Title: CS 851 Presentation: Differentiated Surveillance for Sensor Network
1CS 851 PresentationDifferentiated Surveillance
for Sensor Network
- Presented by Liqian Luo
- Reference
- 1. T. Yan, T. He, and J. A. Stankovic,
Differentiated Surveillance for Sensor
networks, First ACM Conference on Embedded
Networked Sensor Systems (SenSys 2003), Los
Angeles, CA 2003
2Assessment of the Paper
- Pros
- The first algorithm to guarantee different
degrees of coverage for different requirements - Good performance in power conservation and
balancing - Cons
- Pessimistic degree of coverage estimation
- Lack of flexibility
- Require clock synchronization Do not support
mobility work/sleep schedule never changes after
decided Expensive fault tolerance
3Outline
- Problem Statement
- Differentiated Surveillance solution
- Introduction
- Design goals
- Assumptions
- Basic design without differentiation
- Enhanced design with differentiation
- Extensions and Optimizations
- Related Work
- Evaluation
- Conclusion and Discussion
4Roadmap
- Problem Statement
- Differentiated Surveillance solution
- Introduction
- Design goals
- Assumptions
- Basic design without differentiation
- Enhanced design with differentiation
- Extensions and Optimizations
- Related Work
- Evaluation
- Conclusion and Discussion
5Problem statement
- How to provide sensing coverage for a sensor
network in a power-efficient way?
6Problem statement Sensing Coverage
7Problem statement Sensing Coverage
8Problem statement Sensing Coverage
9Problem statement Degree of Sensing Coverage
- Current solutions regard the sensing coverage to
a certain geographic area as a binary. - This paper argues that higher degree of sensing
coverage is desired to obtain high detection
confidence since individual nodes are not
reliable.
10Roadmap
- Problem Statement
- Differentiated Surveillance solution
- Introduction
- Design goals
- Assumptions
- Basic design without differentiation
- Enhanced design with differentiation
- Extensions and Optimizations
- Related Work
- Evaluation
- Conclusion and Discussion
11Differentiated surveillance solution
Introduction
- Degree of coverage (DOC)
- Differentiated surveillance
- Providing different degrees of sensing coverage
for a sensor network according to different
requirements
12Differentiated surveillance solution
Introduction
DOC 1
DOC 2
13Roadmap
- Problem Statement
- Differentiated Surveillance solution
- Introduction
- Design goals
- Assumptions
- Basic design without differentiation
- Enhanced design with differentiation
- Extensions and Optimizations
- Related Work
- Evaluation
- Conclusion and Discussion
14Differentiated surveillance solution Design
Goals
- Provide energy efficient sensing coverage for a
geographic area covered by sensor nodes - extend system life
- Reduce total energy consumption
- Reduce energy consumption variation among nodes
- provide differentiated surveillance
15Roadmap
- Problem Statement
- Differentiated Surveillance solution
- Introduction
- Design goals
- Assumptions
- Basic design without differentiation
- Enhanced design with differentiation
- Extensions and Optimizations
- Related Work
- Evaluation
- Conclusion and Discussion
16Differentiated surveillance solution Assumptions
- Each node knows its own location and nodes are
not moving. - Neighboring nodes are roughly time synchronized.
- The sensing area of a node is a circle with
radius r centered at the location of this node. - Radio radius is larger than 2r
17Roadmap
- Problem Statement
- Differentiated Surveillance solution
- Introduction
- Design goals
- Assumptions
- Basic design without differentiation
- Enhanced design with differentiation
- Extensions and Optimizations
- Related Work
- Evaluation
- Conclusion and Discussion
18Basic design without differentiation Goal
- Goal find a work-sleep schedule for each node
which achieves 100 Sensing coverage guarantee. - Ideally we should consider each point in the area
when do scheduling, but it is impossible because
the number of points is infinite. What can we do?
19Basic design without differentiation 100
sensing coverage
- Solution 100 Grid point sensing coverage
- Divide whole network into grids
- For each grid point x, guarantee that x is
covered by at least one nodes sensing range at
ANY time
20Basic design without differentiation 100
sensing coverage
- 100 Grid point sensing coverage 100 sensing
coverage guarantee? No.
21Basic design without differentiation 100
sensing coverage
- Solution Conservative sensing radius (Rc)
- Rc r d/
- For each grid point x, guarantee that x is
covered by at least one nodes conservative
sensing range at ANY time.
d
22Basic design without differentiation - decide
working schedule
If we want to provide sensing coverage for point
x, we can have either A or B or C awaken.
A scheduling example of A, B and C
100
0
30
70
Node A
10
60
Node B
5
45
Node C
time
Waking
Sleeping
23Basic design without differentiation decide
working schedule
- Challenge For each node, how to coordinate with
other nodes and decide its own schedule? - Solution - Random Reference Point Scheduling
Algorithm
24Basic design without differentiation decide
working schedule
- Concepts
- Initialization Phase
- In this phase, nodes find their own positions,
synchronize time with neighboring nodes and
decide their own working schedule. - Sensing Phase
- Nodes enter this phase after initialization phase
and choose to sense or sleep according to their
schedules. - Sensing Round - T
- Sensing phase is divided into sensing rounds with
equal duration T. A node has the same schedule
for each round.
25Basic design without differentiation decide
working schedule
- Concepts
- A nodes working schedule is determined by Four
parameter tuple (T, Ref, Tfront, Tend) - Ref a random time reference point chosen by a
node within 0, T) - Tfront the duration of time prior to Ref
- Tend the duration of time after Ref.
- By this tuple, A nodes working period is
determined as follows - Tj Ref Tfront , Tj Ref Tend)
- And all the other time the node is sleeping.
26Basic design without differentiation decide
working schedule
- Solution Random Reference Point Scheduling
Algorithm - 1) Each node N chooses a Reference Point (Ref)
randomly from 0, T) and broadcasts its Ref and
position. - e.g. T 100, RefA 40, RefB 90, RefC 20
- 2) For each grid point P in its own sensing area,
N sorts all the Refs from nodes (including N)
which can also sense P in ascending order. - For A according to point P1, we have
- Ref(1) RefC 20, Ref(2) RefA 40, Ref(3)
RefB 90
0
27Basic design without differentiation decide
working schedule
- 3) Assuming RefN is the (i)th Ref, Ns four
parameter tuple is computed as follows - TfronN (Ref(i)- Ref(i-1))/2, 1ltiltM
- TendN (Ref(i1)-Ref(i))/2, 1ltiltM
- TfrontA (Ref(2)-Ref(1))/2 (40-20)/2 10
- TendA (Ref(3)-Ref(2))/2 (90-40)/2 25
- (T, RefA, TfrontA, TendA) (100, 40, 10, 25)
- 4) Ns working period for point P (TwN(P)) is
decided by - Tj RefN TfrontN , Tj RefN TendN), j
0, 1, 2, - TwA(P1) 100j4010, 100j4025) 100j30,
100j65)
28Basic design without differentiation decide
working schedule
- 5) Calculate the union of TwN(Px) for all grid
points within Ns sensing area, choose this union
as the final working period of N (TwN).
29Roadmap
- Problem Statement
- Differentiated Surveillance solution
- Introduction
- Design goals
- Assumptions
- Basic design without differentiation
- Enhanced design with differentiation
- Extensions and Optimizations
- Related Work
- Evaluation
- Conclusion and Discussion
30Enhanced design with differentiation
- Provide different DOC according to different
requirements
31Enhanced design with differentiation
- Goal
- provide sensing coverage with DOC a
- Solution
- Extend 4-parameter tuple to 5-parameter tuple (T,
Ref, Tfront, Tend, a) - Determine a nodes working period as follows
- Tj Ref Tfronta , Tj Ref Tenda)
32Enhanced design with differentiation An example
Schedule for Grid Point P1 (a1)
(T, RefA, TfrontA, TendA) (100, 40, 10, 25) (T,
RefB, TfrontB, TendB) (100, 90, 25, 15) (T,
RefC, TfrontC, TendC) (100, 20, 15, 10) TwA
Tj Ref Tfront , Tj Ref Tend)
100j 30, 100j 65) TwB 100j 65, 100j
105) TwC 100j 5, 100j 30)
33Enhanced design with differentiation An example
Question - Can the algorithm guarantee 100
DOCgt2 sensing coverage by setting a2? Answer -
Yes
Schedule for Grid Point P1 (a2)
(T, RefA, TfrontA, TendA, a) (100, 40, 10, 25,
2) (T, RefB, TfrontB, TendB, a) (100, 90, 25,
15, 2) (T, RefC, TfrontC, TendC, a) (100, 20,
15, 10, 2) TwA Tj Ref Tfront2,Tj Ref
Tend2) 100j 20, 100j 90) TwB
100j 40, 100j 120) TwC 100j -10,
100j 40)
0
A
B
C
30
65
5
34Enhanced design with differentiation An example
Question - Can the algorithm guarantee 100
DOCgt3 sensing coverage by setting a3? Answer -
No
Schedule for Grid Point P1 (a3)
(T, RefA, TfrontA, TendA, a) (100, 40, 10, 25,
3) (T, RefB, TfrontB, TendB, a) (100, 90, 25,
15, 3) (T, RefC, TfrontC, TendC, a) (100, 20,
15, 10, 3) TwA Tj Ref Tfront3,Tj Ref
Tend3) 100j 10, 100j 115)
T TwB 100j 15, 100j 135) T TwC
100j -25, 100j 50)
0
A
B
C
30
65
5
35Enhanced design with differentiation An
extension to guarantee 100 DOCgta
- My Extension to guarantee 100 DOCgta sensing
coverage - Separate the time line into segments by using
Refs and the middle points between Refs - Instead of expanding Tw by its own Tfront or
Tend, expand one segment on both sides when a is
increased by 1.
0
A
B
C
30
65
5
36Roadmap
- Problem Statement
- Differentiated Surveillance solution
- Introduction
- Design goals
- Assumptions
- Basic design without differentiation
- Enhanced design with differentiation
- Extensions and Optimizations
- Related Work
- Evaluation
- Conclusion and Discussion
37Optimizations and Extensions Second Pass
Optimization
A
- Existing Problem
- Taking the union of Tw for all grid points within
sensing range as final Tw will be more than
efficient to provide coverage guarantee - Solution
- make a second pass optimization to reduce the
redundancy
1
B
2
TwA(1)
TwA
TwB(1)
TwB(2)
TwB
38Optimizations and Extensions Second Pass
Optimization
A
- Second Pass Optimization
- 1)After getting the final Tw, each node sends it
to neighbors within the distance of 2r - 2)Within 2r neighbors that have not recalculated
their Tw, the one with the longest Tw
recalculates its Tw and sends it to 2r neighbors - 3) Repeat 2) until everyone has recalculated its
Tw.
1
B
2
TwA(1)
TwA
TwB(1)
TwB(2)
TwB
39Optimizations and Extensions Multi-Round
Extension for Energy Balance
- Existing Problem
- Reference points are selected randomly instead of
uniformly, which results in big variation in Tw
among nodes and big variation in power
consumption. - Solution
- Multi-Round Extension
refC
refB
refA
TwA
TwB
TwC
40Optimizations and Extensions Multi-Round
Extension for Energy Balance
- Multi-Round Extension
- Instead of calculating a single schedule,
calculate M schedules according to M
independently selected random Refs for each node. - At each round T in sensing phase, the nodes
choose one schedule consecutively.
TwA1
TwA1
TwA2
TwA3
TwA2
TwA3
41Roadmap
- Problem Statement
- Differentiated Surveillance solution
- Introduction
- Design goals
- Assumptions
- Basic design without differentiation
- Enhanced design with differentiation
- Extensions and Optimizations
- Related Work
- Evaluation
- Conclusion and Discussion
42Related Work Communication Coverage
- SPAN, ASCENT providing a communication coverage
within an energy conservation context
43Related Work Sensing Coverage 1
- Energy Efficient Robust Sensing Coverage a
probing-based mechanism - After a sleeping node wakes up, use a probing
message to see whether there is another node
working within its sensing area. If no, it takes
the responsibility of sensing until it dies. - Drawbacks
- Overestimate neighbors contribution, so no
guarantee on sensing coverage
44Related Work Sensing Coverage 2
- A Node Scheduling Scheme for Energy Conservation
sponsored coverage scheme - At the beginning of each round, each node
advertises its position to neighbors - After receiving neighbors position advertises,
each node calculates its eligibility for going to
sleep. Here, a back-off scheme is used to avoid
simultaneous actions of multiple nodes.
45Related Work Sensing Coverage 2
- Drawbacks
- Require broadcasting at the beginning of each
round - Underestimate the area that the neighbor nodes
can cover
46Roadmap
- Problem Statement
- Differentiated Surveillance solution
- Introduction
- Design goals
- Assumptions
- Basic design without differentiation
- Enhanced design with differentiation
- Extensions and Optimizations
- Related Work
- Evaluation
- Conclusion and Discussion
47Evaluation - Introduction
- Nodes are distributed with a uniform random
distribution in a 160 X 160 rectangle - Guarantee sensing coverage in the inner 140 X 140
rectangle to eliminate the edge effect - sensing radius 10, communication radius 25
48Evaluation 1 Energy Conservation
Total Energy Consumption per Unit of
Time Sponsored Coverage Basic Design 2nd
Pass Optimization
49Evaluation 1 Energy Conservation
?
Single Node Energy Consumption Standard
Deviation Sponsored Coverage Basic
Design Multiple Round Extension
50Evaluation 1 Energy Conservation
Half-life of the network Sponsored
Coverage Basic Design 2nd Pass Optimization
51Evaluation 2 Sensing Coverage
Actual Degree of Coverage for Differentiated
Surveillance
52Roadmap
- Problem Statement
- Differentiated Surveillance solution
- Introduction
- Design goals
- Assumptions
- Basic design without differentiation
- Enhanced design with differentiation
- Extensions and Optimizations
- Related Work
- Evaluation
- Conclusion and Discussion
53Conclusion and Discussion
- Conclusion
- Novelty - guarantee not only full sensing
coverage to a certain geographic area, but also
sensing coverage with specific degree of
coverage. - Scalability - localized distributed algorithm
- Power management - Good job in energy
conservation and balancing - Robustness - fixed schedule throughout the life
time, expensive fault tolerant extension, can not
work without clock synchronization, can not
support mobility
?
54Conclusion and Discussion
- Discussion 1
- This solution can not guarantee certain degree of
coverage more than 2. - Discussion 2
- Each node chooses its Ref randomly. What if
multiple neighbors have the same Refs? - A simple solution is to order the same Refs by
node ID.
55Conclusion and Discussion
- Discussion 3
- In initialization phase, each node should send
out Ref broadcast and should receive all Refs
from 2r neighbors. It is very hard in high
density sensor network. So there must be some
nodes which are ignored and have not attended the
scheduling algorithm in initialization phase. - An extension, which allows these nodes to attend
the scheduling later, is necessary.
56Conclusion and Discussion
- Discussion 4
- Each node decides its working schedule only based
on sensing coverage. Some other layer protocols
or applications may need a different working
schedule. How to integrate with other working
schedule will be a big problem.
57Conclusion and Discussion
- Discussion 5
- The baseline - Sponsored coverage scheme can
provide fault-tolerance and support certain
mobility since it updates neighbor hood
information every round - DS without the expensive fault tolerance scheme
can not provide fault-tolerance at all - So it is unfair to compare the power consumption
between DS without fault-tolerance and the
baseline with fault-tolerance.
58Thanks!