1. Where to find help when you meet a emergency? HelpMe: a guidance system for self-rescue Kuien Liu Institute of Software Chinese Academy of Sciences

2. Pray for Kunming 2014.3.1

3. Motivation “Please show me the best way to escape”

4. Who needs this service? Run from danger – Getting rubbed when walk alone Go get help – save the wounded, pets and persons. – Search for gas when a car broke down Emergency center – Official, civil or private organizations.

5. What does HelpMe serve us? For persons run from danger – Daytime: run to crowded places Based on population, transportation and POI datasets – At Night: run to bright places Based on nightscape, road network and POI datasets For persons go get help – Route planning: nearest spots + shortest waiting time + fastest transportation

6. What kind of instructions we need? Simpler and shorter – One-click control – Short instruction – Easy to understand Rational – Temporal and/or spatial constrained – Globally optimal, for the wounded – Locally optimal, for the runaways Clearly Induced – Vehicles, e.g., vacant taxi, bus/railway station nearby – Places, e.g., hospital, policy station, gas station, etc.

7. Workflow of HelpMe MobileBack-end Server 1. Patterns of public transit + 2. Navigation on the Map + 3. Optimal route planning current GPS Instructions “northeast, 200 meters, turn left” GPS + movement Emergency Center Rescue plan

8. Scene 1: Go-Get-Help(GGH) For a person meets a wounded man: – T 1 : time to run to the possible place for help – T 2 : time to wait on the possible place for help – T 3 : time to transmit the wounded to a hospital – E.g., the fastest way to hospital by bus find the places with minimum {T 1 +T 2 +T 3 } T1T1 T3T3 T2T2

9. Scene 2: Run-From-Danger(RFD) Where is the best place to run? – Run to crowded areas in daylight, or bright locations at night – Where the crowd/bright locations are at that moment? NOAA knows, NASA knows, we don’t … – Fortunately, we have BIG DATA. Places familiar to individual users Places with a crowd of POIs

10. Technologies (1/3) Our existing works – Algorithms 1. Patterns of public facilities 2. navigation 3. route planning – Techniques on data management To do in the future – Continuously optimizing as running – Person-to-Person, help each others

11. Technologies(2/3) An Example of Route Planning – Main Road Identification  Shortest Path Searching  Evacuation Cost Computing  Best Route Selection

12. Technologies(3/3) The maintenance cost on meta-data and the retrieval cost are not trivial – E.g., the fastest way to hospital by bus Hospitals in BeijingStops in Beijing

13. Prototype on Android It can be implemented as an component on the COS Operating System

14. Q & A