Stuff already at Lagrangia?

> I had looked into this as a possible Ph.D topic many moons ago. (I was
> thinking about infrared searches of Sun-Earth L4&L5.) (It took me a little
> while to find the 17 year old article - and my wife wonders why I keep
> these things! :-)

Very interesting. I've been thinking about Sun-Earth L4&L5 searches, too.
I've tried to come up with novel data analysis tricks for pulling signal
out of noise in such a search based on the techniques I used on my
dissertation. Unfortunately most tricks we used were a bit specific,
so not applicable in this particular area. Anyway, one trick we used
that might be handy here is this:

Instead of averaging the photon counts, we collected a raw time series,
basically a long sequence of blips. Working backward from the expected
dynamics (which were simple in this case), we were able to pull signal out
of noise, based in part on taking the time to really accurately
characterize the noise, and in part on a clear idea of the signal
we were looking for.

Applied to an L4&L5 object search, a generalization of this technique
might yield good results. The key idea is to use high sensitivity IR
detectors, and look right at L4 or L5. Collect the photon counts from
each pixel as a raw time series (memory and bandwidth intensive, but
both are rapidly advancing, and prices are dropping fast). You now have
a set of photon counts resolved in space (angle, really) and time. This
is a huge data set, but again, it's tractable given current (or near-term)
technology (the only thing that I'm not 100% sure of is the readout rate
of the CCD). Now you can go back and digitally reconstruct what the photon
count *would*have*been* had you been tracking a particular volume along a
given path. Obviously you choose a path that corresponds to an orbit about
the lagrange point. Effectively you use the computer to recreate the
photon count of a virtual CCD substantially smaller than the real one, but
moving just right to keep the object centered on a single pixel.
The advantage is that you get to look at lots and lots of different
orbits. Another advantage is that you can integrate the observations from
different times separated by weeks if needed, provided the digital
compensation is handled correctly. For imaging larger regions than are
allowed by the size of your CCD, you can build up a set of data by looking
at regions around the point of interest - all the compensation for the
evolution of the orbits of interest in the time between observations is
handled digitally.

The ability to get really long integration times by superposing
observations over many, many nights also helps.

Anyway, I'm not an Astronomer, so I may be reinventing the wheel, but
it certainly seems to me that this sort of search could work well. I'd
be interested in hearing what you think of this. Actually, if it's got any
merit, I'd be interested in pursuing it further.

......Andrew

Andrew Case
acase@...
Institute for Plasma Research
University of Maryland, College Park |