After two thousand years of GPS, what next?

GPS technology is probably older than you think, probably a few millennia years older. And back then it was more battery efficient. Let’s look at its history and some recent ideas that could make the current system 1000 times more energy efficient.

Longtime OpenSignal user Berthold Horn recently sent me a couple of papers on GPS that set me thinking. I often get frustated by my GPS, particularly when it takes too long to get a fix and then uses too much battery when it does. I realised that for something so important to the way I navigate the world I know surprisingly little about it.

In the beginning

The principle of GPS is to use the location of satellites to work out where you are, which is itself a very old idea. People have navigated by the stars for at least 700 years – tha Carta Pisana is the first solid evidence of this, it’s a maritime “dead reckoning” chart. It’s likely that navigation by stars is millennia old – for example the astrolabe dates back to 2nd century BC.


The Carta Pisana – a prototypical Global Positioning System, with the satellites being the stars.

App idea: At the moment, Google Sky Map uses your GPS to work out what stars you should be seeing and overlays their names on your screen. A reverse app could be built to work without GPS. In this app you would line up the positions of stars on your screen with stars in the night sky, using this information the app could calculate your location. It would be an Android astrolabe or sextant.

The Brick Moon

Man made satellites are a recent invention – just under 150 years old. The idea of an artificial satellite was first described in 1869 in E Edward Everett Hale’s “The Brick Moon”. You can find this story and others by Hale for Kindle on Amazon (we take no commission, it’s free!).

The story describes the building of an artificial moon of 12 million bricks and then flinging it into space using a giant revolving contraption which builds slowly up in momentum. Pretty nuts. But what is most remarkable about the story is that it not only envisions the first man-made satellite but effectively the first GPS satellite, for the Brick Moon is designed to orbit the earth by traveling from the North Pole to the South Pole along the Greenwich meridian and then travel back up along the antimeridian.

The purpose of the Brick Moon is to be “the blessing of all seamen – as constant in its change as it’s older sister has been fickle, and the second cynosure of all lovers upon the waves and of all girls left behind them”.

Incidentally, cynosure is a a gem of a word, it means “one that serves to direct or guide” (Merriam-Webster link) and derives from what the Greeks called Ursa Minor, the “dog’s tail” (and not little bear). To remember this just imagine yourself tripping over a dog’s tail while looking at the North Star. That’s cynosure to guide your memory.

Back to the Brick Moon; Hale’s idea is a definite precursor to the modern GPS satellite, the idea in both cases is to introduce an artificial body into the sky that can be used as a reference point.


Hale’s Brick Moon

In 1869 the Brick Moon was an advanced idea, but let’s take a look at the current state of the art.

Ephemeral Ephemerides – the reason it takes so long to get a first fix

Anyone who’s read the wonderful Longitude by Dava Sobel will understand the importance of knowing the time to calculating (you guessed it) longitude. In brief, it’s not enough to know the position of the Sun/Brick Moon/GPS satellite to know your location. Since obviously those things all change with time, a clock is needed to factor out this time dependence. Think of it like this: with a clock and some theory you can know at what longitude the sun should be directly above at a particular time, and with a sextant you know where the sun is relative to you.

GPS is very similar: satellites broadcast some information to your device telling it where they are and the course they’re on – this is called the orbital information, or ephemeris from the Greek for journal. Sending the ephemeris is one reason why it can take so long to get the first fix. The communication with the satellite takes place at only 50bps and therefore take 30s (unless you have AGPS … more below), but that data is typically valid for around half an hour.

Speeding things up: Assisted GPS

If you have AGPS, the ephemeris can download using your connection to the network (mobile or Wi-Fi), reducing the “time to first fix” to just a couple of seconds. If you’ve used your phone GPS abroad you may have wondered why it seems to take a lot longer to get a lock, and it’s probably because you’ve switched data off. A quick solution can be to turn on your GPS where you have Wi-Fi, and for at least half an hour it should be quicker for you to get a fix even if you turn it off in the intervening period.

One more tip: in the settings section of the GPS Test App by Chartcross there are options to clear an reset the AGPS, running them can troubleshoot many weird issues and ensures you have the latest ephemeris (and I think also almanac data – that’s another story!).

The future of GPS: satellites in the skies above, processing in the clouds

Receiving the ephemeris not only takes time, it also uses battery, and so does the post-processing of the data on the phone’s CPU. In fact these two things use a stunning ~99.9% of the energy used by your GPS. That’s why a group of Microsoft researchers have designed a system to process this data in the cloud Energy Efficient GPS Sensing with Cloud Offloading (Jie Liu, Bodhi Priyantha, Ted Hart, Heitor Ramos, Antonio A.F. Loureiro, and Qiang Wang).

Let’s repeat that again: GPS energy use could be cut to just a thousandth of current levels by just sending the raw data, then later connecting it to the ephemeris (by looking that up based on timestamps) and processing it in the cloud. 1/1000th. Liu et al. provide a great example of what this means in the paper:
“In other words, with a pair of AA batteries (2Ah), CLEO can theoretically sustain continuous GPS sensing (at 1s/sample granularity) for 1.5 years.”

The disadvantage with this is that you wouldn’t get your location in real-time – it’s processed later. But for a lot of purposes – e.g. tracking the locations of animal migration, this could be a huge development. But it could also be really useful to keep a log of one’s own movements or for crowdsourcing applications like OpenSignal.

The future of GPS: Positioning comes back down to Earth

In a future post we’ll discuss some ground-based alternatives to GPS, cell phone trilateration & triangulation, Wi-Fi position, and even using FM radio (which could be the future of indoor location).

Again our thanks to Berthold Horn for setting us thinking.

Read my story The Faraway,  for Kindle on Amazon, to learn how to navigate by the stars even when you cannot see them.

Vote on HN

This entry was posted in Mobile Trends. Bookmark the permalink.

Leave a Reply