Ambient and Cognitive Networks Youn-Hee Han [email protected] May 2009 Korea University of Technology and Education Laboratory of Intelligent Network .

Ambient and Cognitive Networks

Youn-Hee [email protected]

May 2009

Korea University of Technology and EducationLaboratory of Intelligent Network

http://link.kut.ac.kr

2009 미래인터넷 표준기술 워크숍

Ambient Networks

2/29


EU’s FP6 (6th Framework Program) 유럽연합의 6 차 연구개발 프로그램 (2002~2006) Goal & Vision

유럽단일연구공간 (ERA: European Research Area) 의 실현

7 개 중점 연구 분야와 IST, WWI, Ambient Networks 의 관계

EU’s FP6 & Ambient and Cognitive Networks

3/29

연구 분야 예산 (Euro)

1. 생명과학 , 게노믹스 , 생명공학 22 억

2. 정보사회기술(Information Society Technologies, IST)

36 억

3. 나노기술 및 나노과학 , 새로운 생산공정 및 디바이스 13 억

4. 항공우주 10.75 억

5. 식품의 질 및 안전성 6.85 억

6. 지속가능한 발전 , 전지구적 변화 및 생태계 21.2 억

7. 지식기반사회에서의 시민과 통치 2.25 억

기타 13.2 억

계 132.85 억

…

WWI (Wireless World Initiative, 2004~)

yyy

XXX[IST 내의 통합 관리 프로젝트 ]


Ambient Networks (AN) A software-driven dynamic network integration solution

over any access technology and any type of network Phase I (2004-2005)

Fundamental concepts & the overall architecture of ANs Phase II (2006-2007)

The engineering of the overall solution Develop real prototypes

Overall Review Fatna Belqasmi, Roch H. Glitho, and Rachida Dssouli,

“Ambient Network Composition,” IEEE Network, Vol. 22, No. 4, pp. 6-12, July/August 2008.

Overview of Ambient Networks

4/29

[Ambient]: existing or present on all sides: of the surrounding area or environment


Four innovations (design paradigm) of AN

Overview of Ambient Networks

5/29

Network Composition

Enhanced Mobility

Network Heterogeneity Support

Context Awareness

+

To provide common control functions to a wide range of different applications and air interface technologies


Ambient Control Space (ACS)

Technology of Ambient Networks

6/29

Ambient Network Interface (ANI): Standardized single interface to connect the network instead of just connection of nodes: Offer a simple plug & play connection

Ambient Service Interfaces (ASI): Even in a composed Ambient Network, only a single homogeneous control space is visible to external entities : An application or service will always find the same environment


Network Composition (1/2) Network Integration

Involved networks merge intoone common network

E.g. a new PAN creation integrating two different PANs

Control Delegation One AN delegates certain control

functions to the other AN E.g. 3GPP-WLAN interworking:

WLAN delegates authentication, authorization and charging to 3GPP network

Network Interworking Cooperation but no control delegation E.g. dynamic roaming agreements


7/29

Incr

easi

ng c

ontr

ol p

lane

inte

rwor

king


Network Composition (2/2)


8/29

Overlay Control Space

AmbientConnectivity

FE 4FE 3

FE5

CompositionFE

FE1

FE6FE2

Ambient Control SpaceAmbient Control Space

AmbientConnectivity

FE 4FE 3

FE5

CompositionFE

FE1

FE6FE2

Ambient Control SpaceAmbient Control Space

GANS: Generic Ambient Networks Signaling

GANS


Generic Link layer (GLL) for a Multi-Radio Access


9/29

Generic Link Interface (GLI): It provides compatible radio link layers for different radio access technologies: A reconfiguration of the GLL (generic link layer) due to a change of radio access technology will be seamless


Scenario 1

Scenarios of Ambient Networks

10/29


Scenario 2

Scenarios of Ambient Networks

11/29


Active Research and Much Results

Selected Review #1 Instant Media Services for

Users on the Move

Research Results of Ambient Networks

12/29

M. Vorwerk, S. Schuetz, R. Aguero, J. Choque, S. Schmid, M. Kleis, M. Kampmann, M. Erkoc, “Ambient networks in practice - instant media services for users on the move,” 2nd International Conference on TRIDENTCOM, 2006.


Selected Review #2 New Handover Strategy & Business Map


13/29

P. Poyhonen, J. Tuononen, T. Haitao, O. Strandberg, “Study of Handover Strategies for Multi-Service and Multi-Operator Ambient Networks,” 2nd International Conference on CHINACOM, 2007.

DS: Discovered Sets (of Access Networks)CST: Candidate Sets based on Terminal’s policy CSN: Candidate Sets based on Network’s policy AS: Finally selected Active Sets

Business Map


Selected Review #3 Ambient Network Advertising Broker


14/29

L. Ho, J. Markendahl, M. Berg, “Business Aspects of Advertising and Discovery Concepts in Ambient Networks,” IEEE 17th International Symposium on PIMRC, 2006.

Access Broker (Auction-based): Dynamic allocation per Call


Cognitive Networks

15/29


Three motivating problems for Cognitive Networks Complex

Large numbers of highly interconnected, interacting elements and instances of self-organization and emergent behavior

Network need to be able to deal with and adapt to complex environment with minimal or zero user interaction

Motivation

16/29

A school of fish

A termite mound


Three motivating problems for Cognitive Networks Wireless and Its heterogeneity

Large numbers of standards IEEE 802.11, Bluetooth, WiMAX, CDMA2000, UMTS…

Ad-hoc networks are highly dynamic should be capable of self-organization In research papers, simulation is usually used because of the

difficulty in using forms of analysis Difficulty in QoS of Layered Architecture

People wants a sort of end-to-end guarantees It is a very difficult research area because most all

networking stacks do not operate on an end-to-end paradigm.

Current approaches are typically reactive.

Motivation

17/29


Cognitive Network (CN) A network composed of elements that, through learning

and reasoning, dynamically adapt to varying network conditions in order to optimize end-to-end performance

Features Decisions are made to meet the requirements of the network

as a whole (not individual network components) A Cognitive Process

perceive conditions, plan, decide, and act on those conditions

Definition

18/29

Global Internet Map(www.siencedaily.com)

[by Ryan Thomas @ Virginia Tech., 2005]


Similarities Operates in parallel to stack Increases information available to participating layers Optimizes on goals that require multiple layers to achieve

Differences Cognition (as opposed to reactive, localized schemes) Multiple and End-to-end goals (as opposed to single goal at

layer level)

Cognitive Network vs. Cross-layering

19/29

[by Ryan Thomas @ Virginia Tech.]


Basic Decision Model OODA Loop [John Boyd] Decision based on observation of

network environments

Implementation It depends on

Goals, Controllable Network Elements, System Structure, States

Critical Design Issues Behavior: Selfish vs. Cooperation Computational: Level of ignorance Physical: Amount of control

Cognition Process

20/29


분석 및 계획


Input from Requirements Layer End-to-End Goals Cognitive Specification Language

Converts end-to-end goals into cognitive elements goals

Cognitive Elements Adapt and learns to make decisions

that meet end-to-end goals

Software Adaptable Network (SAN) API Configurable Elements

Points of network control for cognitive process Network Status Sensors

Reads status of the network

Cognitive Network Framework

21/29



Cooperative Mobile Robots

Usage Scenarios

Case Study: Mobile Robots & Sensor Network

22/29

[University of Tübingen][USC @ LA]

[Disaster Area]

[Exploring the unknown]

[Robot Army]

Environmental Robotics


How to MOVE? Cognition (Perception) of Obstacles and Other Sensors

Supersonic Wave, Artificial Vision, … Force based on Potential Fields

ForceAccelerationVelocityPosition

Sensor Robot Mobility

23/29


How to expand the covering area? A self-deployment algorithm to achieve the max coverage

level Cognition of coverage level in distributed manner

Coverage Level

24/29

Coverage Level: 28.37% Coverage Level: 76.14% Coverage Level: 98.56%


How to make the network connection robust? A self-deployment algorithm to achieve the max

connectivity level Cognition of connectivity level in distributed manner

Connectivity Level

25/29

Avg. # of Neighbor: 2.6 Avg. # of Neighbor: 3.32

Coverage Level Connectivity Level

trade-off


How to make the overlay level high? An optimized grouping algorithm to achieve the max

energy efficiency

Overlay Level

26/29

70 Active Sensors

Active - Group #1(of 35 Active Sensors)

Sleep - Group #2(of 35 Sleep Sensors)

Active - Group #2(of 35 Active Sensors)

Sleep - Group #1(of 35 Sleep Sensors)


Cognition Scheme in Mobile Sensor Networks

27/29

Sensing Areas,Obstacles, Other Sensors,

Environment Status…

Area Border LocationObstacle Location, Other Sensor Location, Sensing

Range, Communication Range, Current Levels of Coverage, Connectivity, and

Overlay

Optimization Algorithms to

maximize “Coverage Level”,

“Connectivity Level”, and“Overlay Level”

Autonomic Self-deployment of

Sensors

New Position of Sensor Robots

- Heuristic Algorithms (Greedy Algorithm…)- Intelligent Algorithms (Genetic Algorithm…)


A. Howard, M. J. Mataric, and G. S. Sukhatme, “Mobile Sensor Network Deployment using Potential Fields: A distributed, scalable solution to the area coverage problem,” The 6th International Symposium on Distributed Autonomous Robotics Systems (DARS02), June 2002.

Y. Zou and K. Chakrabarty, “Sensor Deployment and Target Localization based on Virtual Forces,” IEEE INFOCOM 2003, Vol. 2, pp. 1293-1303, March 2003.

S. Poduri and G. S. Sukhatme, “Constrained Coverage for Mobile Sensor Networks,” IEEE International Conference on Robotics and Automation, pp. 165–172, May 2004.

G. Wang, G. Cao and T. L. Porta, “Movement-assisted Sensor Deployment,” In Proc. of IEEE INFOCOM 2004, Vol. 4, pp. 2469-2479, March 2004.

B. Liu, P. Brass, O. Dousse, P. Nain and D, Towsley, “Mobility Improves Coverage of Sensor Networks,” ACM MobiHoc 2005, pp. 300-308, May 2005.

J. Wu and S. Yang, “SMART: A Scan-Based Movement-Assisted Sensor Deployment Method In Wireless Sensor Networks,” In Proc. of INFOCOM 2005, pp.2313-2324, March 2005.

G. Wang, G. Cao, T. L. Porta and W. Zhang, “Sensor Relocation In Mobile Sensor Networks,” In Proc. of INFOCOM 2005, pp. 2302-2312, March 2005.

H. Yu, J. Iyer, H. Kim, E. J. Kim, K. H. Yum and P. S. Mah, “Assuring K-Coverage in the Presence of Mobility in Wireless Sensor Networks,” in Proceedings of IEEE GLOBECOM 2006 (selected for best papers), 2006.

D. Wang, J. Liu and Qian Zhang, “Mobility-Assisted Sensor Networking for Field Coverage,” In Proc. of IEEE GLOBECOM '07. pp. 1190-1194, Nov. 2007.

Wang, H. Wu, and N.-F. Tzeng, “Cross-layer Protocol Design and Optimization for Delay/Fault-tolerant Mobile Sensor Networks, IEEE Journal on Selected Areas in Communications, Vol. 26, No. 5, pp. 809-819, June 2008

References of Mobile Sensor Networks

28/29


[Demo]

29/29

Demo animation for mobile sensor network deployment

Ambient and Cognitive Networks Youn-Hee Han [email protected] May 2009 Korea University of Technology and Education Laboratory of Intelligent Network .

Documents