People often ask us about the accuracy of Niu's plotted coordinates. This is a great question with a very lengthy answer. I'll give you the short version and some places to go if you're curious about the longer version. Believe it or not, as long as this is, this is the short version! Read more if you're interested...

We have two ways to compute Niu's coordinates: GPS satellites and Iridium communication satellites. In the ideal world, all of Niu's coordinates would be based on the GPS because the accuracy can be within about 10 meters (30 feet). Sometimes the rough conditions at sea make it impossible to connect with enough GPS satellites to get a good reading, which can reduce the accuracy to about 100 meters (300 feet) or higher. But that's not the whole story...

For a GPS reading, Niu must connect to at least three GPS satellites. If this isn't possible, we use our fallback plan: the Iridium communication satellites. Unfortunately, we've had to use these coordinates more often than we hoped (the next version of Niu, already being designed, will address these issues). When Niu sends its message to these satellites, they compute an estimate as to where the message was coming from (in other words, where Niu is located). These coordinates are provided along with a "computer error probability radius", or CEP radius. For example, this message might tell us, "There's an 80% probability that Niu is somewhere within a circle whose center is at [22.15205, -158.2968] and whose radius is 5 kilometers". The "80% probability" is a fixed number associated with Iridium's calculation, which is pretty complex.

So, we always hope to get the smallest CEP radius possible from the Iridium satellites (1 kilometer) because that gives us the best accuracy. But in reality, these numbers typically range up to about 10 kilometers (about 6 miles). Yikes! If we plotted these coordinates on the map, the errors in the data would make it look like Niu was zig-zagging all over the place (a lot more than what the wind and currents can do). But don't fear, math is here!

With some rather straightforward calculations, we are able to improve the accuracy and show a more realistic path on the map. This is done using a "weighted average" in which each point on the map is actually an average of the several points that preceeded it. Because points with a smaller error radius are more accurate, we give them more "weight" in the average than points with a bigger error radius. Once we've balanced all the numbers so that they have weights that correspond to their accuracy, we then weight them again so that the most recent points have more weight than the older points. We take the average of all these weighted numbers and voila, we have our point.

Of course, the more often we get points, the better the accuracy. We're currently getting points about once an hour, which will help us extend our battery life. When we see Niu's battery getting lower, we will reduce the communication rate to once a day so that we can extend the journey even longer.

What this all means, is that when you look at Niu's position and its path, you're looking at an estimated location that has some small error associated with it. In the near future we will add more details to the map so you know more about what this error is for each point. In the grand scale of the Pacific Ocean though, we have an extremely accurate knowledge of Niu's position.

If you'd like to learn more about these topics, please check out the following links:

Smithsonian: How Does GPS Work?
Wikipedia: Iridium Satellites

- Evan Rapoport

TrackBack URI for this entry

Comments (0)

Subscribe to this comment's feed

Show/Hide comments

Show/Hide comment form