Dr. Salman Durrani

Connectivity of Wireless Ad hoc Networks Dr. Salman Durrani School of Engineering, College of Engineering and Computer Science, The Australian National University, Canberra, Australia. http://engnet.anu.edu.au/DEpeople/Salman.Durrani/ Dec. 2009

Overview

• Introduction

• Research Trends in Wireless Communications
• Open research problems in ad hoc networks

• Connectivity Analysis

• Antenna & System Model
• Analytical framework

• Results & Conclusions

Canberra

ASP Academics

ASP Research Group

• The Applied Signal Processing Group conducts research in the following application areas:

• Physical layer Communications (12 PhD students)
• Telecommunications including Wireless and Mobile Communications
• Space-Time Signal Processing
• Signal Processing (9 PhD students)
• Acoustic and Audio Signal Processing
• Broadband and Near-field Sensor Arrays and Beamforming
• Bio-Signal Processing
• Applied Information Theory (2 PhD students)

My PhD Students

• Xiangyun Zhou (EIPRS Scholarship), Channel estimation in cellular & ad hoc networks. (Jan 2008-present)

• Ali Nasir (ANU International PhD Scholarship), Synchronization in co-operative communication systems. (June 2009-present)

• Zubair Khalid (EIPRS Scholarship) & Rimla Javaid (ANU PhD Scholarship), commencing 2010.

Research Trends

• IEEE ICC 2009 in Dresden, Germany:-

• 3600 paper submissions
• 1046 accepted papers after peer review (29%)
• Selected papers presented in oral sessions.
• On average, 5 to 6 papers being presented in every session.

• We looked at the number of oral sessions presented for different research topics in GlobeCom and ICC for last 3 years.

Research Trends

Research Trends

• Research topics in steady state or decline ?

Research Trends

• Research topics in steady state or decline ?

Research Trends

• Other Research Topics

Research Trends: Ad Hoc Networks

Basic Principles of an Ad hoc Network

• Formed by wireless nodes which may be mobile.

• No need (necessarily) for any pre-existing infrastructure.

• Decentralized operation.

• Multi-hop communication.

Applications: Sensor Networks

• Networks of typically small, battery-powered, wireless devices .

• Wide range of applications:-

• Environmental monitoring,
• Military surveillance
• Medical care
• Home appliance management
• Industrial monitoring

Applications: VANETS

• Vehicular Ad Hoc Networks comprise vehicle-to-vehicle and vehicle-to-infrastructure communications based on WLAN technologies. (IEEE P1609 WAVE Standards)

Applications: Telecomms Networks

Research Challenges in Ad hoc Networks

• Capacity

• What are the fundamental performance limits (in terms of reliable data rate) for ad hoc networks?

• Routing

• How do you efficiently select paths in a network along which to send information?

• Connectivity

• If you select any pair of nodes in an ad hoc network, what is the probability they connected?

• Co-operation

• 3 node Source, Relay, Destination scenario. Synchronization

Connectivity Definitions

• Connectivity from view-point of a single node:-

• Average Node Degree: Average no. of direct links that any given node has to other nodes.
• Probability of node isolation: Probability that a randomly selected node in an ad hoc network has no connections to any other node.

Connectivity Definitions

• Connectivity from view-point of an entire network:-

• 1- connectivity: Probability that every node pair in the network has at least one path connecting them.
• Critical Node density: Node density that yields an almost surely connected network [P (1-con)=0.99].
• Path Probability: Probability that two randomly chosen nodes are connected either via a single hop or a multi-hop path.

Research Challenges in Modelling Ad hoc network Connectivity

• Mobility

• Dynamic network topology
• Realistic models for mobility

• Node distribution

• Uniform
• Clustered

• Channel

• Wireless links subject to shadowing and fading
• Interference from simultaneous transmissions elsewhere

• Multiple Antennas

• Adopted in 3GPP (Release 6), IEEE802.11n, IEEE802.20
• How does beamforming affect the connectivity of ad hoc networks?

Prior Work

• Closed-form analytical results for connectivity are available for special case of:-

• Node locations are Poisson,
• Negligible interference,
• No mobility,
• Channel: Path loss & Shadowing,
• Omni-directional antennas

• Open Research Problem:

Overview

• Introduction

• Research Trends in Wireless Communications
• Open research problems in ad hoc networks

• Connectivity Analysis

• Antenna & System Model
• Analytical framework

• Results & Conclusions

System Model

• Nodes are distributed in 2D according to Poisson point process.

• All nodes are equipped with beamforming antennas.

Antenna Model

• Antenna Model is characterized by associated antenna power pattern.

• Uniform Linear Array or Uniform Circular Array ?

Power Pattern Demo (N=8)

Antenna Model

• UCA configuration is chosen:-

• It has single main lobe.
• 3 dB beamwidth is constant & independent of main beam direction.

• For UCA, the directivity G is given by

Random Beamforming

• Core Idea: Each node randomly selects a main beam direction without co-ordination with other nodes.

• Advantage:-

• MAC is un-coordinated.
• Minimal communication overhead and hardware complexity.

• Disadvantage:-

• May not be optimal strategy

Channel Model

• Received Signal Power

• Shadowing affects only the randomness and not the average value of the channel gain.

Communication Range

• Two nodes can communicate with each other if their distance apart is smaller than a given communication range R.

Effective Coverage Area

• With beamforming, the communication range is

• Effective coverage area is

Effect of Shadowing

• Shadowing factor:

• Depends on:-

• path loss  and
• Shadowing log-normal standard deviation 

• Key Insight:- Shadowing reduces the effective coverage area (for > 2).

Effect of Shadowing

Effect of Beamforming

• Beamforming factor:

Effect of Beamforming

• Beamforming factor:

• Depends on

• path loss 
• Number of antenna elements in the UCA M

• Does not depend on

• Shadowing log-normal standard deviation 

Effect of Beamforming

• Beamforming factor values:

• Key Insight:- For <3, random beamforming increases the effective coverage area.

Overview

• Introduction

• Research Trends in Wireless Communications
• Open research problems in ad hoc networks

• Connectivity Analysis

• Antenna & System Model
• Analytical framework

• Results

• Average Node degree
• Probability of node isolation
• 1-connectivity
• Critical node density
• Path probability

• Conclusions

Connectivity Metrics - P(iso)

• Probability of node isolation

• Main Result:-

• shadowing always increases the probability of node isolation.
• beamforming, compared to omnidirectional antennas, reduces the probability of node isolation when <3.

Connectivity Metrics - P(iso)

• Verification (effect of shadowing, M=1):-

Connectivity Metrics - P(iso)

• Verification (effect of beamforming, M=1,4 & = 4 dB):-

Connectivity Metrics - c

• Critical node density: node density that yields an almost surely connected network, that is, the density at which P (1-con) = 0.99.

• Main Result:-

• shadowing increases the critical node density.
• c can be reduced by using beamforming when <3.

Connectivity Metrics - c

• Effect of shadowing

• No of antennas M=1

Connectivity Metrics - c

• Effect of beamforming

• No of antennas M=1,4

Conclusions

• We have presented an analytical model to characterize the effect of random beamforming on the network connectivity.

Publications

• X. Zhou, S. Durrani and H. Jones, "Connectivity Analysis of Wireless Ad Hoc networks with Beamforming," IEEE Transactions on Vehicular Technology, vol. 58, no. 9, pp. 5247-5257, Nov. 2009.

• S. Durrani, X. Zhou and H. Jones, "Connectivity of Wireless Ad Hoc Networks with Random Beamforming: An Analytical Approach," Proc. IEEE PIMRC, Cannes, France, Sep. 15-18, 2008.

• Paper PDFs available at:-http://engnet.anu.edu.au/DEpeople/Salman.Durrani/papers.html

• Thank you for your attention