Context-Aware Converged Communications B.H. Juang Motorola - PowerPoint PPT Presentation

Context-Aware Converged Communications B.H. Juang Motorola Foundation Chair Professor & Georgia Research Alliance Eminent Scholar School of Electrical & Computer Engineering Georgia Institute of Technology

Communication Services Today – IN IN Perspective Web Server App Server SCE Mngmt SMS Billing Internet IP / Data Network SIP ISP Freephone * SCP Prepaid / Card HLR Call Center * LNP, LTF/LIDB INAP Call Screening * SS7 Network Messaging * VPN ISUP Call Routing * INAP Web based Provisioning *... SN/IP VMS Transport Circuit Switch Access After Kozik

Towards Converged Voice/Internet Services VoiceXML, Web Server SCE App Server Mngmt WML, HTML, SIP, ... SMS Billing Internet IP / Data Network SIP, LDAP, XML, IN, ISP SIP SCTP/TALI, Corba, Jain, Parlay SCP Packet Call Center * HLR Online Comm. Center * Application Virtual Kiosk * Server INAP Mediation Voice Portals * SS7 Network Message Gateway * Voice / Data VPN * ISUP . . . INAP SN/IP VMS Transport Circuit Switch Access Only an intermediate solution!! After Kozik

Example – Multiparty Communication Services Business model: Service provider = IP server + feature server Media Bridge Controller Call Cont. Server MGCP ISP site MEGACO Interest Media Bridge Call group Chat Session RTP Room server client RTP RTP MPC MPC MPC MPC Applet Applet Applet Applet Suite Suite Suite Suite RTP Whiteboard, Text, Peer-to-peer (side-bar) Audio (EC/Codec), audio whisper, copy files Video, Signaling/directory Control Stream file, share files Media data

mostly IN NTT’s Resonant Communication Networks

Context-Aware Communications Context User/Application Context Context Representation Context Sensing Inference and Translation Determination Media Content Information Communication Processing Management Chain Communication Network Architecture & Networking

Architectural Evolution Service AIN management Operations Systems SCP systems SM CCS Service Intelligent STP (SS7) Adjunct Peripheral Network SSP: Service � � switching point CCN (Context-sensitive Converged Networks) To other CSCs Context-Sensitive Information Service CSC Management Systems Management Systems CSC Multimedia ( Context & Service Complex Service Control) Network Digital Digital Terminal Terminal CSSP: Converged Wired or wireless Service switching point

Network Architecture for CCN Application Layer Session Control Flow Registration Authentication Authentication Connection Connection Address for Status Negotiation Processing Authentication Other Party Check Context Service Control Layer Network Control Layer CSC Encryp- Context Access Control Control Control Quality Control tion IP Transport Layer session End-to-end Session Control Modified from Tachimoto & Morita of NTT

Dimensions of Context CONTEXT user environment applications Personal Local Physical Service Expectations Preferences Attributes Resources Information Agreements & Goals Spatial Network Temporal/history Computing Display Activity-Related Physiological

Sensors in the Labs at Gatech Video Wireless Audio

EKG from Control EKG from GTWM Devices for Aging in Place

Multi-channel Hi-Fi Telecommunication Traditional telephony trunk � � switch switch New multi-channel network for information sharing Mic’s Spkr’s Multi- Multi- channel channel Packet Aggregate Aggregate Network Projectors Processor Processor Cameras MIMO MIMO or for intelligent assistance

Spatialization & The Perceptual Dimension Sound spatialization makes talker-tracking easier in multi-party conferencing environments, resulting in improved effectiveness in communication Binaural Hearing & Cocktail Party Effect Spatial separation plays a role. � Compare mono with stereo Stream segregation also plays a role. � Compare one talker (m1+m2) with two (m1+f2) (m1 m2 f2 ) Stereophonic Conferencing Demonstration

Multi-channel Communications Human communication is a spatial-temporal event. It is important to track the source objects.

ROOM 2 Stereophonic Acoustic Echo Cancellation g 1 talker g 2 NL NL h 2 ~ + - h 1 ~ h 2 h 1 ROOM 1

Multi-channel Source Separation ∗ + ∗ s h s h 1 11 2 21 s 1 H 11 x 1 W 11 s’ 1 H 12 W 12 H 21 W 21 s 2 s’ 2 x 2 H 22 W 22 mixing un-mixing ∗ + ∗ (room impulse responses) s h s h 1 12 2 22 One possible approach (Ikram of Gatech and Morgan of Bell Labs) : R’ = <x x H > x = H s s’ = Wx Find un-mixing filter matrix W such that Λ s’ = W R’ W H is diagonalized by minimizing the squared Frobenius norm of the off-diagonal matrix of Λ s’

Sound Source Localization 1. Time Delay Estimation 2. Source Location Estimation Various methods: • triangulation - solve a set of hyperbolic equations • spherical intersection - solve a set of linearized spherical equations • spherical interpolation - similar to SI, but with reduced constraint • one-step-least-squares – transforms the problem into an estimation/minimization problem; works the best Applications: • Conferencing with participant tracking Developed at Bell Labs • Improved sound and sight pickup & Georgia Tech talker Further challenge

The Content Processing Chain CONTENT CONTENT CONTENT ACQUISITION ENCODING & RETRIEVAL Conte x t USER CONTENT INTERFACE Determination DISTRIBUTION CONTENT CONTENT PRESENTATTION DECODING

Managing Content for CAC – Semantic Metadata Back End Applications Produce Catalog/ Integrate Interactive Personalize Aggregate Index Syndicate Marketing What is Where is What What is the What is the this best way to the other right content content utilize and content? content is for this user? interact with about? Whose is it related this content? it? to? Broadcast, Wireline, Wireless, Semantic Metadata Interactive TV

Communication with Region-of-Interest ROI Algorithms for: Lossless Coding Progressive Transmission Electronic Zooming Integration with VoIP, VUI, TIVO and Array Processing Technologies

Context-Aware Content Distribution Multimedia Transport Protocol - MMTP UDP + FEC+ error concealment UDP+ MMTP + selective retransmission Over-provisioning / Packet Loss Rate for Perfect Q: 7 for UDP 2 for MMTP [Theoretical Limit is 1]

Embedded and Layered Coding telephone 3 2 1 NETWORK Voice packet Switch 4 Switch 2 Switch 1 3 2 Switch 3 Cell site Layered bitstream 3 2 1 cell-phone Channel coding

Information Management & Support Layers MetaServices Adaptive and Customized Information Services Context Management

Context-Aware Convergent Communications Research Vision � Imbue Converging Digital Services with the attributes and perception of personalization, flexibility, richness and privacy; � Incorporate Context , broadly defined, into the deep fabric of wired and wireless networks for information, transactions and entertainment; � Harness the capabilities of multimedia processing, embedded computing and pervasive broadband for Context-Aware Telecommunications; � Spawn a new generation of applications of interest to the end user, the communications industry and society, using the exemplary test domains of tele-health, distributed education and entertainment.

Summary Convergent & Converged Communication Networks need research advances in: • Infrastructure for User-Centric Converged Services • Context-Aware Media Processing • Context-Aware Information Management • The Science and Engineering of Context