KR20120079741A - Method and apparatus of selecting backoff counter in wireless mesh network having plural relay nodes and computer program product thereof - Google Patents

Abstract

PURPOSE: A method for selecting a back-off counter, an apparatus thereof, and a computer program product on the wireless mesh network are provided to dynamically change a back-off counter of each relay node according to a communication environment of a relay node. CONSTITUTION: A contention window determining unit(930) determines the size of a contention window in order to reduce the size of the contention window of the relay node. A transmission collision sensing unit(940) senses whether to fail transmission attempt by transmission collision. A controlling unit(950) retransmits data traffic until a back-off counter terminates.

Description

Method and apparatus of selecting backoff counter in wireless mesh network having plural relay nodes and computer program product

The present invention relates to a technique for relaying data traffic in a wireless mesh network, and more particularly, to a technique for selecting a backoff counter waiting for retransmission in a relay node when a transmission collision occurs in a relay node of the wireless mesh network. will be.

In recent years, as well as the mobile communication network, IEEE 802.11, wireless LAN network, Bluetooth network and related technologies are rapidly developed, the use of the wireless Internet is expanding. Accordingly, Mobile Internet Protocol (Mobile IP) technology is in the spotlight as one of the next generation technologies. Mobile IP is a technology that guarantees a connection at all times even when a terminal having an IP address moves. Mobile IP supports mobility through a dual address system.

In addition, a mobile ad hoc network has a high flexibility as a wireless communication network between moving nodes. Ad-hoc network is very difficult to implement network configuration and protocol due to the characteristics of mobile node. For example, the movement pattern, type of traffic, link quality or power margin of the mobile node changes the phase. On the other hand, the mobile node is difficult to give a constant identification information, such as static IP (static IP) due to mobility, and the communication transmission range is limited due to communication band and power constraints. Thus, multi-hop communication is used to transmit data to nodes outside the transmission range.

In general, the capacity (capacity) of the wireless mesh network is O (1 / n), and the fixed capacity of the ad-hoc network is ^{O (1 / n 1/2} ). See Non-Patent Document [2], which is incorporated herein by reference.

1 is a diagram illustrating a bottleneck that is one of problems of a distributed network.

In FIG. 1, mobile stations are distributed across a mesh topology. In this way, when the mobile stations are distributed and distributed, the communication resources of each mobile station can be used flexibly. However, when any one of the mobile stations is deactivated, relaying through the corresponding mobile station becomes impossible, which causes a serious problem in total throughput.

The main difference between a Wireless Mesh Network (WMN) and an ad-hoc network is that in a wireless mesh network all traffic passes through a gateway, whereas in an ad-hoc network traffic flows between any pair of nodes. Therefore, in general, a decrease in transmission rate due to traffic concentration occurs in a wireless mesh network rather than an ad hoc network, and a transmission collision phenomenon occurs frequently. Of course, fewer bottlenecks exist in ad-hoc networks than in wireless mesh networks.

The solutions to bottlenecks include table-driven or proactive management and demand-based management (on-demand or reactive). It is a method of maintaining route information by broadcasting information. While maintaining the route information, there is an advantage that communication can be performed without delay of the route search, but there is a disadvantage that the broadcasting overhead of the control message for managing the route information is too large. In addition, the request-based method is a method of searching for a path at the time of traffic generation, and does not require broadcasting of periodic path information such as a table management method.

In general, the conventional routing protocol is limited to a method for searching the shortest path from the source node to the destination node, and randomly selects a backoff window size within a window of a predetermined size when a transmission collision occurs. Therefore, it is impossible to efficiently use the resources of the distributed network because all relay nodes are waiting for the same contention window. Therefore, there is an urgent need for a technique for changing the relay node's backoff counter appropriately based on dynamic information such as the activation state of the relay node and the number of relay nodes subordinate to it.

1] KR 2004-0060716, "Distributed Wireless Network System and Its Routing Method"

[2] "The nominal capacity of wireless mesh networks", IEEE Wireless Communications. 2003 [3] "Avoiding Bottlenecks Due to Traffic Aggregation at Relay Nodes in Multi-hop Wireless Networks", Binh Ngo and Steven Gordon, Institute for Telecommunications Research, University of South University of South Australia [4] "A Distributed CSMA Algorithm for Throughput and Utility Maximization in Wireless Networks", Libin Jiang, Jean Walrand, UC Berkeley, Allerton Conference, Sep. 26, 2008.

Summary of the Invention An object of the present invention was derived to solve the above-mentioned problem. To provide.

It is also an object of the present invention to provide a data traffic transmission apparatus that can be actively adapted to changes in a communication environment by dynamically changing a backoff counter of a relay node according to an operation mode of a previous relay node of the relay node. will be

과 같은 선형 확률 분포를 가지도록 모델링하고(단, r은 릴레이 카운트를 나타내고, CW는 경쟁 윈도우의 최대 크기를 나타내며, k는 f(x)의 합이 1이 되도록 결정되는 상수를 나타냄), E(x)는 다음

과 같이 연산한다. 또는, 연산 단계는 f(x)를 다음

과 같은 지수 확률 분포를 가지도록 모델링하고(단, r은 릴레이 카운트를 나타내고, CW는 경쟁 윈도우의 최대 크기를 나타내며, k는 f(x)의 합이 1이 되도록 결정되는 상수를 나타냄), E(x)는 다음

과 같이 연산한다. One aspect of the present invention for achieving the above object relates to a method for selecting a backoff counter until transmission after transmission collision in a wireless mesh network comprising a plurality of relay nodes. . According to an aspect of the present invention, in the wireless mesh network, on the transmission path through which data traffic is relayed through a relay node, the number of relay nodes existing before each relay node is determined as a relay count of the corresponding relay node. A relay count determination step of determining a contention window of the relay nodes such that a contention window of the relay nodes is reduced in size as the relay count of each relay node is increased, and a relay in which a transmission collision occurs; Retransmitting the data traffic after the node has waited until the backoff counter corresponding to the determined contention window size has expired. In particular, the relay count determination step is to subtract the number of relay nodes having an empty queue from the relay count when there are no data packets to transmit in the queue of relay nodes that exist before each relay node on the transmission path. It further comprises the step. In addition, the relay count determining step may further include subtracting from the relay count the number of relay nodes in which an operation mode is inactive among relay nodes existing before each relay node on a transmission path. Furthermore, the relay count determining step includes adding a relay count to the relay count of the number of relay nodes on the transmission path whose operation mode is changed from inactive mode to active mode among relay nodes existing before each relay node and the transmission path. The method may further include subtracting the number of relay nodes whose operation mode of the relay nodes existing before each relay node from the activation mode to the deactivation mode to the relay count. In addition, the contention window determination step models f (x) such that the probability density function f (x) of the size x of the contention window follows a non-uniform probability distribution. And calculating the expected value E (x) of the contention window using the modeled f (x) . And, the operation step is f (x)

Linear probability of to model a distribution such as is (where, r denotes a relay count, CW represents the maximum size of the contention window, k represents a constant, the sum of f (x) is determined to be 1), E (x) is

Calculate as Or, the operation step follows f (x)

And is modeled so as to have the same exponential probability distribution (where, r denotes a relay count, CW represents the maximum size of the contention window, k represents a constant, the sum of f (x) is determined to be 1), E (x) is

Calculate as

Another aspect of the present invention for achieving the above objects is directed to an apparatus for selecting a backoff counter until transmission after transmission collision in a wireless mesh network including a plurality of relay nodes. According to another aspect of the present invention, an apparatus determines, on a transmission path in which data traffic is relayed through a relay node in a wireless mesh network, the number of relay nodes existing before each relay node as a relay count (RC) of the corresponding relay node. A relay count determining unit configured to determine a contention window size such that a contention window size of a corresponding relay node is reduced as a relay count of each relay node is increased, and detecting whether a transmission attempt fails due to a transmission collision A transmission collision detection unit, and a control unit for retransmitting data traffic after waiting until the backoff counter corresponding to the determined contention window size expires when the transmission collision is detected. In addition, the relay count determining unit is adapted to subtract the number of relay nodes having an empty queue from the relay count when there is no data packet to be transmitted to the queue of relay nodes existing before each relay node on the transmission path. Furthermore, the relay count determining unit is adapted to subtract from the relay count the number of relay nodes in which the operation mode is inactive among the relay nodes existing before each relay node on the transmission path. In particular, the relay count determining unit adds, to the relay count, the number of relay nodes whose operation mode is changed from inactive mode to active mode among relay nodes existing on each transmission node in the transmission path, and before each relay node on the transmission path. It is further adapted to subtract the number of relay nodes whose operation mode is changed from the activation mode to the deactivation mode among the relay nodes present in the relay count. In addition, the competitive window determiner models f (x) such that the probability density function (PDF) f (x) of the size x of the competitive window follows a non-uniform probability distribution, and uses the modeled f (x) to compete. Calculate the expected value of E (x) . Furthermore, relay nodes are arranged in the form of one of a line topology and a tree topology.

According to the present invention, the transmission rate of data traffic can be optimized by dynamically changing the backoff counter of each relay node of the wireless mesh network according to the communication environment of the corresponding relay node.

In addition, according to the present invention, by dynamically changing the backoff counter of a relay node according to the operation mode of the previous relay node of the corresponding relay node, it is possible to actively adapt to changes in the communication environment and increase the efficiency of network resource operation. Can be.

1 is a diagram conceptually illustrating a wireless mesh network according to the prior art.
2 is a flowchart conceptually illustrating a method for selecting a backoff counter according to an embodiment of the present invention.
3 is a diagram illustrating the flow of data traffic in a wireless mesh network to which the present invention is applied.
4 is a graph illustrating the expected value E (x) of the contention window.
5A and 5B are graphs showing the probability density function f (x) and the expected value E (x) of the contention window, respectively, when the probability density function f (x) has a linear distribution.
6A and 6B are graphs showing the probability density function f (x) and the expected value E (x) of the contention window, respectively, when the probability density function f (x) has an exponential distribution.
FIG. 7 illustrates a method of selecting a backoff counter when an inactive node exists.
8 is a diagram illustrating a method of selecting a backoff counter according to a change in an operation mode of a preceding node.
9 is a block diagram conceptually illustrating a backoff counter selection apparatus according to another aspect of the present invention.
10A and 10B are diagrams illustrating relay nodes arranged according to a line topology and a tree topology, respectively.
11A and 11B are graphs showing simulation results when relay nodes are arranged according to a line topology.
12A and 12B are graphs showing simulation results when relay nodes are arranged according to a tree topology.

In order to fully understand the present invention, operational advantages of the present invention, and objects achieved by the practice of the present invention, reference should be made to the accompanying drawings and the accompanying drawings which illustrate preferred embodiments of the present invention.

Hereinafter, the present invention will be described in detail with reference to the preferred embodiments of the present invention with reference to the accompanying drawings. However, the present invention can be implemented in various different forms, and is not limited to the embodiments described. In order to clearly describe the present invention, parts that are not related to the description are omitted, and the same reference numerals in the drawings denote the same members.

Throughout the specification, when a part is said to "include" a certain component, it means that it may further include other components, without excluding the other components unless otherwise stated. In addition, the terms "... unit", "... unit", "module", "block", etc. described in the specification mean a unit for processing at least one function or operation, which means hardware, software, or hardware. And software.

2 is a flowchart conceptually illustrating a method for selecting a backoff counter according to an embodiment of the present invention.

First, a relay operation of data traffic through a relay node in a wireless mesh network is performed (S210). When a relay node monopolizes communication resources or uses communication resources allocated to it by scheduling, a transmission collision does not occur during transmission. On the other hand, when two or more relay nodes attempt to occupy the same communication resource at the same time, a transmission collision occurs, and in this case, the transmission attempt by one or more relay nodes fails. Therefore, it is determined whether a transmission collision occurs while the data traffic is relayed (S220). The transmission collision may be determined by each relay node attempting to transmit data traffic, but may be determined by a separate scheduler or gateway.

Various techniques have already been proposed to detect and avoid transmission collisions. For example, the prior art may use carrier sense (CSMA / CA) for collision avoidance. Carrier sense (CS) checks whether communication resources are empty. In this case, the situation where two relay nodes cannot hear each other is called a hidden node problem. In this case, it can be solved by using virtual carrier sense (VCS). Generally, IFS (InterFrame Space, IFS) is defined for frame transmission. There are three types of IFS as follows. Short interframe space (SIFS) has a high priority to ACK received. In point-coordination interframe space (PIFS), an access point is used to control a network, and distributed-coordination interframe space (DIFS) is used for data frame transmission and has a lower priority than SIFS. In addition, in the present invention, a backoff counter is a counter set to occupy a communication resource, and the node waits for the time as much as the backoff counter and attempts the next transmission. In this case, a time for listening to transmit the next frame after a collision occurs is called a contention window (CW). For example, in a distributed coordination function (DCF), a backoff counter of 0 to 31 is set in an initial NAV (Network Allocation Vector) and enters a contention window state. The node then starts transmitting frames when the backoff counter reaches zero. If frame transmission fails again, the node randomly sets the backoff counter back to a value between 0 and 127 and attempts to retransmit when the backoff counter expires. However, in the method of setting a backoff counter according to the present invention, the utilization of network communication resources is maximized by setting the backoff counter to be related to the relay count, rather than randomly setting the backoff counter. This will be described later in detail with reference to FIG. 2.

First, a relay count required for setting a backoff counter is determined (S230). In the present specification, the relay count indicates the number of active nodes that a relay node relays. That is, the number of active nodes existing before any particular relay node on the transmission path of data traffic is determined as the relay count. The relay count will be described with reference to FIG. 3.

3 is a diagram illustrating the flow of data traffic in a wireless mesh network to which the present invention is applied.

Referring to FIG. 3, the first to seventh nodes 1-7 each relay data to the destination 8. In FIG. 3, there are four data traffic flows. The first flow is the path through which data traffic is relayed from 1-> 3-> 7-> 8, and the second flow is the path from which data traffic is relayed from 2-> 3-> 7-> 8. Likewise, there are paths of 4-> 6-> 7-> 8 and 5-> 6-> 7-> 8. In each data traffic flow, the number of relay nodes (hereinafter referred to as 'prior nodes') that existed before a particular relay node in the transmission direction of the data traffic is the relay count (RC) of the relay node. Is defined as For example, node 3 has a preceding node of node 1 and node 2, so RC equals 2. Similarly, because RC6 has two preceding nodes, Node4 and Node5, RC becomes 2. In addition, RC is 6 since node 7 has preceding nodes of node 1 to node 6.

In FIG. 3, DR represents a data rate and effective data rate (EDR) represents an effective data rate. DR represents the generation rate of data generated at each relay node. The EDR represents a data generation rate of a specific relay node in consideration of the data generation rate of preceding nodes of the relay node. For example, the EDR of Node 3 is 3 which is the sum of 1, which is its DR, and 1, which is DR of Node 1 and Node 2. Similarly, the node 7's EDR adds up to 7, which is 1, its DR, and 6, which is the sum of the DRs of the previous nodes 1-6. In FIG. 3, the DRs of the respective relay nodes are assumed to be the same for convenience of description, but it should be noted that this is not a limitation of the present invention.

As such, each relay node has a different relay count. In the present invention, the size of the contention window and the backoff counter are determined using the relay count of each relay node.

Referring to FIG. 2 again, the contention window size is determined according to the determined relay count (S240). At this time, the size of the contention window of the relay nodes is determined so that the size of the contention window of the relay node decreases as the relay count of each relay node increases. In particular, the model f (x) to follow the size of the probability density function of x (probability density function, PDF) f (x) is non-uniform probability distribution (non-uniform probability distribution) of the contention window, and the model f (x ) Is used to calculate the expected value E (x) of the contention window. 4 is a graph illustrating the expected value E (x) of the contention window.

4 is a graph illustrating the expected value E (x) of the contention window.

The thick solid line in FIG. 4 represents the result by the prior art, and the thin solid line and the dotted line illustrate the expected value E (x) of the contention window when RC is 1 and 2, respectively. The vertical axis of the graph represents the transmission attempt probability, and the horizontal axis represents the size of the contention window and the probability density function f (x) of the contention window.

In FIG. 4, it can be seen that the probability density function f (x) according to the present invention has a non-uniform distribution that is significantly different from the prior art. That is, the probability density function of the contention window size of the corresponding relay node is selected by using the relay count value of the specific node. In this case, the larger the RC value, that is, the larger the number of preceding nodes, the greater the probability that the CW value becomes smaller, so that the RTS is calculated. Will transmit faster. Therefore, in the backoff counter selection method according to the present invention, the larger the RC, the better the competition, and thus, the faster the data can be transmitted.

The probability density function f (x) is modeled to have a non-uniform probability distribution. There are two modeling methods.

First, f (x) may be modeled to have a linear probability distribution as in Equation 1 below.

In Equation 1, r denotes a relay count, CW denotes a maximum size of a contention window, and k denotes a constant determined so that the sum of f (x) is one. This is because the result of adding the probability density function f (x) in the meantime of [0, CW-1] should be 1.

The value of k satisfying the condition of the probability density function f (x) according to Equation 1 is obtained by Equation 2 below.

Then, E (x) is determined as in Equation 3 below.

Equation 3 holds when r is not zero.

As shown in Equations 1 and 3, the results of modeling the probability density function f (x) and the expected value E (x) of the contention window are shown in FIGS. 5A and 5B, respectively.

5A and 5B are graphs showing the probability density function f (x) and the expected value E (x) of the contention window, respectively, when the probability density function f (x) has a linear distribution.

First, referring to FIG. 5A, the horizontal axis of FIG. 5A represents values of a contention window and a probability distribution, and the vertical axis represents a transmission attempt probability. The thick solid line represents the prior art and the thin solid lines represent the values modeled by the present invention. The slope of the thin solid line in FIG. 5A depends on the relay count. In other words, as the relay count increases, the slope increases. The thick solid line representing the prior art shows that the transmission attempt probability is the same regardless of the value of the contention window.

5B shows the expected value E (x) of the contention window depending on the size of the relay count. Referring to FIG. 5B, it can be seen that as the size of the relay count increases, the expected value E (x) of the contention window decreases, which means that the relay node waits for a smaller contention window, so it may try to transmit more often. To indicate that it is.

A second method of modeling the probability density function f (x) to have a non-uniform probability distribution is to model the probability density function f (x) to have an exponential probability distribution as in Equation 4 below.

As described above, r represents the relay count, CW represents the maximum size of the contention window, and k represents a constant that is determined such that the sum of f (x) is one.

The value of k satisfying the condition of the probability density function f (x) according to Equation 4 is obtained as shown in Equation 5 below.

Then, E (x) is determined as in Equation 6 below.

Equation 6 holds when r is not zero.

The results of modeling the probability density function f (x) and the expected value E (x) of the contention window as shown in Equations 4 and 6 are shown in FIGS. 6A and 6B, respectively.

6A and 6B are graphs showing the probability density function f (x) and the expected value E (x) of the contention window, respectively, when the probability density function f (x) has an exponential distribution.

First, referring to FIG. 6A, the horizontal axis of FIG. 6A represents values of a contention window and a probability distribution, and the vertical axis represents a transmission attempt probability. The thick solid line represents the prior art and the thin solid lines represent the values modeled by the present invention. In FIG. 6A the thin Y-intercept depends on the relay count. In other words, the larger the relay count, the larger the Y intercept. As described above, it can be seen that in the prior art, the transmission attempt probability is the same regardless of the value of the contention window.

6B shows the expected value E (x) of the contention window depending on the size of the relay count. Referring to FIG. 6B, it can be seen that as the size of the relay count increases, the expected value E (x) of the contention window decreases, which means that the relay node will try to transmit more often because it waits for a smaller contention window. To indicate that it is.

Referring back to FIG. 2, the relay node determines a backoff counter corresponding to the size of the contention window determined as described above, and waits for the determined backoff counter (S250).

Thus, it is determined whether the backoff counter expires (S260), and if it expires, retransmits the data traffic (S270). This process is repeated until there is no more data traffic to transmit (S280).

In order to understand the backoff counter selection technique according to the present invention, basic knowledge in related fields is required.

First, looking at the IEEE 802.11e EDCA protocol, the protocol defines four access categories in consideration of traffic quality (QoS). In the case of using a binary exponential backoff, the size of the contention window (backoff counter) is defined as twice the size of the previous contention window and the smaller of the maximum value of the contention window size. In addition, when randomly selecting the backoff counter, the size of the contention window is uniformly distributed between 0 and the size of the previous contention window. In this technique, the size of the backoff counter is set independently of the bottleneck caused by the relay.

In addition, a rate balance scheme adjusts a receiving rate and a forwarding rate by using a balance rate. See the non-patent literature [3], which is incorporated herein by reference for speed balancing techniques.

In the rate balancing technique, the rate of transmission can be controlled by responding to a request to send (RTS) only when the balance condition is satisfied, but power is wasted because the source node continuously transmits the RTS. In addition, there is a problem that it is impossible to control the destination, that is, nodes close to the gateway.

Finally, the dynamic CSMA algorithm dynamically changes the contention window size according to the queue length, but likewise does not consider the bottleneck caused by the relay. For a dynamic CSMA algorithm, see Non-Patent Document [4], which is incorporated herein by reference.

In addition, in the method for selecting a backoff counter according to the present invention, a contention window and a backoff counter are dynamically adapted according to a network environment. This will be described with reference to FIGS. 7 and 8.

FIG. 7 illustrates a method of selecting a backoff counter when an inactive node exists.

In FIG. 7, relay node R1 has three preceding nodes A1, IA, A2, and relay node R2 has a unique preceding node A3, as well as the preceding nodes A1, IA, A2 of R1. It can be seen that all five leading nodes up to R1. At this time, it is assumed that A1 to A3 are activated nodes or IA is inactive node. It is to be understood that this is for convenience of description and not for limiting the present invention. In FIG. 7, since R1 has a relay count of 3 and a relay count of R2 is 5, R2 has a transmission opportunity more frequently than R1. However, if the number of relay nodes belonging to R1 is five or more, R1 has a transmission opportunity more frequently than R2. However, since R1 must relay data traffic through R2, no matter how often R1 relays data, if R2 does not relay the data as much as it can, bottlenecks will occur. As such, if the preceding nodes are unconditionally considered when determining the relay count, since R1 may have more transmission opportunities than R2, a degradation of the transmission rate due to R2 may occur.

Therefore, the backoff counter selection algorithm according to the present invention efficiently calculates an RC value by distinguishing between an active node and an inactive node. In the present invention, an inactive node may mean a node that is clearly implemented on the network but is not currently operating, or may be a node that is operating but does not have a data packet to transmit to its queue. In FIG. 7, the relay count of R1 should be 3, but the relay count of R1 becomes 2 since the inactive node IA of the three preceding nodes A1, IA, A2 is excluded from the relay count. As such, since the relay count is determined by considering only the preceding node currently operating, the contention window can be optimized according to the network environment.

8 is a diagram illustrating a method of selecting a backoff counter according to a change in an operation mode of a preceding node.

All leading nodes A, B, C, and D of the destination node De in FIG. 8 (a) are all active nodes, and in (b) two of these nodes C and D are referred to as inactive nodes. The operation mode is changed. In addition, in (c), the operation mode of one inactive node D is changed back to the active mode. However, it is to be understood that this is illustrated for convenience of description and not to limit the present invention.

In FIG. 8, the relay count of A is zero, and the relay count of B is two. At this time, if nodes C and D are changed to inactive nodes, the relay count of node B is changed to zero. When the node D is changed back to the active node, the relay count of the node B is changed back to one. As described above, in the backoff counter selection algorithm according to the present invention, when the operation mode of the preceding node is changed from the active state to the inactive state, for example, when the packet transmission is terminated, the RC value of the relay node is reduced. At this time, if one of the preceding nodes tries to transmit data again, the operation mode of the corresponding preceding node is changed from the inactive state to the active state. The relay count of the relay node then increases again. As described above, in the algorithm according to the present invention, the size of the backoff counter can be selected by actively adapting to changes in the network environment.

9 is a block diagram conceptually illustrating a backoff counter selection apparatus according to another aspect of the present invention.

The backoff counter selection apparatus 900 according to the present invention includes an input buffer 910, an output buffer 990, and a backoff counter 980 connected to the controller 950. In addition, the relay count determiner 920, the contention window determiner 930, and the transmission collision detection unit 940 are controlled by the controller 950.

The relay count determiner 920 determines the number of preceding nodes as a relay count RC of the corresponding relay node. At this time, as described above, it may or may not be reflected in the relay count depending on the operation mode of the preceding node.

Then, the contention window determination unit 930 determines the size of the contention window so that the size of the contention window of the corresponding relay node decreases as the relay count of each relay node increases. As described above, a non-uniform probability density function f (x) is used to determine the size of the contention window, and f (x) can be modeled as having a linear probability distribution and an exponential probability distribution, respectively.

When data traffic is received through the input buffer 910, the controller 950 attempts to transmit the received data through the output buffer 990. Of course, the person skilled in the art can also easily generate the data to be directly transmitted based on the data generation rate DR, although the backoff counter selection device 900 does not necessarily transmit only the data received through the input buffer 910. Will be understood.

The transmission collision detection unit 940 detects whether a transmission attempt fails due to a transmission collision. Then, when a transmission collision is detected, the controller 950 retransmits the data traffic after waiting until the backoff counter 980 corresponding to the determined contention window size expires. That is, the controller 950 determines the backoff counter 980 according to the size of the contention window determined by the contention window determiner 930, and waits until the determined backoff counter 980 expires.

As such, by determining the backoff counter 980 based on the relay count that is dynamically changed according to the network environment, since the waiting time of the backoff counter selection device 900 is dynamically changed, it is adapted to the changed environment. An optimal relay environment is created.

Hereinafter, a result of applying the present invention to various topologies in which relay nodes may be arranged in a wireless mesh network is illustrated.

10A and 10B are diagrams illustrating relay nodes arranged according to a line topology and a tree topology, respectively.

In the line topology shown in FIG. 10A, six relay nodes are arranged to forward data traffic to the gateway GW in the order 6-> 5-> 4-> 3-> 2-> 1, respectively. In the line topology, it can be seen that the RC of node 6 is 0, but the RC of node 1 is 5. Therefore, it can be seen that Node 1 has a relatively smaller contention window size than Node 6, and correspondingly has a smaller backoff counter.

In addition, in the tree topology shown in FIG. 10B, nodes 5 and 6 relay data traffic through node 2 to gateway GW, and nodes 3 and 4 relay data traffic through node 1 to gateway GW. do. In the tree topology, it can be seen that Node 1 and Node 2 have an RC value of 2, but the RC of the other nodes 3, 4, 5, 6 is zero. Therefore, it can be seen that Node 1 and Node 2 will have a smaller contention window size and a smaller backoff counter compared to the other nodes 3, 4, 5, 6.

11A and 11B are graphs showing simulation results when relay nodes are arranged according to a line topology.

11A shows the number of successful transmissions of each node when ten nodes are arranged according to a line topology. In FIG. 11A, the symbol 'O' denotes a case of the prior art, and the '+' and '*' marks indicate a case where the probability density function f (x) has a linear probability distribution and an exponential probability distribution, respectively.

In FIG. 11A, the simulation result for each node is performed for 5 seconds, and assumes that node 10 generates infinite data packets (that is, DR of node 10 is infinite). However, this is for the purpose of illustrating the present invention and should not be construed as limiting the present invention. As can be seen in FIG. 11A, it can be seen that the number of successful transmissions increases in the case of having an exponential probability distribution ('*') and a linear probability distribution ('+') as compared with the conventional technique. . Node 10 has a large number of successes because a collision cannot occur, but it can also be seen that the number of successes decreases closer to the gateway GW. As can be seen in Figure 11a, according to the backoff counter selection algorithm according to the present invention it can be clearly seen that the probability of transmission success of data traffic increases compared to the prior art.

11B shows the results of simulating the packet generation rate of node 10 in a line topology. The horizontal axis of FIG. 11B represents the packet generation rate DR, and the horizontal axis represents the number of packets that were successfully transmitted for 5 seconds. As can be seen in FIG. 11B, it can be seen that the number of successful transmissions increases in the case of having an exponential probability distribution ('*') and a linear probability distribution ('+').

12A and 12B are graphs showing simulation results when relay nodes are arranged according to a tree topology.

12A shows the number of successful transmissions of each node when six nodes are arranged according to the line topology as shown in FIG. 10B. In FIG. 12A, the symbol 'O' denotes a case of the related art, and the symbol '+' and '*' denotes a case where the probability density function f (x) has a linear probability distribution and an exponential probability distribution, respectively.

In FIG. 12A, simulation results for each node are performed for 5 seconds, and assume that nodes 3, 4, 5, and 6 generate infinite data packets (that is, DR of nodes 3, 4, 5, and 6 Infinity). However, this is for the purpose of illustrating the present invention and should not be construed as limiting the present invention. In FIG. 12A, it can be seen that in the case of having a uniform probability distribution 'O' according to the prior art, bottlenecks occur at Node 1 and Node 2, and the number of successful transmissions is significantly reduced. On the other hand, in the case of an exponential probability distribution ('*') and a linear probability distribution ('+'), these bottlenecks do not occur at node 1 and node 2. Rather, node 1 and node 2 have a higher RC value than nodes 3, 4, 5, and 6, thus increasing the number of successful transmissions. This feature is evident in the case of a linear probability distribution ('+') compared to the case of an exponential probability distribution ('*'). That is, as can be seen in Figure 12a, according to the backoff counter selection algorithm according to the present invention, the bottleneck is significantly reduced compared to the prior art.

12B illustrates the results of simulating the packet generation rate of relay nodes in a tree topology. The horizontal axis of FIG. 12B represents the packet generation rate DR, and the horizontal axis represents the number of packets that were successfully transmitted for 5 seconds. As can be seen in FIG. 12B, it can be seen that the number of successful transmissions increases in the case of having an exponential probability distribution ('*') and a linear probability distribution ('+'). In particular, it can be seen that as the DR increases, the number of successful transmissions of the exponential probability distribution '*' and the linear probability distribution '+' increases significantly compared to the case of the uniform probability distribution 'O'.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is evident that many alternatives, modifications and variations will be apparent to those skilled in the art. For example, in FIG. 10A and FIG. 10B, relay nodes are arranged according to a line topology and a tree topology, but the present invention is not limited thereto.

Therefore, the true technical protection scope of the present invention will be defined by the technical spirit of the appended claims.

The present invention can be applied to a technique for adaptively selecting a backoff counter of relay nodes in a wireless mesh network according to a network environment.

Claims (18)

A method for selecting a backoff counter until transmission after transmission collision in a wireless mesh network including a plurality of relay nodes, the method comprising:
A relay count determining step of determining the number of relay nodes existing before each relay node as a relay count of the corresponding relay node on a transmission path through which data traffic is relayed through the relay node in the wireless mesh network;
Determining a contention window size of the contention window of the relay nodes such that the contention window size of the relay node decreases as the relay count of each relay node increases; And
Retransmitting the data traffic after the relay node having a transmission conflict has waited until the backoff counter corresponding to the determined contention window has expired. How to choose. The method of claim 1, wherein the determining of the relay count comprises:
Subtracting the number of relay nodes having an empty queue from the relay count when there are no data packets to transmit in a queue of relay nodes existing before each relay node on the transmission path. And selecting a backoff counter in a wireless mesh network. The method of claim 2, wherein the determining of the relay count comprises:
And subtracting, from the relay count, the number of relay nodes whose operation mode is inactive among the relay nodes existing before each relay node on the transmission path from the relay count. Method for choosing. The method of claim 3, wherein the determining of the relay count comprises:
Adding, to the relay count, the number of relay nodes of the relay nodes existing before each relay node in which the operation mode is changed from an inactive mode to an active mode on the transmission path; and
And subtracting, by the relay count, the number of relay nodes whose operation mode is changed from an activation mode to an inactive mode among the relay nodes existing before each relay node in the transmission path. Method for selecting a backoff counter. The method of claim 1, wherein the determining the contention window comprises:
The probability of the magnitude x of the contention window density function (probability density function, PDF) f (x) the model f (x) to follow the non-uniform probability distribution (non-uniform probability distribution) and the modeled f (x) And calculating an expectation value E (x) of the contention window using the < RTI ID = 0.0 >.≪ / RTI > The method of claim 5, wherein the calculating step,
f (x) then

Modeled to have a linear probability distribution such that r denotes a relay count, CW denotes the maximum magnitude of the contention window, and k denotes a constant that is determined such that the sum of f (x) is equal to 1),
E (x) is

And selecting the backoff counter in the wireless mesh network. The method of claim 5, wherein the calculating step,
f (x) then

Modeled to have an exponential probability distribution such that r denotes a relay count, CW denotes the maximum size of the contention window, and k denotes a constant that is determined such that the sum of f (x) is 1;
E (x) is

And selecting the backoff counter in the wireless mesh network. The method of claim 1,
And said relay nodes are arranged in the form of one of a line topology and a tree topology. A computer program product comprising computer instructions for implementing a method according to any one of claims 1 to 8, stored on a computer readable medium. An apparatus for selecting a backoff counter from transmission transmission to retransmission in a wireless mesh network including a plurality of relay nodes, the apparatus comprising:
A relay count determining unit determining a number of relay nodes existing before each relay node as a relay count (RC) of a corresponding relay node on a transmission path through which data traffic is relayed through a relay node in the wireless mesh network;
A contention window determining unit configured to determine a size of the contention window so that the size of the contention window of the corresponding relay node decreases as the relay count of each relay node increases;
A transmission collision detection unit detecting whether a transmission attempt fails due to a transmission collision; And
And a control unit for retransmitting the data traffic after waiting until the backoff counter corresponding to the determined contention window expires when a transmission collision is detected. Device for selecting. The method of claim 10, wherein the relay count determination unit,
In the transmission path, when there is no data packet to be transmitted to the relay node's queue before each relay node, the radio count is adapted to subtract from the relay count the number of relay nodes having an empty queue. Device for selecting a backoff counter in the mesh network. The method of claim 10, wherein the relay count determination unit,
And the relay count is subtracted from the relay count the number of relay nodes whose operation mode is deactivated among the relay nodes existing before each relay node on the transmission path. The method of claim 12, wherein the relay count determination unit,
On the transmission path, the number of relay nodes whose operation mode is changed from inactive mode to active mode among relay nodes existing before each relay node is added to the relay count,
And further adapted to subtract from the relay count the number of relay nodes of the relay nodes existing before each relay node from the active mode to the inactive mode on the transmission path to the relay count. Device for selecting a counter. The method of claim 10, wherein the contention window determination unit,
The expected value of the contention window, a probability density function (PDF) of the size x of the contention window f and (x) a model f (x) to follow the non-uniform probability distribution, using the model f (x) E ( x) computing an apparatus for selecting a backoff counter in a wireless mesh network. The method of claim 14, wherein the contention window determination unit,
f (x) then

Modeled to have a linear probability distribution such that r denotes a relay count, CW denotes the maximum magnitude of the contention window, and k denotes a constant that is determined such that the sum of f (x) is equal to 1),
E (x) is

And a backoff counter in a wireless mesh network characterized in that it is adapted to compute. The method of claim 14, wherein the contention window determination unit,
f (x) then

Modeled to have an exponential probability distribution such that r denotes a relay count, CW denotes the maximum size of the contention window, and k denotes a constant that is determined such that the sum of f (x) is 1;
E (x) is

And a backoff counter in a wireless mesh network characterized in that it is adapted to compute. The method of claim 10,
And said relay nodes are arranged in one of a line topology and a tree topology. A computer program product comprising computer instructions for operating an apparatus according to any one of claims 10 to 17 and stored on a computer readable medium.

Images