the output of the module 1006 has the highest accuracy at a certain time but is older than the information of the sensor node 1001. Therefore, the module 1004 has a problem as to which information can be used to estimate the most reliable meta information. Therefore, in this embodiment, a time stamp and information reliability are added to the meta information output by the module.

  FIG. 5 is a diagram showing the added time stamp and the reliability of the information together with the meta information. In FIG. 5, the time stamp indicates the time when the meta information is generated. The reliability of information indicates the reliability of the information (for example, likelihood, certainty, error distribution, etc.). The meta information (or context) to be output (Published) is target meta information.

  When the same kind of information is fed back as input, the following processing is performed. In the case of the module 1004 in FIG. 4, the user coordinate xy is input from the sensor node 1001 and the module 1006. First, the time stamps are compared, and when the fed back information is older than the current time by a threshold or more, the feedback information is not used and the information of the sensor node 1001 is used. On the other hand, when the fed back information is less than the threshold from the current time, the reliability of the information is compared and the one with the higher reliability is adopted, or the weighted average is performed using the reliability as a weight. .

  The generated meta information is stored in the context (or meta information) database 105 or used by the application 106 that uses the meta information (context). Also, the meta information stored in the context database 105 is subjected to data mining processing in the data mining engine 112 to extract another meta information (or context), and the extracted meta information (or context) is used by the subscriber. To do. Meta information (or context) extracted by the data mining engine 112 may be accumulated in the mining data database 113.

  Next, an example of detecting the position of the office worker, the behavior such as having a conversation, the status of the office worker's business during a meeting, etc. will be described for the office worker. In the following description, an office worker is called a user. FIG. 6 is a diagram for explaining an example of the configuration of sensor nodes and meta information generation apparatuses in an office and the layout of an office desk. In FIG. 6, (a) shows the configuration of the sensor nodes and meta information generation apparatus in the office, and (b) shows the layout in the office.

  In FIG. 6A, the user 200 carries the mobile terminal 201. The mobile terminal 201 has a communication function such as Wi-Fi (Wireless Fidelity, registered trademark) and Bluetooth (registered trademark), and incorporates a motion sensor such as an acceleration sensor and a gyro sensor corresponding to a sensor node, and a microphone. The portable terminal 201 is connected to the server 206 via a wireless local area network (LAN) access point 211 and a network 210. As shown in FIG. 6B, it is assumed that the user A uses the PC 202 at his or her desk for business, and the installation location of the user A's PC 202 is fixed and known. The PC 202 is connected to the server 206 via the network 210. In addition, Bluetooth anchors 205 are installed in the conference room 215 and the conference table, respectively.

  FIG. 7 is a diagram illustrating a meta information generation process related to the office worker in the office of FIG. When the user A is operating the PC 102 at his / her seat in FIG. 6B, information that the PC key logger has reacted is published from the sensor node 301 in FIG. The module 305 subscribes to the published information, and converts the information that the key logger has reacted into position information that the user A is at his / her seat. Further, the portable terminal 201 carried by the user detects the Bluetooth adapter of the PC 102 and outputs a measured value of RSSI (Received Signal Strength Indicator). At this time, the human sensor corresponding to the sensor node connected to the PC 102 also reacts and outputs position information indicating that there is a person in front of the PC 102.

  Modules 305 to 307 convert the sensor information into the position of user A, which is meta-information, and outputs it in accordance with each logic. In the module 308, position information having a high likelihood is selected from the outputs of the modules 305, 306, and 307. When the user A is at his / her seat but does not operate the PC 102, only the module 306 outputs position information. As a result, the module 308 determines that the user A is present. When the user A moves to the conference room 215, the portable terminal 201 carried by the user A detects the Bluetooth anchor 205 of the conference room 215, and the module 307 determines that the user A is in the conference room 215. In this way, the user's position, that is, meta information can be estimated from available sensors according to the user's situation.

  In the example of FIG. 7, the output of the module 305 indicates that the position of the user A is his / her seat and the likelihood is 1.0. The output of module 306 indicates that user A's position is his seat and the likelihood is 0.8. The output of module 307 indicates that the location of user A is conference room 215 and the likelihood is 0.7. The output of module 308 outputs information indicating the position and likelihood of user A having the highest likelihood.

  The module 310 uses the position information of each of the users A and B, and determines that both are in the conference if both are in the conference room 215. Thus, further meta information can be generated using meta information (position information) generated from sensor information.

  In the module 311, it is determined whether or not the two are talking from the position information of the users A and B and the detection information of the utterance. As a result, it is possible to estimate a state of “whether or not talking” that cannot be observed directly by the sensor.

  The module 309 generates and outputs the location information class of the user B based on the sensor information from the sensor node. The module 309 determines the latest position information with the highest likelihood as the position information of the user B.

  The output information of the module 310 or the module 311 may be stored in the user context database 312. The position information and the like in the user context database 312 may be read out and returned (Reply) with respect to a search (Query) from the position visualization application 313.

  In the above embodiment, the message communication unit (Pub / Sub) of the Publish-Subscribe method is used, but it goes without saying that, for example, a client-server communication unit may be used in the modified example. FIG. 8 is a diagram illustrating an example of a meta information generation apparatus using a client-server communication unit. In FIG. 8, sensor nodes 2001 to 2003 correspond to the sensor nodes 301 to 303 in FIG. 7, and modules 2004 to 2007 correspond to the modules 305 to 308 in FIG. As shown in FIG. 8, since the communication unit Client / Server is provided instead of the communication unit Pub / Sub, it is necessary to provide one Client, that is, one unique communication port for one Server. Therefore, when the additional module 2008 is additionally provided in FIG. 8, it is necessary to provide the module 2007 with a unique (that is, dedicated) communication port (Client) for the additional module 2008.

  On the other hand, in the above embodiment using the Pub / Sub message communication unit, addition of a new sensor and addition of a meta information estimation module are easier than in the above modification, and the extensibility is excellent. A comparison between the case of using Pub / Sub and the case of configuring the meta information generation apparatus without using Pub / Sub will be described below.

(1) Ease of network configuration between modules
In the case of the Client / Server method, after deciding on a logical network, it is necessary to make a physical connection accordingly. In other words, it is necessary to configure a Client / Server between two modules. In addition, it is necessary to perform programming by grasping the position (for example, server address) of the server on the network.

  On the other hand, in the case of the Pub / Sub method, it is only necessary to define a message that each Subscriber subscribes, and it is not necessary to be aware of what message each Publisher outputs. In other words, after deciding the logical connection relationship such as cascade / feedback connection, it is only necessary to define messages that each module pub / sub publishes. Subscribers can also receive messages without worrying about the Publisher network location. That is, since it is not necessary to know the address of a computer with Publisher, it is easy to construct a distributed system without worrying about the address and number of computers.

(2) Module extensibility
In the case of the Client / Server method, one unique communication port needs to be assigned to one Server. For this reason, as the number of modules increases, port management becomes complicated. In addition, each time a module (Server) is added, it is necessary to determine the communication protocol with the Client connected to it. For example, when a module 2008 is added in FIG. 8, it is necessary to newly determine a communication protocol with the module 2007.

  On the other hand, in the case of the Pub / Sub method, when adding a sensor node, a message (Publish information) flowing from the node to the network may be added. When adding a meta information generation module, a message subscribed to by the module may be defined, and a message published from the module may be added. Therefore, the system designer or developer does not need to manage port resources like Client / Server.

(3) Setting message priority and expiration date
In the case of Client / Server method, Client receives messages in the order sent by Server. Developers need to write their own programs if they want to prioritize and process messages from multiple servers or set expiration dates.

  On the other hand, in the case of the Pub / Sub method, the Subscriber can set the priority and expiration date of the message to be processed. For example, processing such as not using information older than a certain threshold from the current time can be performed only by setting in the Subscriber.

  Next, an example of the meta information generation device in another embodiment of the present invention will be described with reference to FIG. In this example, in the hospital, it is detected that the patient has fallen from the bed, and a warning is sent to a nurse who is close to the patient and has a free hand.

  FIG. 9 is a diagram for explaining an example of the configuration of a meta information generation apparatus in a hospital and a layout of beds and the like. In the hospital 600, a server (computer) 602 and a communication device 603 are connected via a network 601. The communication device 603 can communicate with the sensor information uploader 607, the portable terminal 612 of the nurse 613 (A), and the portable terminal 616 of the nurse 617 (B). Hospital patient 109 (α) is hospitalized in hospital room 610. In the hospital room 610, a plurality of (only two shown) ranging sensors 604, a mat sensor 605, a motion detection human sensor 606, a sensor information uploader 607, and a hospital bed 608 are provided. In this example, there is a nurse A who carries a portable terminal 612 in a place 614 where a Bluetooth anchor 611 is installed, and a portable terminal 616 is carried in a place 618 where an infrared (IR) communication anchor 615 is installed. Assume that nurse B is present.

  FIG. 10 is a diagram illustrating an example of a meta information generation apparatus according to another embodiment of the present invention. A meta information generation apparatus 700 shown in FIG. 10 can be formed by a general-purpose computer system including a processor such as a CPU and a storage unit. When the CPU executes the program stored in the storage unit, the computer system functions as a meta information generation device, and meta information generation processing is executed. The storage unit can be formed by a computer-readable storage medium, and may include various databases described later. Needless to say, the storage unit may be formed of a plurality of storage devices. In this example, a publish-subscribe (Publish-Subscribe) message communication unit is used for message communication between modules constituting the meta-information generating apparatus 700, but the Client / Server method is also used. good.

  In FIG. 9, it is assumed that the patient α falls from the bed 608. In this case, the human sensor 606 is turned on in response to the movement when the bed 608 falls. At this time, the distance measuring sensor 604 installed under the bed 608 outputs the distance to the patient α. The above sensor information is published from each of the sensors (ie, sensor nodes) 701 to 704, and the fall detection module 713 subscribes. The sensor nodes 701 and 702 publish information on the distance measuring sensor 604, the sensor node 703 publishes information on the mat sensor 605, and the sensor node 704 publishes information on the human sensor 606. The fall detection module 713 determines whether or not the patient α has fallen from the bed 608, for example, according to the procedure shown in FIG.

  FIG. 11 is a flowchart for explaining an example of the operation of the fall detection module. In FIG. 11, step S1 determines whether or not the human sensor 606 is on, and if the determination result is YES, step S2 determines whether or not the mat sensor 609 is on. If the decision result in the step S2 is YES, a step S3 decides whether or not the continuous on time of the mat sensor 609 is greater than or equal to a threshold value. If the decision result in the step S3 is YES, a step S4 decides whether or not any measured value of the plurality of distance measuring sensors 604 is equal to or less than a threshold value. If the decision result in the step S4 is YES, a step S5 decides whether or not the continuous reaction time of the distance measuring sensor 604 is a threshold value or more. If the determination result of any of steps S1 to S5 is NO, the process returns to step S1. On the other hand, if the decision result in the step S5 is YES, a step S6 publishes meta information that the patient has fallen from the bed, and the process is ended.

  Note that the rules used in the processing of the fall detection module 713 are not limited to the rules of FIG. An output pattern of sensor information from the sensor nodes 701 to 704 at the time of falling is accumulated in the sensor log 718, a sensor output pattern at the time of falling is extracted from the sensor log 718 by the falling pattern analysis module 719, and the falling based on the sensor output pattern The determination method may be implemented in the fall detection module 713. For example, a sensor output at the time of falling may be modeled by a Bayesian network or a Hidden Markov Model (HMM) to learn a state transition probability.

  The meta information that the patient published by the fall detection module 713 has fallen from the bet is subscribed to the context database 720. The context database immediately notifies the fall warning application 721 when notifying the meta information that the patient has fallen from the bet (Notify). When the fall warning application 721 receives notification that the patient has fallen from the bed, the fall warning application 721 searches the context database 720 for a nurse who is free and closest to the fallen patient α (Query).

  The portable terminal 612 carried by the nurse A or the portable terminal 616 carried by the nurse B is in the detection area of the Bluetooth anchor 611 or the IR anchor 615 where the nurse A or B is installed in the environment. When entering, it has a function of detecting an anchor. The portable terminal 612 includes a sensor node 705 on which a publisher that publishes anchor detection information is mounted. The installation position of each anchor is known, and the position estimation modules 714 and 716 of the nurses A and B can refer to information on the known installation positions. The mobile terminals 612 and 616 have built-in sensors that capture the movements of the nurses A and B, such as an acceleration sensor and a gyro sensor, and the degree of activity (activity) of the nurses A and B from these sensor information. And the function of measuring the amount of movement due to walking motion by dead reckoning (inertial navigation). The portable terminal 612 includes a sensor node 706 on which a publisher that publishes a movement amount due to dead reckoning is mounted, and a sensor node 707 on which a publisher that publishes the activity amount of the nurse A is mounted.

  In this embodiment, the positions of nurses A and B are estimated as follows. The position estimation module 714 subscribes the anchor detection event from the sensor node 705 and the movement amount from the sensor node 706. The position estimation module 714 has a function of referring to anchor information whose installation position is known. Usually, since the movement amount obtained by dead reckoning is a relative movement amount, the position estimation module 714 obtains the current position by adding the relative movement amount by dead reckoning from the position when the anchor detection event is detected. The estimated position is subscribed to the context database 720, and the position information of the nurse A is registered in the database.

  Whether the nurse A's hand is available (Busy / Available) is determined by the state determination module 715. Here, nurse A will be described. The state determination module 715 subscribes the position of the nurse A from the position estimation module 714, the activity amount of the nurse A from the sensor node 707, and the schedule of the nurse A from the scheduler 708. The state determination module 715 subscribes from the position estimation module 714 that the position of the nurse A is the patient's β room, subscribes from the schedule 708 that nursing is scheduled for the patient β at the current time, and the sensor node Suppose from 707 that the activity amount is high. In this case, the state determination module 715 determines that the nurse A is nursing the patient β, that is, that the hand is not available, and publishes that the state of the nurse A is Busy. The context database 720 subscribes to this information and sets the status of nurse A as Busy.

  If nurse B is currently moving in the corridor and no nursing schedule is scheduled, state determination module 717 determines that nurse B is free and states that nurse B is available. Publish. The context database 720 subscribes to this information and sets the status of nurse B as Available.

  When the fall warning application 721 receives notification of an event that the patient has fallen from the context database 720, the fall warning application 721 immediately queries the context database 720 for the position and status of the nurse (Query). The context database 720 responds to the fall warning application 721 with the available nurses and their positions (Reply). The fall warning application 721 selects a nurse closest to the fallen patient from among the nurses who responded, and notifies the nurse's mobile terminal of the fall alarm.

  The portable terminal 616 of the nurse B includes a sensor node 710 on which a publisher that publishes the movement amount due to dead reckoning is mounted, and a sensor node 711 on which a publisher that publishes the activity amount of the nurse B is mounted. The schedule for nurse B is published from scheduler 712. Since the sensor nodes 710 to 711 and the scheduler 712 related to the nurse B may have the same configuration as the sensor nodes 706 to 708 related to the nurse A, the description thereof is omitted.

The following additional notes are further disclosed with respect to the embodiment including the above examples.
(Appendix 1)
A plurality of meta information generation modules that receive sensor information from a plurality of sensor nodes and generate the same type of meta information by processing the sensor information according to different algorithms;
Meta information from the plurality of meta information generation modules is input, a meta information integration module that generates the most likely meta information by processing the meta information according to a certain algorithm,
A storage unit for forming a database for storing the most likely meta information;
The meta information generating apparatus according to claim 1, wherein the sensor information includes information on a monitoring target, and the meta information includes position information on the monitoring target.
(Appendix 2)
The plurality of meta information generation modules include a first meta information generation module that generates first meta information to which first sensor information is input and a time stamp and reliability of information are added, and second sensor information includes A second meta information generation module for generating second meta information to which the time stamp and the reliability of information are added;
A determination module that generates and outputs information about the confirmed position of the monitoring target from the first and second meta information;
The first meta information generation module compares the time stamp of the current time added to the first meta information with the time stamp of the information fed back from the determination module, and the information is older than the current time by a threshold or more Outputs the first sensor information without using the fed back information, and if the fed back information is less than the threshold from the current time, it is added to the reliability of the information and the first meta information The meta information generation apparatus according to appendix 1, wherein the reliability is compared and the higher reliability is adopted, or the two reliability levels are weighted averaged and output.
(Appendix 3)
The type of sensor information includes on / off information including a combination of a plurality of types of on or off events, continuous information, a priori information, past data and / or information mined from past data, and combinations of these information The meta information generation device according to appendix 1 or 2, characterized by including at least one of the above.
(Appendix 4)
A first meta information generation module to which the sensor information is input;
A second meta information generation module cascaded with the first meta information generation module;
The meta information generating apparatus according to appendix 1, wherein the upper meta information generated by the second meta information is stored in the database.
(Appendix 5)
The meta information according to any one of appendices 1 to 4, wherein the most likely meta information estimates an event that cannot be directly observed by a sensor that outputs the sensor information by combining a plurality of sensor information. Generator.
(Appendix 6)
The plurality of sensor nodes, the plurality of meta-information generation modules, and the meta-information integration module each use a publish-subscribe message communication unit for message communication between modules. The meta information generation device according to any one of appendices 1 to 5.
(Appendix 7)
A program that causes a computer to generate meta information,
A plurality of meta information generation modules that receive sensor information from a plurality of sensor nodes and generate the same type of meta information by processing the sensor information according to different algorithms;
Meta information from the plurality of meta information generation modules is input, and the computer functions as a meta information integration module that generates the most likely meta information by processing the meta information according to an algorithm,
Accumulating the most likely meta information in a storage unit forming a database;
The program according to claim 1, wherein the sensor information includes information on a monitoring target, and the meta information includes position information on the monitoring target.
(Appendix 8)
The plurality of meta information generation modules include a first meta information generation module that generates first meta information to which first sensor information is input and a time stamp and reliability of information are added, and second sensor information includes A second meta information generation module for generating second meta information to which the time stamp and the reliability of information are added;
Causing the computer to further function as a determination module that generates and outputs information on the confirmed position of the monitoring target from the first and second meta information;
The first meta information generation module compares the time stamp of the current time added to the first meta information with the time stamp of the information fed back from the determination module, and the information is older than the current time by a threshold or more Outputs the first sensor information without using the fed back information, and if the fed back information is less than the threshold from the current time, it is added to the reliability of the information and the first meta information 8. The program according to appendix 7, wherein the reliability is compared and the higher reliability is adopted, or the weighted average is calculated with the two reliability levels as weights.
(Appendix 9)
The type of sensor information includes on / off information including a combination of a plurality of types of on or off events, continuous information, a priori information, past data and / or information mined from past data, and combinations of these information The program according to appendix 7 or 8, wherein at least one of the programs is included.
(Appendix 10)
A first meta information generation module to which the sensor information is input;
Further causing the computer to function as a second meta information generation module cascaded with the first meta information generation module;
The program according to appendix 7, wherein upper meta information generated by the second meta information is stored in the database.
(Appendix 11)
The program according to any one of appendices 7 to 10, wherein the most likely meta information estimates an event that cannot be directly observed by a sensor that outputs the sensor information by combining a plurality of pieces of sensor information.
(Appendix 12)
The plurality of sensor nodes, the plurality of meta-information generation modules, and the meta-information integration module each use a publish-subscribe message communication unit for message communication between modules. The program according to any one of appendices 7 to 11.

  Although the disclosed meta information generating apparatus and program have been described with the embodiments, it is needless to say that the present invention is not limited to the above embodiments, and various modifications and improvements can be made within the scope of the present invention. Yes.

1-1 to 1-N Sensor nodes 2-1 to 2-M Meta information generation module 104 Meta information integration module 105 Context database 112 Data mining engine

Claims (5)

A plurality of meta information generation modules that receive sensor information from a plurality of sensor nodes and generate the same type of meta information by processing the sensor information according to different algorithms;
Meta information from the plurality of meta information generation modules is input, a meta information integration module that generates the most likely meta information by processing the meta information according to a certain algorithm,
A storage unit for forming a database for storing the most likely meta information;
The meta information generating apparatus according to claim 1, wherein the sensor information includes information on a monitoring target, and the meta information includes position information on the monitoring target. The plurality of meta information generation modules include a first meta information generation module that generates first meta information to which first sensor information is input and a time stamp and reliability of information are added, and second sensor information includes A second meta information generation module for generating second meta information to which the time stamp and the reliability of information are added;
A determination module that generates and outputs information about the confirmed position of the monitoring target from the first and second meta information;
The first meta information generation module compares the time stamp of the current time added to the first meta information with the time stamp of the information fed back from the determination module, and the information is older than the current time by a threshold or more Outputs the first sensor information without using the fed back information, and if the fed back information is less than the threshold from the current time, it is added to the reliability of the information and the first meta information 2. The meta information generating apparatus according to claim 1, wherein the reliability is compared and the higher reliability is adopted, or the two reliability levels are used as weights and a weighted average is output.   The meta information generation apparatus according to claim 1, wherein the maximum likelihood meta information estimates an event that cannot be directly observed by a sensor that outputs the sensor information by combining a plurality of sensor information.   The plurality of sensor nodes, the plurality of meta-information generation modules, and the meta-information integration module each use a publish-subscribe message communication unit for message communication between modules. The meta information generation device according to any one of claims 1 to 3. A program that causes a computer to generate meta information,
A plurality of meta information generation modules that receive sensor information from a plurality of sensor nodes and generate the same type of meta information by processing the sensor information according to different algorithms;
Meta information from the plurality of meta information generation modules is input, and the computer functions as a meta information integration module that generates the most likely meta information by processing the meta information according to an algorithm,
Accumulating the most likely meta information in a storage unit forming a database;
The program according to claim 1, wherein the sensor information includes information on a monitoring target, and the meta information includes position information on the monitoring target.

Images