CN102740394A

CN102740394A - Center calculation wireless sensor network 2-node disjoint multipath routing algorithm

Info

Publication number: CN102740394A
Application number: CN2012102512950A
Authority: CN
Inventors: 于磊磊; 于玉杰; 陈冬岩
Original assignee: JINAN PUSAI COMMUNICATION TECHNOLOGY Co Ltd
Current assignee: JINAN PUSAI COMMUNICATION TECHNOLOGY Co Ltd
Priority date: 2012-07-19
Filing date: 2012-07-19
Publication date: 2012-10-17
Anticipated expiration: 2032-07-19
Also published as: CN102740394B

Abstract

The invention relates to a center calculation wireless sensor network 2-node disjoint multipath routing algorithm which adopts a central-dispatching type adaptive route maintenance mechanism to enhance the flexibility of the path maintenance. During the data routing, a data package only needs to carry a path bit sequence, and meanwhile, no control cost is required. According to the disjoint multipath routing algorithm, the hop count and the quality of the link are taken as metrics, the approximately optimal 2-node disjoint path from source node to the sink node is calculated by utilizing the whole network information, and then the micro-routing table containing main nodes, auxiliary nodes and the path bit sequence is generated and then transmitted to the nodes.

Description

Central computing 2-disjoint path routing algorithm for wireless sensor network

Technical Field

The invention relates to a 2-disjoint path routing algorithm of a centrally computed wireless sensor network.

Background

In the wireless sensor network, compared with a single-path routing mechanism, the multi-path routing mechanism has obvious advantages in the aspects of improving network reliability, balancing load, fault-tolerant recovery and the like. However, in multi-path routing, a plurality of paths from a source node to a destination node may contain a common node (link), and failure of the common node (link) may cause failure of the plurality of paths therethrough. To avoid the appearance of common nodes (links), researchers have proposed disjoint multipath mechanisms.

Preetha Thussiraman et al propose a distributed 2-disjoint routing algorithm, which first constructs two color trees of a red tree and a green tree in a network, and establishes two disjoint paths from each node to a sink node in a depth-first search manner from the sink node by a path growth manner. The algorithm can meet the disjoint constraint of the two paths, but the path length established by the algorithm is long, the existence of an ultra-long path cannot be avoided, and meanwhile, the broadcasting of control information in the algorithm can cause more energy consumption. Similarly, different colors can be identified for direct neighbors of the sink node, and then nodes meeting with the color identifiers are found through color identifier broadcasting, so that a k-disjoint path from the source node to the sink node is established, the time complexity of the algorithm is only O (L) (L is the number of edges in the network), but the algorithm cannot effectively optimize the path. In addition, a disjoint path routing algorithm based on geographical location information is proposed, the two algorithms are also distributed algorithms, and the shortcomings are that each node must know the geographical location information of the node and the adjacent nodes thereof, and the optimal disjoint multipath from the source node to the target node cannot be obtained.

Among MPR (multipath routing) algorithms, DMPR (discrete multipath routing) algorithms have higher reliability and fault tolerance. The main challenges facing the DMPR algorithm are two: the method comprises the following steps of firstly, optimizing the disjoint paths and secondly, transmitting data packets on the disjoint paths.

Disclosure of Invention

The invention aims to solve the problems, provides a 2-disjoint path routing algorithm of a wireless sensor network for central computing aiming at the characteristics of relatively stable network topology, strong sink node computing and storage capacity and the like in certain industrial applications (such as mine environment monitoring), and adopts a self-adaptive path maintenance mechanism for central scheduling aiming at the problem of slow reaction of a central computing mode to link state change so as to improve the flexibility of path maintenance; the data packet only needs to carry the path bit sequence without any control overhead.

In order to achieve the purpose, the technical scheme of the invention is as follows:

a wireless sensor network 2-disjoint path routing algorithm of central computation, it regards hop count and link quality as the measurement, utilize the information of the whole network to calculate from source node to sink node the approximate optimal 2-node disjoint path, then produce the little routing table comprising master node, auxiliary node and path bit sequence only and download to each node; the method comprises the following steps:

first, collecting node neighbor information, creating and initializing an adjacency matrix M_AParent node matrix M_FLogical gradient matrix M_NPath matrix M_PAnd a path metric matrix M_PLQ(ii) a Wherein,

M_Ainitializing the matrix according to the collected neighbor information, wherein if the node i is a neighbor of the node j, the ith row and jth column element of the matrix is the link quality value from the node i to the node j, and if the node i is not a neighbor of the node j, the maximum value of the element is 255;

M_Fthe elements of the matrix are all initialized to 255; indicating that no parent-child relationship exists between the start nodes;

logical gradient matrix M_NIs an n multiplied by 1 matrix, n is the number of all nodes except the sink in the network; elements in the matrix represent the logic gradient values from the nodes to the sink node, and are initialized to a maximum value of 255;

path matrix M_PIs an nxhtl matrix, wherein n is the number of all nodes except the sink in the network; HTL is the maximum hop count allowed in the network, and the ith row and jth column elements represent jth hop nodes in the path from the node i to the sink node;

path metric matrix M_PLQIs an n multiplied by 1 matrix, and n is the number of all nodes except the sink in the network;

the elements in the matrix are path metric values from the node i to the sink node, the matrix is initialized according to neighbor information, if the sink node is a neighbor of the node i, the link quality value from the node i to the sink node is initialized, and if not, the link quality value is initialized to 255;

secondly, calculating two disjoint paths; utilizing CCDMPR algorithm to form a global parent node matrix M_FEstablishing a disjoint main path from all nodes to a sink node on a corresponding logic graph through a greedy path tracking algorithm, and then generating disjoint auxiliary paths from all nodes to the sink node by using a next-hop competition algorithm with a backoff set;

thirdly, generating a micro-routing table mechanism based on the main father node, the auxiliary father node pairs and the path bit sequence, and downloading the paths generated by calculation to each node by the sink node after the 2-disjoint paths of all the nodes are calculated;

fourthly, node path information is issued to all nodes except the sink node;

and fifthly, updating and maintaining the path.

In the first step, the process of collecting node neighbor information is as follows:

when a non-intersecting path is not established in the network, the network collects node neighbor information in a flooding mode; the nodes in the network broadcast the packaged neighbor information to own neighbor nodes in a broadcasting mode, if the nodes receive rreq data packets sent by other nodes Origineaddr, whether the rreq data packets are received is judged according to Flag attributes and Seqnum attributes, if the rreq data packets are received, the data packets are discarded, and if the rreq data packets are not received, the rreq data packets are broadcast to other neighbor nodes except the Origineaddr nodes, so that the phenomenon that loops are generated in the network to reduce data redundancy is effectively avoided; finally, stopping until the sink node receives the rreq data packet;

when a non-intersecting path is established in the network, rreq data packets of the nodes are sent to the Sink node according to the established optimal path, after the Sink node receives the rreq data packets, the data packets are analyzed to obtain neighbor nodes and corresponding link quality values, the neighbor nodes and the corresponding link quality values are sent to the PC through the serial port, and the PC temporarily stores neighbor information in a neighbor information table of the database according to the received data packets.

In the first step, a global parent node matrix M_FThe establishment procedure is as follows:

1) in M by breadth-first search algorithm_AEstablishing a logic gradient to the sink node; the logic gradient of the sink node is marked as 0, the logic gradient of the node directly adjacent to the sink node is set as 1, the gradient of the node directly adjacent to the node with the gradient i is set as i +1, and searching is carried out in sequence until each node obtains the logic gradient;

2) each non-leaf node takes a self as a root, and takes a neighbor node with a logic gradient equal to or less than the self logic gradient as a direct son to establish a logic two-layer tree;

3) storing the corresponding tree structure in the global father node matrix M_FThe storage mode is as follows: if node i belongs to the logical two-level tree rooted at node j, then order (M)_F)_ij＝(M_A)_ij' else (M)_F)_ij＝255。

In the second step, the generation process of the disjoint main path is as follows:

path metric matrix M_PLQ，M_PLQIs an n multiplied by 1 matrix, n is the number of all nodes except the sink in the network, (M)_PLQ)_iRepresenting the PLQ value of a path from the node i to the sink node, generating an optimal path by adopting a greedy path tracking algorithm, and simultaneously calculating the sum of LQ of all links contained in the path, which is called PLQ; the generated optimal path is stored in a main path matrix as a main path of the disjoint path

In (2), the corresponding PLQ value is stored in the path metric matrix M_PLQIn (1),

Trace_Greedy (M_{F}, M_{P}^{Master}, M_{PLQ})

1) set S = { L,2,3 … n }, set (M)_PLQ)₀=0;

If (M)_F)_i0<255, order (M)_PLQ)_i=(M_F)_i0,NEXT[i]=0,

If (M)_F)_i0=255, order (M)_PLQ)_i=∞,NEXT[i]=-1.

2) Order (M)_PLQ)_j=min{(M_PLQ)_i,i∈S}，S=S-{j}，

Then turn 4);

3) for all i ∈ S, if present (M)_F)_ji<255, then order (M)_PLQ)_i=min{(M_PLQ)_i,(M_PLQ)_j+(M_F)_ji}，NEXt[i]J, then turn 2);

4) let k =2 and let k be equal to,

T = {(M_{P}^{Master})}_{j 1};

while (T > 0) {k + +; T = NEXT [T]; {(M_{P}^{Master})}_{jk} = T};

if S is an empty set, the algorithm is ended, otherwise, the algorithm is switched to 3);

the time complexity of the algorithm for generating the optimal main path from all nodes to the sink node is O (n)²) And the running process of the algorithm ensures that the generated path has no loop.

The generation method of the disjoint auxiliary path comprises the following steps:

calculating the time complexity of the optimal auxiliary path of all nodes to be O (n) by adopting a next hop competition algorithm with a backoff set²) Each node competitively selects a candidate node with the minimum PLQ as a next hop by nodes contained in all the main paths until the sink node is reached; the nodes that all current secondary paths pass through are also added to AvoidSet,

①

NHC_NodeDisjoint (M_{F}, M_{P}^{Master}, M_{P}^{Secondary})

for each source node indexed by i

1) Set S ═ {1,2,3 … n), will all

And adding a backoff node set AvoidSet, and then making S equal to S-AvoidSt, k equal to l, and u equal to i.

2) Order (M)_PLQ)_v＝min{(M_PLQ)_lL ∈ S and F_ul< 255 >, order

{(M_{P}^{Secondary})}_{ik} = v, S = S - {v} .

3) If v =0, the algorithm ends, otherwise, let k be k + l, u be v, go to 2);

②

NHC_LinkDisjoint (M_{F}, M_{P}^{Master}, M_{P}^{Secondary})

for each source node indexed by i

1) Set S ═ {1,2,3 … n }, for all

Order to

Will edge<e，f>Adding a backoff link set AvoidSet, and then setting k to 1 and u to i;

2) order to

Order to

At the same time, the edges are cut<e，f>Adding AvoidSet;

3) if v = O, the algorithm ends, otherwise let k ═ k +1, u ═ v, go to 2).

In the third step, the micro-routing table contains five fields: the ID represents the node identification of the micro routing table, the Primary node domain stores a father node of the current node on the main path, the Secondary node stores a father node of the current node on the Secondary path, the Primary path bit series domain stores a main path bit sequence, the Secondary path bit series stores a Secondary path bit sequence, each bit of the bit sequence represents the father node selection of each hop, the main father node is represented by 0, and the auxiliary father node is represented by 1;

when a source node selects to transmit a data packet to a sink node along a main or auxiliary path, the source node adds a main or auxiliary path ratio specific sequence and a sequence pointer of the source node to the head of the generated data packet, the sequence pointer is 1 at the beginning, and the data packet is transmitted to the sink node along the main or auxiliary path;

after receiving the data packet, the intermediate node firstly analyzes the data packet head, adds 1 to the sequence pointer, and reads the bit element value in the path bit sequence indicated by the current sequence pointer, wherein the bit element value is transmitted to the main father node when being 0 and is transmitted to the auxiliary father node when being 1.

In the fourth step, the sink node is taken as a starting point, the path information of the nodes is sequentially issued according to the increase of the logic gradient of the nodes, a stable network is established as long as the nodes receive the path information, and even if the nodes which do not receive the path information exist in the global network, the stable local network established by the nodes which receive the path information cannot be influenced;

and the PC calculates two disjoint paths according to the neighbor information, packs the path information of the node and the reverse route from the base station to the node into rrep and sends the rrep to the destination node. In the rrep issuing process, the node receives the rrep and firstly judges whether the node is the destination node or not, if not, the node continues to issue the reverse routing information stored according to the array path in the rrep until the node reaches the destination node.

In the fifth step, after the algorithm completes the generation of the disjoint path, the metric value of the main path of each node reaching the base station is recordedAnd secondary path metric value

When the base station receives the rreq data packet sent by the node again, the algorithm is used again for calculationOutgoing path metric matrix M_PLQObtaining the current main path metric value of each node

And secondary path metric value

When the path state satisfies the formula (6) or the formula (7), the path state of the node reaches the update condition;

wherein the expressions (8), (9), (10), (11) are used to calculate

Link quality value of primary path of previous network,

Link quality value of secondary path for previous network,

For the link quality value of the current network primary path,

The link quality value, lambda, for the current network secondary path is the locally updated threshold value (lambda)>0),ΔPLQ^MasterTaking the link quality increment of the current network main path as delta PLQ^Master>Lambda indicates that updating is triggered only when the link quality degradation exceeds a threshold, and otherwise updating is not required; delta PLQ^SecondaryTaking the absolute value of the link quality increment of the current network auxiliary path, namely delta LQ^Secondaty>When the lambda represents the link quality variation, namely the reduction or the improvement exceeds a threshold value, the maintenance is triggered; when the calculated result meets the updating condition, the path information updating network for updating all the nodes is issued,

<math> <mrow> <msup> <mi>ΔPLQ</mi> <mi>Master</mi> </msup> <mo>=</mo> <msubsup> <mi>PLQ</mi> <mi>t</mi> <mi>Master</mi> </msubsup> <mo>-</mo> <msubsup> <mi>PLQ</mi> <mn>0</mn> <mi>Master</mi> </msubsup> <mo>></mo> <mi>λ</mi> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>6</mn> <mo>)</mo> </mrow> </mrow> </math>

or

<math> <mrow> <msup> <mi>ΔPLQ</mi> <mi>Secondary</mi> </msup> <mo>=</mo> <mo>|</mo> <msubsup> <mi>PLQ</mi> <mi>t</mi> <mi>Secondary</mi> </msubsup> <mo>-</mo> <msubsup> <mi>PLQ</mi> <mn>0</mn> <mi>Secondary</mi> </msubsup> <mo>|</mo> <mo>></mo> <mi>λ</mi> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>7</mn> <mo>)</mo> </mrow> </mrow> </math>

<math> <mrow> <msubsup> <mi>PLQ</mi> <mn>0</mn> <mi>Master</mi> </msubsup> <mo>=</mo> <munderover> <mi>Σ</mi> <mrow> <mi>K</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msubsup> <mi>LQ</mi> <mn>0</mn> <mi>Master</mi> </msubsup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>8</mn> <mo>)</mo> </mrow> </mrow> </math>

<math> <mrow> <msubsup> <mi>PLQ</mi> <mn>0</mn> <mi>Secondary</mi> </msubsup> <mo>=</mo> <munderover> <mi>Σ</mi> <mrow> <mi>K</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msubsup> <mi>LQ</mi> <mn>0</mn> <mi>Secondary</mi> </msubsup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>9</mn> <mo>)</mo> </mrow> </mrow> </math>

<math> <mrow> <msubsup> <mi>PLQ</mi> <mi>t</mi> <mi>Master</mi> </msubsup> <mo>=</mo> <munderover> <mi>Σ</mi> <mrow> <mi>K</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msubsup> <mi>LQ</mi> <mi>t</mi> <mi>Master</mi> </msubsup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>10</mn> <mo>)</mo> </mrow> </mrow> </math>

<math> <mrow> <msubsup> <mi>PLQ</mi> <mi>t</mi> <mi>Secondary</mi> </msubsup> <mo>=</mo> <munderover> <mi>Σ</mi> <mrow> <mi>K</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msubsup> <mi>LQ</mi> <mi>t</mi> <mi>Secondary</mi> </msubsup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>11</mn> <mo>)</mo> </mrow> <mo>.</mo> </mrow> </math>

The invention has the beneficial effects that:

by adopting a CCDMPR algorithm in a central computing mode, the algorithm can generate an approximate optimal 2-node (link) disjoint path from a source node to a sink node, so that the average path length can be obviously reduced, and the reliability of the path is improved. The micro-routing table adopted by the algorithm can greatly reduce the storage and communication overhead of the nodes; the self-adaptive path maintenance mechanism of central scheduling obviously improves the flexibility of route maintenance; in data routing, the data packet only needs to carry a path bit sequence without any control message overhead. The algorithm is suitable for application scenarios with stable network topology and strong sink node operation and storage capacity.

Drawings

FIG. 1a is a logical nested two-level tree;

FIG. 1b is a corresponding nested collection;

FIG. 2 is a micro routing table format;

FIG. 3 is a data packet format and path bit sequence;

FIG. 4 is a system workflow;

FIG. 5 is a flooding network topology;

FIG. 6 is a flow chart of a disjoint path algorithm;

FIG. 7 is a flow chart for generating a bit sequence;

fig. 8 is a flow chart of a node processing rrep packets;

fig. 9 shows the success rate of packet transmission in the real experiment.

Detailed Description

The invention is further described with reference to the following figures and examples.

The working principle of the invention is as follows:

1. collection of neighbor information

1.1 rreq packet format

Type data packet Type with RREQ _ MSG value

RREQ _ ID RREQ data packet ID, which takes the value of the node number of the source node generating RREQ

The Flag neighbor information group number is sent in three groups due to the limitation of network bandwidth, and the attribute can be selected as INFO _ FIRST, INFO _ SECOND, and INFO _ THIRD.

A Bit path information Bit value, wherein rreq is sent according to the Bit value and the established optimal path after the path is established

Hop count between Hopcnt from the source node to the node that is processing the rreq packet.

The node number of the node which Origineaddr sends rreq is introduced in 1.2-node flooding transmission

Seqnum the current rreq sequence number, as described in section 1.2

The first Neighbor node number in the Neighbor information table of Neighbor1 node.

Link quality values of Cost1 and Neighbor1

......

The Neighbor and the cost value thereof are Neighbor information to be collected by the sink node.

1.2 Collection of rreq

When the disjoint paths are not established in the network, the network collects node neighbor information in a flooding manner. The nodes in the network broadcast the packaged neighbor information to their own neighbor nodes in a broadcast manner, if the nodes receive the rreq sent by other nodes (Origineaddr), firstly, according to Flag attribute and Seqnum attribute, judging whether to receive the rreq data packet, if the rreq data packet is received, discarding the data packet, if not, broadcasting the rreq data packet to other neighbor nodes except the Origineaddr node, thus effectively avoiding the occurrence of loops in the network and reducing data redundancy. And finally stopping until the sink node receives the rreq data packet.

When a non-intersecting path is established in the network, an rreq data packet of a node is sent to a Sink node and a Sink node according to the established optimal path, the rreq data packet is received, the data packet is analyzed to obtain a neighbor node and a corresponding link quality value, the neighbor node and the corresponding link quality value are sent to a PC through a serial port, and the PC temporarily stores neighbor information in a neighbor information table of a database according to the received data packet.

2. Problem description and model definition

Assuming that the links between nodes are bilaterally symmetric, the wireless sensor network can be abstractly represented by an assigned undirected graph G (V, E), where V is a set of nodes and E is a set of wireless links. 2-disjoint paths from any sensor node S to sink node

And

subject to the constraints of equations (1) or (2).

2-node disjointness

<math> <mrow> <mo>&ForAll;</mo> <mi>v</mi> <mo>&Element;</mo> <msup> <mi>V</mi> <mo>′</mo> </msup> <mrow> <mo>(</mo> <mi>v</mi> <mo>&Element;</mo> <msubsup> <mi>P</mi> <mrow> <mi>S</mi> <mo>-</mo> <mi>to</mi> <mo>-</mo> <mi>Sink</mi> </mrow> <mi>Master</mi> </msubsup> <mo>&RightArrow;</mo> <mi>v</mi> <mo>&NotElement;</mo> <msubsup> <mi>P</mi> <mrow> <mi>S</mi> <mo>-</mo> <mi>to</mi> <mo>-</mo> <mi>Sink</mi> </mrow> <mi>Secondary</mi> </msubsup> <mo>)</mo> </mrow> <mo>,</mo> </mrow> </math>

Wherein V = V- { S, Sink } (1)

2-link disjoint

<math> <mrow> <mo>&ForAll;</mo> <mi>e</mi> <mo>&Element;</mo> <mi>E</mi> <mrow> <mo>(</mo> <mi>e</mi> <mo>&Element;</mo> <msubsup> <mi>P</mi> <mrow> <mi>S</mi> <mo>-</mo> <mi>to</mi> <mo>-</mo> <mi>Sink</mi> </mrow> <mi>Master</mi> </msubsup> <mo>&RightArrow;</mo> <mi>e</mi> <mo>&NotElement;</mo> <msubsup> <mi>P</mi> <mrow> <mi>S</mi> <mo>-</mo> <mi>to</mi> <mo>-</mo> <mi>Sink</mi> </mrow> <mi>Secondary</mi> </msubsup> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>2</mn> <mo>)</mo> </mrow> </mrow> </math>

Definitions 1. Adjacent matrix M_A，M_AIs an n x n matrix, n is the number of all nodes except the sink in the network, (M)_A)_ijThe present invention assumes that the link quality LQ can be known by a certain method, the range of LQ values is 0-255, and a smaller LQ value indicates better link quality.

Define 2. main path matrix

Sum and secondary path matrix

And

are all n × HTL matrixes, n is the number of all nodes except the sink in the network, HTL (hop-to-live) represents the maximum hop number allowed in the network,

represents the j-th hop node on the main path from the source node with index i to the sink node,

represents the j-th hop node on the secondary path from the source node with index i to the sink node,

or

A value of 0 indicates that the path has reached the destination node (sink).

The central calculation model adopted by the invention adopts an adjacency matrix M_AAs input variables, disjoint main path matrices

Sum and secondary path matrix

As output variable. According to the formula (1) and the formula (2), the disjoint path constraint of the central calculation model is converted into the formula (3) and the formula (4).

2-point disjoint Path constraint:

<math> <mrow> <mo>&ForAll;</mo> <mi>i</mi> <mo>&Element;</mo> <mo>{</mo> <mn>1,2,3</mn> <mo>·</mo> <mo>·</mo> <mo>·</mo> <mi>n</mi> <mo>}</mo> <mo>&ForAll;</mo> <mi>p</mi> <mo>,</mo> <mi>q</mi> <mo>&Element;</mo> <mo>{</mo> <mn>1,2,3</mn> <mo>·</mo> <mo>·</mo> <mo>·</mo> <mi>HTL</mi> <mo>}</mo> </mrow> </math>

(3)

2-link disjoint path constraint:

<math> <mrow> <mo>&ForAll;</mo> <mi>i</mi> <mo>&Element;</mo> <mo>{</mo> <mn>1,2,3</mn> <mo>·</mo> <mo>·</mo> <mo>·</mo> <mi>n</mi> <mo>}</mo> <mo>&ForAll;</mo> <mi>p</mi> <mo>,</mo> <mi>q</mi> <mo>&Element;</mo> <mo>{</mo> <mn>1,2,3</mn> <mo>·</mo> <mo>·</mo> <mo>·</mo> <mrow> <mo>(</mo> <mi>HTL</mi> <mo>-</mo> <mn>1</mn> <mo>)</mo> </mrow> <mo>}</mo> </mrow> </math>

(4)

CCDMPR algorithm design

The operation process of the CCDMPR algorithm is divided into four phases: the method comprises the steps of global father node matrix establishment, disjoint path generation and downloading, data routing and adaptive path maintenance.

3.1 Global parent node matrix building

When the network is initialized, the algorithm firstly passes through the settingCollecting neighbor table information of all sensor nodes and LQ between the nodes in a diffusion mode, and storing the LQ in an adjacency matrix M_A. In order to prevent paths with link quality that are too poor, a threshold η is set as an LQ threshold, and links greater than η (indicating poor link quality) are discarded. Then, the algorithm is at M_ABased on the logic gradient, a global father node matrix M is established_F。

Definition 3. Global parent node matrix M_F，M_FIs an n × n matrix, n is the number of all nodes except the sink in the network, and if the node i is a child node of a logical two-level tree with the node j as the root, let (M)_F)_ij=(M_A)_ijOtherwise (M)_F)_ij=255。

The establishment process of the global parent node matrix is as follows: first, search algorithm is first applied to M through breadth first_AA logical gradient to the sink node is established. The logic gradient of the sink node is marked as 0, the logic gradient of the node directly adjacent to the sink node is set as 1, the gradient of the node directly adjacent to the node with the gradient i is set as i + l, and searching is carried out in sequence until each node obtains the logic gradient; secondly, each non-leaf node takes a self node as a root, and takes a neighbor node with a logic gradient equal to or less than the self logic gradient as a direct son to establish a logic two-layer tree;

thirdly, storing the corresponding tree structure in the global father node matrix M_FThe storage mode is as follows: if node i belongs to the logical two-level tree rooted at node j, then order (M)_F)_ij＝(M_A)_ij' else (M)_F)_ij＝255。

The logical representation corresponding to the global parent node matrix is shown in fig. 1(a), and the corresponding logical nested set is shown in fig. 1 (b).

The establishment of the global parent node matrix can enable each sensor node to join a plurality of trees, establish an infrastructure for the generation of the following 2-disjoint paths, and simultaneously prune some redundant edges and edges which possibly generate loops.

3.2 disjoint Path Generation

CCDMPR algorithm on global father node matrix M_FAnd establishing disjoint main paths from all nodes to the sink node on the corresponding logic graph through a greedy path tracking algorithm, and then generating disjoint auxiliary paths from all nodes to the sink node by using a next-hop competition algorithm with a backoff set.

3.2.1 disjoint Main Path Generation

Define 4. Path metric matrix M_PLQ，M_PLQIs an n multiplied by 1 matrix, n is the number of all nodes except the sink in the network, (M)_PLQ)_iThe PLQ value represents the path from node i to the sink node.

The algorithm adopts a greedy path tracking algorithm to generate an optimal path, and simultaneously calculates the sum of LQs of all links contained in the path, which is called PLQ (Path Link quality). The generated optimal path is stored in a main path matrix as a main path of the disjoint pathIn (2), the corresponding PLQ value is stored in the path metric matrix M_PLQIn (1). The algorithm is described as follows:

Trace_Greedy (M_{F}, M_{P}^{Master}, M_{PLQ})

1) set S = {1,2,3 … n }, set (M)_PLQ)₀＝0;

If (M)_F)_i0<255, order (M)_PLQ)_i=(M_F)_i0,NEXT[i]=0,

If (M)_F)_i0=255, order (M)_PLQ)_i=∞,NEXT[i]=-1.

2) Order (M)_PLQ)_j=min{(M_PLQ)_i,i∈S}，S=S-{j}，Then turn to 4)

3) For all i ∈ S, if present (M)_F)_ji<255，

Then order (M)_PLQ)_i=min{(M_PLQ)_i,(M_PLQ)_j+(M_F)_ji}，NEXT[i]J, however, turn 2).

4) Let k be 2 and k be equal to 2,

T = {(M_{P}^{Master})}_{j 1};

while (T > 0) {k + +; T = NEXT [T]; {(M_{P}^{Master})}_{jk} = T};

if S is empty set, the algorithm endsOtherwise go to 3).

3.2.2 disjoint Secondary Path Generation

After the main path of the disjoint path is generated, the auxiliary path of the disjoint path can be obtained by deleting the corresponding node (link) and then applying the greedy algorithm, but in this way, each node needs to be subjected to one-time full-algorithm operation, and the time complexity of the disjoint auxiliary paths of all the nodes is calculated to be O (n)³) And is not suitable for a larger-scale wireless sensor network. The invention uses a Next Hop Competition (NHC) algorithm with a backoff set, the algorithm can obtain an approximate optimal solution, and the time complexity of the optimal auxiliary path of all nodes is calculated to be O (n)²). Each node adds all nodes (links) included in the main path to the backoff set AvoidSet, all parent nodes (links) not included in the AvoidSet constitute a candidate node (link) set, and then selects a candidate node with the smallest PLQ as a next hop through competition until the sink node is reached. The competition mechanism of the NHC algorithm cannot guarantee the occurrence of the loop, so that in the running process of the algorithm, all nodes (links) passed by the current auxiliary path are also added into the AvoidSet, and the mechanism can effectively avoid the occurrence of the loop.

The NHC algorithm is described as follows:

①

NHC_NodeDisjoint (M_{F}, M_{P}^{Master}, M_{P}^{Secondary})

for each source node indexed by i

1) Set S ═ {1,2,3 … n }, and compare all

And adding a backoff node set AvoidSet, and then making S equal to S-AvoidSt, k equal to 1 and u equal to i.

2) Order (M)_PLQ)_v=min{(M_PLQ)_lL ∈ S and F_ul< 255 >, order

{(M_{P}^{Secondary})}_{ik} = v, S = S - {v} .

3) If v is 0, the algorithm ends, otherwise let k be k +1, u be v, go to 2).

②

NHC_LinkDisjoint (M_{F}, M_{P}^{Master}, M_{P}^{Secondary})

For each source node indexed by i

1) Set S ═ {1,2,3 … n), for all

Let us orderWill edge<e，f>For joining the backoff link set AvoidSet, then let k be 1 and u be i.

2) Order toOrder toAt the same time, the edge is moved<e，f>AvoidSet was added.

3) If v =0, the algorithm ends, otherwise let k = k +1, u = v, go 2).

3.3 micro-routing tables and data routing

And when the 2-disjoint path calculation of all the nodes is completed, the sink node downloads the calculated path to each node. Considering that the complete routing table is large, the downloading will bring large communication overhead, and meanwhile, the storage space of the sensor node itself is limited, so the whole routing table downloading scheme is not feasible. The invention designs a micro-routing table mechanism based on < main father node, auxiliary father node > pair and path bit sequence, and the format of the micro-routing table is shown in figure 2.

The micro-routing table contains five fields: the ID represents the node identification of the micro-routing table, the Primary node field stores the father node of the current node on the Primary path, the Secondary node stores the father node of the current node on the Secondary path, the Primary path bits series field stores the bit sequence of the Primary path (for example: 00000000), the Secondary path bits series field stores the bit sequence of the Secondary path (for example: 01010010), each bit of the bit sequence represents the father node selection of each hop (0 represents the Primary father node, and 1 represents the Secondary father node). The advantage of using the path bit sequence is that each hop on the micro-routing table only occupies one bit, which can greatly reduce the size of the routing table, thereby saving communication and storage overhead.

When the source node selects to transmit the data packet to the sink node along the main (auxiliary) path, the source node adds the own main (auxiliary) path bit sequence and sequence pointer (the sequence pointer is 1 at the beginning) to the head of the generated data packet, and transmits the data packet to the sink node along the main (auxiliary) path. The format of the data packet generated by the source node is shown in fig. 3.

The Path flag is a Path flag, Path flag =0 represents a main Path, and Path flag =1 represents an auxiliary Path; the Series pointer is a sequence pointer and is used for indicating the current bit position in the sequence; the Pathbit series is the carried bit sequence; the LQ is the current value of the LQ from the source node to the primary (secondary) parent node, and is mainly used for path maintenance.

After receiving the data packet, the intermediate node firstly analyzes the data packet head, adds 1 to the sequence pointer, and reads the bit element value in the path bit sequence indicated by the current sequence pointer, wherein the bit element value is transmitted to the main father node when being 0 and is transmitted to the auxiliary father node when being 1. By adopting the routing strategy, when a failure node (link) occurs, data can effectively avoid the failure node (link) only by transmitting along the auxiliary path of the current node without backtracking to the source node for resending, thereby improving the reliability and saving the communication overhead.

Issue of path information

4.1 Path information packet rrep packet Format

Type

Path

Pri_parent

Pri_bit

Sec_parent

Sec_bit

Destiaddr

Seqnum

Type packet Type RREP _ MSG

Path node array, reverse routing Path node

Pri _ parent Primary parent node number

Bit sequence of the Pri _ bit main path

Sec _ parent Secondary parent node number

Bit sequence of Sec _ bit auxiliary path

Destination node address of Destiaddr, node number of node to which path information is to be transmitted

Sequence number of Seqnum rrep

4.2 routing information packet rrep

4.2.1 determining the routing information issuing sequence

As shown in the figure, the stable network can be established more quickly by sequentially issuing the path information of the nodes according to the increase of the logic gradient of the nodes by taking the sink node as a starting point, the stable network can be established as long as the nodes receive the path information, and even if the nodes which do not receive the path information exist in the global network, the stable local network established by the nodes which receive the path information cannot be influenced.

4.2.2 reverse routing of rrep packets

5. Maintenance update of paths

5.1 update Condition

When the algorithm completes the generation of the disjoint path, the main path metric value of each node reaching the base station is recorded

And secondary path metric value

When the base station receives the rreq data packet sent by the node again, the path metric matrix M is calculated by using the algorithm again_PLQObtaining the current main path metric value of each nodeAnd secondary path metric value

And when the path state satisfies the formula (6) or the formula (7), the path state of the node reaches the update condition. Wherein the expressions (8), (9), (10), (11) are used to calculateLink quality value of primary path of previous network,

Link quality value of secondary path for previous network,

For the link quality value of the current network primary path,

The link quality value, lambda, for the current network secondary path is the locally updated threshold value (lambda)>0),△PLQ^MasterTaking the link quality increment of the current network main path as delta PLQ^Master>λ indicates that the update is triggered only when the link quality degradation exceeds a threshold, otherwise no update is required. Delta PLQ^SecondaryTaking the absolute value of the link quality increment of the current network auxiliary path, namely delta LQ^Secondary>When the link quality variation (decrease amount or increase amount) exceeds a threshold value, the maintenance is triggered. And when the calculated result meets the updating condition, issuing and updating the path information of all the nodes to update the network.

Or

<math> <mrow> <msubsup> <mi>PLQ</mi> <mi>t</mi> <mi>Secondary</mi> </msubsup> <mo>=</mo> <munderover> <mi>Σ</mi> <mrow> <mi>K</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>N</mi> </munderover> <msubsup> <mi>LQ</mi> <mi>t</mi> <mi>Secondary</mi> </msubsup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>11</mn> <mo>)</mo> </mrow> </mrow> </math>

The real experimental environment is deployed in a corridor from 2 to 5 floors of an office building, all nodes are distributed relatively uniformly, and the sink node is located in the center of the corridor from 4 floors. The node used in the experiment mainly consists of two parts: MSP430F5438 and TI radio frequency chip CC2420, the operating system adopts multitask real-time system FreeRTOS.

For convenient experiments, the applicant adopts upper computer software based on C + + for calculation of the disjoint paths and statistics of the transmission success rate according to the sequence numbers of the received data packets. All nodes in the network transmit collected data to the sink node every 1 second, an ACK retransmission mechanism is not adopted in the transmission process, and the sink node receives the data and then sends the data to upper computer software for processing. In the experiment, the CCDMR algorithm is compared with the classic distributed single-path routing algorithm, namely the minimum-hop algorithm. We collect the data transmission success rates when the number of the data packets successfully received by the sink node is 100, 200, 300, 400, 500 respectively, as shown in fig. 9.

As can be seen from the figure, under the condition that an ACK retransmission mechanism is not adopted, compared with the minimum hop single path algorithm, the primary path and the secondary path generated by the CCDMPR algorithm both have a higher path transmission success rate, and the transmission success rate of the primary path can reach more than 65%. Under the condition of simultaneously using 2-disjoint paths, the success rate of path transmission of the data packet can reach more than 75%, and meanwhile, the stability of the path transmission is improved to a certain extent.

Although the embodiments of the present invention have been described with reference to the accompanying drawings, it is not intended to limit the scope of the present invention, and it should be understood by those skilled in the art that various modifications and variations can be made without inventive efforts by those skilled in the art based on the technical solution of the present invention.

Claims

1. A wireless sensor network 2-disjoint path routing algorithm of central computation is characterized in that hop count and link quality are used as measurement, the full-network information is used for computing an approximate optimal 2-node disjoint path from a source node to a sink node, and then a micro routing table only comprising a main node, an auxiliary node and a path bit sequence is generated and is downloaded to each node; the method comprises the following steps:

first, collecting node neighbor information, creating and initializing an adjacency matrix M_AParent node matrix M_FLogical gradient matrix M_NMoment of travelMatrix M_PAnd a path metric matrix M_PLQ(ii) a Wherein,

fourthly, node path information is issued to all nodes except the sink node;

and fifthly, updating and maintaining the path.

2. The centrally computed 2-disjoint path routing algorithm of a wireless sensor network as defined in claim 1, wherein in said first step, the process of collecting node neighbor information is:

when the non-intersecting paths are not established in the network, the network collects node neighbor information in a flooding mode; the nodes in the network broadcast the packaged neighbor information to own neighbor nodes in a broadcasting mode, if the nodes receive rreq data packets sent by other nodes Origineaddr, whether the rreq data packets are received is judged according to Flag attributes and Seqnum attributes, if the rreq data packets are received, the data packets are discarded, and if the rreq data packets are not received, the rreq data packets are broadcast to other neighbor nodes except the Origineaddr nodes, so that loops are avoided in the network, and data redundancy is reduced; finally, stopping until the sink node receives the rreq data packet;

3. The centrally computed 2-disjoint path routing algorithm in a wireless sensor network as defined in claim 1 wherein in said first step, a global parent node matrix M_FThe establishment procedure is as follows:

1) by breadth-first search algorithm at M_AEstablishing a logic gradient to the sink node; the logic gradient of the sink node is marked as 0, the logic gradient of the node directly adjacent to the sink node is set as 1, the gradient of the node directly adjacent to the node with the gradient i is set as i +1, and searching is carried out in sequence until each node obtains the logic gradient;

4. The centrally computed 2-disjoint path routing algorithm of a wireless sensor network as defined in claim 1, wherein in said second step, the generation of the disjoint main path is performed by:

path metric matrix M_PLQ，M_PLQIs an n multiplied by 1 matrix, n is the number of all nodes except the sink in the network, (M)_PLQ)_iRepresenting the PLQ value of a path from the node i to the sink node, generating an optimal path by adopting a greedy path tracking algorithm, and simultaneously calculating the sum of LQ of all links contained in the path, which is called PLQ; the generated optimal path is stored in a main path matrix as a main path of the disjoint pathIn (2), the corresponding PLQ value is stored in the path metric matrix M_PLQIn (1),

Trace_Greedy (M_{F}, M_{P}^{Master}, M_{PLQ})

1) set S = {1,2,3 … n }, set (M)_PLQ)₀=0;

If (M)_F)_i0<255, order (M)_PLQ)_i=(M_F)_i0,NEXT[i]=0,

If (M)_F)_i0=255, order (M)_PLQ)_i=∞,NEXT[i]=-1.

2) Order (M)_PLQ)_j=min{(M_PLQ)_i,i∈S}，S=S-{j}，

Then turn 4);

3) for all i ∈ S, if present (M)_F)_ji＜255，

Then order (M)_PLQ)_i＝min{(M_PLQ)_i，(M_PLQ)_j+(M_F)_ji}，NEXT[i]J, then turn 2);

4) let k be 2 and k be equal to 2,

T = {(M_{P}^{Master})}_{j 1};

while (T > 0) {k + +; T = NEXT [T]; {(M_{P}^{Master})}_{jk} = T};

the time complexity Of the algorithm for generating the optimal main path from all nodes to the sink node is Ofn²) And the running process of the algorithm ensures that the generated path has no loop.

5. The centrally computed wireless sensor network 2-disjoint path routing algorithm of claim 1, wherein said disjoint secondary paths are generated by a method comprising:

①

NHC_NodeDisjoint (M_{F}, M_{P}^{Master}, M_{P}^{Secondary})

for each source node indexed by i

1) Set S ═ {1,2,3 … n }, and compare all

And adding a backoff node set AvoidSet, and then setting S to S-AvoidSet, k to 1 and u to i.

2) Order (M)_PLQ)_v＝min{(M_PLQ)_lL ∈ S and F_ul＜255}，

Order to

{(M_{P}^{Secondary})}_{ik} = v, S = S - {v} .

3) If v is 0, the algorithm ends, otherwise, k is k +1, u is v, go to 2);

②

NHC_LinkDisjoint (M_{F}, M_{P}^{Master}, M_{P}^{Secondary})

for each source node indexed by i

1) Set S ═ {1,2,3 … n }, for all

Order to

2) order to

Order to

At the same time, the edge is moved<e，f>Adding AvoidSet;

3) if v is 0, the algorithm ends, otherwise let k be k + l, u be v, go to 2).

6. The centrally computed 2-disjoint path routing algorithm of a wireless sensor network as recited in claim 1, wherein in said third step, the micro-routing table comprises five fields: the ID represents the node identification of the micro-routing table, the Primary node domain stores the father node of the current node on the main path, the Secondary node stores the father node of the current node on the Secondary path, the Primary path bit series domain stores the bit sequence of the main path, the Secondary path bit series stores the bit sequence of the Secondary path, each bit of the bit sequence represents the father node selection of each jump, the main father node is represented by 0, and the auxiliary father node is represented by 1;

7. The centrally-computed 2-disjoint path routing algorithm for a wireless sensor network according to claim 1, wherein in the fourth step, the sink node is used as a starting point to sequentially issue the path information of the nodes according to the increase of the logical gradient of the nodes, and a stable network is established as long as the nodes receive the path information, and even if there are nodes that do not receive the path information in the global network, the stable local network established by the nodes that receive the path information is not affected;

8. The centrally computed wireless sensor network 2-disjoint road of claim 1A path routing algorithm, wherein in the fifth step, after the algorithm completes the generation of the disjoint path, the metric value of the main path from each node to the base station is recorded

And secondary path metric value

When the base station receives the rreq data packet sent by the node again, the path metric matrix M is calculated by using the algorithm again_PLQObtaining the current main path metric value of each node

And secondary path metric value

wherein the expressions (8), (9), (10), (11) are used to calculateLink quality value of primary path of previous network,Link quality value of secondary path for previous network,

For the link quality value of the current network primary path,The link quality value, lambda, for the current network secondary path is the locally updated threshold value (lambda)>0),ΔPLQ^MasterTaking the link quality increment of the current network main path as delta PLQ^Master>λ indicates that only when the amount of link quality degradation exceeds a thresholdThe update is triggered, and the update is not needed in other situations; delta PLQ^SecondaryTaking the absolute value of the link quality increment of the current network auxiliary path, namely delta LQ^Secondary>When the lambda represents the link quality variation, namely the reduction or the improvement exceeds a threshold value, the maintenance is triggered; when the calculated result meets the updating condition, the path information updating network for updating all the nodes is issued,

or