Crosslayer Energy and Delay Optimization in Smallscale Sensor Networks

1
Cross-layer Energy and Delay Optimization in
Small-scale Sensor Networks

• Authors Shuguang Cui, Ritesh Madan,
• Andrea J, Goldsmith and Sanjay Lall
• Presenter Zhong Zhou

2
Outline

• Introduction and related work
• Contributions
• System model
• Cross layer optimization
• Applications and special cases
• Numerical results
• Delay analysis
• Conclusions

3
Introduction and related work

• Energy characteristics of sensor networks
• Short transmission range circuit energy
  consumption can not be neglected any more
• Transmission power and circuit processing power
  need to be jointly considered
• Cross layer design in sensor network
• Much work focus on throughput maximization
• For work on energy minimization scheme
• No circuit power is considered
• No data rate adaptation

4
Contributions

• Propose a convex optimization framework to
  minimize the total energy consumption
• Quantify the best trade-off curve between the
  delay and energy consumption.
• Derive a scheduling algorithm to minimize the
  worst-case packet delay

5
System model

• Interference-free TDMA MAC
• Synchronization among nodes is available
• Joint design routing, MAC, and physical layer
• Power path-loss model
• if signal is transmitted from node i to node j,
  the received signal power at node j Pr(i,j) is
• Pr(i,j)Pt(i,j)/(Gdijk)
• Pt(i,j) transmitting power at node I
  G loss factor
• dij distance between node i and node
  j.

6
Physical layer

• M-ary Quadrature Amplitude Modulation (MQAM) is
  used
• the constellation size Mij
• Mij2bij bij is the number of bits per
  symbol
• Transmitting Power and data rate is adjustable.

7
MAC layer

• A slotted synchronous TDMA MAC scheme
• Frame length T, it is divided into multiple slots
  t
• The link schedule is periodic, i.e. if link i-gtj
  transmits in slot n, it also transmit in slot
  nT/t
• Only one link is assumed to transmit in each slot
• then, we have
• Each link may be given a different number of
  slots

8
Traffic flow

• One sink node (denoted as node n) is in the
  network and all traffic will be routed to this
  sink node, And we can get

Ri the generated bit rate of node i

• data are available for transmission at each
  node at time T, 2T,. Thus, no random delay
  is considered

9
Traffic flow

• If data is transmitted from node i to node j at
  rate bij bps for time tij,, the number of
  transmitted bits Wij will be
• The flow conservation equations are satisfied in
  every frame, which means that

Ri the generated bit rate of node I Wij the
total transmitted bits from node I to node j in
one frame
10
Energy consumption model

• Each node operate in two model
• Active mode including transmitting/Receiving
• Sleep mode all circuit is turned off
• Neglect transient modes, which is usually in the
  order of micro-seconds

11
Energy consumption model

• For uncoded MQAM, the total energy consumption to
  transmit Wij bits in tij is as

• the first term is the transmission energy and
  second term is the total transceiver circuit
  energy
• Xij, tij are system constants , which is
  determined by the distance, et.al

12
Cross layer optimization

• To minimize the total energy consumption

(4)

• Variables are Wij and tij
• The first constraint is the TDMA constraint
• The second constraint is the flow conservation
  constraint
• The third constraint is the data rate constraint
• The forth constraint model that the transmitting
  time of every node can only take integer values

13
Cross layer optimization

• Relax this integer programming problem to the
  following standard convex problem,

(5)
This convex problem can be easily solved by the
interior point methods
14
Cross layer optimization (cont)
15
Applications and special cases(1)

• Case 1 MAC Optimization with link adaptation
• Single-hop transmission, each node transmits data
  directly to the sink node

16
Applications and special cases(2)

• (a) Minimization of the total transmission
  energy (ignore the circuit energy)

And the solution is
17
Applications and special cases (3)

• (b) Minimization of the total transmission and
  circuit energy

And the solution is
18
Applications and special cases (3)

• Case 2 Routing and MAC optimization with fixed
  Links (note fixed links here means the rate of
  these links are fixed)

This is a linear optimization problem and can be
solved in polynomial time
19
Numerical results (1)

• Case 1 MAC Optimization with link adaptation
• One hop network
• System parameters

20
Numerical results (2)
21
Numerical results (3)

• Routing and MAC Optimization with Fixed links

22
Numerical results (4)
23
Numerical results (5)
24
Numerical results (6)

• Routing and MAC optimization with link adaptation

25
Delay analysis (1)
26
Delay analysis

• Delay and energy tradeoff
• The less the frame length T, the more the
  consumed energy, but the less the end-to-end
  delay
• To show the tradeoff between delay and energy, a
  weighing factor is imported to formulate an
  optimization problem

27
Delay analysis (2)
28
Delay analysis (3)
29
Conclusions

• A Jointly optimization scheme across routing
  ,MAC, and physical layer
• Optimal transmission scheme can be a mix of
  single-hop and multi-hop routing
• Link adaptation is significant in energy saving