Putting global positioning on the spot

September 01, 2016 // By EDN Europe
Tim Bonnett, Director at Alpha Micro Components
We’ve become used to knowing roughly where we are on the planet when driving or walking thanks to GPS, first made available to the general public in vehicle satnavs and walkers’ handsets and later in smartphones.

But is it possible to use the same sort of satellite positioning systems to help put us more literally ‘on the spot’, with a positional accuracy of centimetres rather than metres? With some careful engineering, it turns out it is. Anyone who has been walking in the hills with a specialist GPS will know the growing sense of relief that comes with getting out the satnav, watching it wake up and then produce an increasingly accurate map position as it acquires signals from more of the satellites overhead. When GPS was the only global navigation satellite system (GNSS) available, positional accuracy relied to a large extent on the number of GPS satellites your receiver could ‘see’ from its position, and its ability to pick their signals out of the background noise and process them into useful data.

GNSS has moved on since then, with the introduction of GPS alternatives such as GLONASS, the Russian system; BeiDou, the Chinese system; and Galileo, Europe’s alternative. GNSS receivers that can track more than one of these systems at once, and so have access to a larger pool of satellites, can offer better positional accuracy than those limited to a single GNSS system. They’re especially useful in areas where antenna positioning is difficult or some satellite signals are weak or even blocked.

More satellites in sight is obviously an advantage, but what else can we do to squeeze greater accuracy out of GNSS? One useful trick is differential GPS, in which one receiver is fixed in a position that is well known and used as a reference for a mobile receiver nearby.

Many of the sources of positional errors, such as satellite clock bias, orbital error, ionospheric and tropospheric delay, are relatively constant within a reasonable distance from the basestation, so if the receivers are relatively close, they will be subject to similar errors. Since the position of the fixed receiver is known, it can be compared with the position the GNSS calculates to reveal any error being introduced. This can then be used to correct the positional errors of the nearby mobile receiver. Real-time kinematics is a further refinement of differential GPS, using a fixed basestation, roving receivers, and a communications channel between them to share error data in real time.

One form of error that does vary over relatively small distances is the impact of signals travelling over multiple paths to the receiver, which tends to make it more difficult to work out exactly what the received signal should be. In surveying, one of the early adopters of differential GNSS techniques, multipath error issues are addressed by using expensive, high-performance antennae. But what if we want to bring centimetre-level accuracy to the positioning of robot lawn mowers, automated tractors, low-cost drones and similar applications?

u-blox has done some work on this, setting up a trial at an office block and car park in Tampere, Finland. The trial compared survey-quality antennae mounted on tripods to low-cost patch antennae with and without local ground planes on tripods, and a further patch antenna on a