Within the field of intelligent transport systems (ITS) the accurate positioning of vehicles via satellite is crucial. According to the Satellite Positioning Performance Assessment for Road Transport (SaPPART) COST Action, GNSS is “expected to deliver many benefits, including reducing congestion, increasing capacity and improving safety. The road sector is estimated to represent more than 50% of the GNSS market and 75% when we consider the mobility services on smartphones“. There are, however, challenges which remain to be solved.
Portal met with the chair of this Action, François Peyret, at the conference ‘Intelligent Transport Systems: a Tool or a Toy?’ in Žilina, Slovakia, to discuss some of the ways in which satellite-based positioning systems will play a role in the transport systems of the future and some of the challenges involved.
To begin, why is the standardisation and certification of GNSS for road transport important?
Firstly, GNSS will be the heart of all positioning systems used in the road transport sector. This is true not only for navigation purposes, because when it comes to any advanced driver assistance system, the accurate positioning of both the vehicle and indeed the other vehicles on the road will be vital.
Indeed, positioning is really at the centre of ITS, and GNSS will be in any positioning system, even if it is used alongside other sensors and technologies – which will be necessary due to some of the limitations that GNSS has, such as a degradation of signal in so-called ‘urban canyons’ in cities. Thus, there is a need to hybridise GNSS technology with other sensors (which are now available) and which can even work independently from GNSS – such as LIDAR. This will be crucial in order to ensure an accurate position.
GNSS is a quite complex system. To characterise the position you obtain via a GNSS system, it is necessary to have a knowledge of that system, which will provide you with your position, speed and time; you also need to have good metrics to measure those features so as to be able to assess their performance; and you need to have complex tests for the specific conditions you are operating in. Indeed, one cannot generalise the accuracy of any given GNSS system; it is only possible to say that it is accurate to whatever degree in a specific condition and at a specific time.
However, there is currently nothing available to be able to define the metrics and tests and to check them, because this is an area that remains under-emphasised in the transport and ITS fields. There is thus an urgent need to investigate this properly, because in order to use a vehicle which will use a positioning system, the performance of that system will have to be certified; to do that, standards will need to be developed in order to know which feature to test and which metric to use to measure it. We are currently working on that within the standardisation group at CEN-CENELEC.
Given such endeavours as Galileo, what is the current state of GNSS and other satellite positioning services in Europe? Does there need to be more improvement, more work to enhance capacity, performance, or coverage?
Yes, there is a need for improvements, and work is already happening to help their realisation. When we first started looking at positioning, GPS – the American system – was the only system available. Now, of course, there are others: GLONASS (Russia), BeiDou (China) and, of course, Galileo in Europe.
The Galileo constellation has experienced several problems and has been delayed considerably – although this has tended to come down to politics rather than technology – and there is now a very real chance that BeiDou will be finished first. This could cause significant problems for Galileo, particularly when it is considered that the electronic chips it uses are not compatible with their GPS or GLONASS.This could prove to be problematic because the manufacturers of receivers may not see it as being worth their while to adapt their technologies. Nevertheless, Galileo will be a good system, but from the point of view of the end users and the manufacturers, the main concern is the efficiency of the system, and that could mean that Galileo will not be used if it doesn’t bring anything new to the table.
What could help Galileo is the inclusion of functionalities that are unavailable for the other systems. Such services, of course, would be in addition to the basic, free civil service equivalent to that provided by GPS. These could include commercial services, such as the provision of a very secure system with authentication capabilities, which allow the user to know that the signals they are using come from Galileo (this would be of use to the security and military sectors). A further commercial service could be the provision of so-called ‘precise point positioning’ (PPP) which uses a phase of the satellite signal to provide centimetre-level positioning accuracy. Galileo is in a prime position to do this as it will be utilising numerous frequencies.
In Europe there are those who are working on PPP, and progress is being made, which is imperative because it could prove to have important applications in the field of ITS and autonomous vehicles, where signature accuracy is of paramount importance.
What are the biggest scientific challenges when it comes to autonomous vehicles and the utilisation of GNSS services?
Perhaps the most significant challenges concern safety, and there is a need to have what is known in the civil aviation sector as positioning ‘integrity’. That is, you need to know that you are where you think you are – if your system is inaccurate by, for instance, a metre, this can be quite dangerous.
Safety very much depends on the integrity of the positioning, and integrity is the confidence you have in your position, which is generally measured by the maximum error you think you have – known as ‘the protection level’, which is a concept borrowed from the air transport sector. As long as it can be ensured that the actual error is below this threshold, then safety can be maintained. Achieving this level of integrity can be tricky, and using the best of the available information necessarily means combining different sources of information.
As smart city concepts evolve, will the necessary hybridisation of systems for automated driving evolve alongside it?
These two areas are indeed linked, but, for the moment at least, they are evolving in parallel. Smart cities will, of course, involve a network of sensors, but such networks are not typically built to support autonomous vehicles; they have other objectives.
In the future, vehicles – whether they are autonomous or not – will be connected (that is, they will have the capacity to talk to one another) and be co-operative (what they do with this connectivity); they will be connected to the infrastructure in particular. In cities, vehicles are almost always surrounded by objects, and so to support satellite positioning systems, these objects can become ‘active’. Sensors located in certain places can transmit their position and these can be used to complement the GNSS positioning.
How will the positioning issues come to be addressed moving forwards?
I think we will come to see prototypes emerge which are able to use GNSS to navigate precisely – and not only those which make use of PPP. Indeed, I am involved in some European projects which are already working in this domain.
When it comes to autonomous vehicles and ITS more generally, the future is uncertain. There is a distinct sense that automation is an area that has now become very popular. Following the Google initiative, many different car manufacturers have turned their attention to the field and are proposing partial automation systems, but they are still far from being commercially available. Indeed, when Uber showcased its autonomous vehicles they were equipped with €50,000 worth of sensors. That, clearly, is a long way from being anything close to a level which could see a wider rollout.
Before this happens, vehicles will go through various degrees of connectivity and co-operation, but, of course, these things are not as attention grabbing as vehicles which drive themselves, and so tend not to feature in the media spotlight.
Dr François Peyret
SaPPART COST Action
This article first appeared in issue 13 of Horizon 2020 Projects: Portal, which is now available here.