[keynote3]詹景堯博士

Transcript

1. Ching-Yao Chan California PATH (Partner for Advanced Transportation Technology) University of California at Berkeley The 18th Mobile Computing Workshop (MC2013) Miaoli, Taiwan August 28, 2013 Big Data For Transportation
2. • Why Big Data? – Connectivity – Data Availability – Computing Power – Making a Case for Transportation • All about Data – Generating and Acquiring Data – Managing Data – Shaping Data – Using Data • Case Studies of Big Data Applications • Looking Ahead Presentation Outline
3. The Pieces are Coming Together!
4. > 2.4 Billions Internet Users http://www.internetworldstats.com/stats.htm, update 2012 Approximately 35% of World Population (~ 7.1 Billions) http://en.wikipedia.org/wiki/World_population
5. 6.8 Billions Cellphone Subscribers Source: UN's International Telecommunications Union (ITU), update 02/2013
6. We Live in a Connected World! • The Unquenchable Thirst for Connectivity – Connectivity has been growing non-stop as the mechanism of connectivity evolves from telephone, fax, e-mail, PDA, laptop, smart- phone, tablets, etc. • The Next Frontier - Cars – Intel Capital creates a $100 million Intel Capital Connected Car Fund to accelerate technology innovation in the automotive industry. • (Intel Newsroom, 02/2012)
7. Difficult to Hide in the Digital Age - Data Availability - Somewhere and somehow, someone is reading and receiving some information from you, whether you know it and like it or not.
8. It will Only Take a Nanosecond! - Computing Power - “the average smart phone today has more computing power than Apollo 11 did when it journeyed to the moon (1969)” (Time Magazine, 08/27/2012)
9. Travel faster, safer, & more efficiently? - Intelligent Transportation Systems - Goodness of Transportation System Travel Timely Good Match of Capacity & Demand Efficient and Reliable Services User Comfort & Convenience User Experience of Safety Minimum losses due to Accidents & Incidents And more …
10. I am here. Are you there? Source: VSC GM Presentation Start small, then become bigger: • One car to one car • One to many • Many to one • Signal control • Many intersections • Coordination • Traffic management • Incident detection • Emergency handling • And more ..
11. ★ Success Story at 参宮橋カーブ 首都 高速4号新宿線 ★ 60% of rear-end type collisions reduced at Sangubashi Curve. I am here. What is there? Start small, then become bigger: • Alert to one • Alert to many • Add road surface • Expand to network • Add traffic • Add weather • Incident detection • Traffic rerouting • Emergency handling • Demand management • And more ..
12. A Connected World for Travellers - Graphic from Drive C2X, Europe -
13. Making the Case for Big Data - Premises and Hypotheses - • Informed road users can behave in a safer manner and make more efficient choices. • Connected infrastructure and vehicles can facilitate better management of travel network. • Integrated data are necessary inputs to intelligent transportation systems.  Big Data can be a powerful enabler.
14. It is all about Data.
15. Big Data for Connected Vehicles - Excerpt from CVTA Workshop - • Generating and Acquiring Data • Managing Data • Shaping Data • Using Data
16. Data Generation and Acquisition Data Sources: • Roadways • Vehicles • People • And more
17. Data Processing & Management Managing the Data: • Compatibility • Evolution • Incompleteness • Scalability • Distribution: local or cloud? • Cost?
18. Data Extraction Shaping the Data: • Usefulness • Applicability • New knowledge • Discovery
19. Data Utilization Using the Data: • Useful business intelligence • Decision support • Value proposition to customers • New services
20. Case Studies Transportation Applications
21. Traffic Data and Traveler Information (Older Model) Infrastructure Sensing (camera, radar, loop, etc.) Live reporting (police, helicopter, phone calls) Radio Stations Message Signs Traffic Management Center Users and Customers • Data Model: • Limited Data Set • Minimal Processing & Integration • Business model: • Tax (public) • Advertisement (private)
22. Traffic Data and Traveler Information (Newer Model) Infrastructure Sensing (camera, radar, loop, etc.) Live reporting (police, helicopter, phone calls) Radio Stations Message Sign Traffic Management Center User and Customers Fleet Data (probe data) Subscriber Vehicle data + reporting Data Provider In-vehicle Display Portable Device
23. Functional Needs (Newer Model) Infrastructure Sensing (camera, radar, loop, etc.) Live reporting (police, helicopter, phone calls) Radio Stations Message Sign Traffic Management Center User and Customers Fleet Data (probe data) Subscriber Vehicle data + reporting Data Provider In-vehicle Display Portable Device Data Collection Data Integration Data Dissemination
24. Required Contents and Supports (Newer Model) Data Collection Data Integration Data Dissemination • Fleet Data Providers - trucks, taxis, company fleets, etc. • Vehicle-based data from subscriber • Cloud Computing – data centers, database, etc. • Processed and integrated data • Mobile Computing - Client application development • Subscriber & user interaction
25. Business Cases and Market Players (Newer Model) Data Collection Data Integration Data Dissemination • Data collection platforms HW and SW providers • Sale by data owner • Cloud computing system, HW and SW providers • Data reseller • Network providers • Mobile Computing – HW and SW providers • Advertisements • Used-based subscription fees
26. http://www.inrix.com/default.asp INRIX uses GPS data from over 100 million trucks, cameras, road sensors, in-car navigation systems, and the other app users in your driving community to create the … traffic data.
27. Google Maps can help you identify which route to choose and how much time your trip might take based on current traffic conditions. … Live and historic data is refreshed regularly ... Traffic data is available in more than 600 areas in over 50 countries.
28. Traffic Data and Traveler Information (Newer Business Models) • What is New? – Enhanced services with content enrichment – Ubiquitous yet targeted delivery • Who is paying? – Public (tax), same as before – Private (advertisement), same as before but larger and targeted audience – Data reseller & subscriber services (new) • Paradigm Shift via Big Data? – User is an integral part of data structure – Customer connectivity lends greater data mining opportunities and potentially enhanced services
29. Needs for Vehicle Repair & Maintenance Vehicle Owner/Driver Service (Older Model) Customers Call for Assistance Service Centers Repair Shops Roadside Assistance Provision of Services to Customers • Data Model: • Limited Data Set • No Data Integration • Business model: • Service-based fees • Ad hoc interaction
30. Needs for Vehicle Repair & Maintenance Vehicle Owner/Driver Service (Newer Model) Customers Call for Assistance Service Centers Repair Shops Roadside Assistance Provision of Services to Customers Subscriber Vehicle-based data Telematics Provider Online Reporting & Diagnostics Telematics Services Call Centers & Service Portal
31. Call Centers & Service Portal Needs for Vehicle Repair & Maintenance Vehicle Owner/Driver Service (Newer Model) Customers Call for Assistance Service Centers Repair Shops Roadside Assistance Provision of Services to Customers Subscriber Vehicle-based data Telematics Provider Online Reporting & Diagnostics Telematics Services Data Collection Data Integration Data Dissemination
32. Required Contents and Supports (Newer Model) Data Collection Data Integration Data Dissemination • Vehicle-Based Data (e.g. vehicle states, location updates) • Subscriber requests by phones or internets • Cloud Computing, e.g. vehicle diagnostics and monitoring • Integration with weather, traffic, tourism, entertainment, etc. • Maintenance and online repair • Localized data transmission • Remote access and control
33. Business Cases and Market Players (Newer Model) Data Collection Data Integration Data Dissemination • Data collection platforms HW and SW providers • Online reporting and subscription fees • Cloud computing system, HW and SW providers • Data mining for customized Services • Insurance and monitoring • Mobile Computing – HW and SW providers • Location-based services • Advertisements
34. Services on the Go! Toyota G-Book Ford Sync BMW Assist GM OnStar
35. Insurethebox - A New Way of Doing Business - • Assessing Driver Behaviors to Determine Insurance Premium. • Telematics Device sends information from the vehicle:  Number of journeys  Distances travelled  Types of roads used  Speed  Time of travel  Levels of acceleration and braking  Accidents Octo Telematics
36. Vehicle Owner/Driver Services (Newer Business Models) • What is New? – Enhanced services and new functions – Less effort and burden on customers • Who is paying? – Private users, same as before but more targeted audience – Subscriber services (new customers) • Paradigm Shift via Big Data? – Customer connectivity lends greater data mining opportunities – New services evolve out of data availability (such as innovations in insurance models)
37. Looking Ahead Big Data in Our Daily Life
38. Industries are Moving Forward! - Telematics Services - • Microsoft and Toyota Announce Strategic Partnership on Next-Generation Telematics (April 2011) • CES 2012: Mercedes launches mbrace2 telematics system (January 2012) • Verizon acquires Hughes Telematics for > 612 M (June 2012) • Nissan North America teams up with SiriusXM for telematics services (September 2012) • Sprint powers Chrysler’s Uconnect Telematics (November 2012) • GM OnStar Telematics Has Estimated Revenue of $1.5 Billion (May 2013) • VW Launches Car-Net Telematics with Safety Features (August 2013)
39. Cloud Computing & Data Management Telematics Service Providers Telecom & Network Providers Automakers & Suppliers ICT System Providers Contents Providers An Expanded Network of Players in A Connected World
40. Your Car is invited to be LinkedIn with My Car? Cars’ Social Networking Join CarBook, Anyone? You car have six new friends! • Cars will “do things” for us when we are on the move. • We used to tell them what to do • Now they are more likely to tell us what to do. • Cars are extended icons of our identities.
41. • Data Management  Data Scalability  Legacy Database & Data Evolution  Intelligence Distribution (Cloud vs. Local) • Institutional and Inter-party Conflicts  Needs of Individual Companies  Public Agencies vs. Private Investors • Ownership and Privacy  Data Sharing for Public Good?  Transparency • Business Model  Sustainable Returns  Cost vs. Benefits Challenges and Issues - To Name a Few -
42. Questions? Ching-Yao Chan, California PATH, UC Berkeley cychan@berkeley.edu