Sunday, February 04, 2007

Search and Rescue, the US Coast Guard & Web 2.0

Doctor Gray, a renown technologist, disappeared while sailing off Northern California. The Coast Guard did a thorough search before calling off the effort; most search and rescue stories would have unfortunately ended here. However, Doctor Gray's colleagues and the public mounted an impressive effort that put technology to work to continue the search. Those still hoping need only look to the pantheon of "survival at sea stories" for inspiration.

So, how is this effort actually helping the search? By working on some of the components of the search and rescue equation. Although there is some art to search and rescue, there is a lot of science as well. There is a basic equation that drives all search planning which I have paraphrased. More info is available.

The Probability that you are searching in the right area
The Probability that you will spot the survivor
The Probability of finding a survivor
The civilian effort supporting Doctor Gray improves The Probability that you are searching in the right area. It's not easy to find the right spot to search in a giant ocean. A Search and Rescue planner has to make a best estimate of where the mariner might be so that he/she can focus limited resources. By posting satellite images on Amazon's mechanical Turk that anyone can search, more bandwidth is being thrown at the problem; now you don't have to be constrained by searching a small area. This is what people are calling the Web 2.0 portion of this. Perhaps closer to open source, the modular and self contained nature of each satellite picture allows the overall effort to be divided up between many participants.

However, the civilian effort is still constrained by The Probability that you will spot the survivor. Although you might be looking at a satellite photo of the right area, cloud cover, inexperience, carelessness could prevent spotting the survivor. By having multiple people look at the same image, carelessness can be screened out, but the clarity of the photo is still a limiting factor. A survivor once told us a Coast Guard helicopter had flown right over him. Night, high seas and rain, though, lowered the The Probability that you will spot the survivor.

So, is there potential here for the Coast Guard? SETI is a proven distributed effort. Crowdsourcing is being considered as a viable solution. If the Coast Guard wants to extend and formalize this effort, there are a few next steps starting with proof of concept. I'm not sure if there is opportunity here, but I hope it is evaluated to find out.
  • Validate the ability to search using satellites photos. Can the average person spot a sailboat? Could a computer program be made to do it? How often will weather prevent a suitable picture?
  • Gauge the interest of the public to help more anonymous mariners. Would civilians mobilize to help out a mariner without a large public persona?
  • Evaluate how quickly people would mobilize. Help needs to begin immediately as the chance for survival is a function of time.

One more comment on this story. I've read a few comments that the Coast Guard gave up and that simply gives the wrong impression. Sometimes the duration of disappearance, weather conditions and other factors make survival unlikely. In those cases, the Coast Guard has to weigh the extreme unlikelihood of finding a survivor against the danger to its own crews and the need to perform other missions, which include search and rescue for someone else. Believe me, as someone who has had to inform people that the Coast Guard was suspending a search, you do everything possible. It's not an easy decision and it certainly is not giving up.

No comments: