JP5943476B2 - Sensor network system and data acquisition method in sensor network system - Google Patents

Description

  The present invention relates to a sensor network system and a data acquisition method in the sensor network system.

  Sensor network systems capable of acquiring various environmental information have been developed. Such a sensor network system includes a plurality of small devices called sensor nodes each having one or more sensors and having a function of transmitting sensing data acquired by the sensors to the outside using wireless or wired communication.

  Sensing data emitted from each sensor node is sent to a node that collects sensing data called a base station using wireless or wired communication. Thereby, the sensor network system which can acquire environmental information in a base station is constructed | assembled.

  In this way, the current sensor network technology focuses on a method that unilaterally aggregates data acquired by each sensor on the user (synonymous with user) side.

  However, in a large-scale sensor network system, the sensing data becomes enormous and it is difficult to accurately and quickly acquire the environmental information desired by the user.

  In order to solve this, the inventions described in Patent Document 1 and Patent Document 2 have been proposed.

  The invention described in Patent Document 1 is a configuration in which a management database is provided in which a distributed database (DB), data held by each distributed DB, and an identifier (ID) of the distributed DB are grouped and managed.

  In such a configuration, the user sends a search key to the management DB. The management DB searches the management table with the search key sent from the user. As a search result, the ID of the distributed DB that owns the desired data is known, and the data is requested from the corresponding distributed DB. Then, the data acquired from the distributed DB is presented to the user as a search result via the management DB.

  On the other hand, in the invention described in Patent Document 2, a script manager that interprets a script is provided in each sensor node or base station. A script that describes the user's expected behavior is sent from the base station to the sensor node, and the script manager interprets the script to dynamically change the behavior of the sensor node, and the desired sensing data or related events. To the base station.

  Thereby, the user does not only acquire sensing data transmitted from the sensor node. That is, it is possible to generate and receive an event notifying that a certain event has occurred by changing the sensing method of the sensor node or the size of the acquired sensing data. Therefore, even in a large-scale sensor network, the data amount can be suppressed and the sensor network can be customized.

JP 2011-118950 A JP 2011-198379 A

N. Otsu, “Towards Flexible and Intelligent Vision Systems-From Thresholding to CHLAC-,” MVA2005 IAPR Conference on Machine Vision Applications, 12-1, May 16-18, 2005 Tsukuba Science City, Japan. Http: //www.mva -org.jp/Proceedings/CommemorativeDVD/2005/papers/2005430.pdf (accessed 2012.7.26) Sylvia Ratnasamy, Paul Francis, Mark Handley, Richard Karp, Scott Schenker (2001). “A scalable content-addressable network”. Proceedings of the 2001 conference on Applications, technologies, architectures, and protocols for computer communications (New York, NY, USA: ACM Press): 161-172.doi: 10.1145 / 383059.383072. T. Hayashi, H. Fukuhara, R. Fuijta, T. Miyazaki, and S. Saito, "A Messaging Network to Realize an SOA-Based System,". In Proceedings of the 7th IEEE international Conference on Computer and information Technology (October 16-19, 2007). CIT. IEEE Computer Society, pp. 1083-1088.

  However, in the invention described in Patent Document 1, the sensing data acquired by the search is a simple numerical value acquired by the sensor. For this reason, the user terminal needs to perform conversion into a format that can be understood by the user. If the acquired data is enormous, the data processing load on the user terminal increases.

  In addition, whenever there is a change in the data type to be acquired, etc., the data processing program of the user terminal must be changed. If there are many users, the program stored in the terminal owned by those users must be changed. There is a problem that it takes enormous effort to change everything.

  In the invention of Patent Document 2, the upper node interprets a script set in advance in the client and extracts a script for the lower node. Next, the extracted script is sequentially distributed to the lower nodes. Therefore, it is difficult for a user to write a script unless he / she knows which node is connected next to which node to form a hierarchical relationship between the nodes.

  Furthermore, it is not easy to define a complex action with a script, and it is not efficient to dynamically interpret and execute it at each sensor node. Each script is assigned a unique identifier (ID) and managed, and even if multiple scripts are submitted at the same time, it is described how to process without confusion, but the script manager's memory management method is single memory. Is implied.

  This makes it relatively easy for a malicious user to erode the memory area used by another person's script, to perform actions that the original user did not intend, and to intercept the acquired data. The method of mounting a script manager on each sensor node proposed in Patent Document 2 has a security vulnerability.

  Accordingly, an object of the present invention is a sensor network system that solves the problems in the inventions disclosed in Patent Documents 1 and 2 above, taking into consideration the surrounding situation and the requests made by the user in the sensor network itself. An object of the present invention is to provide a sensor network system that can dynamically change functions and acquire desired sensing data.

  The sensor network system according to the present invention that achieves the above object is characterized in that, as a first aspect, the function can be changed according to each task set with at least one sensor, and sensing data detected by the sensor is used as the function. A plurality of sensor nodes having information processing function units to be generated according to the above, a means for generating a task corresponding to a scenario based on a user request given by a user, and the generated task corresponding to the plurality of sensor nodes It has a communication means to transfer to a sensor node.

  In the sensor network system according to the present invention that achieves the above object, in the first aspect, the information processing function unit includes a CPU, logic variable hardware including an FPGA (Field Programmable Gate Array), or both. The task is set as a program executed by the CPU, logic circuit information realized by logic variable hardware, or both.

  Furthermore, it has at least one base station, and the task generated by the means for generating the task is transferred from the base station to the plurality of sensor nodes by radio.

  Furthermore, each of the plurality of sensor nodes is detected by a sensor and transfers information processed by the function of the information processing function unit to the base station by a multi-hop communication method.

In the sensor network system according to the present invention for achieving the above object, in the first aspect, the means for changing the function of the information processing function unit is a task library in which basic operations executable by the plurality of sensor nodes are registered. And a scenario compiler that decomposes the operation specifications given by the user into tasks to be executed on each sensor node with reference to the task library, and generates a task that operates on the sensor node,
The communication function unit causes the task generated by the scenario compiler to be downloaded to a corresponding sensor node, and the information processing function unit of the sensor node is dynamically changed.

  The sensor network system according to the present invention that achieves the above object is further characterized in that, in the first aspect, the information processing function unit detected and changed in function by the sensor based on an operation specification first given by a user. A scenario generator for generating a next operation specification based on the processed information and the user's final goal; and changing the function of the information processing function unit based on the generated operation specification, and It is characterized by customizing the user.

  The sensor network system according to the present invention that achieves the above object has, as a second aspect, a plurality of sensor nodes, each sensor node having at least one sensor, and information on information detected by the sensors. A plurality of sensor networks, other existing information systems, the plurality of sensor networks, and the like, each having a processing function unit and a communication function unit for transmitting and receiving data, and capable of changing functions of the information processing function unit The demand addressable network has a demand addressable network connected to existing information systems. The demand addressable network breaks down information acquisition requests given by users into sub-requests consisting of simple parameter combinations. And a sub-request expanded by the means for expanding into the sub-request or its sub-request A means for obtaining a sensor network or an existing information system having the corresponding information in the network based on the matching, and a means for providing the user by integrating the data obtained by the sensor network or the existing information system in the network It is characterized by having.

  The sensor network system according to the present invention that achieves the above object, in the second aspect, has a sensor network that holds corresponding information in the network based on a sub-request or a combination thereof developed by the means for developing the sub-request. Alternatively, when there is no existing information system, an operation specification of the sensor network is generated so as to respond to an information acquisition request given by the user, and the information processing function unit is based on the operation specification of the generated sensor network. The function is changed.

  According to the present invention, since it is not necessary to process and display data acquired at each user terminal, it is not necessary to make any changes to each user terminal even when there is a change in the data to be acquired.

  Also, in the sensor network itself, a task composed of independent programs executed by each sensor node or logic circuit information of logic variable hardware (hereinafter, configuration data) is generated according to an operation scenario defined by the user. By sending the task to each sensor network, the system operation can be customized.

  Unlike the invention described in Patent Document 2, the present invention sends the task itself that operates independently, so that not only the above-described security vulnerability due to memory infringement can be eliminated, but also a non-volatile memory is mounted on the sensor node. However, if a task once sent is saved in the nonvolatile memory, the task itself is retained even if the power is turned off. Therefore, if the power is turned on again, the task before the power is turned off without sending the task from the outside again. The movement can be reproduced immediately.

  On the other hand, in the embodiment of the so-called interpreter system in which the script is interpreted each time described in Patent Document 2, even if the same task is executed every time the power is turned on, for example, before the power is turned off again Therefore, the problem remains that the script must be sent.

  According to the present invention, there is a possibility that one sensor network system can be shared in time according to the application and used for another application. This means that various services that realize a ubiquitous computing environment can use a sensor network system installed in a wide area as a common infrastructure.

It is a figure which shows the example structure of one Embodiment of the sensor network system according to this invention. It is a figure which shows the structural example of the sensor node which is a component of a sensor network. Data of one task generated by the scenario compiler (specifically, binary or character data). It is a figure showing the packet structure of a task transfer packet when sending the task data of FIG. 3 to a sensor node. The procedure (sequence diagram) in the case of transmitting task data from a task server to a sensor node using a task transfer packet is shown. It is a figure showing the structure of a task transfer start packet. It is a figure showing the structure of a task transfer response packet. It is a figure showing the structure of an execution start command packet. FIG. 1 is an example w of scenario description that is input to the scenario compiler. It is a figure which shows how each sensor node collects neighboring sensor node information. It is a figure showing an example of an embodiment in the present invention at the time of using sensor node 2 carrying a surveillance camera. It is a block diagram in the case of comprising a large-scale sensor network using two or more sensor networks of FIG. FIG. 13 is a diagram for explaining integration (mashup) of data in XML format with reference to FIG. 12. It is a figure which shows an example of the application constructed | assembled with the network according to this invention. It is a figure explaining another example of an embodiment, and is a figure showing the composition which added the scenario generator to the sensor network of Drawing 1.

  Embodiments of the present invention will be described below with reference to the drawings. However, the embodiments are for the purpose of understanding the invention, and the scope of protection of the present invention is not limited to such embodiments.

  FIG. 1 is a diagram showing the configuration of an embodiment of a sensor network system according to the present invention.

  In the following description, when each of a plurality of similar elements shown in the figure is indicated, a reference number is attached with a subscript, and when referring to the whole of the same element, only the reference number is attached without a subscript. To do.

In FIG. 1, it has a plurality of base stations 1 ₁ to 1 _p (shown from 1 ₁ to 1 _{3 in the} figure) and sensor nodes 2 ₁ to 2 _q (shown from 2 ₁ to 1 _{10 in the} figure). The sensor network 100 is configured by connecting each other by wireless communication.

The sensor nodes 2 _{1 to} 2 _q transfer the sensing data acquired by each sensor node 2 to one or more base stations 1 by the multi-hop communication method.

Sensing data that arrive to the base station 1 from the sensor node 2 via the network 3 for connecting the base station to each other, are aggregated to the user terminal 4 connected to one base station 1 ₃ in the final. Here, the base station network 3 for connecting the base stations 1 can be constructed using a standard wireless LAN such as IEEE802.11b. If a dynamic routing method such as the AODV (Ad hoc On-Demand Distance Vector) protocol, which is standardized as a routing protocol, is adopted, even if several base stations 1 die in the middle, communication between base stations is not possible. It can be maintained continuously.

Therefore, if a user terminal 4 having a wireless function that can participate in the same base station network 3 is introduced, the sensing data received by each of the base stations 1 ₁ to 1 _p is transferred to the terminal 4 so that all sensing data Can be aggregated in the user terminal 4 and the local database (local DB) 5.

  On the other hand, the sensor node network 6 composed of the sensor nodes 2 can be constructed by using a standard wireless LAN in the same manner as the base station 1 or constructed by using other standardized wireless methods such as the ZigBee method. it can.

  Similarly to the base station network 3, the sensor node network 6 dynamically secures routes from each sensor node 2 to a plurality of base stations 1 by using a dynamic routing protocol corresponding to multi-hop communication. be able to.

Further, by installing a plurality of base stations 1 ₁ to 1 _p , if the physical position of each base station 1 is known, the position estimation of each sensor node 2 whose position is unknown is based on that. Is possible. That is, since the plurality of base stations 1 ₁ to 1 _p constitute the base station network 3 and can exchange information with each other freely, the number of communication hops from a certain sensor node 2 to each base station 1 is set to the user terminal 4. Can be aggregated.

  The user terminal 4 estimates the position of each sensor node 2 by combining the information. As a specific position estimation method, a known method such as a three-point survey or a method based on the number of hops can be used. Further, in order to know the absolute position of each base station 1, it is possible to mount a GPS receiver on each base station 1. Of course, when the base station 1 is installed, if its position is known, it can be realized by setting the value directly in the base station.

  Further, a gateway device (GW) 7 is provided for connecting the sensor network 100 to the Internet or a wide area network described later.

  FIG. 2 shows a configuration example of the sensor node 2 that is a component of the sensor network 100. The sensor node 2 includes a wireless module 20, a logic variable hardware 21, a CPU 22, an ADC (analog / digital converter) 23, a sensor 24, a working memory 25, a program storage memory 26, a configuration memory 27, and a configuration circuit 28. These elements can exchange information with each other via the bus 29.

  The wireless module 20 transmits numerical information acquired by the sensor 24 and digitized through the ADC 23 as a wireless packet to the outside via the CPU 22 or the logic variable hardware 21 as control means. The wireless module 20 receives a wireless packet that arrives from the outside.

  The CPU 22 controls other components mounted on the sensor node 2 and controls the entire sensor node 2. The working memory 25 is a memory that stores a program currently running on the CPU 22, data acquired from the sensor 24, and the like, and is realized using a DRAM or the like.

  The program storage memory 26 is composed of a rewritable nonvolatile memory such as a flash memory and has a bank structure. In the example shown in FIG. 2, a 5-bank configuration is used, but a configuration having more banks may be used. Bank 0 has a “start bank storage area” for storing an operating system (OS) operating on the CPU 22, a common application program, and a bank number currently used.

  When the sensor node 2 is turned on or reset, the contents in the bank 0 are mechanically copied to the working memory 25, and the program counter of the CPU 22 is set to the top address in which they are copied in the working memory 25, and the OS And start application programs.

  Thereafter, the CPU 22 accesses the “start bank storage area” in the memory bank 0 of the program storage memory 26, knows the bank number in which the task program to be activated at present is stored, and from the program storage memory 26 into the bank. The stored task program is copied to the working memory 25. Next, the task is executed.

  Further, the CPU 22 can access an arbitrary bank of the program storage memory 26 at an arbitrary time, copy the contents to the working memory 25, and activate the task.

  The configuration memory 27 also has a bank configuration, and the logic circuit information of the logic variable hardware 21, that is, configuration data is stored in each bank. In FIG. 2, the configuration memory 27 has a five-bank configuration from bank 11 to bank 15, but the number of banks may be arbitrarily set according to the application. The configuration memory 27 is composed of a nonvolatile memory capable of writing and reading such as a flash memory.

  The configuration circuit 28 has a “start bank number register” 281 inside, and implements (configures) a logic circuit in the logic variable hardware 21 using configuration data in the bank indicated by the register 281.

  The configuration circuit 28 can be activated not only when the sensor node 2 is powered on or reset, but also at an arbitrary time according to an instruction from the CPU 22.

  Further, when only the CPU 22 does not use the logic variable hardware 21, the power supply to the logic variable hardware 21 may be stopped and logically disconnected from the bus 29 in order to reduce power consumption. it can. The contents of the “start bank number register” 281 in the configuration circuit 28 can be rewritten from the CPU 22 at an arbitrary time.

  If the configuration of the sensor node 2 in FIG. 2 is used, the program storage memory 26 holds a program operating on the CPU 22, and the behavior and function of the sensor node 2 can be changed by rewriting the contents.

  Even if the power is turned off, the program contents are retained in the program storage memory 26. When the power is turned on again, the programs stored when the power is turned off from the program storage memory 26 to the working memory 25 are copied, Start operation. As a result, the function is restored to the state before the power was turned off. Similarly, the function of the logic variable hardware 21 is also configured by the configuration data stored in the configuration memory 27 via the configuration circuit 28 and indicated by the start bank number storage register 281 in the logic variable hardware 21. Will be automatically reproduced.

  Here, the logic variable hardware 21 is used to efficiently execute a part of the processing performed by the CPU 22 with the hardware logic. For example, error correction encoding / decoding processing of a wireless packet, processing for determining whether a certain dangerous situation occurs using values of a plurality of acquired sensing data, and the like are included. The logic variable hardware 21 is configured using a device such as an FPGA (Field Rogrammable Gate Array) that can change a logic circuit function from the outside.

  Further, without providing the above-described configuration circuit 28, a program for configuring the variable hardware is started in the CPU 22, and the configuration data stored in the configuration memory 27 is used as a logic variable hardware. The logic function of the logic variable hardware 21 may be changed by loading the software 21.

  Further, each sensor node 2 has the following functions.

  (1) Control and control the own sensor 24 to determine and execute when and what type of sensor 24 is activated, or when and where the sensing data acquired by the sensor 24 is mounted and transmitted in a wireless packet And (2) a function of determining and executing whether to transfer a wireless packet received from the adjacent sensor node 2 to another sensor node 2 or the base station 1.

The entire sensor network 100 is configured as an aggregate of sensor nodes 2 ₁ to 2 _q and base stations 1 ₁ to 1 _p that autonomously execute the above-described operation. Therefore, by changing the function of the sensor node 2 in FIG. 2, even if the configuration of the sensor network 100 is physically the same, a method for acquiring environmental information and a method for transferring the acquired sensing data to the user terminal 4 are possible. Therefore, different functional operations can be realized.

  The present invention provides a function customization technique in order to actively release customization of such function operations to users. As shown in FIG. 1, a task server 10 as a means for changing the function of each sensor node 2 is connected to a sensor network 100.

  The task server 10 comprises a general-purpose computer such as a PC, and has a scenario compiler 11, a task library 12, and a scenario description 13 as a function customization technique, and a task group 14 including various tasks is obtained.

  The user accesses the task server 10 from the user terminal 4 through the sensor network 100, and gives the operation specifications to be fulfilled by the sensor node 2 to the scenario compiler 11 as the scenario description 13. The scenario compiler 11 interprets the scenario description 13 and generates a task group 14 to be executed by each sensor node 2 while referring to the task library 12.

  Here, the task library 12 is a collection of tasks that can be executed by each sensor node 2, and the scenario compiler 11 combines tasks registered in the task library 12 along the scenario description 13. Thus, a task group 14 to be executed in each sensor node 2 is generated.

The substance of the task is a program executed by the CPU 22 in the sensor node 2 described above or configuration data for logic variable hardware. The generated task, the base station the task server 10 to which the function customization technique has been implemented is connected wirelessly, for example, via the base station 1 _1, each sensor node using wireless communication from the base station 1 ₁ 2 and stored in the memory of each sensor node 2. As a result, each sensor node 2 executes a program newly stored in the memory 26 by the CPU 22 or uses configuration data newly stored in the configuration memory 27 to change the logic variable hardware. By changing the logic circuit 21, the function of the sensor node 2 is changed, and it becomes possible to operate so as to realize the scenario defined by the user.

  FIG. 3 shows data of one task group 14 generated by the scenario compiler 11 (specifically binary or character data, hereinafter referred to as task data) 30, which will be described later k (k>). This shows a breakdown when transmitting to the sensor node 2 in the task transfer packet of 0).

  In this example, the task data 30 is composed of k task transfer packets 1 to k, and it is assumed that the amount of data loaded in each is n bytes (n is an integer of n> 15). Therefore, the size of the entire task data is kn + 15 bytes. However, only the last k-th packet is 15 bytes.

  FIG. 4 shows a packet structure of the task transfer packet 40 when the task data 30 of FIG. 3 is sent to the sensor node 2.

  The communication packet header 41 is a portion used when data is transmitted / received between the sensor nodes 2 and the base station 1 in the sensor network 100 shown in FIG. 1. For example, when using TCP / IP communication, the IP header And corresponds to the TCP header. After the communication packet header 41 is a payload, that is, data. When sending task data, as shown in FIG. 4, it has a structure composed of a header 41 of a task transfer packet and subsequent task data 42, that is, a part of a program or configuration data.

  Next, the contents of each field of the header 41 of the task transfer packet will be described. A plurality of types of fields included in the header 41 are distinguished from each other by adding a suffix. The same applies to the description of the header configuration in the following embodiment.

In the header 41 of the task transfer packet, "Task transfer packet identifier" field 41 _1, the packet is a numerical value uniquely determined by the overall system to indicate that a task transfer packet. That is, the packet recipient, the "Task Transfer Packet Identifier" field 41 ₁ of the region, if the value of the corresponding, recognizes that the packet is a task transfer packet.

"Target sensor node ID" field 41 ₂ is the ID of the final destination sensor node of the packet. This packet is also assumed to reach the final destination sensor node via another sensor node or base station, that is, multi-hop. Therefore, to compare the present target sensor node ID 41 ₂ and his ID, if the same, the recipient will recognize that this task transfer packet is addressed to itself.

The "target memory" field 41 ₃ 0 or 1 is written at the transmission side. In the case of 0, it indicates that this packet carries task data to be stored in the program storage memory 26 of FIG. In the case of 1, it indicates that the task data to be stored in the configuration memory 27 in FIG. 2 is being carried.

"Memory bank number" field 41 ₄ illustrates a bank number or memory bank number in the configuration memory 27 in the program storage memory 26 of FIG. Based on the contents of the “target memory” field 41 ₃ and the “memory bank number” field 41 ₄ , which bank of which memory of the sensor node 2 should store a part of the task data carried by the packet is determined. I understand.

Here, "memory bank number" field 41 _4, leave promise starting with 1 or more, if when the memory bank number is 0, the sensor node 2 which has received, the packets to any bank unused Indicates that task data carried by can be stored.

  The memory bank that the receiving sensor node 2 actually stores is stored in the task transfer response packet whose structure is shown in FIG. 7 and sent back to the task server 10 side. I can know that.

"Sequence number" field 41 _5, the corresponding number of the task data in FIG. 3, the integer value of 1~k enters in this embodiment. "Total number of packets" field 41 _6, in this example a k, "data size" field 41 _7, n enters in this embodiment. However, only for the last k-th packet, the total number of packets is 15 instead of n.

  FIG. 5 shows a procedure (sequence diagram) when task data is transmitted from the task server 10 to the sensor node 2 using the task transfer packet 40.

  The task server 10 first sends a task transfer start packet 60 shown in FIG. 6 to the sensor node 2 (step S1).

Task transfer start packet 60 shown in FIG. 6, of the header 41 of the task transfer packet 40 in FIG. 4, except for changing the "Task Transfer Packet Identifier" field ₄₁₁ to "Task transfer start packet identifier" field 61 _1, It has the same header structure and there is no task data part.

"Target sensor node ID" field 61 ₂ opponent sensor node ID, "target memory" field 61 ₃ 0 or 1 as described above is set. "Memory bank number" field 61 ₄ specifies the usually 1 or more. If this is set to 0, it means that the bank number to be stored is left to the sensor node 2 as described above. "Sequence number" field 61 ₅ 0, "the total number of packets" field 61 _6, in sending task data illustrated in FIG. 3 is k, "data size" field 61 ₇ stores kn + 15.

  Upon receiving the task transfer start packet 60, the sensor node 2 confirms the internal state, and if task data can be accepted, sends a task transfer response packet 70 shown in FIG. 7 to the task server 10 (step S2).

Task transfers the response packet 70, of the header structure 41 of the task transfer packet 40 in FIG. 4, "the task transfer packet identifier" field 41 ₁ of the task transfer response packet to change to "Task transfer response packet identifier" field 71 ₁ header And has a structure in which the task data part 42 is changed to an unreceived packet number 72.

  The task transfer response packet 70 is used not only as a response to the task transfer start packet 60 in FIG. 6 but also as a response packet to the task transfer packet 40 in FIG. 4 as will be described later.

In the response to the task transfer start packet 60, each field of the header in the task transfer response packet 70 copies the information loaded in the task transfer start packet 60 as it is (71 ₂ to 71 ₆ ). However, the data size (71 ₇ ) is set to 0, and it is clearly indicated that the unreceived packet number area 72 is 0 bytes.

Further, the task transfer start packet 60, when "memory bank number" field 61 ₄ is 0, i.e., if an instruction of the task server 10, the "bank to be used is left to the sensor node 2" was, The sensor node 2 searches for a vacant bank. If there is a vacant bank, the sensor node 2 selects it, and if it does not exist, selects a bank that is allowed to be overwritten with new task data.

Then, store the number of the selected bank "memory bank number region" of the header field 71 _4. Thereafter, a task transfer response packet 70 is returned to the task server 10. If, if not possible the transfer of the task data 10, replies "memory bank number field" 71 ₄ 0.

Next, the task server 10 receives the task transfer response packet 70, other than the memory bank number 71 ₄ is 0, that is if it was possible to receive the task data, using the task transfer packet 40 in FIG. 4, the sensor The task data is transferred to the node 2 (steps S3 _{1 to} S3 _k ).

  The sensor node 2 waits to receive k packets, checks the sequence number of the task transfer packet 40 sent so far, and if there is a jump in the sequence number, the packet has not been received. This is notified to the task server using the task transfer response packet 70 of FIG. 7 (step S4).

Specifically, the sequence number of the unreceived packet is stored in the unreceived packet number area 72 of FIG. When representing one non-receipt packet number in 2-byte, the "data size" field 71 ₇ writes 2x bytes. Where x is the number of unreceived packets.

The sequence diagram of FIG. 5 shows an example in which two task transfer packets with sequence numbers 4 and 7 have not been received. Upon receiving this response packet, the task server 10 retransmits the task transfer packets with sequence numbers 4 and 7 (steps S5 ₁ and 5 ₂ ). As a result, the sensor node 2 receives all task transfer packets.

Next, the sensor node 2 returns a response indicating that it has been received to the task server 10 using the task transfer response packet 70 of FIG. 7 (step S6). The task transfer response packet 70, specifically, in the packet structure of Fig. 7, "data size" field 71 ₇ 0, non-reception packet number field 72 is zero bytes.

  Finally, when an execution start command packet 80 shown in FIG. 8 described later is sent from the task server 10 to the sensor node 2 (step S7), execution of a new task is started on the sensor node 2 side. After the start, a response packet indicating that the execution has started is returned from the sensor node 2 to the task server 10 (step S8).

  In this embodiment, the example in which retransmission processing is performed after receiving k task transfer packets 40 has been shown. However, each time one or several task transfer packets 40 are received, the sensor node 2 side In FIG. 7, when a time-out time is provided and a time-out occurs, a retransmission request for the task transfer packet 40 can be made using the task transfer response packet 70 of FIG.

FIG. 8 shows a packet structure of the task execution start instruction packet 80. The execution start instruction packet 80 can start a plurality of tasks with one packet. A “target memory” field 81 _{1 to} 81 _s (0 indicates the program storage memory 26, 1 indicates the configuration memory 27) and a “memory bank number” field 82 _{1 to} 82 _{s in} which a task to be started is stored Specify in pairs.

  Each field is 1 byte, and an integer value 2s is written in the data size field 83 when instructing activation of s (s> 0) tasks. The task execution start instruction packet 80 from the sensor node 2 is responded by sending the task execution start instruction packet 80 to the task server 10 as it is. However, the memory bank number 82 corresponding to the task that failed to start is rewritten to 0 and returned.

  The sequence diagram of FIG. 5 shows an example in which execution of a task in the bank 3 in the program storage memory in which the task just transferred from the task server 10 is stored (steps S7 and S8).

  In the configuration of the sensor node 2 in FIG. 2, a hard disk (HDD) is provided instead of the program storage memory 26 and the configuration memory 27, and tasks stored in the banks of the program storage memory 26 and the configuration memory 27 are stored. A similar function can be achieved even if the HDD is given a unique name and stored as a file, and the file is accessed as necessary.

In this case, when task data transmission / reception between the task server 10 and the sensor node 2 described above or a task stored in the sensor node 2 is activated, the processes shown in FIGS. 4, 6, 7, and 8 are performed. In the packet structure, the “target memory” field (41 ₃ , 61 ₃ , 71 ₃ , 81) and the “memory bank number” field (41 ₄ , 61 ₄ , 71 ₄ , 82) have unique names stored in the HDD. Each packet structure is used as a storage field.

  FIG. 9 is an example of a scenario description that is input to the scenario compiler 11 in FIG. FIG. 9 shows, “If there is no sensor node 2 measuring the temperature within 2 hops, turn on its own temperature sensor and measure the temperature every second. "Turn on and measure temperature every 10 seconds."

  The first line neighborCount (2, “temperature”) is a built-in function that returns the number of sensor nodes that satisfy the condition given by the argument. The first argument represents the number of hops, and the second argument represents the sensor type. Also, on (“temperature”, “rate = 1s”) used in the second line of FIG. 9 is a built-in function, and the sensor given to the first argument (in this case, temperature) is turned ON. The second argument is optional and can describe parameters according to the sensor. “Rate = 1s” in this example indicates that temperature sensing is performed every 1s, that is, every second.

  Note that neighborCount (2, “temperature”) is a function 10 that returns the number of sensor nodes 2 that meet the conditions. This function is a function that each sensor node 2 obtains the neighboring sensor node information acquired by the method described below. The value is obtained by reference.

  FIG. 10 shows how each sensor node 2 collects neighboring sensor node information. Here, for the sake of simplification, a case where the IDs of neighboring sensor nodes are known is shown as an example.

  FIG. 10A shows the initial state of each sensor node 2. In the figure, for example, n (a) is an adjacent node list indicating information of the sensor node 2 adjacent to the sensor node a, that is, the sensor node 2 one hop ahead. In the state of FIG. 10A, the adjacent node list of each sensor node 2 is (), that is, empty. Next, each sensor node 2 broadcasts its own adjacent node list to adjacent sensor nodes. As a result, each sensor node 2 obtains information shown in FIG.

  When all the adjacent node lists received from each adjacent sensor node 2 are arranged, as shown in FIG. 10C, each sensor node 2 can acquire information of its own adjacent node. Next, when this adjacent node list is broadcast again to adjacent nodes and the adjacent node list received from the adjacent sensor nodes is arranged, information on the sensor node that is two hops away from itself can be obtained.

  By repeatedly executing the above steps, the information on the network topology shown in FIG. Here, an example is described in which only the IDs of adjacent sensor nodes are exchanged. However, each sensor node can be exchanged by exchanging information such as the type of sensor possessed by each sensor node 2 and information of the currently operating sensor together with the ID. 2, information on all other sensor nodes can be obtained.

  Thus, the information acquired from the adjacent node is stored and stored in a memory area that can be referred globally from each task operating in the sensor node 2. Since the information exchange between the adjacent sensor nodes described above is periodically performed between the sensor nodes without the user being aware of it, the information regarding the adjacent sensor nodes is updated every moment.

  Therefore, the neighborCount () function described above can immediately know the number of adjacent sensor nodes that meet the conditions by accessing the information in the memory area.

FIG. 11 shows an embodiment of the present invention when a sensor node 2 equipped with a monitoring camera is used. Sensor node 2 ₃ having a sensor node 2 ₁ and the camera 3 with camera 1 is connected directly to the base station 1, the sensor node 2 ₂ having a camera 2 via the sensor node 2 ₁ with camera 1, information Transfer to base station 1.

  In this embodiment, for example, when a task for realizing a scenario of “notify when a person wearing a red shirt is detected” is sent to each sensor node 2 as a sensing target, the sensing target that has entered the imaging range of the camera of the sensor node Can be sent to the base station 1 from the corresponding sensor node 2 if there is a sensing target that meets the conditions.

  Here, image recognition can be performed by each camera, or each camera itself sends image data directly to the base station 1 and a program in the gateway (GW) 7 connected to the base station 1 performs image recognition. , Check whether there is a sensing target that meets the conditions. It is also possible to construct a form in which the result is returned to the user terminal 4.

  In the former case, the image recognition program is prepared as one task in the task library 12 shown in FIG. That is, the scenario compiler 11 generates a task to be mounted on each camera so as to give the above-mentioned scenario “notify when a person wearing a red shirt is found” as a parameter of the task, and generates a corresponding sensor node. The system of the present embodiment is constructed by sending it to the camera.

  In the latter case, that is, when image recognition is performed on the gateway (GW) 7 side, the scenario for each camera is simply “transfer image every second”. The gateway (GW) 7 realizes the above system by inputting the image data sent from the camera into an image recognition program realized in advance and recognizing “a person wearing a red shirt”. Become.

  Note that the image recognition itself can be realized by using existing technology, for example, Non-Patent Document 1 described above. In addition, the task server 10 described above includes a configuration shared with the gateway (GW) 7 and a method for realizing all of them in the user terminal 4.

FIG. 12 is a configuration diagram in a case where a large-scale sensor network is configured by using a plurality (three sensor networks 100 _{1 to} 100 _{3 in} the figure) of the sensor network 100 of FIG. 1 of the present invention. The center of this configuration is a demand addressable network (hereinafter abbreviated as DAN). When the sensor networks 100 _{1 to} 100 ₃ are connected to the DAN, the user terminal 4 connected to the sensor network 100 in FIG. 1 is not necessary, and a gateway device (GW) is provided in one base station.

  The DAN includes the following functional means.

  (a) Demand interpreter 200: Instantly interprets user requests from the abstract user terminal 201 and develops them into specific sub-request combinations. User requests are decomposed into elements consisting of (When, Where, What), and become sub-requests consisting of simple parameter combinations. For example, the user request “What is the possibility of a fire within 10 km from here?” Is expanded as ((When, Now), (Where, Here <10 km), (What, Temperature = Rapid-Up)).

  In addition, the user request “Are there a possibility of a fire within 2km within 5km north?” ((When, Now> 2 day), (Where, ((Here <5 km) & North), (What, Temperature = Rapid-Up)) The parameter and reserved word information included in the expanded request is prepared in a table format so that it can be read when the interpreter 200 is started, and the contents of the table are changed. Can be expanded at any time.

  (b) Demand addressing: Resources such as sensors or local databases (DB) that hold desired sensing data from expanded sub-requests or combinations thereof can be dynamically changed without referring to a centrally managed directory. Discover.

  This part can be realized, for example, by a technique using the distributed hash table (DHT) 202 used in the Peer-to-Peer (P2P) technique (Non-Patent Document 2 described above).

  (c) Data mashup network: The acquired data is integrated (mashed up) in the network and presented to the user. For this, the messaging network (MN) technology of Non-Patent Document 3 described above can be used as a construction base.

  The MN is a technology for constructing an Overlay-network using the XML router 203 and a gateway device (GW). In the MN, a data transfer function based on content is realized on the Overlay-network in units of messages described in XML.

  The GW dynamically converts the received local data into an XML format message. The message is transferred to the target terminal by the XML router 203 having an XML interpretation function. The XML router 203 performs integration (mashup) of XML format data as necessary.

  FIG. 13 is a diagram for explaining the integration (mashup) of data in the XML format with reference to FIG.

The mashup programs 1 and 2 operate on the XML routers 203 ₁ and 203 ₂ , respectively. Each mashup program determines in advance which events and data to mash up. It may be when the mashup program is created or a method of giving which event or data should be mashed up from the outside at the start of the operation or after the start of the operation.

In the configuration example shown in FIG. 13, mashup program 1 mashup data obtained from thermal event 1 and the temperature event 2, and the existing information system (1) 210 obtained from the sensor network system 100 ₁ described above. Then, as the information 1 that is mashup, over a network, and sends the mashup program 2 of XML router 203 _2.

In mashup program 2, in addition to the information 1 that is mashup, sensor network system 100 ₂ and or 100 ₃ temperature events 3 obtained from the data obtained from another existing information systems, not shown (2), further not shown Map Mash up the map information obtained from the information service site. Next, the mashed information 2 is sent to the user terminal 201 via the network.

  When the existing information system 210 is a weather information database site, the mashup information 2 is displayed only on the received mashup information 2 without performing mashup on the user terminal 201. Thus, a desired result can be obtained.

  FIG. 14 is an example of an application constructed by the network according to the present invention. When the user requests from the user terminal 201 “Where has the temperature exceeded 31 degrees in Fukushima Prefecture in the past month?”, This request is sent to the demand interpreter 200 ((Where, “Fukushima”), (When, [now, 1 month]), (What, Temperature> 31 degree), (Display, Map)) and sub-requests. Then, a search request is passed to the corresponding temperature sensor network 130, the weather information data site 131 storing past weather observation data, and the map information site 132. Therefore, the hit information is converted into XML data when passing through each GW and sent to the XML router 203.

  In the XML router 203, the information is mashed up and delivered to the user terminal 201. As a result, the user terminal 201 obtains the information 204 including the map shown in the upper part of FIG. 14 by simply displaying the information on the screen without integrating any information.

  Here, when the temperature sensor network is searched in FIG. 14, if no sensor network observes the temperature and the sensor network is the sensor network 100 shown in FIG. 1 of the present invention, A scenario of “measure temperature and notify if it is 31 degrees or more” is sent to the sensor network 100 via the task server 10 of FIG. 1 to cause temperature measurement. Thereafter, the sensing data is returned from the sensor network 100 to the demand addressable network (DAN) that requested it. Thereby, in the observation area which the sensor network 100 is in charge, the current temperature information that meets the user's request can be obtained.

  Here, the above-mentioned scenario “measure temperature and notify if it is 31 degrees or more” can be generated dynamically in the task server 10 each time. It may be prepared in the server 10 and called. Furthermore, when the same scenario is frequently used, the scenario compiler 11 is started only once, the generated task group itself is stored in the task server 10, and it is stored in each sensor node 2 as necessary. It may be sent in.

  Alternatively, if there is room in the program storage memory 26 (or configuration memory 27) of each sensor node 2, if these tasks are inserted in the sensor node 2 in advance, the corresponding task is started to execute the task shown in FIG. The task can be driven immediately by simply sending the packet 80.

  FIG. 15 is a diagram for explaining still another embodiment. This is a configuration in which a scenario generator 140 is added to the sensor network 100 of FIG. A task to be downloaded to each sensor node 2 is generated from the initial scenario given by the user by using the task server 10, downloaded to each sensor node 2 constituting the sensor network 100 via the base station 1, and the sensor network 100 To customize.

  Next, the sensing data acquired from the sensor network 100 is sent to the scenario generator 140, and the scenario generator 140 receives a final target separately given in advance by the user and the sensing data as a new sensor network 100. Automatic operation scenario is generated. It is input to the task server 10 again. By repeating the above series of processing, the operation is performed so as to achieve the final goal given by the user.

  For example, when the user's final goal is “find a fire” and the initial scenario is “find rapid temperature information”, the task server 10 first “measures the temperature and the measured temperature is once per second. A task “notify the event when it rises above” is generated, and the task is downloaded to each sensor node 2.

  When each sensor node 2 senses an event that satisfies the condition, it sends it to the scenario generator 140 as sensing data. In response to this, the scenario generator 140 automatically generates a scenario “detect a flame” and sends it to the task server 10. In response to this, the task server 10 activates the illuminance sensor, generates a task “notify if the value is L or more”, and downloads it to each sensor node 2.

  Here, L is numerical data representing the reaction intensity of the illuminance sensor, and it is assumed that the larger the value, the stronger the reaction to the flame (light).

  When a certain sensor node 2 senses the event, it returns it to the scenario generator 140. In the scenario generator 140, since there has been a sudden change event of the previous temperature and a reaction event of the illuminance sensor this time, it is assumed that a fire has occurred in the observation range of the sensor network, and this is notified to the user as a final result. return. In the configuration of the present embodiment, the scenario generator 140 can be configured in the same physical PC as the task server 10 or configured as a separate device and connected to each other via a communication line.

  Here, in the above description of the embodiment, it has been described that the sensor network 100 of FIG. 1 is wireless communication, but of course, a sensor network using wired communication may be used. In FIG. 1, a plurality of base stations 1 are provided. However, even if one base station is used or the base station is not used, the sensor node 2 is directly connected to the user terminal 4, the task server 10, and the like. Even a network configuration capable of communication is within the scope of the present invention. In the case of wired communication, the sensor node 2 shown in FIG. 2 is configured using a wired module instead of the wireless module.

  2 has been described as being sent from the task server 10 shown in FIG. 1. However, instead of the task server 10, the CPU 22 of the sensor node 2 sends the task shown in FIG. A program on the task server side that executes a procedure in accordance with the sequence diagram may be mounted so that the task held by the sensor node 2 may be sent to another sensor node.

  By using the configuration of the present invention described above, (1) a sensor network that can customize functions and roles that operate according to the environment and user requirements on each sensor node according to a “scenario” that can be defined by the user; (2) A function of presenting necessary and sufficient information to a user in real time according to a user request can be realized.

  With regard to the former (1), by replacing the “scenario” according to the situation at the observation site, a sensor network meeting the needs can be quickly constructed.

  The latter (2) is characterized by a technique of actively acquiring data, mashing it up in the network, and providing it to the user. In order to obtain desired data, resources in the network are searched without using a central management server or a directory service, so a large-scale network can be easily constructed.

  Further, in conjunction with the former (1) sensor network technology, it is characterized in that a role of the sensor node is dynamically changed and a series of operations for actively acquiring sensing data satisfying a user request is automatically performed.

  Further, normally, a router apparatus arranged in a network performs mashup work of data of different formats acquired from a plurality of systems executed by a server or a terminal together with data transfer. As a result, not only can large-scale data be handled without imposing a load on the terminal, but even when new data needs to be mashed up, the user side software can be updated without modification. It can correspond only by. This is a great advantage in constructing a large-scale sensor network. In addition, the present invention proposes a new sensor network construction method and operation method that can be realized by merging the two functions described above and has no other concept of active sensing according to user requirements. is there.

1 ₁ to 1 _p base station 2 ₁ to 2 _q sensor node 3 base station network 4 user terminal 5 local database 6 sensor node network 7 gateway device 10 task server 11 scenario compiler 12 task library 13 scenario description 14 task group 100 sensor network

Claims (9)

A plurality of sensor nodes each having at least one sensor and an information processing function unit capable of changing the function according to a set task and generating sensing data detected by the sensor according to the function;
Means for generating a task corresponding to a scenario describing the operation specifications given by the user;
And a communication means for transferring the generated task to a corresponding sensor node of the plurality of sensor nodes.
A sensor network system characterized by the above. In claim 1,
The information processing function unit includes a CPU, a logic variable hardware including an FPGA (Field Programmable Gate Array), or both, and the task is realized by a program executed by the CPU or a logic variable hardware. Set as logic circuit information or both
A sensor network system characterized by the above. In claim 1,
And at least one base station,
From the base station, the task generated by the means for generating the task is wirelessly or wiredly transferred to the plurality of sensor nodes.
A sensor network system characterized by the above. In claim 3,
The plurality of sensor nodes respectively transfer information obtained by processing sensing data detected by the sensor by the function of the information processing function unit to the base station by a multi-hop communication method.
A sensor network system characterized by the above. In claim 1,
The means for generating the task is:
A task library in which basic operations that can be executed by the plurality of sensor nodes are registered;
A scenario compiler for generating a task that operates on the sensor node by decomposing a user request given by the user into tasks to be executed on each sensor node with reference to the task library;
The communication means causes the task generated by the scenario compiler to be downloaded to the corresponding sensor node, and dynamically changes the information processing function part of the sensor node.
A sensor network system characterized by the above. In any one of Claims 1 thru | or 5,
Furthermore, based on the user request first given by the user, the following operation specifications are determined based on the information detected by the sensor and processed by the information processing function unit whose function has been changed, and the user's final goal. A scenario generator for generating, changing the function of the information processing function unit based on the generated operation specifications, and customizing the user;
A sensor network system characterized by the above. It has a plurality of sensor nodes, each sensor node has at least one sensor, and has an information processing function unit that performs processing on information detected by the sensor. Possible multiple sensor networks,
With other existing information systems,
A demand addressable network connecting the plurality of sensor networks and other existing information systems;
The demand addressable network is:
Means for decomposing an information acquisition request given by a user into sub-elements consisting of simple parameter combinations;
Means for obtaining a sensor network or an existing information system having the relevant information in the network based on the sub-requests or combinations thereof developed by the means for developing the sub-requests;
Means for providing the user with the data obtained by the sensor network or the existing information system integrated in the network;
A sensor network system characterized by the above. In claim 7,
Responds to an information acquisition request given by the user when there is no sensor network or existing information system that holds the corresponding information in the network based on the sub-request developed by the means for developing the sub-request or a combination thereof. So that the operation specifications of the sensor network are generated,
Based on the operation specifications of the generated sensor network,
Change the function of the information processing function unit;
A sensor network system characterized by the above. A data acquisition method in a sensor network system having a plurality of sensor networks, another existing information system, and a demand addressable network connecting the plurality of sensor networks and another existing information system,
In the demand addressable network,
Decomposing information acquisition requests given by users into elements and expanding them into sub-requests consisting of simple parameter combinations;
Obtaining a sensor network or an existing information system that holds relevant information in the network based on the deployed sub-request or a combination thereof;
Integrating the data obtained by the sensor network or the existing information system in the network and providing the user with the data;
A data acquisition method in a sensor network system.