Labs‎ > ‎Protei‎ > ‎Oil Spill‎ > ‎3. Inspirations‎ > ‎

1. Sense


One of the biggest challenge is to sense oil remotely and directly on site. We are focusing on surface oil for this project. Oil is not easy to observe

Satellite

Frequency : once a day
Accuracy : GPS precision
Parsable : no, simple image, bitmap

Planes

Frequency : x times a day
Accuracy : no precision, reportage, sea-sky liaison
Parsable : no, simple image / video

Boats

Frequency : x times a day
Accuracy : no precision / GPS precision when EXIF data
Parsable : no, simple image / video / but can be mashed up (see below)

People

20100529labucketbrigade
Frequency : x times a day
Accuracy : Relative precision depending on the device and the amount of reports (more reports = more accurate)
Parsable : Yes, RSS | KML | JSON | CSV (table)

Institutions
NOAA

gomex.erma.noaa.gov
Frequency : x times a day
Accuracy : Relative precision depending on the device and the amount of reports (more reports = more accurate)
Parsable : ?

Robots
Underwater

Centralized + Decentralized system

energy_animal_system_diagram

- So what do we use for centralized information system (global) machine to central point?
- So what do we use for decentralized system (local) machine to machine?
- How do these two imputs feed a database => map in a 2 way communication?

Swarming application

Status
Routine
Privileges and override



Pattern recognition

swarm trajectory prediction
- Thick red is satellite image.
- green is prediction based on computer models of currents and winds.
- thin red is where the oil really is, where the robots can sense it on site.

Routine

 Stop    
 Scan    
 Join    
 Follow    
 Eat    
 Return-full    
 Return-error    
 Hide    
 Avoid    
 Search    

eating oil for dinner
Human Override
Parameters :
    - emergency
    - greater density somewhere else
Comments