Các thuật toán xấp xỉ tìm đường xâm nhập có khả năng bị phát hiện nhỏ nhất trong mạng cảm biến dây (approximate algorithms for solving the minimal exposure path problems in wireless sensor networks) tt tiếng anh

MINISTRY OF EDUCATION AND TRAININGHANOI UNIVERSITY OF SCIENCE AND TECHNOLOGYNGUYEN THI MY BINH APPROXIMATE ALGORITHMS FOR SOLVING THE MINIMAL EXPOSURE PATH PROBLEMS IN WIRELESS SENSOR NE

MINISTRY OF EDUCATION AND TRAININGHANOI UNIVERSITY OF SCIENCE AND TECHNOLOGY

NGUYEN THI MY BINH

APPROXIMATE ALGORITHMS FOR

SOLVING THE MINIMAL EXPOSURE PATH PROBLEMS IN WIRELESS SENSOR NETWORKS

Major: Computer Science Code: 9480101

SUMMERY OF COMPUTER SCIENCE DISSERTATION

Hanoi − 2020

This dissertation was finalized at Hanoi University of Science

and Technology

Supervisors:

1 Assoc Prof Dr Huynh Thi Thanh Binh

2 Assoc Prof Dr Nguyen Duc Nghia

Date and time:

This dissertation could be found at:

1 Ta Quang Buu Library- Hanoi University of Science and Technology

2 National Library

WSNs have great potential for various applications in many scenarios frommilitary target tracking and surveillance, natural disaster relief, biomedical healthmonitoring, hazardous environment exploration, seismic sensing, industrial diagno-sis, and so on In military target tracking and surveillance, a WSN can assist inintrusion detection and identification Specific examples include spatially-correlatedand coordinated troop and tank movements With natural disasters, sensor nodescan sense and detect the environment to forecast disasters before they occur Inbiomedical applications, surgical implants of sensors can help monitor a patient’shealth For seismic sensing, ad hoc deployment of sensors along the volcanic areacan detect the development of earthquakes and eruptions

To evaluate the efficiency of a WSN, coverage measurement is considered

as one of the fundamental problems By this means, the coverage measurement iscalled coverage problems which can be classified into three categories based onthe coverage type: point-coverage, area coverage and barrier coverage In pointcoverage, the sensor nodes are deployed to cover all given specified points, whereasthe goal of area coverage is that the whole region of interest is covered by the WSN.Unlike these coverage types, in barrier coverage, the subjects to be covered are notfixed or known before node deployment The main focus of barrier coverage is ondynamic targets tracking and surveillance Due to the mobility of targets, barriercoverage is considered to be complicated as compared to the static ones such asarea coverage and point coverage Moreover, to detect intruders penetrating theROI, it is not necessary to guarantee that every point in the ROI is covered by one

or multiple sensor nodes Therefore, full area coverage model is not suitable forintruder detection anymore In contrast, barrier coverage was proposed specificallyfor intruder detection in WSNs where sensing regions of sensor nodes form one ormultiple barriers so that every intruder crossing the ROI will be detected Compared

to full area coverage, barrier coverage can efficiently detect intruders with muchfewer sensor nodes

Figure 1 Examples of target coverage (a), area coverage (b), and barrier coverage(c)

Motivation

This dissertation investigates the minimal exposure path (MEP) problem inWSNs which is the fundamental barrier coverage problem The MEP is a good per-formance metric, which can be used to measure the quality of surveillance systems

or coverage quality of the sensor networks The MEP problem aims at finding apath on which an intruder can penetrate through the sensing field with the lowestprobability of being detected Through the found MEP, the defenders or networkdesigners can estimate the worst-case coverage provided by a sensor network, be-cause objects moving through a sensor field along this path is the most difficult to

be detected Therefore, the knowledge of MEP can be used in managing, ing and maintaining WSNs Furthermore, the applications of MEP are not onlyrestricted to WSNs, but also in many other fields

optimiz-Figure 2 Illustration of a general barrier coverage problem

The MEP problem in WSN strongly depends on various factors such as type ofsensors, sensing coverage model, deployment strategies, deployment environments,approach methods solving the problem, etc.

It is certain that almost the related works did not regard to several aspects

as follows:

ˆ Moving ability of sensor nodes or mobile wireless sensor networks

ˆ Realistic sensing coverage model

ˆ Heterogeneous wireless networks

ˆ Deployment environment with obstacles

These problems have become a motivation for this dissertation to do further researchand come up with a more efficient solutions for the MEP problem

Methodology

The methodology of this dissertation are as follows:

ˆ Theoretical study of barrier coverage problems as optimal problems is done

ˆ Analyses the related works to the barrier coverage problems Especially, theminimal exposure path problem is comprehended and considered

ˆ Propose the practical and useful models of the minimal exposure path problem

in wireless sensor networks

ˆ Proffer efficient algorithms to solve the proposed minimal exposure path lems

prob-Scope of research

The scope of dissertation is to solve the MEP problem which is a typicaltype of the barrier coverage problems in WSNs The MEP problem is a practicaloptimization problem with high dimension, non-differentiation and non-linearity

To efficient cope with these characteristics, the dissertation mainly investigates thestrength of metaheuristic algorithms Besides, the MEP problem becomes moreuseful when it is considered under some challenging scenarios as follows:

ˆ Mobile wireless sensor networks

ˆ Practical sensing coverage model

ˆ Heterogeneous wireless sensor networks

ˆ Deployment environments with obstacles

This dissertation explores the barrier coverage problems in WSNs, especially,the minimal exposure path problem which is a well-known method for evaluatingthe coverage quality of WSNs, and is ultimately useful for sensor network designers.Our main contributions in this dissertation are as follows:

ˆ Presenting a rigorous theoretical analysis, and devise the formulation of theMEP problem in mobile wireless sensor networks, called MMEP Based on thecharacteristics of the MMEP problem, two efficient meteheuristic algorithms(GAMEP, GPSO-MMEP) are proposed to solve it

ˆ Formulating a minimal exposure path problem under practical sensing age model, i.e the probabilistic coverage model with noise in a WSN, calledPM-based-MEP A new definition of exposure measure for this model is alsointroduced We then propose the GB-MEP algorithm to obtain the solutionbased on the traditional grid-based method incorporated with several improve-ments To enhance the search space and more efficiently solve the problem,

cover-we design a new individual representation, an efficient crossover and a able mutation operator to form a genetic algorithm Conduct experiments invarious scenarios to examine the proposed algorithms Quality of solutionsand computation time are compared with existing methods and analyzed togive insights into the use of each algorithm in the PM-based-MEP

suit-ˆ Establishing mathematical models to represent the MEP problem in neous wireless multimedia sensor networks, called HM-MEP We also proposetwo efficient meta-heuristic algorithms: HEA - a hybrid evolutionary algo-rithm in combination with local search and GPSO - a novel particle swarmoptimization based on the gravity force theory Analysis, evaluate and com-pare the experimental results and show that our proposed algorithms out-perform the previous methods for most cases regarding quality solution andcomputation time

heteroge-ˆ The dissertation investigates a systematic and generic MEP problem underreal-world deployment environment networks with presenting obstacles calledOE-MEP Based upon its characteristics, we devise an elite algorithm namelyFEA for solving the OE-MEP An extension to a custom-made simulationenvironment to integrate a variety of network topologies as well as obstaclesare created Experimental results on numerous instances indicate that theproposed algorithm is suitable for the converted OE-MEP problem and per-forms better in both solution accuracy and computation time than existingapproaches

Dissertation organization

The dissertation is organized as follows:

Chapter 1 presents background knowledge about the barrier coverage lem such as wireless sensor networks, optimal problems, approximate algorithmsespecially heuristics and metaheuristics

prob-Chapter 2 focuses on the MEP problems in omni-directional sensor networks.Chapter 3 highlights the MEP problem in heterogeneous directional sensornetworks

Chapter 4 investigates the MEP problem under real-world deployment ronment sensor networks with presenting obstacles

envi-Finally, the summary and evaluation of the achieved results of the dissertationare included Additionally, the future work is also described briefly

CHAPTER 1

BACKGROUND 1.1 Wireless Sensor Networks

Wireless Sensor Networks have revolutionized the IoTs industry by building

a reliable and efficient communication system With the rapidly growing technology

of sensors, WSNs play an important role in implementing IoTs Recently, wirelesscommunications, computing and sensor technology have enabled the rapid develop-ment of low-cost, small-size sensor nodes that integrate sensing, data processing andwireless communication Although sensor nodes are usually resource limited, such

as limited battery, memory and computation capacities, they can collaborate witheach other to accomplish big tasks efficiently A typical WSN consists of thousands

of sensor nodes deployed in the region of interest, which can be used to monitorphysical phenomena of the ROI

1.1.1 Sensors

A sensor is a device, module, machine, or subsystem whose purpose is todetect events or changes in its environment (such as heat, light, sound, pressure,magnetism, etc.) and send the information to other electronics, frequently a com-puter processor

1.1.2 Sensor nodes

A sensor node is a node in a sensor network that is capable of ing some processing, gathering sensory information and communicating with otherconnected nodes in the network A typical of a sensor node consists of sensor unit,communication unit, micro-controller unit, and memory and power unit

perform-1.1.3 Sensor coverage model

Sensor coverage models are used to reflect the sensing capability and quality

of sensors Commonly, most sensing functions share two aspects in common

ˆ Sensing ability decreases as distance increases

ˆ Due to diminishing effects of noise bursts in measurements, sensing ability canimprove as the allotted sensing time (exposure) increases

Assume sensor si is deployed at point (xi, yi) For any target point at locationl(x, y), the Euclidean distance between siand the target point as:

d(si, l) =

q(xi− x)2+ (yi− y)2 (1.1)

The Boolean disk coverage model

a Boolean disk coverage model

The truncated attenuated coverage model

The coverage measure becomes very small when the distance between a space pointand a sensor becomes very large In such cases, the coverage measure might beignored, and some approximations can be made by truncating the coverage measurefor larger values of distance in Equation 1.3

f (d(si, l)) =

(

Ce−αd(s,z)if d(si, l) ≤ r

where α is a parameter representing the physical characteristics of the sensor, and

r the sensing range Figure 1.1(b) demonstrates the truncated attenuated coveragemodel

Figure 1.1 Illustration of (a) the Boolean disk coverage model in which the redstars are the target points respectively belonging inner and outer the green sensingarea of a sensor, (b) the truncated attenuated coverage model

1.1.4 Sensing field intensity models

The sensing field intensity models specify the collaboration of sensors in thesensing field There are two types of sensor field intensity are usually applied:

The all sensing field intensity

bound-to be a two-dimensional belt region that is bounded by two parallel lines The ROIcan be a closed belt or an open belt region

Penetration path

A penetration path, or a crossing path is a path that connects two the posite sides in the ROI, where the entry point and the exit point reside on twoopposite sides of the region For a two dimensional belt, orthogonal crossing pathsare straight lines, whose length is equal to the width of the belt

op-Static sensor node

A static sensor node has the ability to collect sensed data, send or receivemessages, process data and messages, and do other types of computation in staticWSNs Typically, these sensor nodes do not move once they are deployed.Mobile sensor node

A mobile node not only has all the characteristics of the static sensor nodes,but also has some mobility A mobile sensor node can move act as a router when it

is in a low or even no coverage area, and accomplish the recovery task

1.1.6 Wireless sensor network scenarios

Sensor nodes are deployed in a ROI to monitor some physical phenomena.Such a field is called a sensor field, and the sensor nodes form a sensor network

Diverse scenarios exist in the architecture and management of sensor networks.

ˆ Homogeneous versus heterogeneous wireless sensor networks

ˆ Static versus mobile wireless sensor networks

ˆ Single-Hop versus multi-Hop wireless sensor networks

1.2 Optimization problems

Applications of optimization are everywhere countless Every process has thepotential to be optimized The optimization problems are often classified into severalcategories as follows:

ˆ Continuous optimization: An optimization problem is called a continuousoptimization if the variables which demonstrate the objective function must

be continuous A continuous variable is a variable which is chosen from a set

of real values

ˆ Combinatorial optimization: An optimization problem is called a binatorial optimization if the variables which decides the objective functionmust be discrete

1.3.2 Population-based metaheuristics

Population-based metaheuristics (P-metaheuristic) maintain and improve tiple candidate solutions, often using population characteristics to guide the search.P-metaheuristic algorithms provide a natural, intrinsic way for the exploration ofthe search space They start from an initial population of solutions Then, theyiteratively apply the generation of a new population and the replacement of thecurrent population

mul-1.4 Conclusion

This chapter has presented background knowledge related to the barrier erage problems in wireless sensor networks such as wireless sensor networks, sensors,sensor nodes, the barrier-covering problems in WSNs, and common terms in thebarrier coverage problems In addition, the concepts of optimal problems, approx-imate algorithms for solving the optimal problems, especially the theoretical basis

cov-of heuristic/metaheuristic algorithms are described These concepts will be usedthroughout this dissertation

The MMEP problem can be briefly stated as follow: given N mobile sensors

S = {s1, s2, , sN} are randomly deployed in the ROI < The sensor si has theinitial location si0and the location at time t is denoted as sit Each sensor moves

in a preset trajectory at constant speed Any object penetrating the monitoringregion always starts at the source point and targets to reach the destination point.Time interval for which the object stays inside the sensing field is [0, T ] Speed ofthe object is limited at a maximum value and we assume that the object alwaysmoves at its max speed since exposure will get higher if the object stays longer inthe sensing field The goal of the MMEP problem is to find out a penetration path

℘ from the beginning point B to the ending point E such that the exposure value

of the path ℘ is minimum More precisely, the MMEP problem is formulated asfollows:

Input

ˆ W , L: the width and the length of the ROI <

ˆ N: the number of mobile sensors

ˆ {s1, s2, , sN}: set of mobile sensor nodes deployed in <

ˆ si0: the initial location of sensor

ˆ Ri: trajectory of sensor si

ˆ vs: the speed of sensors, all sensor nodes have the same speed

ˆ vI: the max speed of the intruder

ˆ (0, yB): the coordinates of the source point B

ˆ (L, yE) : the coordinates of the destination point E

Output:

A path ℘ in < connects from the source point B to the destination point E