Mobile Filtering for Error-Bounded Data Collection in Sensor Networks

1
Mobile Filtering for Error-Bounded Data
Collection in Sensor Networks

• Dan Wang
• Hong Kong Polytechnic Univ.
• Jianliang Xu
• Hong Kong Baptist Univ.
• Jiangchuan Liu, Feng Wang
• Simon Fraser University

Int'l Conference on Distributed Computing Systems
(ICDCS 2008)
2
Outline

• Previews
• Introduction
• Mobile Filtering Problem
• System Model
• Mobile Filtering Algorithms
• Chain Topology
• Cross Chain Topology (Multi-Chain Trees)
• General Trees
• Simulations
• Conclusion

3
Previews

• Precision Allocation Problem
• X. Tang and J. Xu, Extending Network Lifetime
  for Precision-Constrained Data Aggregation in
  Wireless Sensor Networks (INFOCOM06) .
• Optimizing Lifetime for Continuous
  DataAggregation With Precision Guarantees in
  Wireless Sensor Networks. (IEEE Trans. on
  Networking)

PowerPoint
EZLMS
4
Data Collection with Precision-Guarantee
Total Error bound 3 oC
Error bound of each sensor 0.5 oC
Approximate 20.15 Real 20.16
20
20
20.3
20
20
18
22
18
22
20
22
19
20
20
22
18
20
20-gt20.2
22
18-gt19
20-gt19.8
20
22
18
20
Time t
Time t1
AVERAGE
5
Precision Allocation-- Stationary

• The problem is
• How to allocate user-specified precision among
  sensor nodes such that the network lifetime is
  maximized?

Given the error bound E
Objective
Maximize network lifetime
Subject to
e1
e2
e3
e4
e5
e6
6
Mobile Filtering Problem
7
System Model

• No data Aggregation
• -- target at collecting raw data.
• Error measurement

8
System Model
1.5
3
2
2, 3
3
s1
s0
s3
s2
3
1.5
1.5
9
Network Model
(c) cross multi-chain topology
(a) chain topology
(b) multi-chain topology
10
Operations of a Sensor Node

Listening state
Processing state
Sleeping state
11
(No Transcript)
12
Optimal Solution for Chain Topology--- Dynamic
Programming

• Assume all data changes are known a priori.
• Offline version

13
(No Transcript)
14
Greedy Online Heuristic

• Thresholds
• TR decide if the filter is sent or not.
• TS decide if a data change is sent or not.

15
Multi-chain Topology
Given the error bound E

1
2
3
n
e1
e2
en-1
en
Objective
Maximize network lifetime
Subject to
e12
e8
e4
16
Cross Multi-chain Topology-- General Tree
17
Simulations

• Setup
• Two data sets
• Synthetic data trace
• Real data trace (dew-point), LEM
• Three topologies
• Chain topology
• Cross multi-chain topology
• Grid topology

18
Chain Topology
19
Cross Multi-chain Topology
20
Filter Re-allocation Period
21
Precision-- Grid Topology
22
Conclusion

• This paper proposed a novel mobile filtering
  scheme for error-bounded non-aggregate data
  collection in sensor networks.
• By exploring the migration of filters, a mobile
  filter extracts and relays unused filters in the
  network to suppress as many data update reports
  as possible.