EnergyEfficient, CollisionFree Medium Access Control for Wireless Sensor Networks

1
Energy-Efficient, Collision-Free Medium Access
Control for Wireless Sensor Networks

• Venkatesh Rajendran, Katia Obraczka, J.J.
  Garcia-Luna-Aceves
• Department of Computer Engineering
• University of California, Santa Cruz
• Sensys03

2
Outlines

• Introduction
• Background
• TRAMAs Protocols
• Neighbor Protocol
• Schedule Exchange Protocol
• Adaptive Election Protocol
• Simulation Results
• Conclusions

3
Introduction

• Wireless Sensor Networks
• Wireless, battery-powered
• The deployment is usually done in ad-hoc manner.
• -gt A major challenge is the scheduling of
  transmissions among nodes
• Self-adaptive to changes in traffic, node state,
  or connectivity
• Prolongs the battery life of each node.
• TRAMA
• The TRaffic-Adaptive Medium Access protocol
• Energy-efficient collision-free channel access

4
MAC Protocols

• Contention-based MAC
• Choose transmitter by contention.
• CSMA, PAMAS, IEEE 802.11 DCF, S-MAC
• The probability of collisions increases with the
  offered load.
• Degrades channel utilization and reduces battery
  life.

Contention
Send/Receive (or Sleep)
Contention
Send/Receive (or Sleep)

time
5
MAC Protocols (contd)

• Schedule-Based MAC
• Choose transmitter by scheduling.
• NAMA(the Node Activation Multiple Access)
• Uses a distributed election algorithm to achieve
  collision-free transmissions
• Does not address energy conservation.

Send/Receive (or Sleep)
Send/Receive (or Sleep)
Send/Receive (or Sleep)

time
6
NCR (Neighborhood-aware Contention Resolution)

• A node i derives itself as the winner based on
  NCR
• during the contention context t
• Contention set Mi ? i

7
NAMA (Node Activation Multiple Access)

• The contender set nodes within a two hop
  neighborhood
• It is sufficient for collision-freedom if nodes
  within two hops
• do not transmit at the same time.
• Based on NCR, a node decides whether itself is
  the transmitter at a time slot in a part.

Time Division in NAMA
8
TRAMA

• vs. NAMA
• Identifiers of the nodes within a two-hop
  neighborhood are used to give conflict-free
  access to the channel.
• TRAMA addresses energy efficiency by having nodes
  going into sleep mode and uses traffic
  information to influence the schedules.
• vs. S-MAC
• TRAMA is collision-free (schedule-based vs.
  contention-based)
• TRAMA uses an adaptive, dynamic approach to
  switch nodes to low power mode.

9
Protocol Overview

• The Neighbor Protocol (NP)
• Propagates one-hop neighbor information
• The Schedule Exchange Protocol (SEP)
• Collision-free data exchange
• Schedule propagation
• The Adaptive Election Algorithm (AEA)
• Selects transmitters and receivers to achieve
  collision-free transmission

10
Access Modes
Time slot organization

• Random access mode
• Signaling packets
• TRAMA starts in random access mode
• To permit node additions and deletions
• Time synchronization
• Contention-based
• Schedule access mode
• Schedule information and data packets

11
The Neighbor Protocol

• Gathers neighborhood information by exchanging
  signaling packets
• during the random access period
• To maintain connectivity between the neighbors
• A node knows the one-hop neighbors of its one-hop
  neighbors.
• -gt two-hop neighborhood information

12
Schedule Exchange Protocol

• A nodes schedule information
• periodically broadcast to the nodes one-hop
  neighbors
• In last winning slot
• Schedule generation
• The interval t, tSCHEDULE_INTERVAL for which
  it has the highest priority among its two-hop
  neighbors
• -gt Winning slots
• will be selected as the transmitter
• Announces the intended receivers for these slots
• If a node does not have enough packets to
  transmit, it announce that it gives up the slots
• Other nodes can make use of these vacant slots
• Nodes announce their schedule via schedule
  packets

13
Schedule packet format
14
Schedule packet format an example

• Source addr 1, Timeout 3, Width 4,
  numSlots 6

(Unicast)
(multicast)
(broadcast)
15
Adaptive Election Algorithm

• A node is selected to transmit if it has the
  highest priority among its contending set
• Contending set all nodes that are in the nodes
  two-hop neighborhood.

( the pseudo-random hash of the concatenation of
node us identity and time t )

• Possible States
• Transmit (TX)
• Receive (RX)
• Sleep (SL)
• Each nodes executes AEA to decide its current
  state based on current node priorities and also
  on the announced schedules from one-hop neighbors

16
Adaptive Election Algorithm (contd)

• Notations and terminologies

PTX(u)
(node y is a noe-hop neighbor of node u)
17
Adaptive Election Algorithm (contd)

• Decision a node us state
• (1) a node u is the Absolute Winner ( TX )
• (2) one-hop neighbor of a node u is the Absolute
  Winner ( RX, SL )
• (3) if not 1 or 2, one-hop neighbor of a node u
  is the Alternative Winner hidden from the
  Absolute Winner ( RX, SL )
• (4) if a node u is in 1,2 or 3 but the
  transmitter have no data to send, check Need
  Transmitter ( ntx(u) ) ( TX, RX, SL )

Case (3) Node A is hidden from Node D
18
(1)
(2)
(3)
(4)
19
Simulation

• Setup
• Simulation platform Qualnet
• Average power consumption
• transmit 24.75mW
• receive 13.5mW
• Sleep 17µW
• 50 nodes are uniformly distributed over 500m x
  500m area.
• Transmission range 100m
• 6 one-hop neighbors on average ( 17 in two-hop )
• Protocol parameters
• SCHEDULE_INTERVAL 100 transmission slots
• Max size of a signaling packet 128 bytes
• Transmission slots are 7 times longer than the
  signaling slots
• Random access period 72 transmission slots (
  and is repeated once every 10000 transmission
  slots)
• Application
• Data gathering application

20
Simulation Results (1)
lt Average packet delivery ratio for synthetic
traffic gt Schedule-based MACs achieve better
delivery than contention-based MACs -gt because of
collision freedom guaranteed at all times during
data transmissions
21
Simulation Results (2)
lt Average queuing delay for synthetic traffic
gt Schedule-based MACs incur higher average
queuing delays -gt but, deliver more packets than
contention-based MACs and this will reduce the
retransmissions at the higher layers
22
Simulation Results (3)
lt Energy savings and average sleep interval for
synthetic traffic gt A higher value of average
sleep length is preferred because this implies
less radio-mode switching and hence more savings.
23
Conclusions

• TRAMA a new energy-aware channel access
  protocol for sensor networks.
• Traffic-based scheduling to avoid wasting slots
  (when nodes do not have data to send) and to
  switch nodes to a low-power mode
• As simulation result, TRAMA can achieve
• significant energy savings depending on the
  offered load (nodes can sleep for up to 87 of
  the time)
• higher throughput. (around 40 over S-MAC and
  CSMA , around 20 for 802.11)