Towards Optimal Sleep Scheduling in Sensor Networks for Rare-Event Detection

1
Towards Optimal Sleep Scheduling in Sensor
Networks for Rare-Event Detection

• Information Processing in Sensor Networks, 2005.
  IPSN 2005. Fourth
• speaker Chang Lung Huang

2
Introduction

• Sensor networks energy supply is one fundamental
  bottleneck
• Large number of sensing nodes
• Unfriendly environment in which these nodes
• Redundancy nodes go sleeping
• Trade-off between energy savings and coverage

3
Delay

• Detection delay
• Each point in the environment is sensed within
  some finite interval of time
• Packet delivery latency
• The latency of packet delivery to a base station

event
Wake up
Time line
4
Duty Cycle

• Duty-cycle Td
• Each node in the primary subset sleeps for Td
  then wakes up for Ton
• TonTwTsTp
• Tw Warm up time
• Ts Sample time
• Tp Processing time
• Especially interested in very low duty-cycle
• operation Td ltlt Ton

5
First Level (Primary Subset)

• Using two-level sleep scheduling
• First level
• Selects a minimal subset of all nodes sensing
  coverage is maintained using the fewest primary
  nodes
• This process is repeated periodically at a fairly
  large period (e.g.,of the order of tens of hours)

6
Second level

• The second level savings can be made arbitrarily
  large by decreasing the duty cycle of primary
  nodes
• Random
• Synchronized
• Near-optimal

7
Average Detection Delay

• If the number of primary nodes within a sensory
  radius is a
• A point is sensed on average no more than once
  every Td/a time units
• Average detection delay no lower than
• Td/2a

Duty cycle
8
State Transition of Optimization Algorithm
Our overall algorithm for minimizing detection
delay is a three stage transition process.
9
Quadratic expression of detection delay

• Given inter-sample separations x1,,xn.
• The average detection delay of xj is xj/2
• Delay time
• Equal to

S
10
Find the time t to lower delay time

• To determine the wakeup time of node O ,The
  average of point A is

11
POLYNOMIAL FUNCTION TABLE
12
Finding the min value

• We obtain a continuous piecewise quadratic
  equation.
• The optimality curve
• Point A shown in Figure 2
• is given in Figure 3.

13
Min value of weighted
14
End-to-End Delay Optimization

• We need some nodes synchronize awake to deliver
  the packet.
• We consider communication range is relatively
  large than sensory range.
• After the 2-level scheduling, make a streamlined
  sequence to pipe data.

15
End-to-End Delay Optimization

• 1. Floods the network with hop
  count
• 2 Run the detection delay minimization
  algorithm.
• 3 Finds one neighbor with the closest wakeup
  time of its own
• 4 Overlaps its wakeup interval with that
  neighbors.

16
Simulation

• a

17
Simulation

• a