Hacker News Reader

Hacker News

Comment by pedalpete | Hacker News Reader

Full thread

pedalpete

Can somebody explain to me how "pings narrowed the location of MH370 at that moment to one of two arcs"?

If a ping is picked up, shouldn't it be somewhat definitive of where it came from?

View on HN

Not really. Let's look at what is generally known:

1. We may if we are lucky know the time in transit.

2. We may know the signal strength.

Unless you have the same signal being intercepted by multiple (minimum number is 3) satellites, you can't triangulate. From a single satellite, you can probably extrapolate from either of these to some extent the location, however you can't do this with any precision beyond a full circle arc.

A good parallel might be earthquake detection and the location of an epicenter. Both these involve tracing a location based on the strength of weakening waves reaching detectors. To locate an epicenter, you need detection of the earthquake by three independent detections. One gives you very little information. Two gives you a line of possibilities. Only with three do you get a chance to determine exactly where the waves originated from.

Now if you know signal strength, and you have two detections, you know it is somewhere in a line based on signal strength, but you can trace it to two more precise possibilities if you have time in transit. Based on the information we have it sounds like they must have two satellites picking up on the signal and comparing times in transit.

There's only one satellite, so they only have (as far as we know) the delay time between sending out the ping and hearing a response to determine distance from satellite to aircraft, with no ability to triangulate the position as something like GPS (which uses many satellites) would do. So they can figure out a large curve around the satellite's position, and the aircraft could have been at any point (within the limits of its fuel capacity) when it sent out that final ping.

Distance alone should give you a full circle, but remember you also have processing time which is significant.

If they have two possible arcs, this suggests that the same ping was heard by two satellites (and no more), and that these are the places where the difference in time that the satellites heard the pings would work for both satellites.

You have basically three intersecting spheres (one is the earth and the other is based on the satellites' light cones). They all meet in two locations. If you had a third satellite, you should be able to determine which of the two locations was right.

You're misunderstanding. There is a single circle, which people are calling two arc. That is because they know the airplane can't fly more than a certain distance from its last known position (this reduces the circle to an arc) and because they know where the airplane isn't from radar data (that breaks the one arc into two pieces, which happen to be roughly the same size).

GPS relies on extremely precise timing as well, and extremely precise knowledge of the orbits of the satellites. If you have 10 millisecond timing on signals that narrows things down to a 3000km range of distances.

It seems similar to how GPS is used to find location - you only know the distance from each satellite, not the direction. Contact with one satellite places you on an arc (or circle), two sats put you on one of the two spots where the arc around each intersects, and three isolates your position to a single intersection of three arcs at established distances from each satellite. (And four gives you altitude.)

The satellite just recorded the time the signal took to get from it to earth and back. That gives you a circular arc of places.

The interesting bit in the article is to scroll down to "GOODFELLOW REBUTTAL...Interesting analysis of my analysis." in the comments.

Reply on news.ycombinator.com