The Galileo, Glonass and Compass constellations will improve satellite navigation for everybody - if the receivers can be made to work.
Satellite navigation systems take their location cues from 30 GPS satellites that circle the Earth twice a day transmitting status, date and time, and orbital information. Soon there will be around 100 satellites to lock on to as GPS is joined by global constellations from Europe (Galileo), Russia (GLONASS), and China (Compass).
GPS wasn't built to help us find our way to the shops - it was a Cold War project funded by the US Department of Defense to ensure that nuclear submarines could surface and target their missiles accurately. There are strategic rumblings about the new satellite constellations too, but the current consensus is that civilians have most to gain from more accurate and reliable location and tracking applications. That's if receiver designers can get the power consumption under control.
Russia's GLONASS system used to be famous for its satellites failing faster than they were launched, but since last month it has had 24 functioning satellites in orbit. Meanwhile, Europe's much-delayed Galileo system will have 14 satellites operating by 2014, according to the European Commission, with the full 30 available by 2017. The US GPS system is being modernised to become GPS III by 2013, with additional navigation signals for both civilian and military use. Information about China's Compass system is sketchier - it was going to be a regional system but is now understood to be global.
'All this activity is great news because whatever the application, there will potentially be multiple constellations to get a position fix from, which will help with signal integrity in safety-critical environments such as maritime, aviation or rail, and accuracy for mobile phone users in urban areas,' says Andrew Sage, director of Helios, a consultancy specialising in satellite navigation.
A GPS receiver should be able to 'see' at least four GPS satellites anytime, anywhere on the globe and establish three position coordinates (latitude, longitude, and altitude). But in city streets hemmed in by tall buildings, a receiver is unlikely to be able detect more than two satellites and the signals will often have bounced off structures.
'For the average pedestrian, the position fix can be a long way out and very unpredictable,' says Sage. 'Most users don't see that today because GPS receivers match us to maps and smooth the errors out. But if you are walking around a city and not on a road in a car, multi-path reflections are a problem.'
The more satellites visible from within these 'urban canyons', the easier it is to carry out consistency checks on the received signals. 'Even when you can't isolate the multipath-contaminated signals, the more signals you have, the more your errors average out,' says Dr Paul Groves, lecturer in global navigation satellite systems (GNSS), navigation and location technology at UCL.
Better GNSS integrity would enable new applications, such as road-user charging, enforcing bail conditions and pay-as-you-drive insurance. 'Clearly, if position information might be used as legal evidence, it has to be reliable,' says Groves.
The delayed arrival of Galileo and the resurrection of GLONASS have complicated matters for receiver makers. Galileo was designed to offer the simplest possible upgrade path from GPS to a dual-constellation system. Agreements were made to put the carrier frequencies of the main open services in the same part of the spectrum as GPS, at around 1575MHz, so receivers could share the same radio, analogue components and antenna. Both systems also send their signals using a spread-spectrum code-division multiple-access (CDMA) approach. GLONASS uses a frequency-division multiple-access coding technique (FDMA) and a main open-service carrier frequency of 1602.2MHz.
STMicroelectronics, which makes satnav chipsets, has a working GLONASS/GPS prototype codenamed GGG implemented as a field programmable gate array (FPGA), which it expects to offer as a dedicated chip. Because the GLONASS carrier frequency is offset by 30MHz and each GLONASS satellite uses a different frequency, ST has developed a wide-band receiver with a second path for the GLONASS signals. 'We have a double down-conversion in the GLONASS part of the receiver and then we pass the GPS and GLONASS signals into the base-band via two interfaces. We process the signals separately because we have to do the channel selection between all the GLONASS satellites,' says Philip Mattos, chief engineer of GPS and navigation.
GPS at your fingertips
However, supporting multiple GNSS constellations in battery-powered systems, presents a dilemma. 'Today when you launch GPS on a handset and watch it track, within an hour the battery is dead,' says Nick Taluja, head of connectivity at ST-Ericsson, who has tested multiple handsets.
Multi-GNSS systems with clever software will be able to resolve the multi-path problem in urban canyons, says Taluja. But even with new receiver architectures and a shared baseband, he questions the practicality of multi-GNSS support in phones. 'Phones are not getting any bigger and neither are batteries,' he says. 'Every time you power-on GPS and GLONASS and Galileo and Compass, a 45-minute battery time will drop to 20 minutes.'
Kanwar Chadha, founder of the GNSS chip company SiRF (now part of Cambridge Silicon Radio) says: 'Even when GNSS systems are somewhat harmonised, you still have slightly different processing in terms of the modulation schemes and error correction. To support multi-GNSS systems you need a powerful processing engine.'
Firms developing chips and IP for battery-powered systems are making plans to support multi-GNSS with hybrid schemes that use sensors already in phones and cameras such as compasses, accelerometers and gyros. These approaches also help to bring location technology indoors, where GNSS doesn't work.
'If you go into an urban canyon or indoors and lose your satellites, the GPS knows where you were last, the on-board accelerometer knows what speed you are travelling, and the compass in what direction,' says Taluja. 'You can estimate where you are, but using power consumed by the mobile phone.'
An EPSRC-funded project called Innovative Navigation using new GNSS Signals with Hybridised Technologies (iNsight) may set the tone for multi-GNSS navigation and tracking. The four-year collaboration between Imperial College, UCL, the University of Nottingham, the University of Westminster and industrial partners including EADS Astrium, the Ordnance Survey, Leica Geosystems, STMicroelectronics, Thales, QinetiQ, the Civil Aviation Authority and Nottingham Scientific Ltd aims to extend use and improve efficiency of positioning. The project is designing GNSS sensors and radio systems, modelling data error sources and looking at the integration of GNSS with other position-related inputs such as inertial sensors.
Paul Groves is developing algorithms for picking out multi-path signals. Even with additional constellations, he says, urban multi-path effects can limit positional accuracy to more than 100m. New approaches to signal processing will be required to get to within 1-3m. (The best a consumer GPS receiver might currently achieve is 5-10m). One idea is to use the extra civil frequencies that next-generation constellations will offer. 'The relative strength of a reflected signal can vary from frequency to frequency so each time you add a frequency, you add one unknown and three measurements: the code, the carrier phase and the signal-to-noise ratio. With a second or third frequency, it is easier to calibrate out the multi-path errors,' he says.
By the end of the project, the partners hope to have a software-defined GNSS receiver platform capable of converting all the different constellations' frequencies and handling any complex numerical processing tasks in the digital domain.
'Then you are you not limited to Galileo, GPS, GLONASS or Compass signals one at a time. You could acquire them all simultaneously, as well as receiving WiMAX, Wi-Fi and GSM signals, so that in future projects we could integrate all these signals to aid the positioning,' says Professor Izzet Kale, who runs the applied DSP and VLSI research group in the school of electronics and computer science at the University of Westminster.
Kale's group is working on various receiver architectures, one of which was the subject of a paper given at the IGNSS2009 conference last December and which is now undergoing extensive tests in the lab. 'We have also considerably reduced the complexity of acquiring signals in the frequency domain in a further work package linked to the iNsight project,' he says.
Helios' Sage predicts that by early 2012, mobile phones will be on sale supporting both GPS and GLONASS and based on sub-dollar chipsets. But the most interesting opportunities, he believes, will be for a new generation of software-defined receivers, along the lines iNsight is exploring, that can be upgraded as constellations become available.