The Chinese Univ. of Hong Kong Energy-Conserving Coverage
Configuration for Dependable Wireless Sensor Networks Chen Xinyu
Term Presentation 2004-12-14
Slide 2
Dept. of Computer Science and Engineering Outline Motivation
Coverage configuration with Boolean sensing model Coverage
configuration with general sensing model Performance evaluations
with ns-2 Conclusions and future work
Slide 3
Dept. of Computer Science and Engineering Wireless Sensor
Networks Composed of a large number of sensor nodes Sensors
communicate with each other through short-range radio transmission
Sensors react to environmental events and relay collected data
through the dynamically formed network
Slide 4
Dept. of Computer Science and Engineering Applications Military
reconnaissance Physical security Environment monitoring Traffic
surveillance Industrial and manufacturing automation Distributed
robotics
Slide 5
Dept. of Computer Science and Engineering Requirements
Maintaining coverage Every point in the region of interest should
be sensed within given parameters Extending system lifetime The
energy source is usually battery power Battery recharging or
replacement is undesirable or impossible due to the unattended
nature of sensors and hostile sensing environments
Slide 6
Dept. of Computer Science and Engineering Requirements (contd)
Fault tolerance Sensors may fail or be blocked due to physical
damage or environmental interference Scalability High density of
deployed nodes Each sensor must configure its own operational mode
adaptively based on local information, not on global
information
Slide 7
Dept. of Computer Science and Engineering Approach: Coverage
Configuration Coverage configuration is a promising way to extend
network lifetime by alternately activating only a subset of sensors
and scheduling others to sleep according to some heuristic schemes
while providing sufficient coverage in a geographic region
Slide 8
Dept. of Computer Science and Engineering Concerns A good
coverage-preserved and fault-tolerant sensor configuration protocol
should have the following characteristics: It should allow as many
nodes as possible to turn their radio transceivers and sensing
functionalities off to reduce energy consumption, thus extending
network lifetime Enough nodes must stay awake to form a connected
network backbone and to preserve area coverage Void areas produced
by sensor failures and energy depletions should be recovered as
soon as possible
Slide 9
Dept. of Computer Science and Engineering Two Sensing Models
Boolean sensing model (BSM) Each sensor has a certain sensing
range, and can only detect the occurrences of events within its
sensing range General sensing model (GSM) Capture the fact that
signals emitted by a target of interest decay over the distance of
propagation Exploit the collaboration between adjacent sensors
Slide 10
Dept. of Computer Science and Engineering Problem Formulation
for the BSM Each sensor node N i knows its location (x i, y i ),
sensing radius r i, communication radius R Sensors are deployed in
a two-dimensional Euclidean plane Responsible Sensing Region (RSR)
i = { p | d(N i,p) < r i } A point is covered by a sensor node
when this point is in the sensor's RSR The one-hop neighbor set of
N i N(i) = { N j | d(N i, N j ) R, j i }
Slide 11
Dept. of Computer Science and Engineering Some Definitions NiNi
NjNj Sponsored Sensing Arc (SSA) ij Sponsored Sensing Region (SSR)
Sponsored Sensing Angle (SSG) ij Covered Sensing Angle (CSG)
ij
Slide 12
Dept. of Computer Science and Engineering Special Cases of SSR
and SSA d(N i, N j ) r i + r j NiNi NjNj
Slide 13
Dept. of Computer Science and Engineering Special Cases of SSR
and SSA d(N i, N j ) r i r j NiNi NjNj SSG ij =2 CSG ij is not
defined Completely Covered Node (CCN) of N i
Slide 14
Dept. of Computer Science and Engineering Special Cases of SSR
and SSA d(N i, N j ) r j - r i NiNi NjNj Complete-Coverage Sponsor
(CCS) of N i Degree of Complete Coverage DCC i = | CCS(i) | SSG ij
is not defined CSG ij =2 CCS(i)
Slide 15
Dept. of Computer Science and Engineering Minimum Partial
Arc-Coverage (MPAC) The minimum partial arc-coverage (MPAC)
sponsored by node N j to node N i, denoted as ij, The number of N i
's non-CCSs covering the point on the SSA ij that has the fewest
nodes covering it.
Slide 16
Dept. of Computer Science and Engineering Derivation of MPAC ij
0 22 ij jl jm ij = 2 ij = 1 Covered Sensing Angle (CSG) Sponsored
Sensing Angle (SSG) ij
Slide 17
Dept. of Computer Science and Engineering MPAC and DCC Based
k-Coverage Sleeping Candidate Condition K-coverage Every point in
the deployed area is covered by at least k nodes Theorem A sensor
node N i is a sleeping candidate while preserving k-coverage, iff i
k or N j N(i) - CCS(i), ij > k - i.
Slide 18
Dept. of Computer Science and Engineering Extended Sleeping
Candidate Condition Constrained deployed area
Slide 19
Dept. of Computer Science and Engineering Node Scheduling
Protocols Round-based Divide the time into rounds Approximately
synchronized In each round, every live sensor is given a chance to
be sleeping eligible Adaptive sleeping Let each node calculate its
sleeping time locally and adaptively
Slide 20
Dept. of Computer Science and Engineering Round-Based Node
Scheduling Protocol on sleeping ready-to- sleeping ready-to-on
uncertain T round eligible / STATUS ineligible T round T wait
eligible / STATUS ineligible / STATUS on-sleeping decision phase
1.Set a backoff timer T hello, a window timer T win, a wait timer T
wait, and a round timer T round 2.Collect HELLO messages from
neighbors 3.After T hello times out, broadcast a HELLO message to
all neighbors 4.After T win expires, evaluate the sleeping
eligibility according to sleeping candidate conditions
Slide 21
Dept. of Computer Science and Engineering An Example of
Sleeping Eligibility Evaluation
Slide 22
Dept. of Computer Science and Engineering Connectivity
Requirement Considering only the coverage issue may produce
disconnected subnetworks Simple connectivity preservation If a
sensor is sleeping eligible, evaluating whether its one-hop
neighbors will remain connected through each other when the
considered sensor is removed
Slide 23
Dept. of Computer Science and Engineering Adaptive Sleeping
Node Scheduling Protocol A node may suffer failures or deplete its
energy loss of area coverage Round-based: timer T round is a global
parameter and not adaptive to recover a local area loss Letting
each node calculate its sleeping time locally and adaptively
Slide 24
Dept. of Computer Science and Engineering Adaptive Sleeping
Node Scheduling Protocol 1.Set a timer T sleeping 2.When T sleeping
times out, broadcast a PROBE message 3.Each neighbor receiving the
PROBE message will return a STATUS message to the sender 4.Evaluate
sleeping eligibility. If eligible, set T sleeping according to the
energy information collected from neighbors
Slide 25
Dept. of Computer Science and Engineering Discussions for the
BSM Each sensor has a deterministic sensing radius Allow a
geometric treatment of the coverage problem Miss the attenuation
behavior of signals Ignore the collaboration between adjacent
sensors in performing area sensing and monitoring
Slide 26
Dept. of Computer Science and Engineering Problem Formulation
for the GSM The sensibility of a sensor N i for an event occurring
at an arbitrary measuring point p is defined by : the energy
emitted by events occurring at point p : the decaying factor of the
sensing signal
Slide 27
Dept. of Computer Science and Engineering All-Sensor Field
Sensibility (ASFS) Suppose we have a background distribution of n
sensors, denoted by N 1, N 2, , N n, in a deployment region A
All-Sensor Field Sensibility for point p With a sensibility
threshold , the point p is covered if S a (p)
Slide 28
Dept. of Computer Science and Engineering Discussions for the
ASFS Need a sink working as a data fusion center Produce a heavy
network load in multi- hop sensor networks Pose a single point of
failures
Slide 29
Dept. of Computer Science and Engineering Neighboring-Sensor
Field Sensibility (NSFS) Treat each sensor as a sensing fusion
center Each sensor broadcasts its perceived field sensibility Each
sensor collects its one-hop neighbors messages Transform the
original global coverage decision problem into a local problem
Slide 30
Dept. of Computer Science and Engineering Responsible Sensing
Region Voronoi diagram Partition the deployed region into a set of
convex polygons such that all points inside a polygon are closet to
only one particular node The polygon in which sensor N i resides is
its Responsible Sensing Region i If an event occurs in i, sensor N
i will receive the strongest signal Open RSR and closed RSR
Slide 31
Dept. of Computer Science and Engineering NSFS-Based
Pessimistic Sleeping Candidate Condition
Slide 32
Dept. of Computer Science and Engineering NSFS-Based Optimistic
Sleeping Candidate Condition
Slide 33
Dept. of Computer Science and Engineering Sensibility-Based
Sleeping Configuration Protocol (SSCP) on sleeping ready-to-
sleeping ready-to-on T round eligible / STATUS ineligible T round T
wait eligible / STATUS ineligible / STATUS uncertain II uncertain
I
Slide 34
Dept. of Computer Science and Engineering Performance
Evaluation with ns-2 ESS: extended sponsored sector Proposed by
Tian et. al. of Univ. of Ottawa, 2002 Consider only the nodes
inside the RSR of the evaluated node Mpac: round-based protocol
with elementary MPAC condition MpacB: round-based protocol with
extended MPAC condition in constrained area MpacBAs: adaptive
sleeping protocol with MpacB SscpP: Sscp with the pessimistic
sleeping condition SscpO: Sscp with the optimistic sleeping
condition
Slide 35
Dept. of Computer Science and Engineering Bridge between BSM
and GSM Ensured-sensibility radius
Slide 36
Dept. of Computer Science and Engineering Default Parameters
Setting The deployed area is 50m x 50m = 1, = 3, = 0.001 (r = 10m)
R = 12 m The number of deployed sensor: 120 Power Consumption: Tx
(transmit) = 1.4W, Rx (receive) = 1W, Idle = 0.83W, Sleeping =
0.13W
Slide 37
Dept. of Computer Science and Engineering Performance
Evaluation (1) Sleeping sensor vs. communication radius
Slide 38
Dept. of Computer Science and Engineering Performance
Evaluation (2) Network topology
Slide 39
Dept. of Computer Science and Engineering Performance
Evaluation (3) Sleeping sensor vs. sensor number
Slide 40
Dept. of Computer Science and Engineering Performance
Evaluation (4) Sleeping sensor vs. sensibility threshold
Slide 41
Dept. of Computer Science and Engineering Performance
Evaluation (5) Network lifetime vs. live sensor when the MTBF is
800s, R is 12m
Slide 42
Dept. of Computer Science and Engineering Performance
Evaluation (6) -coverage accumulated time The total time during
which or more percentage of the deployed area satisfies the
coverage requirement
Slide 43
Dept. of Computer Science and Engineering Approaches to Build
Dependable Wireless Sensor Networks Decreasing the communication
radius or increasing the coverage degree is equivalent in providing
fault tolerance Detecting sensor failures and recovering the area
loss as quick as possible: adaptive sleeping configuration
Exploiting the cooperation between neighboring sensors: general
sensing model
Slide 44
Dept. of Computer Science and Engineering Conclusions Develop
MPAC-based node sleeping eligibility conditions for the BSM achieve
k-coverage degree can be applied with different sensing radii
Develop SSCPs for the GSM exploit the cooperation between adjacent
sensors Suggest three effective approaches to build dependable
sensor networks
Slide 45
Dept. of Computer Science and Engineering Future Work Exploit
algorithms to identify node redundancy without location information
Study the network behavior with node failures Build dependable
sensor networks both on area coverage and network connectivity
Slide 46
Dept. of Computer Science and Engineering Performance
Evaluation (5) Distribution of coverage degree
Slide 47
Dept. of Computer Science and Engineering Performance
Evaluation (4) MpacBCa sleeping sensor