Multi-channel Wireless Networks with Infrastructure Support:

Capacity and Delay

Hong-Ning Dai (Macau University of Science and Technology, Macau)

Raymond Chi-Wing Wong (Hong Kong University of Science and Technology, Hong Kong)

Qinglin Zhao (Macau University of Science and Technology, Macau)

Presented by Hong-Ning Dai

Comparison with existing networks

Pure Ad Hoc Ad Hoc with Infrastructure

Single Channel

SC-AH networks

[1]

SC-IS networks

[2] [3]

Multiple Channels

MC-AH networks

[4]

MC-IS networks

our proposed

frequ

ency d

om

ain

Ad Hoc hybridity

Our proposed MC-IS networks have more benefits than the existing networks

MC-IS Network

Base station

Common node

Wired network

Ad hoc communications withinH hops

uplink

downlink

Infrastructure communications

No capacity constraint and delay constraint within wired networks

Our contributions

Formally identify an MC-IS network.

Derive both the upper bounds and the constructive lower bounds of the capacity of an MC-IS network.

Find that the capacity of an MC-IS network contributed by ad hoc communications is mainly limited by four requirements.

Four requirements

1( )

AC O F

2( )

AC O F

1( )

AC F

and 2( )

AC F

1( )o GH

1( )GH

2( )o GH

2( )GH

3( )o GH

3( )GH

Sub-case 1 Sub-case 2 Sub-case 3 Sub-case 4 Sub-case 5 Sub-case 6(Interface-bottleneck)

(Connectivity) (Interference) (Interface-bottleneck)

(Destination-bottleneck)

(Interface-bottleneck)

Summary of our major results

MC-IS networks

SC-IS networks

MC-AH networks

SC-AH networks

DelayDelayCapacity Capacity NetworksNetworks

Our proposed MC-IS networks have the optimal capacity and the minimal delay!

Methodology overview

Derivation of Network Capacity Contributed by Ad Hoc Communications Upper bounds Lower bounds (constructive)

Derivation of Network Capacity Contributed by Infrastructure Communications Upper bounds Lower bounds (constructive)

To prove these bounds is non-trivial!

Constructive lower bounds

Tessellation

Bound the number of interfering nodes Apply channel models (path loss, interference,

shadowing) Determine the interference range

The size of cell needs to be satisfied

with connectivity constraintconnectivity constraint

Cell lattice

We must ensure the interference constraintinterference constraint is satisfied.

Design proper routing algorithms

Connect each source destination pair Si-Di

Design shorted routing path (almost straight lines)

Assign the proper number of routes at each node so that

(i) not too many routes

(ii) not too few routes

Design TDMA scheme to schedule interfering cells to be active one by one

Other contributions

Our proposed MC-IS network offers a more general theoretical framework than other existing networks.

Other networks such as an SC-AH network, an MC-AH network, and an SC-IS network can be regarded as special cases of our MC-IS network.

More details about our results

Please refer to our paper in Proc. of IEEE International Conference on Communications (ICC) 2014 - Ad-hoc and Sensor Networking SyAd-hoc and Sensor Networking Symposiummposium

More complete version: H.-N. Dai, R. C.-W. Wong, and Q. Zhao, “Multi-chan

nel wireless networks with infrastructure support: Capacity and delay,” Tech. Rep., Feb. 2014. [Online]. Available: http://arxiv.org/pdf/1310.1530v2.pdf or