Unified Power Management in Wireless Sensor Networks

1
Unified Power Management in Wireless Sensor
Networks
Guoliang Xing Department of Computer Science and
Engineering Washington University in St.
Louis Advisor Prof. Chenyang Lu
2
Wireless Sensor Networks
Structural monitoring (Golden gate bridge, SF)
Industrial automation (Intel fabrication plant)
Healthcare (Boston Medical Center)

• Challenges
• Limited power supplies batteries, small solar
  panels
• Long lifetime requirements months to tens of
  years
• Must minimize network power consumption

3
Limitations of Existing Power Management
Approaches

• Separate treatment of sensing and communication
• Comm. power mngmt sleep scheduling, power
  control
• Sensing power mngmt coverage maintenance
• Power minimization of individual radio states
• Sleep scheduling only reduces idle listening
  power
• Power control only reduces transmission power
• Unrealistic communication and sensing models
• E.g., deterministic sensing/communication range

4
Unified Power Management

• Minimum power configuration
• Minimizes total power consumed by all radio
  states
• Relationship btw sensing and communication
• Enables integrated power management
• Two realistic power management protocols
• Coverage maintenance for target detection
• Configurable topology control for lossy WSNs

• Minimum power configuration
• Minimizes total power consumed by all radio
  states
• Relationship btw sensing and communication
• Enables integrated power management
• Two realistic power management protocols
• Coverage maintenance for target detection
• Configurable topology control for lossy WSNs

5
Understanding Radio Power Cost
Power consumption of CC1000 Radio in different
states

• Sleeping consumes much less power than idle
  listening
• Motivate sleep scheduling Polastre et al. 04, Ye
  et al. 04
• Transmission consumes most power
• Motivate transmission power control Singh et al.
  98,Li et al. 01,Li and Hou 03
• None of existing schemes minimizes the total
  energy consumption in all radio states

6
Minimizing Total Radio Energy Consumption an
Example

• a sends to c at normalized rate of r
  Data Rate / Band Width
• Source and relay nodes remain active
• Configuration 1 (topology control)
• a ? b ? c
• Configuration 2 (Sleep scheduling)
• a ?c, b sleeps

c
b
a
7
Average Power Consumption
c

• Configuration 1 a ? b ? c

as avg. power
cs avg. power
bs avg. power
b
rx
a
idle
activity of bs radio
time
tx

• Configuration 2 a ? c, b sleeps

8
Power Control vs. Sleep Scheduling
Transmission power dominates use low
transmission power
Power Consumption
3Pidle
2PidlePsleep
1
r0
Idle power dominates use high transmission power
since more nodes can sleep
9
Minimum Power Configuration (MPC)

• a transmits to c at rate r, b sleeps

ignore sleeping power
group rate related terms
Ca,cPtx(a,c)Prx-2Pidle
edge (a,c) has a cost of Ca,c per unit of data
each active node has a cost of Pidle
Pidle
r Ca,c
c
Pidle

• Assumption
• Total workload lt bandwidth

a
b
rate r
10
Minimum Power Configuration (MPC)

• Given traffic demands I( si , ti , ri ) and
  G(V,E), find a sub-graph G(V, E) minimizing

sum of edge cost from si to ti in G
independent of data rate!
node cost

• Sleep scheduling

• Sleep scheduling
• Power control

• Sleep scheduling
• Power control
• MPC is NP-Hard

11
Solutions

• Matching Based Algorithm (MBA) has an approx.
  ratio of lgk, k is num of sources Meyerson et
  al. 00
• Only works for one-sink case of MPC
• Distributed implementation is expensive
• Cannot handle dynamic data flows
• Developed two new distributed and online
  algorithms
• Incremental Shortest-path Tree Heuristic (ISTH)
• Known approx. ratio is k
• Similar average-case performance as MBA
• Adapt to dynamic network workloads
• Minimum Steiner Tree Heuristic
• Approx. ratio is 1.5(PrxPtx-Pidle)/Pidle ( 5
  on Mica2 motes)

12
Incremental Shortest-path Tree Heuristic (ISTH)

• Initially, all nodes are labeled as asleep
• For each traffic demand (si, ti, ri)
• Find the shortest path from si to ti under cost
  functions He(u,v, ri) and Hn(u)
• Label all nodes on the found path as active

Edge cost
Node cost
13
Illustration of ISTH

• Cu,v2, Pidle1
• Find a new path in each iteration

sink
0
1
2
cost reduction!
2
1
1
1
0
2
2
New cost
2
1
1
1 3 0.2 2 23.8
2
1
2
2
Source 2 r2 0.2
1
1 0.2 2 21.8
0
1
Source 1 r1 0.2
14
Minimum Power Configuration Protocol (MPCP)

• Routing
• Extends DSDV with routing metrics He and Hn
• Sleep scheduling
• Turns on radio if on a route, runs a duty cycle
  otherwise
• Power control
• Determines transmission power to different
  neighbors
• Cross-layer coordination
• Data rate, transmission power, and state of the
  node (active/asleep) jointly determines the
  routing cost

15
Simulation Environment

• Prowler simulator extended by Rmase project
• Prowler http//www.isis.vanderbilt.edu/projects/n
  est/prowler/
• Rmase http//www2.parc.com/spl/projects/era/nest/
  Rmase/
• Implemented USC model Zuniga et al. 04 to
  simulate lossy links of Mica2 motes
• Baseline protocols
• MT Extended DSDV that minimizes num of Txs
• MTP Extended DSDV that minimizes Tx Power
• Data rate per flow 0.3 Kbps, 100 nodes

16
Network Energy Consumption
Energy Cost of All Nodes (J)
Energy Cost of Non-Source Nodes (J)
MPCP saves up to 30 energy
MPCP saves up to 80 energy
17
Unified Power Management

• Minimum power configuration
• Relationship btw sensing and communication
• Network connectivity vs. sensing coverage
• Impact of coverage on geographic routing
• Two realistic power management protocols
• Configurable topology control for lossy WSNs
• Coverage maintenance for target detection

18
Power Management under Performance Requirements
sleeping node
base station
active node

• Minimize num of active nodes under requirements
• Any target within the region must be detected
• ? K-coverage every point is monitored by at
  least K active sensors
• Report the target to the base station promptly
• ? N-connectivity network is still connected
  if N-1 active nodes fail
• Routing performance short route length
• Focus on fundamental relations between the
  requirements

19
Assumptions

• Simplistic disc model for analysis
• A point p is covered by a node v if pv lt Rs
• Rs sensing range
• Nodes u and v are connected if uv lt Rc
• Rc communication range

20
Connectivity vs. Coverage Analytical Results

• Network connectivity does not guarantee coverage
• Connectivity only concerns with node locations
• Coverage concerns with all locations in a region
• If Rc/Rs ? 2
• K-coverage ? K-connectivity
• Implication given requirements of K-coverage and
  N-connectivity, only needs to satisfy max(K,
  N)-coverage
• Solution Coverage Configuration Protocol (CCP)
• If Rc/Rs lt 2
• CCP SPAN chen et al. 01

21
Greedy Forwarding

• Always forward to the neighbor closest to
  destination
• Simple, local decision based on neighbor
  locations
• Fail when a node cant find a neighbor better
  than itself
• Always succeeds in presence of coverage when
  Rc/Rs gt 2
• Hop count from u and v is upper-bounded by

shortest Euclidean distance to destination
Rc
A
destination
B
22
Geometric Properties of Coverage

• The Voronoi Diagram of a set of nodes partitions
  the plane into a set of Voronoi regions, one for
  each node
• Voronoi region of node u, Vor(u) a point lies
  inside Vor(u) if and only if u is the closest
  node to the point.
• The sensing range of u fully covers Vor(u)
• Neighbors in Voronoi diagram are connected when
  Rc/Rs gt 2

uv lt 2Rs lt Rc
v
p
23
Bounded Voronoi Greedy Forwarding (BVGF)

• A neighbor is a candidate only if the line
  joining source and destination intersects its
  Voronoi region
• Greedy choose the candidate closest to
  destination

x and y are candidates
Rc
x
y
u
z
v
not a candidate
24
One-Hop Projected Progress

• Route from u to v lies in a uv 2Rs rectangle

route lies inside uv x 2Rs rectangle
Rc
exist a candidate due to coverage
Rs
x
y
u
2Rs
v
w
distance to line uv less than Rs
minimal projected progress in one hop
25
Analytical Results
Dilation
GF bound is high when Rc/Rs ? 2
Both performs well for high Rc/Rs
BVGF bound
Dilation
26
Unified Power Management

• Minimum power configuration
• Relationship btw sensing and communication
• Two realistic power management protocols
• Coverage maintenance for target detection
• Configurable topology control for lossy WSNs

27
Coverage for Target Detection

• Probability of detecting any target gt 95,
  false alarm rate lt 2
• Probabilistic sensing model
• Accounts for signal decay and Gaussian noise
• Target detection
• Each sensor uses Likelihood Ratio Test
• Combine decisions from multiple nodes

28
Co-Grid Efficient Coverage Maintenance

• Activates a small set of nodes to achieve
  coverage, schedule other nodes to sleep
• Overlapping grid layout
• Multi-sensor fusion within each grid
• Efficient inter-grid coordination

Co-Grid 129 active nodes
Separate-grid 224 active nodes
29
Topology Control for Lossy WSNs

• Wireless links are inherently lossy
• Higher transmission power improves link quality
• New topology control problem achieve required
  topology quality using minimum transmission power

Transmission power (dbm)
Son et al. 2004
Cerpa et al. 2004
30
Configurable Topology Control

• Defined a new topology quality metric that
  accounts for lossy links
• Developed a set of topology control algorithms
• Localized each node computes its transmission
  power based on 2-hop neighborhood information
• Configurable trade-off btw required quality and
  power consumption

31
Contributions

• Minimum power configuration
• First to minimize total radio energy consumption
• Relationship btw sensing and communication
• Laid foundation for unified power management
  under both sensing and comm. requirements
• Two new power management protocols
• Incorporated realistic sensing and communication
  models in protocol designs

32
Publications

• ACM/IEEE Transaction Papers
• Design and Analysis of Just-in-Time Prefetching
  Protocols for MobiQuery, G. Xing, S.
  Bhattacharya, C. Lu, O. Chipara, C. Fok, G.
  Roman, submitted to ACM Transactions on Sensor
  Networks.
• Minimum Power Configuration for Wireless
  Communication in Sensor Networks, G. Xing C. Lu,
  Y. Zhang, Q. Huang, R. Pless, ACM Transactions on
  Sensor Networks, subject to minor revisions.
• Integrated Coverage and Connectivity
  Configuration for Energy Conservation in Sensor
  Networks, G. Xing X. Wang Y. Zhang C. Lu R.
  Pless C. D. Gill, ACM Transactions on Sensor
  Networks, Vol. 1 (1), 2005
• Impact of Sensing Coverage on Greedy Geographic
  Routing Algorithms, G. Xing C. Lu R. Pless Q.
  Huang. IEEE Transactions on Parallel and
  Distributed Systems (TPDS),17(4), 2006
• Conference Papers
• Minimum Power Configuration in Wireless Sensor
  Networks, G. Xing C. Lu Y. Zhang Q. Huang R.
  Pless, The Sixth ACM International Symposium on
  Mobile Ad Hoc Networking and Computing
  (MobiHoc), 2005,acceptance ratio 40/28114
• Dynamic Wake-up and Topology Maintenance
  Protocols with Spatiotemporal Guarantees, S.
  Bhattacharya, G. Xing, C. Lu, G. Roman, B.
  Harris, O. Chipara, International Symposium on
  Information Processing in Sensor Networks (IPSN),
  2005, acceptance ratio 44/21320.6.
• Real-time Power-aware Routing in Wireless Sensor
  Networks, O. Chipara, Z. He, G. Xing, Q. Chen, X.
  Wang, C. Lu, J. Stankovic, T. Abdelzaher, 4th
  IEEE International Workshop on Quality of Service
  (IWQoS), 2006, acceptance ratio 27/15217.
• A Spatiotemporal Query Service for Mobile Users
  in Sensor Networks, C. Lu G. Xing O. Chipara
  C. Fok S. Bhattacharya, International Conference
  on Distributed Computing Systems (ICDCS),
  Columbus, OH, 2005, acceptance ratio 76/54014.
  
• On Greedy Geographic Routing Algorithms in
  Sensing-Covered Networks, G. Xing C. Lu R.
  Pless Q. Huang. The Fifth ACM International
  Symposium on Mobile Ad Hoc Networking and
  Computing (MobiHoc), May, 2004, Tokyo, Japan,
  acceptance ratio 24/2759
• Co-Grid An Efficient Coverage Maintenance
  Protocol for Distributed Sensor Networks, G.
  Xing C. Lu R. Pless J. A. O'Sullivan, The 3rd
  International Symposium on Information Processing
  in Sensor Networks (IPSN), 2004, acceptance
  ratio 50/14534.4
• Integrated Coverage and Connectivity
  Configuration in Wireless Sensor Networks, X.
  Wang G. Xing Y. Zhang C. Lu R. Pless C. D.
  Gill, First ACM Conference on Embedded Networked
  Sensor Systems (SenSys), 2003, acceptance ratio
  24/13517.8
• Middleware Specialization for Memory-Constrained
  Networked Embedded Systems, V. Subramonian G.
  Xing C. Gill C. Lu R. Cytron, 10th IEEE
  Real-Time and Embedded Technology and
  Applications Symposium (RTAS), 2004, acceptance
  ratio 62/20530.2

33
Acknowledgements

• Dr. Chenyang Lu
• Dr. Robert Pless, Dr. Gruia-Catalin Roman, Dr.
  Joseph OSullivan
• Dr. Yixin Chen, Dr. Shirley Dyke
• Dr. Ying Zhang, Dr. Qingfeng Huang, Dr. Markus
  Fromherz at Palo Alto Research Center
• Fellow students
• Sangeeta Bhattacharya, Octav Chipara, Chien-Liang
  Fok, Kevin Klues, Xiaorui Wang, Yuanfang Zhang,
  Venkita Subramonian

34
Research Summary

• Systems
• UPMA unified power management architecture in
  TinyOS Tech. report 06
• MobiQuery spatiotemporal query service for
  mobile users in mostly sleeping sensor networks
  ICDCS 05, IPSN 05
• nORB light-weight real-time middleware for
  networked embedded systems RTAS 04
• Algorithms, protocols, and analyses
• Minimum power configuration TOSN 06, MobiHoc 05
• Integrated coverage and connectivity
  configuration TOSN 1(1), SenSys 03
• Impact of sensing coverage on geographic routing
  TPDS 17(4), MobiHoc 04
• Configurable topology control for lossy WSNs
  Tech. report 06
• Real-time power-aware routing in sensor networks
  IWQoS 06
• Data fusion for target detection IPSN 04

35
Topology Control Problem

• Metrics
• Tx count expected num of attempts before a
  successful transmission on a link
• Dilation of tx count (DTC) max ratio of the tx
  count between any two nodes to that in the
  max-power case
• Given a DTC, minimize total power of all links
  while satisfying the DTC

Tx count 1.17 Power 6
Dilation 2.8/1.7 1.65
36
Related Work

• Power management
• Sleep scheduling Polastre et al. 04, Zheng et
  al. 03, Hohlt et al. 04, Ergen 02, Lu et al. 05,
  Dam and Langendoen 03, Heidemann et al. 02
• Topology control Alzoub et al. 03, Ramanathan
  and Hain 00, Rodoplu and Meng 99, Narayanaswamy
  et al. 02, Li et al. 01, Kawadia and Kumar 03, Li
  and Hou 04, Li and Hou 03, Hajiaghayi et al. 03,
  Jia et al. 05
• Power aware routing Singh et al. 98, Li et al.
  01, Sankar and Liu 04, Chang and Tassiulas 00,
  Dong et al. 05, Doshi et al. 02, Li et al. 04
• Lifetime maximization
• Surveillance Liu et al. 05, Rosenberg et al. 04,
  Lee et al. 04, Mhatre et al. 04
• Routing Chang and Tassiulas 99, 00, 04, Sankar
  and Liu 2004
• Sensor network architectures
• SNA UC Berkeley, Tenet USC

37
Performance of Centralized Algorithms

• TMST Min spanning tree Li et al. 2003
• TSPT Shortest path tree singh et al. 1998
• MBA the Matching Based Algorithm Meyerson et
  al. 00
• MBA-opt MBA with our optimization

• 200 nodes distributed in a 500m X 500m region
• Each data flow has a rate of 0.2 Kbits/s
• Radio uses Mica2 mote setting

38
Delivery Rate and Delay

• MPCP/MASP cause slightly higher network
  contention due to more path reuse

39
Overhead

• Most MPCP route updates are local ? overhead
  remains roughly constant as num of flows grows

40
Architectural Requirement I Flexibility

• Allow users to choose different PM polices
• Example
• Sync sleep scheduling high throughput, suitable
  for periodic data collection
• Async sleep scheduling low overhead, suitable
  for sporadic event monitoring

01010101010101
sender
sender
active window for communication
receiver
receiver
Async Sleep Scheduling
Sync Sleep Scheduling
41
Architectural Requirement II Composibility

• Allow users to combine multiple PM policies
  together
• Example Delay Tolerant Networks
• Sleep scheduling when no connections
• Sleep scheduling power control for
  opportunistic transfers
• Support cross-layer optimization

42
Unified Power Management Architecture
Protocol 2
Protocol 1
Protocol 3
Protocol 0

interfaces of sleep schedulers
SyncSleep
AsyncSleep
Other Interface

OnTime
Mode
Param 0
OffTime
Preamble
Param 1
DutyCycling Table
LPL Table
Other Table
parameters specified by upper-level protocols
Power Management Abstraction

• Consistency check
• Aggregation

Power Manager

Async Listening
Others
Sync Scheduler
sleep scheduling protocols
MAC
PreambleLength
ChannelMonitor
On/Off
interfaces with MAC
PHY
43
Summary of Impact of Coverage on Routing

• Greedy forwarding
• Always succeeds if Rc/Rs ? 2
• Long routes when Rc/Rs ? 2
• New geographic routing algorithm BVGF
• Bounded route length for any Rc/Rs ? 2
• Simple geographic routing algorithms work well
  with coverage

44
Network Topology Coverage Connectivity
SPAN
CCP
SPANCCP

• Rc 1.5Rs
• Combination of SPAN CCP is necessary for
  desired coverage and connectivity when Rc lt 2Rs

45
Wireless Sensor Network Platform Motes

• Sensors acoustic, magnetic, light,
  temperature..
• Radio lt 250 Kbps
• Memory 4-10KB data, 128-256 KB program
• Limited power source 2AA batteries
• Only sustain days if continuously active

Mica mote
Mica2 dot mote
Mica2 mote
Telos mote
Intel Mote2
46
Issues with Current Architecture

• PM is coupled with other system functions
• Sleep scheduling is in MAC, power control is in
  routing
• Complicates the design of MAC
• Lack of flexibility
• Non-standard interfaces
• Conflicts between different PM policies

Protocol 0
Protocol 1
Protocol 2
Protocol 3
S-MAC
802.15.4
B-MAC
TDMA Scheduling
Sync Sleep Scheduling
Async Sleep Scheduling
PHY
Sleep scheduling in TinyOS
47
Two/Four-Hop Projected Progress

• Greedy Euclidean distance between any two
  non-adjacent nodes on a route is longer than Rc
• Bounded BVGF routes are inside uv x 2Rs
  rectangle

x
y
u
2Rs
v
xz gt Rc
z
48
Unified Power Management Architecture (UPMA)

• Interfaces between sleep scheduling and MAC
• Different types of MAC CSMA, TDMA, Hybrid
• Power management abstraction
• Allows a user to specify desirable sleeping
  strategy
• Power manager
• Coordinates the use of multiple strategies
• Implemented UPMA in TinyOS 2.0
• Separated the sleep scheduler from B-MAC
• Implemented two new schedulers on top of B-MAC