CN109005113B - Low-delay routing method of intertidal zone sensor network based on machine learning - Google Patents

Low-delay routing method of intertidal zone sensor network based on machine learning Download PDF

Info

Publication number
CN109005113B
CN109005113B CN201810725803.1A CN201810725803A CN109005113B CN 109005113 B CN109005113 B CN 109005113B CN 201810725803 A CN201810725803 A CN 201810725803A CN 109005113 B CN109005113 B CN 109005113B
Authority
CN
China
Prior art keywords
node
delay
nodes
sensor network
link
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201810725803.1A
Other languages
Chinese (zh)
Other versions
CN109005113A (en
Inventor
徐文渊
冀晓宇
周歆妍
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhejiang University ZJU
Original Assignee
Zhejiang University ZJU
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Zhejiang University ZJU filed Critical Zhejiang University ZJU
Priority to CN201810725803.1A priority Critical patent/CN109005113B/en
Publication of CN109005113A publication Critical patent/CN109005113A/en
Application granted granted Critical
Publication of CN109005113B publication Critical patent/CN109005113B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00—Routing or path finding of packets in data switching networks
    • H04L45/12—Shortest path evaluation
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L45/00—Routing or path finding of packets in data switching networks
    • H04L45/12—Shortest path evaluation
    • H04L45/121—Shortest path evaluation by minimising delays

Abstract

The invention discloses a low-delay routing method of a intertidal zone sensor network based on machine learning, which comprises the following steps of 1) modeling the delay of the intertidal zone wireless sensor network, 2) calculating the delay caused by data packet blocking by each node, 3) calculating the link delay from each node to a neighbor node, and 4) calculating a delay optimal path by using a shortest path algorithm aiming at delay optimization. The method comprises the steps of modeling each component of the time delay by analyzing the main reason of the time delay in the intertidal zone sensor network, quantizing each time delay component, and finally selecting the route with the optimal time delay as a data transmission path with the aim of time delay optimization. The invention takes time delay as guidance, and greatly relieves the time delay problem of the intertidal zone sensor network.

Description

Low-delay routing method of intertidal zone sensor network based on machine learning
Technical Field
The invention belongs to the field of wireless sensor networks, and particularly relates to a low-delay routing method of an intertidal zone sensor network based on machine learning.
Background
The intertidal zone is an area along the coastline, between the high tide line and the low tide line. The intertidal zone is a habitat for many marine organisms to live on, however, the zone has complicated physical environment changes, such as water level, temperature, oxygen content and the like. Therefore, the deployment of the wireless sensor network in the intertidal zone has important significance for environmental monitoring and ecological protection.
Deploying wireless sensor networks in the intertidal zone faces a number of challenges. The first is the choice of communication medium. Usually for land-based sensor networks, it is common to use radio frequency signals as a carrier for data transmission. However, the penetration capacity of the radio frequency signal to water, especially seawater, is poor, and is only about 12 centimeters, so that data cannot be uploaded to the base station in real time during the period that the sensor node is submerged by seawater. On the other hand, even in a low tide period, normal communication is seriously affected by tidal water due to a harsh physical environment in the intertidal zone, and the efficiency of data transmission is very low. The intuitive manifestation of these problems is that the transmission delay of data is much higher than that of other terrestrial sensor networks of the same scale.
According to experimental statistics, the time delay of the intertidal zone wireless sensor network is far higher than that of an indoor wireless sensor network under the same scale. Taking the delay caused by the link quality as an example, the link delay of the indoor wireless sensor network is usually within one system period, while in the intertidal wireless sensor network, the value is 15 system periods. In addition, the accumulation of data packets due to link congestion also becomes a major cause of data transmission delay.
In order to solve the problems, the invention provides a low-delay routing method of an intertidal zone sensor network based on machine learning. The method comprises the steps of firstly modeling the delay of the intertidal zone wireless sensor network, quantizing the model by using a machine learning method, and finally designing a routing method taking low delay as guidance.
Disclosure of Invention
The invention provides a low-delay routing method of a intertidal zone sensor network based on machine learning.
The technical scheme of the invention is as follows:
a low-delay routing method of a intertidal zone sensor network based on machine learning comprises the following steps:
1) modeling intertidal zone wireless sensor network delay;
2) each node calculates the delay caused by the data packet blockage;
3) each node calculates the link delay to the neighbor node;
4) and calculating the optimal delay path by using a shortest path algorithm aiming at delay optimization.
In the above technical solution, the step 1) is specifically as follows:
the latency of the intertidal zone wireless sensor network is divided into link latency and congestion latency, and the total latency of a selected route is:
Figure BDA0001719684170000021
where n is the number of nodes passing through a route, ldi,i+1For the link delay, nd, of nodes i to i +1j+1The congestion delay for node j + 1.
The step 2) is specifically as follows: the congestion delay caused by the data packet congestion of the node is obtained according to the following formula:
Figure BDA0001719684170000022
wherein, Bi(t) is the number of data packets cached by the node i in the t-th cycle, which is directly obtained by reading the register of the cache region, Ci(t) is the number of data packets that can be processed by the node i in one cycle, and is obtained from the average number of data packets processed by the node i in the previous 10 cycles, and is calculated as follows:
Ci(t)＝Avg.{Ci(t-λ)},λ＝1,2,…,10
Si' (t + τ) is the future state of node i, i.e., whether the node is on water or submerged in seawater at the t + τ th cycle, estimated using a machine learning algorithm.
Future state S of the node ii' (t + τ) was obtained by the following method: so as to: 1) time D when node has been on waterawAnd 2) the number N of the current neighbor nodes of the nodenbAnd 3) the number N of currently available links of the nodelinkAnd selecting a logistic regression model as a machine learning classification algorithm to estimate the state of the node in the future period of tau.
The step 3) is specifically as follows:
first, whether communication can be established between any two nodes is judged
Figure BDA0001719684170000023
When there is an available communication link between two nodes, i.e. /)ijWhen (t) is 1, node i broadcasts a beacon estimate with CTP routing algorithm and the link quality ETX between node jij(t)；
Node i andlink delay between nodes j, ldij(t) is calculated by the following formula:
Figure BDA0001719684170000024
the step 4) is specifically as follows: after the nodes calculate the corresponding congestion delay and the link delay between the nodes and other neighbor nodes, the current shortest path of the node and the congestion delay of the node are broadcasted to all the neighbor nodes in a beacon form, and finally all the nodes in the network use a shortest path algorithm to calculate iteration to obtain the shortest delay route from each node to a base station, wherein the shortest path algorithm adopts a dijkstra method.
The invention has the following beneficial effects: the method firstly models the time delay of the intertidal zone wireless sensor network, divides the time delay in the wireless sensor network into node congestion delay and link delay, and respectively quantizes the node congestion delay and the link delay. In the link delay quantification process, the method uses a machine learning algorithm. The method finally takes time delay as guidance, designs a data forwarding route, and finally effectively shortens the time delay of the intertidal zone sensor.
Drawings
FIG. 1 is a schematic diagram of routing implemented by the present invention.
FIG. 2 is a flow chart illustrating an embodiment of the present invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
Fig. 1 shows the routing method of the present invention. The invention selects a path based on a shortest path algorithm, and mainly contributes to providing a routing matrix which aims at the intertidal zone sensor network and takes time delay as guidance, and each time delay component in the intertidal zone sensor network needs to be modeled and quantized in the process. As shown in fig. 1, the present invention mainly divides the delay components in the intertidal sensor network into link delay (numbers on the solid line) and congestion delay of the node (numbers in the circle), and the larger the number is, the larger the delay for selecting the link or the node as the route is. The circles in fig. 1 represent a node, the dotted lines represent the available links, and the solid lines represent the final selected path.
The low-delay routing method of the intertidal zone sensor network based on machine learning comprises the following steps:
the method comprises the following steps: modeling the intertidal zone wireless sensor network delay. Through data analysis of a wireless sensor network deployed in an intertidal zone, the main reasons for causing large delay of the intertidal zone wireless sensor network are found to be as follows: 1) poor link quality due to the environment and link delay due to tidal fluctuations; and 2) congestion delay of the buffer area, which is caused by the network congestion and the data packets are accumulated continuously. For convenience of explanation, the symbols used in the following description are set forth in table 1.
TABLE 1 symbolic meanings description used in summary of the invention
Figure BDA0001719684170000031
Figure BDA0001719684170000041
The latency of the intertidal wireless sensor network is divided into link latency and congestion latency, and the total latency of a selected route is defined by the following equation:
Figure BDA0001719684170000042
where n is the number of nodes traversed in a route, i ∈ {1,2, …, n }, j ∈ {1,2, …, n-1.
Step two: and each node calculates congestion delay of each node. Because data packets are continuously accumulated in the buffer area of the node, according to the FIFO (first in first out) principle, a newly generated data packet needs to wait for the existing data in the buffer area to be transmitted before being transmitted. The time from when a packet is pushed into the buffer until it is sent is called the congestion delay. For a general sensor network, network connectivity and link quality are better, so the congestion delay is usually shorter. But for the wireless sensor network in the intertidal zone, the congestion delay is very serious. According to experimental data, the number of data packets in the cache region of the intertidal wireless sensor network node is often hundreds. The congestion delay of a node may be calculated by:
Figure BDA0001719684170000043
wherein B isi(t) can be obtained directly by reading the register of the buffer, Ci(t) is derived from the average number of packets processed by node i in the first 10 cycles, as follows:
Ci(t)＝Avg.{Ci(t-λ)},λ＝1,2,…,10
in the calculation process of congestion delay, the future state of a node, namely S, is usedi' (t + τ). We have found that the future state of the node needs to be taken into account when estimating congestion delay. If a node will be flooded with seawater after a period of τ in the future, the congestion delay of the node should be set to a higher value to avoid that the node is selected as parent by other nodes. To accurately estimate the future state of a node, we look at: 1) time D when node has been on waterawAnd 2) the number N of the current neighbor nodes of the nodenbAnd 3) the number N of currently available links of the nodelinkCharacteristically, a machine learning algorithm is used to estimate the state of the node for the future τ cycles. The node future state can be predicted because the node state is directly related to the tidal water cycle, and the tidal water has obvious periodicity, so that a machine learning algorithm can be used for predicting the possible future state of the node. The performance of three common machine learning algorithms including logistic regression (logistic regression), Adaboost classifier, Gaussian naive Bayes algorithm and the like is compared on actual experimental data, the overhead of each algorithm is considered comprehensively, and a logistic regression model is selected as the classification algorithm of machine learning finally.
Step three: each node estimates the link delay of its available links. It is first determined whether communication can be established for any two nodes according to the following equation.
Figure BDA0001719684170000051
When there is an available communication link between two nodes, i.e. /)ijWhen (t) is 1, node i broadcasts a beacon estimate with CTP routing algorithm and the link quality ETX between node jij(t) of (d). At this time, we define the link delay between node i and node j to be calculated by:
Figure BDA0001719684170000052
step four: and calculating the optimal delay path by using a shortest path algorithm aiming at delay optimization. After the node calculates the congestion delay of the node and the link delay between the node and other neighbor nodes, the current shortest path and the congestion delay of the node are broadcasted to all the neighbor nodes in a beacon form. Finally, all nodes in the network use a shortest path algorithm to calculate and iterate to obtain the delay shortest route from each node to the base station, the process is shown in fig. 2, and the shortest path algorithm can directly adopt a dijkstra algorithm.

Claims (2)

1. A low-delay routing method of an intertidal zone sensor network based on machine learning is characterized by comprising the following steps:
1) modeling intertidal zone wireless sensor network delay;
2) each node calculates the delay caused by the data packet blockage;
3) each node calculates the link delay to the neighbor node;
4) calculating a delay optimal path by using a shortest path algorithm with the aim of delay optimization;
the step 1) is as follows:
the latency of the intertidal zone wireless sensor network is divided into link latency and congestion latency, and the total latency of a selected route is:
Figure FDA0002357847950000011
where n is the number of nodes passing through a route, ldi,i+1For the link delay, nd, of nodes i to i +1j+1Congestion delay for node j + 1;
the step 2) is specifically as follows: the congestion delay caused by the data packet congestion of the node is obtained according to the following formula:
Figure FDA0002357847950000012
wherein, B1() The number of the data packets cached for the node i in the t-th cycle is directly obtained by reading a register of a cache region Ci() The number of data packets that can be processed by the node i in one cycle is obtained from the average number of data packets processed by the node i in the previous 10 cycles, and is calculated as follows:
Ci(t)＝Avg.{Ci(t-λ)},λ＝1,2,…,10
S′i(t + τ) is the future state of the node i, namely whether the node is positioned on water or submerged by seawater in the t + τ period, and is estimated by using a machine learning algorithm;
the step 3) is specifically as follows:
first, whether communication can be established between any two nodes is judged
Figure FDA0002357847950000013
When there is an available communication link between two nodes, i.e. /)ijWhen (t) is 1, node i broadcasts a beacon estimate with CTP routing algorithm and the link quality ETX between node jij(t)；
Link delay ld between node i and node jij(t) is calculated by the following formula:
Figure FDA0002357847950000014
the step 4) is specifically as follows: after the nodes calculate the corresponding congestion delay and the link delay between the nodes and other neighbor nodes, the current shortest path of the node and the congestion delay of the node are broadcasted to all the neighbor nodes in a beacon form, and finally all the nodes in the network use a shortest path algorithm to calculate iteration to obtain the shortest delay route from each node to a base station, wherein the shortest path algorithm adopts a dijkstra method.
2. The method of claim 1, wherein the future state S 'of node i is a low latency routing method for machine learning based intertidal sensor networks'i(t + τ) was obtained by the following method:
so as to: 1) time D when node has been on waterawAnd 2) the number N of the current neighbor nodes of the nodenbAnd 3) the number N of currently available links of the nodelinkAnd selecting a logistic regression model as a machine learning classification algorithm to estimate the state of the node in the future period of tau.
CN201810725803.1A 2018-07-04 2018-07-04 Low-delay routing method of intertidal zone sensor network based on machine learning Active CN109005113B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810725803.1A CN109005113B (en) 2018-07-04 2018-07-04 Low-delay routing method of intertidal zone sensor network based on machine learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810725803.1A CN109005113B (en) 2018-07-04 2018-07-04 Low-delay routing method of intertidal zone sensor network based on machine learning

Publications (2)

Publication Number Publication Date
CN109005113A CN109005113A (en) 2018-12-14
CN109005113B true CN109005113B (en) 2020-04-10

Family

ID=64598145

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810725803.1A Active CN109005113B (en) 2018-07-04 2018-07-04 Low-delay routing method of intertidal zone sensor network based on machine learning

Country Status (1)

Country Link
CN (1) CN109005113B (en)

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101895952A (en) * 2010-07-16 2010-11-24 山东省计算中心 Multi-route establishment method and parallel data transmission method of wireless sensor network
CN102196527A (en) * 2011-05-28 2011-09-21 东华大学 Route recovery method and recovery protocol for mobile Sink wireless sensor network
CN102497440A (en) * 2011-12-20 2012-06-13 山东大学 Low-latency data aggregation algorithm
CN103974442A (en) * 2014-04-24 2014-08-06 东南大学 Low-delay scheduling method suitable for wireless sensor network
CN107453801A (en) * 2017-08-28 2017-12-08 西安电子科技大学 A kind of Layered Multipath method for routing towards satellite network
CN108112048A (en) * 2017-12-28 2018-06-01 重庆邮电大学 Recognize a kind of efficiency route protocol based on stability in sensor network

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9531658B2 (en) * 2014-07-16 2016-12-27 International Business Machines Corporation Routing messages based on geolocation information associated with both the messages and with subscribers
US10477359B2 (en) * 2014-12-08 2019-11-12 International Business Machines Corporation Publishing messages based on geographic area

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101895952A (en) * 2010-07-16 2010-11-24 山东省计算中心 Multi-route establishment method and parallel data transmission method of wireless sensor network
CN102196527A (en) * 2011-05-28 2011-09-21 东华大学 Route recovery method and recovery protocol for mobile Sink wireless sensor network
CN102497440A (en) * 2011-12-20 2012-06-13 山东大学 Low-latency data aggregation algorithm
CN103974442A (en) * 2014-04-24 2014-08-06 东南大学 Low-delay scheduling method suitable for wireless sensor network
CN107453801A (en) * 2017-08-28 2017-12-08 西安电子科技大学 A kind of Layered Multipath method for routing towards satellite network
CN108112048A (en) * 2017-12-28 2018-06-01 重庆邮电大学 Recognize a kind of efficiency route protocol based on stability in sensor network

Also Published As

Publication number Publication date
CN109005113A (en) 2018-12-14

Similar Documents

Publication Publication Date Title
CN108075975B (en) Method and system for determining route transmission path in Internet of things environment
CN112188583B (en) Ocean underwater wireless sensing network opportunistic routing method based on reinforcement learning
CN113163466B (en) Self-adaptive fish school routing packet routing method based on fuzzy decision tree
CN104113855A (en) Channel-based routing algorithm of wireless self-organizing network
Domingo Marine communities based congestion control in underwater wireless sensor networks
CN106211187A (en) A kind of water sound sensor network dynamic gateway node deployment method based on prediction
CN109005113B (en) Low-delay routing method of intertidal zone sensor network based on machine learning
Sivakumar Minimizing transmission loss using inspired ant colony optimization and Markov chain Monte Carlo in underwater WSN environment
US20130242746A1 (en) Wireless communication device and route selection method in wireless network
CN108632785B (en) Ant colony self-adaptive Internet of vehicles routing method based on link quality
CN109842888B (en) Underwater acoustic channel quality dynamic evaluation and prediction method and system for underwater sensor network
JP5716836B2 (en) Wireless multi-hop network monitoring method
KR102006292B1 (en) Apparatus and method for transporting multimedia using FEC, and DNN model selecting apparatus for determining FEC parameter
CN109996171B (en) Amorphous positioning method for genetic-tabu search optimization of wireless sensor network
Alhumyani et al. Deployment strategies for underwater sensing and processing networks
Umar et al. On enhancing network reliability and throughput for critical-range based applications in uwsns
Liu et al. Fusion of state estimates over long-haul sensor networks under random delay and loss
Alhumyani et al. Efficient surface-level gateway deployment using underwater sensing and processing networks
Deepak et al. Secured route selection using E-ACO in underwater wireless sensor networks
Matsuura Maximizing lifetime of multiple data aggregation trees in wireless sensor networks
CN109831264B (en) Time sequence underwater sound channel quality prediction method and system based on nearest neighbor regression
Najeeb et al. A new advanced energy efficient routing protocol for UWSNs
CN113938976B (en) Intelligent utility planning passive perception routing algorithm of Internet of things
Ruan et al. A self-adaptive spatial-temporal correlation prediction algorithm to reduce data transmission in wireless sensor networks
Lou et al. A new anti-jamming reliable routing protocol for tactical MANETs

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant