KR100760041B1 - Routing path setup method based on virtual sink and its rotation for the wireless sensor networks - Google Patents

Abstract

The present invention discloses a sensor network communication system and a routing path establishment method used therein. The routing path setting method of the present invention is a method for establishing a routing path in a sensor network including a plurality of sensor nodes collecting predetermined information and a sink receiving information from the sensor nodes. Selecting a virtual sink from among; The virtual sink broadcasting a tree configuration message to a plurality of sensor nodes, and the plurality of sensor nodes receiving the tree configuration message rebroadcast the tree configuration message to construct a virtual sink tree; And setting the virtual path from the virtual sink to the sink by transmitting the virtual path setting message to the virtual sink through the sensor node adjacent to the sink. The present invention can extend the life of the entire sensor network by distributing the energy consumption of the sensor network, it is possible to implement a good scalable sensor network system at a low cost.

Description

Routing path setup method based on virtual sink and its rotation for the wireless sensor networks}

1A to 1D are diagrams illustrating a sensor network routing path setting method according to the prior art.

2 is a flowchart illustrating a routing path establishment method according to an exemplary embodiment of the present invention.

3A and 3B illustrate an example of configuring a virtual sink tree according to a preferred embodiment of the present invention.

4 is a flowchart illustrating a virtual path establishment process and an information transfer process when a new sink is added to a sensor network according to a preferred embodiment of the present invention.

FIG. 5 is a flowchart illustrating a process of removing a virtual path and resetting a virtual path when the sink moves in the sensor network, and FIG. 6 is a diagram illustrating a configuration of the sensor network when the sink moves. .

7 is a flowchart illustrating a method of repairing and managing a virtual sink tree.

8A to 8C illustrate the average message delay time (FIG. 8A), the average energy consumed (FIG. 8B), the message transmission efficiency (FIG. 8C), and the number of sinks from 1 to 16 in a sensor field environment of 400 sensor nodes. It is a figure which showed the measured result graphically.

9A-9C compare the average message delay time (FIG. 9A), average energy consumed (FIG. 9B), and message transmission efficiency (FIG. 9C) of the present invention as the number of sources varies from 1 to 16. FIG. One drawing.

10A to 10C measure the average message delay time, average energy consumed, message transmission efficiency, etc. when the sink speed is increased from 0 m / s to 20 m / s in a sensor field environment with 400 nodes. Each figure shows the result.

FIG. 11 is a diagram showing the results of message transmission efficiency experimented by partially breaking any node for 30 seconds out of 100 seconds of simulation time.

12A and 12B show the average consumed energy when increasing the number of sensor nodes.

13 is a diagram illustrating a performance measurement result of the present invention according to the virtual sink selection method.

The present invention relates to a sensor network communication system and a method for establishing a routing path in the sensor network communication system. Specifically, the present invention provides a sensor network communication system and routing used therein that can increase the efficiency of message transmission between each sensor node and a mobile sink and extend the life of the network by establishing a virtual sink in the sensor network communication system. A route setting method.

A general mobile communication system transmits and receives data between a mobile communication terminal and a base station. That is, the mobile terminal and the base station directly transmit and receive data without passing through other mobile communication terminals / nodes. However, since such a conventional mobile communication system depends on an infrastructure called a base station, there is a problem in that communication cannot be performed in an environment and an application area in which the infrastructure is not provided.

In order to solve this problem, Ad-hoc network method and sensor network method have been developed as a technology for configuring a communication network using only mobile communication nodes without relying on infrastructure.

1A is a diagram illustrating a basic concept of a general sensor network. Referring to FIG. 1A, a sensor node collects information about a target set by a designated user. Target data collected by sensor nodes include ambient temperature, object movement, seismic measurements, tsunami measurements, and unattended monitoring of battlefields. The sensor node transmits the collected information to the sink node. The sink node receives data transmitted by the sensor nodes forming the sensor network. The sensor node located within a certain distance from the sink node directly transmits data to be transmitted to the sink node. However, the sensor node not located within a certain distance transmits the collected data to the sensor nodes adjacent to the sink node instead of transmitting the collected data directly to the sink node.

As described above, the reason why a node that is not located within a certain distance transmits data using adjacent nodes is to minimize power consumption due to data transmission. That is, the power consumed by the sensor node to transmit data to the sink node is generally inversely proportional to the square of the distance between the sink node and the sensor node. Therefore, the sensor node that is not located within a certain distance from the sink node can minimize the power consumption due to the data transmission by transmitting the collected data using a plurality of sensor nodes. Of course, the relay node that receives data from the other sensor node and delivers the data to the sink also transmits the collected data to the sink node using another relay node or directly.

Since the sensor nodes used in such a sensor network are severely constrained by energy, computing power, and storage capacity, a protocol for saving energy should be applied. Scalability over network size is important.

1B and 1C are diagrams illustrating an example of a routing protocol of a sensor network according to the first prior art. The first prior art shown in FIGS. 1B and 1C is “Intanagonwiwat, C., Govindan, R. and Estrin, D., Directed Diffusion: A Scalable and Robust Communication Paradigm for Sensor Networks, In Proceedings of the Sixth Annual International Conference on Mobile Computing and Networks, pp. 2 16, Vol. 11, Issue 1, 2000 ".

Directed Diffusion (hereinafter referred to as "DD") is a data-driven routing scheme based on the query of the sink and has characteristics suitable for query distribution and processing application. In the DD scheme, the sink broadcasts an interest message corresponding to the query to the entire network, and each source detects an event that matches the query and transmits data to the sink through all paths where the interest message was initially received. A sink receiving data through a plurality of paths selects one path having the least delay, and data transmitted from a later source is transferred to the sink through the selected path.

However, in the DD scheme, as the number of sources and sinks increases, overhead due to flooding and path selection of a broadcast message due to broadcasting increases rapidly, and thus, it is effective for a sensor network having a small number of nodes. In addition, when the sink moves, since the position of the sink must be transmitted to all nodes every time the sink moves, there is a problem in that it is difficult to apply to a mobile sink that can freely move in the sensor field due to the rapid increase in overhead.

1D is a diagram for explaining an example of a routing protocol of a sensor network according to the second conventional technology. The second prior art shown in FIG. 1D is " Ye, F., Luo, H., Cheng J., Lu, S., Zhang, L., A Two-Tier Data Dissemination Model for Large-Scale Wireless Sensor Networks, In Proceeding of Mobile Computing and Networks, pp. 148 159, Sep. 2002 ". Referring to FIG. 1D, in a Two-Tier Data Dissemination (hereinafter, referred to as "TTDD") method, when a source detects an event to be delivered to a sink, a grid structure is broadcast by broadcasting the data announcement to the entire network. When the sink selects an adjacent dissemination node and notifies its location, the sink's location information is transmitted to the source through the grid, and the source transmits the detected information to the sink through the grid.

Since the TTDD scheme forms a grid around the source, there is less control overhead according to the number of sinks. In addition, even if the sink moves, since the sink position in the cell is flooded in the cell and connected to the dissemination node, the sink position is not known to the entire network, thereby reducing the overhead of the sink movement.

에 해당하는 전송지연을 더 가지게 된다. However, in the TTDD scheme, as the number of sources increases, a grid structure must be formed for each source, and the overhead for maintaining the grid of each source becomes large. In addition, in terms of transmission delay, TTDD has a grid structure, so it is more than unicast transmission.

It will have more transmission delay.

In addition, the performance of TTDD is greatly affected by the size of the cell. If the size of the cell is too small, the overhead caused by the new dissemination node selection is increased due to frequent cell-to-cell movement of the mobile sink. If the cell size is too large, the flooding overhead inside the cell is increased due to the mobile sink.

In particular, it is assumed that both of the above-described DD scheme and TTDD scheme obtain position information through GPS. However, GPS is not only available externally, but due to the cost of the GPS receiver, it is not suitable for many sensor nodes. Therefore, the routing protocol of the conventional mobile sensor network with GPS is difficult to apply in practical terms.

The present invention has been made to solve the problems of the prior art as described above, the technical problem to be achieved by the present invention, to provide a low power high performance routing path setting method that efficiently supports a plurality of sinks and a plurality of sources. It is to provide a way to increase the life of the sensor network by distributing energy consumption to the entire sensor network.

In addition, another technical problem to be achieved by the present invention is to establish a routing path with scalability and economy that can be used for large-scale sensor networks consisting of low-cost sensor nodes or Active RFID tags and future ubiquitous networks without using GPS. To provide a way.

The routing path setting method of the present invention for achieving the above-described technical problem, a method for establishing a routing path in a sensor network including a plurality of sensor nodes for collecting predetermined information and a sink receiving information from the sensor nodes. A method comprising: (a) selecting a virtual sink among a plurality of sensor nodes; (b) the virtual sink broadcasting a tree configuration message to a plurality of sensor nodes, and the plurality of sensor nodes receiving the tree configuration message rebroadcast the tree configuration message to construct a virtual sink tree; And (c) the sink transmitting a virtual path setting message to the virtual sink through the sensor node adjacent to the sink, thereby establishing a virtual path from the virtual sink to the sink.

In addition, in the above-described virtual path setting method, step (a) selects a new virtual sink when the energy level of the sensor node selected as the virtual sink falls below a predetermined reference level, and (b) and (c) for the new virtual sink. Steps may be performed.

In addition, in step (a), the virtual sink may be randomly selected from among the plurality of sensor nodes.

In addition, when the sink is added to the sensor network, the above-described virtual path setting method may include: (d) selecting a sensor node having a minimum hop number among sensor nodes adjacent to the sink added to the sensor network; And (e) setting the virtual path by transmitting the virtual path setting message to the virtual sink by the selected sensor node.

In addition, when the sink is moved, the above-described virtual path setting method includes: (f) the sink transmitting a virtual path removal message to the virtual sink, and the virtual sink receiving the virtual path removal message to remove the virtual path; And (g) after the sink is moved, transmitting the virtual path setting message to the virtual sink through the selected sensor node adjacent to the sink to reset the virtual path from the virtual sink to the sink. .

In addition, the step (f) described above, (f1) the sink transmits a virtual path removal message to the sensor node on the virtual path connected to the sink; (f2) sensor nodes on the virtual path receiving the virtual path removal message from the lower node and transmitting the received virtual path removal message to the upper node, and returning information transmitted to the sink to the virtual sink; And (f3) the virtual sink receiving the path removal message to remove the virtual path and storing the returned information.

In addition, the above-mentioned step (g), (g1) after the sink is moved, selecting a node having the minimum number of hops among the adjacent sensor nodes, and transmitting a virtual path setting message to the virtual sink through the selected sensor node; And (g2) the virtual sink receiving the virtual path setup message to reset the virtual path and transmitting the stored information to the sink through the reset virtual path.

In addition, when any sensor node is broken in the virtual sink tree, a child node of the broken sensor node selects an adjacent parent node as a parent node; If the adjacent parent node does not exist, the child node selects an adjacent peer node that is not a sibling node as a parent node; And selecting a child node as a parent node if there is no neighboring peer node other than the sibling node, and if the child node is a node on the virtual path, sending a virtual path setting message to the newly selected parent node. The method may further include transmitting.

On the other hand, the sensor network communication system of the present invention for achieving the above technical problem, a sensor network communication system including a plurality of sensor nodes for collecting predetermined information and the sink receiving the information, of the plurality of sensor nodes A virtual sink selected to collect information from the plurality of sensor nodes and transmit the information to the sink through the sensor nodes; A plurality of sensor nodes which collect predetermined information and transmit the predetermined information to the virtual sink, and transmit the information transmitted from the virtual sink to the sink; And a sink connected to an adjacent sensor node among the plurality of sensor nodes to receive information transmitted from the virtual sink.

In addition, the above-described plurality of sensor nodes may receive a tree setting message broadcast from a virtual sink from an adjacent sensor node, set an adjacent sensor node as its parent node, and rebroadcast the tree setting message to another adjacent sensor node. By doing so, it is preferable to be hierarchized in a tree structure.

In addition, when the energy level of the above-described virtual sink is below a predetermined reference level, it is preferable that a new virtual sink is selected.

In addition, the above-described virtual sink may be randomly set among a plurality of sensor nodes, and an optimal node may be selected in consideration of other factors such as the energy level of the node and the distance to the sink as needed. The rotation of the virtual sink is intended to extend the life of the network by distributing the energy consumption of the sensor network.

In addition, the sink added to the above-described sensor network communication system selects a sensor node having the minimum number of hops among adjacent sensor nodes, transmits a virtual routing message to the virtual sink through the selected sensor node, and the virtual sink is a virtual routing message The virtual path from the virtual sink to the sink may be set by receiving the.

In addition, when the sink moves, the sink transmits a virtual path removal message to the virtual sink through the connected sensor node, and after the move, the virtual path setting message is transmitted to the virtual sink through the selected sensor node among adjacent sensor nodes. The virtual sink may remove the virtual path by receiving a virtual path removal message and set a new virtual path from the virtual sink to the sink by receiving a virtual path setting message from a newly selected sensor node.

In addition, when the sink moves, the plurality of sensor nodes on the virtual path receive the virtual path removal message from the lower node and transmit the received message to the upper node, and return the information transmitted to the sink to the virtual sink, and the virtual sink returns the returned information. The stored information may be transmitted to the sink through the new virtual path.

In addition, when a plurality of sensor nodes is damaged, a plurality of sensor nodes select an adjacent parent node of the same layer as the parent node as a parent node, and if an adjacent parent node does not exist, an adjacent peer that is not a sibling node. If a node is selected as a parent node and there is no neighbor peer node that is not a sibling node, a child node can be selected as a parent node.

In addition, when the above-described sensor node is a node on the virtual path, the sensor node may transmit a virtual path setting message to the newly selected parent node.

Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings.

2 is a flowchart illustrating a routing path establishment method according to a preferred embodiment of the present invention. Referring to FIG. 2, in the routing path setting method of the present invention, first, a virtual sink is selected from a plurality of sensor nodes (S200), a virtual sink tree is configured based on the selected virtual sink (S210), and a virtual sink. The virtual path from the virtual sink to the actual sink is set in the virtual sink tree (S220). When the routing path between the virtual sink and the actual sink is established by performing steps S200 to S220, the virtual sink collects information from the sensor nodes (S230) and transmits the collected information to the virtual sink through the virtual path ( S240).

Hereinafter, each step will be described.

First, step S200 of selecting a virtual sink will be described.

The virtual sink of the present invention is a sensor node representing the role of all sinks. The virtual sink is selected from among a plurality of sensor nodes constituting the sensor network according to a predetermined rule described later. In the virtual sink of the present invention, it is preferable that a new virtual sink is selected when the energy level of the virtual sink falls below a reference level, or a new virtual sink may be set at a predetermined time period. That is, the virtual sink of the present invention is not fixed to one sensor node, but different sensor nodes are cyclically selected.

An important consideration in sensor network protocol design is that the energy consumption of the sensor nodes around the sink is very high. The sink communicates with several sensor nodes around it. This many-to-one communication causes a lot of traffic around the sink, so that nodes around the sink consume more energy than nodes that are relatively remote from the sink. . This phenomenon also occurs in the virtual sink peripheral nodes of the present invention, and this energy consumption phenomenon is more serious because the present invention selects a general sensor node as the virtual sink.

Therefore, step S200 according to the preferred embodiment of the present invention selects a new virtual sink when the energy of the sensor node selected as the virtual sink falls below a predetermined reference level in order to distribute energy consumption throughout the network. In addition, it will be appreciated by those skilled in the art that the virtual sink may be changed to be selected every predetermined time period.

Referring to the method of selecting a virtual sink among the sensor nodes, if there is only one sink in the sensor field and there is no change in the location, it is preferable that one of the neighboring nodes of the sink becomes the virtual sink. Also, when a new sink is added to the sensor field, the position of the virtual sink may be relatively far away depending on the previous sink position. In order to minimize the transmission path of information, it is desirable to position the virtual sink so that the communication distance between each sink and the virtual sink is minimized. Therefore, the position of the virtual sink is most preferably at the center of gravity of all the sinks, assuming that the source as the sensor node is distributed evenly over the entire sensor field. However, since the present invention assumes that the addition and movement of each sink in the sensor field is free, it is preferable to adopt the method described below rather than this virtual sink selection method.

The preferred embodiment of the present invention extends the life of the entire sensor network by distributing the energy consumed in the sensor network by selecting the virtual sink in consideration of the energy of the entire sensor node, rather than considering the position of the virtual sink in the sensor field. .

The most preferable virtual sink selection method in terms of this energy distribution is to select the node with the highest residual energy among the sensor nodes as the virtual sink. Considering the structure of the virtual sink tree, which will be described later, the sensor node selected as the virtual sink will be the lowest end node of the virtual sink tree.

However, finding the sensor node with the maximum energy requires additional overhead, so the preferred embodiment of the present invention adopts a method of randomly selecting a virtual sink as a simpler method. Since the present invention assumes that the number of sensor nodes is large enough, a good effect can be obtained.

Hereinafter, operation S210 of configuring a virtual sink tree with respect to the selected virtual sink will be described.

The sensor node selected as the virtual sink collects information from the source and delivers the collected information to the actual sink.

The virtual sink constructs an efficient spanning tree called a virtual sink tree in the present invention to deliver the collected information to the actual sink. The spanning tree contains all the nodes in the field, collecting and merging information from the source. The virtual sink tree is constructed as the sensor nodes are sown, and the virtual sink is changed or reconfigured according to the appearance of a new node due to the depletion of the virtual sink or the breakdown of the sensor node, which is a virtual sink.

Once the virtual sink tree is constructed, the virtual sink acts as the information delivery center of the sinks. The virtual sink configures an information delivery path centered on the virtual sink to deliver information to the sink. If there are multiple sinks, you can create a multicast tree from the virtual sink to the actual sink. Virtual Paths and Information Delivery The multicast tree is embedded within the virtual sink tree.

According to the present invention, since only the virtual sink needs to know the path to the actual sink by introducing the virtual sink, it is possible to support a sink having a plurality of mobility without requesting information on the location of the continuous sink of the sensor nodes. In addition, when a virtual sink is selected, the actual sink must transmit its location information to the virtual sink. The present invention is different from the conventional method in which the location information of the sink is transmitted locally or throughout the network. The location of the sink can be easily transmitted to the virtual sink by using an information collection path (report path) from the existing sensor node to the virtual sink.

In detail, the present invention configures a virtual sink tree by classifying nodes in a sensor field into layers. Nodes in the same layer have the same hop count from the virtual sink. Nodes in Layer 1 are one hop away from the virtual sink. Layer 2 nodes are two hops away from the virtual sink.

3A and 3B illustrate an example of configuring a virtual sink tree, and FIG. 3A illustrates an example in which 26 sensor nodes are classified into four layers to form a virtual sink tree. As shown, all sensor nodes in the sensor field form a spanning tree rooted in the virtual sink.

The process of configuring the virtual sink tree is initiated by the virtual sink broadcasting the tree setup message. A node broadcasting a tree configuration message transmits its own address, parent address, and hop count information to a virtual sink in neighboring nodes in a tree configuration message. Nodes receiving the tree configuration message directly from the virtual sink become layer 1 nodes. Layer 1 nodes retransmit the tree configuration message to the sensor nodes connected to the node 1 and transmit the tree configuration message to the layer 2 nodes, and layer 2 nodes also transmit the tree configuration message to the nodes of the lower layer connected to the node 2 nodes. The upper nodes overhearing the tree configuration message broadcasted by the lower layer node register the lower layer node as their child node if the parent address of the message is their own. This process is repeated until all network nodes have joined the tree.

Since wireless communication uses broadcasting, it is possible to quickly send tree configuration messages from the root to each node. In order to reduce the communication delay time between the sensor node and the virtual sink, the height of the tree should be minimized. The height of each node represents the number of hops to the virtual sink. As a result, the virtual sink tree preferably forms a wide, low tree in the sensor field. The present invention constitutes a spanning tree having a minimum height.

Neighbor nodes in the virtual sink tree are classified according to their proximity to the virtual sink by UP, DOWN, and PEER. If node i is UP to node j then H (i) <H (j). At this time, H (i) means the number of hops of the node i. If node i is DOWN with respect to node j, then H (i)> H (j). Finally, node i, called PEER, is H (i) = H (j) compared to node j. Each node must have at least one UP node to pass information to the virtual sink. Each node selects one node as a parent node among UP nodes, and it becomes a child node of the parent. Nodes on the path (virtual path) for transferring information from the virtual sink to the sink should be able to transfer the information received from the UP node to the DOWN node. In conclusion, each sensor node should have at least one parent node and zero or more child nodes. Therefore, all neighbor nodes are divided into nodes of UP, DOWN, and PEER as shown in FIG. 3B.

Hereinafter, step S220 will be described. If the virtual sink tree is configured in step S210 described above, the virtual sink should perform a function of collecting information from all sources and transferring the information to the sink. To this end, each sensor node has an upstream path to a virtual sink through a parent node called a report path to send detected information to the virtual sink, and the virtual sink is a virtual path for transmitting information collected from a source to a real sink. Must have a downward path called. In order for each sink to configure the virtual path, the actual sink location information must be transmitted to the virtual sink. This process is called virtual path setup. Since the sensor nodes in the virtual sink tree know the report path to the virtual sink as the virtual sink tree is configured, the virtual path setting can be easily configured by the existing report path.

To establish the virtual path, the real sink sends a virtual path establishment message to the connected sensor node (hereinafter referred to as a "forward agent"), and the forward agent receiving the virtual path establishment message passes through its report path to the virtual sink. Pass the virtual routing message. Each node on the virtual path records the direction information of the actual sink therein and transmits the virtual path setting message received from its child node to its parent node along its report path.

When the virtual sink receives the virtual path setting message, the virtual sink sets the virtual path in the reverse direction of the report path where the virtual path setting message is received and records it therein. Then, when collecting information from the sensor nodes, the virtual sink receives the virtual path through the virtual path. Send to the sink.

4 is a flowchart illustrating a virtual path establishment process and an information transfer process when a new sink is added to a sensor network according to a preferred embodiment of the present invention.

As described above, when a virtual sink is selected and a new sink is added to the sensor network in which the virtual sink tree is configured according to the selected virtual sink, the added sink examines adjacent sensor nodes and has a minimum number of hops to the virtual sink. The node is selected as a forward agent (S400).

The actual sink transmits the virtual path setting message to the virtual sink through the report path of the forward agent (S410).

The parent node of the forward agent that receives the virtual routing message from the forward agent remembers the direction of the actual sink therein and transmits the virtual routing message to its parent node. It records therein and transmits a virtual path setting message to a higher node (S420).

The virtual sink that receives the virtual path setup message from the node of the first layer sets the virtual path by recording the direction of the actual sink (S430).

Thereafter, the virtual sink collects information from the sensor nodes and transmits the collected information to the actual sink through the virtual path (S440).

On the other hand, as the number of sinks added to the sensor network increases, virtual paths for multiple sinks may be configured as a multicast tree without additional overhead. If a node of the virtual sink tree belongs to two virtual paths, the two virtual paths will naturally merge. As a result, the tree will be connected to two mobile sinks starting from the virtual sink. As the number of sinks increases, the number of virtual paths increases. As the number of virtual paths increases, the virtual sink becomes the root of the tree, and each sink becomes an end node of the tree.

This tree is called an information delivery multicast tree, and information can be efficiently delivered to the sink through the multicast tree. Multicast trees can take advantage of locality between sinks by concatenating each virtual path. The virtual path and data delivery multicast tree are shown in FIG. 6. The virtual sink is an information delivery multicast tree that serves as an efficient information delivery center for all sinks in the sensor field.

FIG. 5 is a flowchart illustrating a process of removing a virtual path and setting up a virtual path when an actual sink moves within a sensor network, and FIG. 6 is a diagram illustrating a configuration of a sensor network when an actual sink moves. Drawing.

5 and 6, when the sink receiving the information from the virtual sink is moved because the virtual path is set, the actual sink first transmits a virtual path removal message to the forward agent (S500).

The forward agent transmits the virtual path removal message to the virtual sink through the virtual path, and each sensor node on the virtual path recognizes the virtual path removal message and returns information transmitted from the virtual sink to the actual sink to the virtual sink (S510).

 Upon receiving the virtual path removal message, the virtual sink removes the virtual path recorded for the actual sink and temporarily stores information returned from the sensor nodes on the removed virtual path (S520).

On the other hand, the actual sink that has completed the movement selects the neighboring node having the least hop count to the sink among the adjacent sensor nodes as the forward agent as in step S400, and forwards the agent in the same manner as step S410 described above. The virtual path setting message is transmitted to the virtual sink through the sensor nodes on the report path of the forward agent, and the new direction of the actual sink is recorded therein (S530).

Upon receiving the virtual path setting message, the virtual sink establishes a virtual path and actually transmits information, which is returned from the sensor nodes on the conventional virtual path and stored therein, to the sink through the newly set virtual path (S540).

7 is a flowchart illustrating a method of repairing and managing a virtual sink tree. Referring to FIG. 7, the sensor node may be damaged due to physical shock or energy consumption. However, it is not desirable that the routing function of the sensor network be affected by the breakdown of these nodes. Therefore, the preferred embodiment of the present invention adopts the report path and virtual path repair algorithm of the broken node as shown in FIG. If the broken node is an end node of the virtual sink tree, the broken node does not affect the network. Thus, the repair algorithm shown in FIG. 7 is applied when the broken node is a node of the middle layer.

If the node of the virtual sink tree is broken, the child nodes of the broken node cannot communicate with the parent node, and thus the connection with the virtual sink tree is broken (S700).

The child node of the broken node disconnected from the virtual tree checks whether an adjacent parent node exists (S710), and if an adjacent parent node exists, selects the parent node as a parent node (S720). At this time, the selected parent node belongs to the same hierarchy as the existing parent node.

On the other hand, when there is no adjacent parent node, the disconnected child node checks whether there is an adjacent peer node that is not a sibling node (S730), and if present, selects an adjacent peer node as a parent node (S740).

On the other hand, if there is no peer node that is not an adjacent sibling node, it notifies the child nodes that all uplink paths are broken, and selects a child node that can select a parent other than itself as a new parent node among the child nodes. It becomes an end node of the virtual sink tree (S750). By performing steps S700 to S750 described above, the report path is reset.

On the other hand, if the disconnected child node belongs to the virtual path, the child node sends the virtual path setting message to the newly set parent node, and the newly set parent node delivers the virtual path setting message to the virtual sink through its report path to the virtual sink. A path is set (S760).

Hereinafter, the performance of the routing path establishment method according to the preferred embodiment of the present invention described above will be described in comparison with other conventional routing methods. First, the routing path establishment method of the present invention is described in the ns-2 simulator (http://www.isi.edu/nsman/ns), which is well known to those skilled in the art. Was implemented.

The MAC layer of the routing path establishment method of the present invention is designed as 1.6Mbps 802.11 MAC. MAC protocols previously proposed in sensor networks (Christian C. Enz., Amre, EH, Decotignie, JD, Peiris, V., WiseNET: An Ultralow-Power Wireless Sensor Network Solution, IEEE computer magazine, Vol. 37, No. 8, Aug 2004, and Ye, W., Heidemann, J., Estrin, D., An Energy-Efficient MAC Protocol for Wireless Sensor Networks, In Proceedings of the INFOCOM 2002, pp. 1567 1576, Vol. 3, 2002. ) Has a standby mode during the IDLE period. On the other hand, 802.11 MACs are not suitable for sensor networks because radios of 802.11 MACs consume energy even during the IDLE period. To make the model close to the sensor network, the power consumption during the IDLE period was set to 35mW, the power consumption at reception was set to 395mW, and the power consumption at transmission was set to 660mW. The power consumption is set to the same value to compare with the experiments in the existing sensor network.

In order to analyze the performance of the present invention with respect to the network size, a sensor field having several sizes is generated. The sensor field size is set to have 800 nodes by increasing by 100 from 100 nodes. In general experiments, 400 sensor nodes were randomly located at 2000 ㅧ 2000m. The radio range of the node is 250m. In order to keep the density of the node constant, the experiment was performed by increasing the size of the sensor field without changing the radio range. All sources were randomly selected according to the random source model, and sinks were evenly sprayed onto the sensor field. Each source generates one piece of information to be transmitted per second, and the size of each information is 64 bytes. Each information is modeled so that it is passed to all sinks in the field. Each simulation was run for 100 seconds.

In order to compare the performance of the present invention with the conventional technology, it is analyzed using three characteristics of average transmission delay time, average energy consumed, and message transmission efficiency. The average transmission delay time is the average time taken when information is transmitted from the source to the sink. Average transmission latency refers to the performance in the sensor network for data transmission. Average energy consumed is the average value of energy consumed per node. It represents the average energy of the nodes needed to deliver a certain number of information up to the sink. Energy consumption is also related to network life. Message transmission efficiency is the ratio of transmitted data to information arriving at the sink. In this experiment, these characteristics were compared by varying the number of sinks, the number of sources, the speed of the mobile sink, and the number of sensor nodes.

First, the performance of the present invention was measured according to the change in the number of sinks under the assumption that all the sinks were fixed. 8A to 8C illustrate the average message delay time (FIG. 8A), the average energy consumed (FIG. 8B), the message transmission efficiency (FIG. 8C), and the number of sinks from 1 to 16 in a sensor field environment of 400 sensor nodes. It is a figure which showed the measured result graphically. The simulation of FIGS. 8A to 8C is a case where the number of sources is 5, and the present invention is compared with the Directed diffusion protocol (hereinafter referred to as "DD") and Two-Tier Data Dissemination (hereinafter referred to as "TTDD"). The protocol is designed considering multiple sinks and multiple sinks with mobility.

Referring to FIG. 8A, as the number of sinks increases, the average message delay time of DD gradually increases. Directed diffusion protocol performs reinforcement and negative reinforcement from sink to source. As the number of sinks increases, traffic overhead and collisions caused by reinforcement and negative reinforcement that need to be performed increase, resulting in increased latency. The TTDD protocol is not as severe as the DD protocol, but as the number of sinks increases, the number of queries to maintain the grid increases, so the delay time increases.

On the other hand, in the case of the present invention, it can be seen that the delay time is constantly maintained regardless of the change in the number of sinks. Moreover, since the average message delay time is 20 ms or less, the performance is twice that of DD and TTDD except for one sink.

Referring to FIG. 8B, it can be seen that the average consumed energy of the present invention when the number of sinks is one is 3.8J. This value is about 35.8% of DD and 43.2% of TTDD. Since DD has a large overhead such as interest and reinforcement propagation, and TTDD has a packet overhead for maintaining a grid, energy consumption is larger than that of the present invention. It can be seen that the present invention consumes very little energy compared to other protocols.

8C shows the message transmission efficiency according to the change in the number of sinks. Compared to the present invention and the TTDD, the transmission efficiency of the DD is so low that only the TTDD and the present invention are compared. When the number of sinks is 1, 2, 4, and 8, the TTDD shows a transmission efficiency of 76% to 83%, while the present invention shows that almost all packets are successfully delivered to the sink.

9A-9C compare the average message delay time (FIG. 9A), average energy consumed (FIG. 9B), and message transmission efficiency (FIG. 9C) of the present invention as the number of sources varies from 1 to 16. FIG. One drawing. Here, the case of one sink is shown.

Referring to FIG. 9A, since DD finds an optimal path through a reinforcement process, it shows the result of the least delay time in one source compared to the present invention and TTDD. However, DD and TTDD gradually increase the delay time as the number of sources increases. Since DD performs reinforcement and negative reinforcement from sink to source, as the number of sources increases, traffic overhead and collisions caused by reinforcement and negative reinforcement that need to be performed increase, resulting in increased latency. TTDD is not as severe as DD, but as the number of sources increases, the number of grids that must be maintained increases the latency. However, in the case of the present invention, it can be seen that the delay time is constant even if the number of sources changes.

Referring to FIG. 9B, it can be seen that in all cases, the present invention consumes less energy than DD or TTDD. In the present invention, since a path that can be sent from all sources to the sink is already configured as compared to other protocols, the information to be transmitted increases only as the number of sources increases. In contrast, DD and TTDD gradually increase protocol overhead as the number of sources increases. The simulation results for 16 sources consume 25% less energy than DD and 34% less than TTDD. It is surprising to see how much less energy is consumed compared to DD and TTDD, which have data merging and caching capabilities.

Referring to FIG. 9C, while the transmission efficiency of TTDD decreases by 60% as the number of sources increases, the present invention shows that almost all data is transferred to the sink, which shows superior performance compared to TTDD.

10A to 10C measure the average message delay time, average energy consumed, message transmission efficiency, etc. when the sink speed is increased from 0 m / s to 20 m / s in a sensor field environment with 400 nodes. Each figure shows the result. Since DD is not a suitable protocol for mobile sinks, the experiment evaluated the performance of the present invention through comparison with TTDD.

FIG. 10A illustrates an average message delay time according to a change in a moving speed of a sink. Referring to FIG. 10A, the average message delay time of the present invention does not change even when the sink speed increases. This is because, although the sink moves fast, the present invention passes only one virtual path setting with less overhead and fast setting. TTDD, on the other hand, performs a local broadcast in the cell to select a new agent when the sink is more than a certain distance away from the main agent. The old agent also delivers messages that it could not deliver while the sink is moving to the new agent. This increases collisions and traffic, which increases message latency. The average message delay time of the present invention is 20 ms or less, more than twice as fast as TTDD.

10B shows the average consumed energy. Even though the sink speed increases, the average consumed energy of the present invention is almost unchanged to 5J. The results of TTDD similarly show energy consumption around 11J up to medium travel speeds. However, when the sink speed reaches 20m / s, the energy consumed per node is increased to 22.4J due to trajectory forwarding and many local broadcastings. As can be seen from the experimental results, the present invention shows better performance than TTDD in terms of energy and delay time.

10C shows the message transmission efficiency. Referring to FIG. 10C, the present invention shows a transmission efficiency of 91% or more at all moving speeds, whereas TTDD has a transmission efficiency of about 80% at a moving speed between 5m / s and 15m / s. Moreover, TTDD shows 30% transmission efficiency at 20 m / s. TTDD shows very poor performance at sink speeds above 20 m / s.

FIG. 11 is a diagram showing the results of message transmission efficiency experimented by partially destroying any node for 30 seconds during a simulation time of 100 seconds. Referring to FIG. 11, it can be seen that transmission efficiency gradually decreases as node failure increases to 15% or more of all nodes. It can be seen that the transmission efficiency drops drastically below 95% at 15% to 20% node failures. This is because broken nodes cannot create a hole in the sensor field and send it to any sink. However, it should be noted that the message transmission efficiency of the present invention exceeds 94%. This proves that the node maintenance and management algorithm described above is a reliable algorithm that fully considers the failure of a node.

12A and 12B show the average consumed energy when increasing the number of sensor nodes. Figure 12a is a measure of the energy consumed while increasing the density of the node at 2000 ㅧ 2000m size, Figure 12b is a measure of energy consumed while increasing the network size and the number of nodes while maintaining the density of the node as it is. 12A and 12B, it can be seen that the energy consumption of the present invention is almost constant even though the number of sensor nodes is increased. Even if the number of hops for the communication path increases due to the increase in the number of sensors, the average energy consumption per node is constant because the number of nodes participating in the communication increases.

However, in the case of TTDD and DD, as node density increases, energy consumption per node increases because the number of broadcasting due to interest and information transmission increases. On the other hand, since the virtual sink tree of the present invention connects as many child nodes as possible within one hop distance, the overhead due to data transmission is relatively low even though the number of nodes increases. Although omitted in FIG. 12B, similar results were obtained when the size of the sensor field was changed in DD and TTDD.

13 is a diagram illustrating a performance measurement result of the present invention according to the virtual sink selection method. The experiment has 5 sources, 4 sinks, and each selection method is simulated for 500 seconds. 13 illustrates a case in which (a) a virtual sink is fixed, (b) a sensor node having a maximum energy level is selected as a virtual sink, and (c) a virtual sink is arbitrarily selected from a plurality of sensor nodes. As shown, in the case of (b) and (c), there are no nodes that consumed all the energy as compared to the case of (a). As a result of the unexpected, the method of randomly selecting virtual sinks with the ratio of (a) 12.2%, (b) 5.7%, and (c) 2% of nodes with less than 6% of the total energy selects the sensor node having the maximum energy. Better performance. As a result, the method of selecting and circulating virtual sinks has the effect of preventing holes in the sensor node and extending the life of the sensor network.

So far I looked at the center of the preferred embodiment for the present invention. Those skilled in the art will appreciate that the present invention can be implemented in a modified form without departing from the essential features of the present invention. Therefore, the disclosed embodiments should be considered in descriptive sense only and not for purposes of limitation. The scope of the present invention is shown in the claims rather than the foregoing description, and all differences within the scope will be construed as being included in the present invention.

As described above, the present invention selects a virtual sink by circulating one of the plurality of sensor nodes included in the sensor network one by one, constructs a virtual sink tree for the selected virtual sink, and synchronizes the sensor nodes from the sensor nodes in the virtual sink. Collect the information sent. When the location of the mobile sink is determined, the real sink notifies its position to the virtual sink to establish a virtual path from the virtual sink to the real sink, and the virtual sink transmits information to the real sink.

As described above, since the present invention selects a virtual sink while circulating among the sensor nodes, it is possible to improve the performance of the entire network by distributing energy consumption throughout the network.

In addition, in the present invention, since only the virtual sink needs to record the position of the actual sink, each sensor node only needs to store the structure of its own upper node and lower node, and it is necessary to store information about the entire sensor network and its own location information. Since there is no, the mobile sink can be efficiently supported at a low cost, and the determinism is good.

Claims (23)

A method of establishing a routing path in a sensor network, the sensor network comprising a plurality of sensor nodes collecting predetermined information and a sink receiving the information from the sensor nodes. (a) selecting a virtual sink among the plurality of sensor nodes; (b) the virtual sink broadcasting a tree configuration message to the plurality of sensor nodes, and the plurality of sensor nodes receiving the tree configuration message rebroadcast the tree configuration message to construct a virtual sink tree ; And (c) the sink transmitting a virtual path setting message to the virtual sink through the sensor node adjacent to the sink to establish a virtual path from the virtual sink to the sink. . The method of claim 1, Step (a) selects a new virtual sink when the energy level of the sensor node selected as the virtual sink is less than a predetermined reference level, The step (b) and the step (c) is performed for the new virtual sink. The method of claim 1, Step (a) selects a new virtual sink every predetermined time period, The step (b) and the step (c) is performed for the new virtual sink. The method according to any one of claims 1 to 3, wherein step (a) And a virtual sink is randomly selected from the plurality of sensor nodes. The method according to any one of claims 1 to 3, wherein step (a) And selecting the sensor node having the highest residual energy as the virtual sink among the plurality of sensor nodes. The method according to any one of claims 1 to 3, (d) selecting a sensor node having a minimum number of hops among adjacent sensor nodes added to the sensor network by the sink; And and (e) setting the virtual path by transmitting the virtual path setting message to the virtual sink by the selected sensor node. The method according to any one of claims 1 to 3, wherein when the sink is moved (f) the sink transmitting a virtual path removal message to the virtual sink, and the virtual sink receiving the virtual path removal message removes the virtual path; And (g) after the sink is moved, transmitting the virtual path setting message to the virtual sink through the selected sensor node adjacent to the sink to reset the virtual path from the virtual sink to the sink. Routing path setting method comprising the. 8. The method of claim 7, wherein step (f) (f1) the sink transmitting the virtual path removal message to a sensor node on the virtual path connected to the sink; (f2) sensor nodes on the virtual path receiving the virtual path removal message from a lower node and transmitting the received virtual path removal message to an upper node, and returning information transmitted to the sink to the virtual sink; And (f3) the virtual sink receiving the path removal message to remove the virtual path, and storing the returned information. The method of claim 8, wherein step (g) (g1) after the sink is moved, selecting a node having the minimum hop number among adjacent sensor nodes and transmitting a virtual path setting message to the virtual sink through the selected sensor node; And (g2) the virtual sink receiving the virtual path setting message to reset the virtual path, and transmitting the stored information to the sink through the reset virtual path. The method according to any one of claims 1 to 3, If any sensor node in the virtual sink tree is broken, Selecting, by a child node of the broken sensor node, an adjacent parent node as a parent node; If a neighboring parent node does not exist, the child node selects a neighboring peer node that is not a sibling node as a parent node; And And selecting a child node as a parent node if there is no neighboring peer node that is not a sibling node. The method of claim 10, If the child node is a node on a virtual path, And sending a virtual path establishment message to the newly selected parent node. The method according to any one of claims 1 to 3, wherein the tree setting message is And at least one of an address of a sensor node broadcasting the tree configuration message, an address of a parent node, and hop count information to a virtual sink. A sensor network communication system comprising a plurality of sensor nodes collecting predetermined information and a sink receiving the information, A virtual sink selected from the plurality of sensor nodes to collect the information from the plurality of sensor nodes and transmit the information to the sink through the sensor nodes; The plurality of sensor nodes which collect the predetermined information and transmit the information to the virtual sink, and transfer the information transmitted from the virtual sink to the sink; And And a sink connected to an adjacent sensor node of the plurality of sensor nodes to receive the information transmitted from the virtual sink. The method of claim 13, wherein the plurality of sensor nodes are A tree configuration message broadcast from the virtual sink is received from an adjacent sensor node, and the adjacent sensor node is set as its parent node, and the tree configuration message is layered into a tree structure by rebroadcasting the tree configuration message to another adjacent sensor node. Sensor network communication system, characterized in that. The method of claim 13, And a new virtual sink is selected when the energy level of the virtual sink is less than or equal to a predetermined reference level. The virtual sink of claim 13, wherein the virtual sink is The sensor network communication system, characterized in that selected from the plurality of sensor nodes at a predetermined time period. 17. The virtual sink of any of claims 13-16 wherein the virtual sink is Sensor network communication system, characterized in that randomly selected from among the plurality of sensor nodes. 17. The virtual sink of any of claims 13-16 wherein the virtual sink is And a sensor node having the highest residual energy among the plurality of sensor nodes is selected. The method of claim 13, The sink added to the sensor network communication system selects a sensor node having a minimum number of hops among the adjacent sensor nodes, and transmits a virtual routing message to the virtual sink through the selected sensor node, wherein the virtual sink is the virtual sink. And receiving a path setting message to set a virtual path from the virtual sink to the sink. The method of claim 13, wherein when the sink moves, The sink transmits a virtual path removal message to the virtual sink through the connected sensor node, and after the movement, transmits a virtual path setting message to the virtual sink through a selected sensor node among adjacent sensor nodes. The virtual sink removes the virtual path by receiving the virtual path removal message, and sets the new virtual path from the virtual sink to the sink by receiving the virtual path setting message from the newly selected sensor node. Sensor network communication system. The method of claim 20, wherein when the sink moves, The plurality of sensor nodes on the virtual path receive the virtual path removal message from a lower node and transmit the received virtual path removal message to an upper node, and transmits the information transmitted to the sink to the virtual sink, The virtual sink stores the returned information and transmits the stored information to the sink through the new virtual path. The method of claim 13, When the plurality of sensor nodes have their parent nodes broken, Select an adjacent parent node of the same layer as the parent node as a parent node, If no neighboring parent node exists, the neighboring peer node is selected as the parent node, not the sibling node. And a child node is selected as a parent node when there is no neighbor peer node other than the sibling node. The method of claim 22, If the sensor node is a node on the virtual path, And the sensor node transmits a virtual routing message to the newly selected parent node.

Images