context-aware data discovery

11
Context and Application- awareness. Context-Aware Data Discovery Dmitry Namiot Lomonosov Moscow State University Manfred Sneps-Sneppe Ventspils University College

Upload: coldbeans-software

Post on 30-Oct-2014

1.037 views

Category:

Documents


0 download

DESCRIPTION

Our presentation for ICIN-2012

TRANSCRIPT

Page 1: Context-Aware Data Discovery

Context and Application-awareness.Context-Aware Data Discovery

Dmitry Namiot Lomonosov Moscow State UniversityManfred Sneps-Sneppe Ventspils University College

Page 2: Context-Aware Data Discovery

• Agenda:• Introduction• Data sets in Wi-Fi connectivity• SpotEx (Spot Expert) approach• SpotEx examples• Reality mining in SpotEx• SpotEx vs. Foursquare, etc• The future development• Conclusion

8-11 Oct 2012 Context and Application-awareness / Context-aware Data Discovery 2

Page 3: Context-Aware Data Discovery

8-11 Oct 2012 Context and Application-awareness / Context-aware Data Discovery 3

Introduction

• Smart phone & Wi-Fi access point

• How to provide data to mobile subscribers nearby Wi-Fi access point?

• How to use wireless sensors on smart phone for data discovery?

• It is not about the connectivity. It is about data discovery.

Our initial case:

Smart phone as a wireless sensorWi-Fi network is visible on the mobile There is no connection to this AP yet

Page 4: Context-Aware Data Discovery

8-11 Oct 2012 Context and Application-awareness / Context-aware Data Discovery 4

Data set• AP: SSID, MAC-address, channel, signal strength

• Phone: MAC-address, history of using this MAC-address, time

• How can we link the visibility for our data to the above mentioned measurements?

• The simialar questions could be raised for Bluetooth

Wi-Fi network is visible on the mobile There is no connection to this AP yetWhat kind of data can we use?

All this info is available for mobile application

Page 5: Context-Aware Data Discovery

8-11 Oct 2012 Context and Application-awareness / Context-aware Data Discovery 5

SpotEx approach• External database with rules (productions)

• IF IS_VISIBLE (‘SSID1’) and TIME_WITHIN (1pm, 2pm) THEN {show coupon for lunch}

• Expert systems for spots

• Mobile application requests DB via HTTP

• RETE algorithm for making conclusions

Important note: access point, our rules linked to, could be opened on the another mobile phone. It is like Dynamic LBS

Page 6: Context-Aware Data Discovery

8-11 Oct 2012 Context and Application-awareness / Context-aware Data Discovery 6

Proximity marketing Chat for mobile visitors nearby

Page 7: Context-Aware Data Discovery

8-11 Oct 2012 Context and Application-awareness / Context-aware Data Discovery 7

Reality mining • Reality mining: use logged mac-address for dicovering mass behavior indoor

• Extract patterns from logged mac-addresses

• Web statistics analogue for real places

• Use statistics in proximity rules:

IF IS_VISIBLE(‘myshop’) AND FIRST_VISIT() THEN { ... }

Different data for the first time and follow-up visitors

Mobile application can record mac-address from mobile devices and use this information in rules and for the discovery users behavior

Page 8: Context-Aware Data Discovery

8-11 Oct 2012 Context and Application-awareness / Context-aware Data Discovery 8

SpotEx vs. Foursquare, etc • Analogue: automatic check-in

• Mobile application automatically check-in user’s state against wireless environment

• They are customized check-ins. Rules for “badges” (as in Foursquare) are user-defined

• “Places” could be dynamic (just another phone)

• Check-ins in SpotEx are anonymous

Important note: access point our rules linked to could be opened on the another mobile phone. It is like Dynamic LBS

Page 9: Context-Aware Data Discovery

8-11 Oct 2012 Context and Application-awareness / Context-aware Data Discovery 9

The future development • Replace external database with markup in our context: encode rules as custom attributes for HTML5. Web intents for processing

• Add developers API

• Use network proximity principles from SpotEx vs. location for social streams integration (e.g. Twitter, etc.) and augmented reality

• Wi-Fi direct support

Mobile application can record mac-address from mobile devices and use this information for the discovery users behavior

Page 10: Context-Aware Data Discovery

8-11 Oct 2012 Context and Application-awareness / Context-aware Data Discovery 10

Conclusion • A new model for context-aware data discovery for mobile users developed on the ideas of Wi-Fi and Bluetooth proximity.

• Uses smart-phone as a proximity sensor. Service can use any network node as presence trigger for delivering user-defined content right to mobile subscribers.

• Completely software based approach. A novel implementation of context-aware browser for mobile subscribers. Presented as mobile application for Android platform and HTML5 web site.

•This service could be used for delivering commercial information (deals, discounts, coupons) in malls, hyper-local news data, data discovery in Smart City projects, personal news, etc.

Page 11: Context-Aware Data Discovery

11

The EndDmitry Namiot

Manfred Sneps-Sneppe