JP2004343534A - Sensor network evaluating method and designing method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a method for evaluating the operating ratio of a sensor network that includes a plurality of sensor nodes and is connected to the Internet through a gateway. <P>SOLUTION: Data sensed by each sensor node are relayed by the other sensor node and is transferred through the gateway and the network. Information used for evaluating the operating ratio of the sensor network includes the number of sensor nodes included in the sensor network or the area of a sensing area, a transfer route of the data sensed by the sensor nodes and a parameter of each sensor node. In evaluating the operating ratio, these pieces of information are inputted, an arithmetic unit is used to calculate the operating ratio on the basis of a prescribed expression. To optimize the operating ratio, the calculated operating ratio is compared with a target value, and when the operating ratio does not exceed the target value, an arithmetic operation is repeated while changing inputted data. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

（ただし、θ１＝１とし、ｃ_ｉ，１及びｃ_ｉ，２はセンサノードｉに関するパラメータであり下位ノードのセンシングデータを１ビット中継する毎の平均消費電力がｃ_ｉ，１＋ｃ_ｉ，２ｄｉ^２ となるものとする）
に基づいて決定することを特徴とする。
【００１３】
【発明の実施の形態】
入出力装置、記憶装置、演算装置からなるセンサネットワーク設計・評価システム上でのセンサネットワーク設計・評価方法について説明する。評価装置は、各種のデータやパラメータを入力する入力装置、入力したパラメータ等を記憶するデータベース、入力パラメータについて所定の演算処理を実行する演算装置、及び演算結果を出力する出力装置を有するものとする。
各センサノードは、センシング機能、通信機能及びセンシングデータの中継機能を有し、バッテリが内蔵されているものとする。各センサノードにより検出されたデータは、無線リンクのマルチホップにより他のセンサノードを中継してゲートウェイに転送する。当該ゲートウェイはインターネットに接続する。従って、各センサノード１〜ｎは、中間のセンサノードで順次中継されてゲートウェイ１に転送され、インターネット２を経てワークステーション等に送信される。
【００１４】
センサネットワークの稼働率評価は、以下の工程に従って行う。
初めに、設置すべきセンサノードの数ｎ及びセンシングエリア面積（距離）Ｄを決定して入力する。より一般的なネットワークトポロジについては、以下のようにして、一列の単純な構成に分解する。図２は、直線型の網構成を示す。図２において、インターネット１にゲートウェイ２が接続され、ゲートウェイ２には直線状に配列されたｎ個のセンサノード３ａ〜３ｎがライン状に接続される直線型の網とする。
【００１５】
次に、センサネットワークのトポロジ・センサノード間の結合関係を入力する（図３）。さらにその上で、センシングデータが送信される経路を入力する（図４）。ゲートウェイを端点とするツリー構造が通常である。ゲートウェイを根とし、末端のセンサノードまでの一連のノードを、演算装置は取り出す（図５）。ゲートウェイを端点とする１列の構成となる。ｎ個のセンサノードについては、ゲートウェイ２に近い側から順次識別番号を付与する。
【００１６】
センサノードｉは常時毎秒ａ_ｉの電力を消費し、この常時消費電力ａ_ｉに加えて、センシングやデータ転送に際して、以下の電力消費を生ずる。まず、センサノードｉがセンシングする毎に電力ｂ_ｉを消費する。この電力ｂ_ｉは平均１／β_ｉの指数分布に従う確率変数であるとする。そして、下位ノードのセンシングデータを中継する毎に電力ｃ_ｉを消費する。ここで、ｃ_ｉは、平均１／γ_ｉの指数分布に従う確率変数とする。センサノードｉにおけるセンシングは平均λ_ｉ／秒のポアソン分布にしたがって生ずるものとする。また、センサノードｉの初期バッテリ量ｈｉ（時刻０できバッテリ量）は、平均１／η_ｉに従うものとする。尚、以下の説明においてデータの転送遅延は無視する。
【００１７】
次に、入力すべきセンサノードのパラメータについて説明する。ｘ_ｉを、全センサノードは稼動し続ける（バッテリ切れしない）とした上でのセンシングと中継によりセンサノードｉが時刻０以降に消費する電力とする。ｐ_ｔ（ｘ_１，……ｘ_ｎ）を時刻ｔにおけるｘ_１，……ｘ_ｎの同時確率密度関数とする。このとき、ｐ_ｔは以下の式を満たす。
【数５】

ｑ_ｔ（ｓ_１，……ｓ_ｎ）をｐ_ｔ（ｘ_１，……ｘ_ｎ）のＬａｐｌａｓ−Ｓｔｉｅｌｔｊｅｓ ｔｒａｎｓｆｏｒｍ（Ｌ．Ｓ．Ｔ．）とすると、ｑ_ｔは以下の式で表される。
【数６】

【００１８】
ｙ_ｉを、全センサノードは稼動し続ける（バッテリ切れしない）とした上での、センサノードｉが時刻０以降に消費する電力とする。すなわち、時刻ｔにおいて、ｙ_ｉ＝ｘ_ｉ＋ａ_ｉｔ となる。ｒｔ（ｙ_１，………ｙ_ｎ）を時刻ｔにおけるｙ_１，………ｙ_ｎの同時確率密度関数とし、ｕｔ（ｓ_１，……ｓ_ｎ）を、そのＬ．Ｓ．Ｔ．とする。すると、
【数７】

となる。
【００１９】
以上の結果より、時刻ｔのセンサネットワークの稼働率ｗ_ｔは以下のように表される。
【数８】

この式より、稼働率ｗ_ｔはｔに対して指数関数的に減少することが分かる。
特に、各パラメータ（ａ_ｉ，β_ｉ，γ_ｉ，η_ｉ，λ_ｉ）がセンサノード番号ｉによらず同一の場合（すなわち、各ｉについてａ_ｉ＝ａ，β_ｉ＝β，γ_ｉ＝γ，η_ｉ＝η，λ_ｉ＝λ）、消費電力ｗ_ｔは以下の式で表される。
【数９】

この場合、上式から、ｗ_ｔはλに関して、指数関数的に減少することが分かる。
【００２０】
上述した事項に基づき、センサネットワークを評価するためのセンサノードパラメータとして、常時消費電力ａ_ｉ／秒、１センシング平均消費電力１／β_ｉ、下位ノードのセンシングデータ１ビット中継に対する平均消費電力１／γ_ｉ、センシング発生平均頻度λ_ｉ／秒、平均初期バッテリ量（時刻０でのバッテリ量）１／η_ｉを用い、これらのパラメータの入力を行う。尚、全ノード同一パラメータの場合は、全ノード同一パラメータモードにより、常時消費電力ａ／秒、１センシング平均消費電力１／β、下位ノードのセンシングデータ１ビット中継に対する平均消費電力１／γ、センシング発生平均頻度λ／秒、平均初期バッテリ量（時刻０でのバッテリ量）１／ηを入力する。
【００２１】
これらの入力したパラメータやノード数等に関する情報は記憶装置に記憶する。次に、演算装置は、予め定められている上記式により、センサネットワーク緑働率を評価する
【００２２】
予め定められ記憶されているセンサネットワーク稼働率目標値に対して、評価結果と比較し、結果を出力装置に表示する。評価結果が目標値を下回る場合は、センサノードのパラメータまたはセンサノード数の変更要求を出力装置に表示する。一般的なネットワークトポロジの場合は、結合関係の経路の変更、あるいは、センサノードパラメータの変更の要求を出力装置に表示する。これに従って、ユーザは、センサノードのパラメータ等を変更して入力する。入力パラメータは記憶され、再び、演算装置により評価式に基づいて演算が実行される。センサネットワーク稼働率目標値と評価結果を比較し、結果を再び出力装置に表示する。評価結果が目標値を上回るまで、この手順を繰り返し、センサネットワークの稼働率の最適化を行う。
【００２３】
また、最大化オプションを指定すると、センサノードのパラメータ範囲の入力画面がでる。これに従って、パラメータ範囲をユーザは入力する。このパラメータ範囲内の各々の値に対して、評価式を評価し、センサネットワーク稼働率が最大となるパラメータ値と評価値を出力装置に表示することもできる。
【００２４】
次に、センサノードをあるエリア内に設置する場合の、最適な設置方法について説明する。エリアは簡単化するため、長さＤの１次元とし、センサノードｉとｉ−１との間の距離をｄ_ｉとする。センシングや中継による消費電力及び検出頻度は、センサノード間の距離に依存するパラメータｂ_ｉ，１，ｂ_ｉ，２，ｃ_ｉ，_１，ｃ_ｉ，２を用いることができる。これらのパラメータは、センサノードパラメータである１センシング平均消費電力１／β、下位ノードのセンシングデータ１ビット中継に対する平均消費電力１／γ、センシングの発生頻度λと以下の関係がある。１／β_ｉ＝ｂ_ｉ，１＋ｂ_ｉ，２ｄ_ｉ ^ｍ，／γ_ｉ＝ｃ_ｉ，_１＋ｃ_ｉ，２ｄ_ｉ ^ｍ，λ_ｉ＝ｌｄ_ｉ，Ｄ＝Σ_ｉ＝１^ｎｄ_ｉ（ｍ＝２） これにより、ノード数などの変わりに、センサノード間の間隔ｄ_ｉを与え、評価式に基づいて評価することで、あるエリアのセンシングを行うセンサネットワークの稼働率に関する最適化の評価が可能となる。尚、以下の数値例では、各センサノードについての一例として以下の値を用いることができる。ｔ＝１（ｙｅａｒ）＝３１５３６０００秒、ｍ＝２、ａｉ＝１（ｎｊ）、ｂ_ｉ，ｊ＝９５（ｎＪ）、ｂ_ｉ，２＝０．０１（ｎＪ）、ｃ_{ｉ， １}＝１８０（ｎＪ）、 ｃ_{ｉ， ２}＝１８０（ｎＪ）、１／η_ｉ＝１０（Ｊ）、ｌ_ｉ＝０．０００１（ｉ＝１………ｎ）
【００２５】
［均等配置］
各センサノードを均等に配置する例（すなわち、ｄ_ｉ＝Ｄ／ｎ）について検討する。ノード数ｎを、
式ｎ＝Ｄ｛（ｍ−１） ｃ_{ｉ， ２}／ｃ_{ｉ， １}｝^１／ｍ
に基づいて決定する方法を均等配置法と呼ぶことにする。均等配置法は、センサノードを均等に配置するという条件のもとで、中継のための電力消費が支配項である場合の平均消費電力を最小化するようなノード数を与える。均等にセンサノードを配置する場合、センサノード数とネットワーク稼働率との関係は、図６のようになるが、均等配置法によれば、ｎ＝４となり、均等配置法で与えたノード数を用いれば稼働率上も最適なノード数となっている。
【００２６】
［不均等配置］
センサノードを等間隔に設置するという条件をはずして、決められたセンサノード数ｎ（センサノード数は固定する）について任意の設置のなかでセンサネットワーク稼働率を最大化する設置法を検討する。いま、ｄ_ｉ＝θ_ｉＤ／（Σ_ｉ＝１ ^ｎθ_ｉ）とし、θ_ｎ＝１、θ_ｊ（ｊ＝１……ｎ−１）を以下の式に基づいて規定することにより、中継のための電力消費で距離比例項が支配項である場合の平均消費電力を最小化するようにセンサネットワーク設計することができる。すなわち、式
【数１０】

をｄ_ｉ＋…ｄ_ｎ＝Ｄという条件で最適化する。これはｄ_ｉ＝θ_ｉＤ＝１とすることで、直接、数式１１を得る。
【数１１】

【００２７】
上記で与えられる不均等配置法が、センサネットワーク稼働率上も有効であることを、数式７（または８）により評価したものが、図７である。図７では、ノード数を可変として、等間隔にノードを配置した場合と数式１１により間隔を決め配置した場合の稼働率のグラフである。後者の稼働率が前者を上回っていることが分かる。数式１１により決まるセンサノード間隔を図８に図示した。これで分かるように、数式１１による配置法は、ゲートウェイに近いほどノード間距離を短くする配置法となっている。ゲートウェイに近いノードほど下位ノードからのセンシングデータ中継による消費電力が増加するが、これを隣接ノードとの距離を短くすることで制御し、稼働率を上げていると考えられる。
【図面の簡単な説明】
【図１】センサネットワークの例を示す図である。
【図２】センサネットワークを直線型網に展開した図である。
【図３】センサネットワークの結合関係を示す図である。
【図４】センシングデータの転送経路を示すである。
【図５】センサネットワークをゲートウェイを拠点としたツリー構造とした図である。
【図６】ノード数を可変とした場合のセンサノード数とセンサネットワークの稼働率との関係を示す図である。
【図７】ノード数を固定した場合のセンサノード数と稼働率との関係を示す図である。
【図８】ノード数が固定されている場合のゲートウェイからのセンサノード番号とノード間隔との関係を示すグラフである。
【符号の説明】
１ インターネット
２ ゲートウェイ
３ センサノード[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an operation rate evaluation method and a design method of a sensor network including a plurality of battery-equipped sensor nodes having a sensing function and a communication function.
[0002]
[Prior art]
In recent years, with advances in electronics and MEMS technology, many sensors in factories, automobiles, offices, and homes have been miniaturized and have started to be connected in a network. For example, anti-theft vehicles combined with notifications to security companies and mobile phones, houses with intrusion monitoring, combinations of various tank residual quantity sensors and tank distribution systems, landslide sensor reporting systems, and gas supply suspension with earthquake sensors Control is an actual example. In particular, a sensor network including a large number of ultra-small sensor nodes (in the present specification, a sensor having a communication function is referred to as a “sensor node”) that uses a wireless technology, has a built-in battery, and autonomously configures a network. It has become possible. As an example of such a sensor network, a future-type sensor network represented by smart-dust is known (for example, see Non-Patent Document 1). Such a future-type sensor network has features different from the conventional networks as described below (for example, see Non-Patent Document 2).
[0003]
i) The communication function is a low function.
ii) It can be a network of very many nodes. Also, the density of the nodes per unit area may be very high.
iii) The location of the sensor nodes may not be well planned.
iv) The sensor nodes may run out of battery and fail.
V) The sensor nodes may get stuck easily and may be updated all at once.
vi) Sensor nodes may move.
vii) The sensor node has an endpoint function and may also have a relay function.
viii) The sensor node may have an upper layer function such as integration of sensing data sensed by the own node or a lower node or information processing.
As such, such sensor networks have different aspects than previous networks. Therefore, it is necessary to evaluate the characteristics using a different scale.
[0004]
[Non-patent document 1]
I. F. Akyildiz, W.C. Su, Y .; Sankarabramaniam, and E.C. "A Survey on Sensor Networks" by Cayirci, IEEE Communications Magazine, pp. 102-114, August 2002
[Non-patent document 2]
J. M. Kahn, R .; H. Katz, and K.K. S. J. "Next century challenges: Mobile networking for smart dust," by Peter, In Processeds of the Fifth Annual ACM / IEEE International Conference on Membership, Limited. 27-278, Seattle, August 1999
[0005]
FIG. 1 shows an example of a sensor network. The sensor network includes a plurality of sensor nodes and gateways. The broadly defined sensor network includes a database, an address resolution device, and an application system connected to yet another network, but this patent does not particularly refer to a database, an address resolution device, an application system, and the like. Each sensor node may be directly connected to the gateway (FIG. 1A), or may be connected to another sensor node and relay the sensing data of another node (FIG. 1B). The gateway connects the sensor network to another network (usually the Internet). There may be a plurality of gateways. If the gateway is an Internet host, it has an IP address, but the sensor node may not have an IP address. As a special case, there is also a form shown in FIG. 1C in which the sensor node incorporates a gateway function.
[0006]
The data sensed at each sensor node is sent to a database via a gateway and another network. The application system accesses a database or a sensor node in order to obtain sensing data. Usually, the application system accesses using a geographical position or an attribute-base address such as “temperature at an xyz intersection in Tokyo”. Therefore, the address resolution device resolves the IP address of the database and the gateway, and the network address of the sensor node. When the address of the gateway node is resolved, the data acquisition request is transferred to the gateway, and is associated with the network address of the sensor node under the gateway. In this way, the application system that has acquired the request data from the database or the sensor node performs a process such as graphing and notifies the user of the sensing data.
[0007]
As mentioned above, sensor networks are quite different from existing networks. Therefore, it is necessary to consider a measure relating to the performance and quality of the sensor network that is suitable for the sensor network.
First, a feature of the sensor network is that the sensor node is often of a type with a built-in battery, and the function stoppage due to running out of battery may occur. As an example of the measure related to the battery, “sensor network operation rate” (probability that all sensor nodes configuring the sensor network are operating) can be considered.
[0008]
On the other hand, in the existing network, a call loss rate, a connection delay, a packet delay, a transmission error rate, and the like are measured. The network design device of the existing network has a storage device, an arithmetic device, and an input / output device, and provides and stores ground traffic, network topology / path, and node parameters (call processing capacity, etc.) from the input device, According to a predetermined formula, the above scale is evaluated by an arithmetic operation. As a result, a network is designed by finally finding the number of routes and nodes that satisfy the same scale target value.
[0009]
[Problems to be solved by the invention]
Although a sensor network is quite different from an existing network, there is no appropriate measure, and no design method or evaluation method based on the measure has been proposed.
Accordingly, an object of the present invention is to provide a method for evaluating the operating rate of a sensor network that includes a plurality of sensor nodes with a built-in battery and in which sensing data is transferred via a gateway.
Still another object of the present invention is to provide an operating rate design method of a sensor network including a plurality of sensor nodes with a built-in battery and transmitting sensing data via a gateway.
[0010]
[Means for Solving the Problems]
An operation rate evaluation method for a sensor network according to the present invention includes a plurality of sensor nodes having a data sensing function, a communication function, a data relay function, and a battery, and is connected to a network via a gateway, and data sensed by each sensor node. Is a method for evaluating the operation rate of the sensor network transferred via the gateway and the network by relaying another sensor node located in the middle,
As information used to evaluate the operation rate of the sensor network, including the number or sensing area area of the sensor nodes included in the sensor network, a transfer path of data sensed by the sensor nodes, and parameters of each sensor node,
Inputting information used for the operation rate evaluation,
Calculating the operating rate w _t of the sensor network at time t based on the input data using an arithmetic device.
[0011]
The method for designing a sensor network according to the present invention includes a plurality of sensor nodes having a data sensing function, a communication function, a data relay function, and a battery, connected to the network via a gateway, and data sensed by each sensor node. A method for designing a sensor network, which is transferred via the gateway and the network by relaying another sensor node located in the middle,
Defining a sensor network sensing area D and the number n of sensor nodes arranged in the sensing area;
Defining a connection relationship between the sensor nodes;
Comprising the step of defining the distance d _i between the sensor nodes,
The distance d _i between the sensor nodes, and sets smaller than the spacing between the far side of sensor nodes apart gateway between the side of the sensor nodes near gateway.
[0012]
Another sensor network design method according to the present invention includes a plurality of sensor nodes having a data sensing function, a communication function, a data relay function, and a battery, and is connected to the network via a gateway, and data sensed by each sensor node. Is a method for designing a sensor network which is transferred via the gateway and the network by relaying another sensor node located in the middle, wherein the sensor network sensing area D and the number of sensor nodes arranged in the sensing area defining n;
Defining a connection relationship between the sensor nodes;
Comprising the step of defining the distance d _i between the sensor nodes,
The distance di between the sensor nodes is _expressed by the following equation.

(However, it is assumed that θ1 = 1 _, and ci _{, 1} and ci _{, 2} are parameters relating to the sensor node i, and the average power consumption for each one-bit relay of the sensing data of the lower node is ci _{, 1} + ci _, 2di. ² Shall be)
Is determined based on
[0013]
BEST MODE FOR CARRYING OUT THE INVENTION
A sensor network design / evaluation method on a sensor network design / evaluation system including an input / output device, a storage device, and an arithmetic device will be described. The evaluation device has an input device for inputting various data and parameters, a database for storing input parameters and the like, a calculation device for executing a predetermined calculation process on the input parameters, and an output device for outputting a calculation result. .
Each sensor node has a sensing function, a communication function, and a relay function of sensing data, and has a built-in battery. Data detected by each sensor node is transferred to the gateway by relaying the other sensor nodes by multi-hop of a wireless link. The gateway connects to the Internet. Accordingly, each of the sensor nodes 1 to n is sequentially relayed by the intermediate sensor nodes, transferred to the gateway 1, and transmitted to a workstation or the like via the Internet 2.
[0014]
The operation rate evaluation of the sensor network is performed according to the following steps.
First, the number n of sensor nodes to be installed and the sensing area area (distance) D are determined and input. A more general network topology is broken down into a simple row of configurations as follows. FIG. 2 shows a linear network configuration. In FIG. 2, a gateway 2 is connected to the Internet 1, and the gateway 2 is a linear network in which n sensor nodes 3a to 3n arranged in a line are connected in a line.
[0015]
Next, the connection relationship between the topology and the sensor nodes of the sensor network is input (FIG. 3). Then, a path through which the sensing data is transmitted is input (FIG. 4). A tree structure with the gateway as an end point is usually used. The arithmetic unit extracts a series of nodes from the gateway to the terminal sensor node (FIG. 5). It has a one-row configuration with the gateway as the end point. The identification numbers are sequentially assigned to the n sensor nodes from the side closer to the gateway 2.
[0016]
The sensor node i always consumes _ai power every second, and in addition to the constant power consumption _ai , the following power consumption occurs at the time of sensing and data transfer. First, the sensor node i consume power b _i for each sensing. This power b _i is assumed to be random variables according to the exponential distribution with mean 1 / beta _i. Then, power is consumed c _i each time relays sensing data of the lower node. Here, c _i is a random variable that follows an exponential distribution with an average of 1 / γ _i . Sensing at sensor node i shall occur according to a Poisson distribution with an average of λ _i / sec. The initial battery of the sensor node i hi (time 0 can battery amount) is subject to the average 1 / eta _i. In the following description, data transfer delay is ignored.
[0017]
Next, parameters of the sensor node to be input will be described. Let x _{i be} the power consumed by sensor node i after time 0 due to sensing and relaying while all sensor nodes continue to operate (the battery does not run out). _{p t (x 1, ...... x} n) x 1 at the time t, the joint probability density function of ...... _{x n.} At this time, p _t satisfies the following equation.
(Equation 5)

_{q t (s 1, ...... s} n) of _{p t (x 1, ...... x} n) When the Laplas-Stieltjes transform (L.S.T.) , q t is expressed by the following equation.
(Equation 6)

[0018]
Let y _{i be} the power consumed by sensor node i after time 0, with all sensor nodes kept operating (the battery does not run out). That is, at time _t, the _{y i = x i + a i} t. _{rt (y 1, ......... y n} ) y 1 at a time t, the joint probability density function of _{......... y n, ut (s 1} , ...... s n) , and the L. S. T. And Then
(Equation 7)

It becomes.
[0019]
From the above results, the operating rate w _t of the sensor network at the time t is expressed as follows.
(Equation 8)

From this equation, it can be seen that the operating rate w _t decreases exponentially with respect to t.
In particular, when the parameters (a _i , β _i , γ _i , η _i , λ _i ) are the same regardless of the sensor node number i (that is, for each i, a _i = a, β _i = β, γ _i = γ, η _i = η, λ _i = λ) and power consumption w _t are represented by the following equations.
(Equation 9)

In this case, it can be seen from the above equation that w _t decreases exponentially with respect to λ.
[0020]
Based on the above-mentioned items, as sensor node parameters for evaluating the sensor network, constant power consumption a _i / sec, average sensing power consumption 1 / β _i , average power consumption 1 / bit relaying of 1-bit sensing data of lower nodes are used. These parameters are input using γ _i , the average frequency of occurrence of sensing λ _i / sec, and the average initial battery amount (battery amount at time 0) 1 / η _i . In the case of the same parameter for all nodes, constant power consumption a / sec, average sensing power consumption 1 / β, average power consumption for 1-bit relay of sensing data of lower nodes, 1 / γ, sensing The average occurrence frequency λ / sec and the average initial battery amount (battery amount at time 0) 1 / η are input.
[0021]
Information about these input parameters and the number of nodes is stored in a storage device. Next, the arithmetic unit evaluates the sensor network green utilization rate according to the above-mentioned predetermined equation.
A predetermined and stored sensor network operation rate target value is compared with an evaluation result, and the result is displayed on an output device. If the evaluation result is lower than the target value, a request for changing the parameters of the sensor nodes or the number of sensor nodes is displayed on the output device. In the case of a general network topology, a request for a change in a connection-related path or a change in a sensor node parameter is displayed on an output device. In accordance with this, the user changes and inputs parameters and the like of the sensor node. The input parameters are stored, and the calculation is executed again by the calculation device based on the evaluation formula. The sensor network operation rate target value is compared with the evaluation result, and the result is displayed again on the output device. This procedure is repeated until the evaluation result exceeds the target value, and the operation rate of the sensor network is optimized.
[0023]
When the maximizing option is specified, an input screen for the parameter range of the sensor node appears. In accordance with this, the user inputs the parameter range. The evaluation formula is evaluated for each value within this parameter range, and the parameter value and the evaluation value at which the sensor network operating rate is maximized can be displayed on the output device.
[0024]
Next, an optimal installation method when a sensor node is installed in a certain area will be described. Because the area is simplified, and the one-dimensional length D, and the distance between the sensor nodes i and i-1 and d _i. Power consumption and detection frequency by sensing and relaying the parameters _b i which depends on the distance between the sensor _{node, 1, b i, 2,} c i, 1, c i, 2 can be used. These parameters have the following relationship with sensor node parameters: 1-sensing average power consumption 1 / β, average power consumption 1 / γ for relaying 1-bit sensing data of lower nodes, and sensing occurrence frequency λ. _{1 / β i = b i,} 1 + b i, 2 d i m, / γ i = c i, 1 + c i, 2 d i m, λ i = ld i, D = Σ i = 1 n d i (m = a 2) which, instead of such a node number, given the distance d _i between the sensor nodes, to evaluate based on the evaluation formula, the evaluation of the optimization for operating rate of the sensor network for sensing an area It becomes possible. In the following numerical examples, the following values can be used as an example for each sensor node. t = 1 (year) = 31536000 seconds, m = 2, ai = 1 (nj), bi _{, j} = 95 (nJ), bi _{, 2} = 0.01 (nJ), ci _{, 1} = 180 ( nJ), c _{i, 2} = 180 (nJ), 1 / η _i = 10 (J), l _i = 0.0001 (i = 1... n)
[0025]
[Evenly arranged]
Examples of evenly arranging the sensor node _(i.e., d i = D / n) Consider. The number of nodes n is
Formula n = D {(m-1) ci _{, 2} / ci _{, 1} } ^{1 / m}
Will be referred to as a uniform arrangement method. The equal arrangement method gives the number of nodes that minimizes the average power consumption when the power consumption for relay is the dominant term under the condition that the sensor nodes are arranged uniformly. When the sensor nodes are arranged uniformly, the relationship between the number of sensor nodes and the network operation rate is as shown in FIG. 6, but according to the uniform arrangement method, n = 4, and the number of nodes given by the uniform arrangement method is If it is used, the number of nodes is also optimal in terms of the operation rate.
[0026]
[Non-uniform arrangement]
Excluding the condition that the sensor nodes are installed at equal intervals, an installation method that maximizes the sensor network operation rate in a given number of sensor nodes n (the number of sensor nodes is fixed) is examined. Now, by the _{d i = θ i D / (} Σ i = 1 n θ i), defined on the basis of _{θ n = 1, θ j (} j = 1 ...... n-1) following equation, relay The sensor network can be designed to minimize the average power consumption when the distance proportional term is the dominant term in the power consumption for. That is, the equation

The optimization under the condition that _{d i + ... d n = D} . This is obtained by setting d _i = θ _i D = 1 to directly obtain Expression 11.
(Equation 11)

[0027]
FIG. 7 shows that the non-uniform arrangement method given above is also effective in terms of the sensor network operation rate, evaluated by Expression 7 (or 8). FIG. 7 is a graph of the operating rate when the nodes are arranged at equal intervals while the number of nodes is variable, and when the intervals are determined and arranged according to Equation 11. It can be seen that the operating rate of the latter is higher than the former. FIG. 8 shows the sensor node interval determined by Expression 11. As can be seen from this, the arrangement method based on Expression 11 is an arrangement method in which the distance between nodes is shortened as the distance from the gateway increases. The closer the node is to the gateway, the higher the power consumption due to the relay of the sensing data from the lower node, but it is considered that this is controlled by shortening the distance to the adjacent node to increase the operation rate.
[Brief description of the drawings]
FIG. 1 is a diagram illustrating an example of a sensor network.
FIG. 2 is a diagram in which a sensor network is developed into a linear network.
FIG. 3 is a diagram illustrating a connection relationship between sensor networks.
FIG. 4 shows a transfer path of sensing data.
FIG. 5 is a diagram showing a sensor network in a tree structure based on a gateway.
FIG. 6 is a diagram illustrating a relationship between the number of sensor nodes and the operation rate of a sensor network when the number of nodes is variable.
FIG. 7 is a diagram illustrating a relationship between the number of sensor nodes and the operation rate when the number of nodes is fixed.
FIG. 8 is a graph showing a relationship between a sensor node number from a gateway and a node interval when the number of nodes is fixed.
[Explanation of symbols]
1 Internet 2 Gateway 3 Sensor node

Claims (7)

Equipped with a plurality of sensor nodes having a data sensing function, a communication function, a data relay function, and a battery, connected to a network via a gateway, and data sensed by each sensor node is relayed to another sensor node located in between. A method for evaluating an operation rate of a sensor network transferred via the gateway and the network,
As information used to evaluate the operation rate of the sensor network, including the number or sensing area area of the sensor nodes included in the sensor network, a transfer path of data sensed by the sensor nodes, and parameters of each sensor node,
Inputting information used for the operation rate evaluation,
Calculating an operating rate w _t of the sensor network at time t based on the input data using an arithmetic device. 2. The method according to claim 1, wherein the parameter of the i-th (i is a natural number of 1 or more) sensor node is a constant power consumption a _i / sec, and an average consumption required for one sensing. Using the power 1 / β _i , the average power consumption 1 / γ _i for 1-bit relay of the sensing data of the lower node, the average frequency of occurrence of sensing λ _i / sec, and the average initial battery amount 1 / η _i , When the number is n, the expression

Sensor Network uptime evaluation method characterized by calculating the operation rate w _t of the sensor network at time t based on. The method according to claim 2, wherein n sensor nodes having the same parameter are used as the sensor nodes, and the parameters of the sensor nodes are constant power consumption a / sec, and one sensing operation. Using the required average power consumption 1 / β, the average power consumption 1 / γ for 1-bit relay of the sensing data of the lower node, the average frequency of occurrence of sensing λ / sec, and the average initial battery amount 1 / η,

Sensor Network uptime evaluation method characterized by calculating the operation rate w _t of the sensor network at time t based on. Equipped with a plurality of sensor nodes having a data sensing function, a communication function, a data relay function, and a battery, connected to a network via a gateway, and data sensed by each sensor node is relayed to another sensor node located in between. A method of designing a sensor network transferred via the gateway and the network based on an operation rate,
As information for obtaining the operation rate of the sensor network, using the number of sensor nodes or the sensing area area included in the sensor network, a transfer path of data sensed by the sensor nodes, and parameters of each sensor node,
Inputting information used for the operation rate evaluation,
Storing the input information in storage means;
An arithmetic processing step of calculating an operating rate w _t of the sensor network at time t based on the input information using an arithmetic device,
The operation rate obtained as a result of the arithmetic processing is compared with the target operation rate, and if the operating rate is lower than the target value, any of the sensor node number, the data transfer path, or the parameter of the sensor node is changed and the arithmetic processing is performed again. And repeating the arithmetic processing until the value exceeds the target value. A method for designing a sensor network according to claim 4, wherein the arithmetic processing is performed using the expression according to claim 2 or 3. Equipped with a plurality of sensor nodes having a data sensing function, a communication function, a data relay function, and a battery, connected to a network via a gateway, and data sensed by each sensor node is relayed to another sensor node located in between. And a method for designing a sensor network transferred via the gateway and the network,
Defining a sensor network sensing area D and the number n of sensor nodes arranged in the sensing area;
Defining a connection relationship between the sensor nodes;
Comprising the step of defining the distance d _i between the sensor nodes,
Sensor network design method characterized in that the distance d _i between the sensor nodes, is set smaller than the spacing between the far side of sensor nodes apart gateway between the near side of the sensor node to the gateway. Equipped with a plurality of sensor nodes having a data sensing function, a communication function, a data relay function, and a battery, connected to a network via a gateway, and data sensed by each sensor node is relayed to another sensor node located in between. And a method for designing a sensor network transferred via the gateway and the network,
Defining a sensor network sensing area D and the number n of sensor nodes arranged in the sensing area;
Defining a connection relationship between the sensor nodes;
Comprising the step of defining the distance d _i between the sensor nodes,
The distance d _i between the sensor nodes, wherein

(However, it is assumed that θ1 = 1 _, and ci _{, 1} and ci _{, 2} are parameters relating to the sensor node i, and the average power consumption for each one-bit relay of the sensing data of the lower node is ci _{, 1} + ci _, 2di. ² )
A method for designing a sensor network, characterized in that the method is determined based on:

Images