Advanced Driver Assistance Systems: a solution or another problem?
First publishedin ITS International
ADAS have been the subject of much development and study, however new education and training means need to be found for the mass market
Do Advanced Driver Assistance Systems represent a positive step forward for safety, or something of a safety risk? Jason Barnes discusses the issue with leading industry figures.
Advanced Driver Assistance Systems (ADAS) are already common. Anti-lock brakes or electronic stability control are well understood and are either fitted as standard or frequently requested by new vehicle buyers. More advanced ADAS features are appearing on many top-end vehicles and the trickle-down has already started. Adaptive cruise control (ACC) and parking assistance systems, for instance, can now be specified on several mid-range vehicles.
The number and complexity of ADAS is set to increase appreciably and lane departure warning, lane change assistance, blind spot detection, collision avoidance and driver drowsiness detection – to name but a few – will soon be ubiquitous.
A positive or a negative?This would seem to be a good thing: robust, dependable technology will make up for humans failings and result in a safer road environment. That’s the theory: the practice may not be as clear-cut.
The multiplicity and complexity of newer ADAS are giving rise to concerns that the lobbying of technologists and a desire to maintain auto sales is being allowed to take precedence over safety. A key issue is driver distraction, though over the years this has been the subject of significant research resulting in sophisticated Human-Machine Interfaces (HMIs) designed to reduce sensory overload and make the simultaneous use of multiple ADAS intuitive and safe.
The stumbling block is how to translate developers’ familiarity into a more widespread appreciation of ADAS’ features and capabilities. Without that, a driver’s first encounter with more advanced, capable and intrusive systems could be quite startling. It may even introduce a safety risk, to the extent that some are suggesting the need for a moratorium on systems’ introduction until more formalised training is in place.
Gaps in knowledge Demand for ADAS is rising. UK motoring organisation, the RAC, produces an annual Report on Motoring which garners opinion from a representative cross-section of motorists. The 2013 edition notes that ADAS are very firmly on people’s radars, according to Technical Director David Bizley.
“The general opinion is that we ‘ain’t seen nothing yet’; the expectation amongst road users is that over the next quarter of a century we’ll see ACC, parking assist and a greater level of speed control by the car itself. A notable number of people expect to see intelligent collision avoidance and even self-driving cars within that period.
“Motorists recognise the benefits of such technology but a majority also feel that they need to be shown better how to use it. Vehicle manufacturers need to improve clarity in this respect and there is a feeling that driver training should include instruction on new systems’ use.”
Maxime Flament, Ertico – ITS Europe’s head of sector for SafeMobility, notes that automotive manufacturers have worked hard to ensure drivers know intuitively what a vehicle is doing through well-designed HMIs. There is an ISO 26262 standard and a Code of Practice (CoP) covering the development of ADAS, which has been endorsed by the European Automotive Manufacturers Association. The standard describes from a functional perspective how systems must be designed and is intended to prevent systems from catastrophic failure by providing redundancy in operation.
“The CoP is an attempt to protect car manufacturers from liability law suits. The main design guideline is always to have self-explaining interactions with the driver,” says Flament.
“Several manufacturers have already undertaken major programmes of ADAS-related training within their dealership networks, including showing sales staff how to demonstrate the new technology to customers. However, it remains harder to explain ADAS than, for instance, trim options.
“But there isn’t a need for compulsory standards for vehicle handovers. Manufacturers will have their own guidelines and legally the car manufacturers fulfil their duty of care by adopting the CoP and providing exhaustive owners’ manuals.”
A problem, says Mark Fowkes, senior consultant at the Motor Industry Research Association (Mira), is that ‘design for proactive use’ comes up against the fact that few people ever read owners’ manuals.
“Vehicle systems have to conform to regulations in terms of performance but they also have to be understood and valued by the driver - especially if the systems are optional. If the application and delivery are ‘clunky’, how is that going to affect acceptance – and how does acceptance vary according to culture and demographics?
“We have technology roadmaps which consider increasing levels of ADAS all the way up to fully autonomous vehicles but we need to look at the effects of multiple systems being in operation at the same time. The effects can be counter-intuitive.”
Drivers’ perceptions of an ADAS can be quite different to those of technologists. Fowkes points to Intelligent Speed Adaptation (ISA) trials conducted around a decade ago by Mira and Leeds University. These involved simulations, short-term road trials and larger-scale fleet trials.
Maxime Flament, ERTICO – ITS Europe
“The approach used for early trials to control maximum speed was relatively crude – we interrupted the ignition system to govern power outputs so when the ISA system cut in, the engine ‘stuttered’. We, as engineers, weren’t keen on that but users liked it as they perceived that it told them not to drive any faster. That led us, in later trials, to design in a similar tactile feedback as it provided an increased level of user comfort, even though we no longer needed to.
“Just a few years from now we will have to ask questions about systems involving greater intervention in the driving task. Autonomous driving is very much in the news at the moment but what happens when an ADAS reaches the limits of its capabilities and needs to hand back to the driver? Potentially, the person in the driving seat has to change, instantly, from relaxing to controlling a vehicle doing 80mph. It’s not just speed we have to consider – there’s lateral control, acceleration and so on. There are big human factors issues to address.”
Aftermarket devices offer a slightly disruptive influence, too: “Dash-mounted systems which provide sophisticated ADAS-type applications such as ACC and pedestrian detection are already available and we’re already gaining driver feedback on which applications are most appreciated. Headway information is valued over lane departure warnings, for example – drivers might not know that they’re too close to the vehicle in front but they do tend to know if they’re drifting out of lane.”
Piecemeal delivery Andrew Parkes, chief scientist, Behavioural Studies at TRL, highlights the absence of work on policy and legislation for driver testing, licensing and training for ADAS. There’s simply no master deployment plan, he states, with applications arriving in the market piecemeal according to their technical maturity.
Some claims being made for HMIs’ intuitive nature are highly dubious, he adds: “Some, more advanced, HMIs, such as the vibro-tactile ones which are directionally coincident with stimulus which are found in lane departure warning systems, merely provoke an alerting response, not a guided response. Also, many new and emerging ADAS have not been tested with large groups of people including different sections of the population. Given the heterogeneous nature of society, there are presumptions which I don’t think hold true – not all old people struggle with a common set of driving problems, for instance.”
Like Fowkes, Parkes highlights issues such as culture: “Some of the big European ADAS projects have shied away from dealing with this.” He highlights points such as where people undertake their driver training, language – which can be a massive issue at the point of sale training – and even religion. All of these can have big effects on people’s levels of trust of a system or comfort as to where the locus of control lies between the driver and the technology.
“A lot of designers also choose to ignore the huge gap between ease of use and ease of learning. The issues associated with the latter stand in the way of full utilisation of some really quite clever technologies. Modern in-car entertainment and route guidance systems are very simple to use once set up properly but because people don’t learn how to do that, they often barely scratch the surface of features and capabilities. We cannot simply rely on owners and drivers reading vehicle manuals. A lot more evidence is needed of how drivers react to ADAS.”
Back to learning So, how to address the education issue? Maxime Flament says that there are already significant efforts in place: “The FIA already has wider campaigns ongoing on new features such as emergency brake assist and lane departure warning. Also, the European New Car Assessment Programme star rating system has been revised from this year and will make it more difficult to get that much-cherished five-star rating if ADAS are not included in a new vehicle’s design; that’s going to do a lot to raise awareness of what ADAS are and do. More attention would always be welcome depending on national policy on road safety but, at a European level, there already are some pretty good public-awareness mechanisms.”
A good vehicle handover at the point of sale is essential, adds David Bizley, but the person receiving the brand-new vehicle may not always be the driver and, in the case of fleet or rental vehicles, for example, a car may be sold on rather quickly.
He sees social media as a means of getting the necessary information out, not least because in many countries we no longer see the types of big, government-funded public information campaigns that were common a few decades ago.
Bizley envisages something akin to the European Union’s waste management legislation, whereby each stakeholder along the value chain has a percentage responsibility for the overall solution.
“A car manufacturer selling to an individual has a certain level of obligation to that individual in terms of making the right information freely available. But it isn’t reasonable to expect that manufacturer to take responsibility for subsequent owners or fleet drivers.
“I see responsibility throughout the value chain – hire companies, for instance, have to do rather better than simply ejecting customers from a bus having delivered them to their rental vehicles. In the case of second-hand vehicles the manufacturers should at least facilitate access to information. The internet and websites make that easy to do.”
Overall, says Bizley, drivers’ generally positive attitude to new ADAS solutions needs to be supported by manufacturers making knowledge of the specifics of systems less reliant on sales staff. Perhaps, he thinks, we need to take a fresh look at the whole car-buying experience.
“People will travel rather farther to look at and buy a vehicle than they will to get a vehicle serviced. Perhaps more centralised sales centres are an answer, with servicing and parts provision given over to a greater number of smaller, more widely dispersed facilities which are not necessarily in premium retail locations. Training could then be given at the sales location by those who are trained to teach, not just to sell.
“In Germany, it’s not uncommon, particularly among the premium brand manufacturers and tuning houses, to allow customers to pick up a new vehicle direct from the factory. The experience is turned into something of an occasion which ends with the new customer driving away but it also provides an opportunity for familiarisation and training.
“But this all needs to be supported by greater awareness of ADAS’ utility. Without that, buyers won’t look to specify them. Manufacturers’ advertising messages have shifted in recent years – we’re seeing greater emphasis on vehicles’ intrinsic intelligence – but there’s more that could be done. The big motoring organisations need to start flagging this as an issue in the mainstream press.”
Thinking more widely Mark Fowkes highlights the absence of formal strategy, especially where second-hand vehicles are concerned: “New drivers can be taught at the point of first learning but with existing drivers – many of whom have several decades of engrained habits – access is a problem. The association of education with social time via social media may prove more acceptable than a more traditional academic book-learning approach. The internet can certainly help but, ultimately, finding a single champion may prove difficult.”
Andrew Parkes wonders whether we need a much wider range of vehicle categories on drivers’ licences, reflecting the current situation with HGVs. So, for example, if an individual learns to drive on a vehicle equipped with ADAS to a given standard, he or she might then only be licensed to drive a vehicle of a similar standard. There might also be, he notes, a need for minimum standard of test vehicle but this could prove difficult, as many people still learn and are tested in their own cars.
“People trot out the use of social media and viral marketing as some form of saviour but I fail to see any good examples of where this has yet worked in this area,” he says. “The motoring organisations continue to punch above their weight in terms of getting the message across [to the public] but I think that governments and their agencies are notable in their absence.
“The EU’s Horizon 2020 framework for research will include work items to look at usability issues but with some of these technologies starting to arrive on 2014-15 model year vehicles that’s a little late. I also worry that international standards committees are ill-designed to cope with the situation. It took three years to harmonise on a time-to-contact distance of 1.8 seconds for ACC, for example. That’s ludicrous – especially when one considers that at the same time OEMs were already working towards production systems. The standards definition process runs the constant risk of simply being overtaken by technology.”
Training and legislation Many people consider the FIA to be a potential saviour.
In 2008, it conducted a mystery shopper study of 500 car dealerships during which buyers specifically requested more information on eSafety and ESC. That study demonstrated, says Jacob Bangsgaard, director general of FIA Region 1 (EMEA), that car dealerships are woefully under-prepared to demonstrate and sell ADAS. Only 18% of the 500 car dealers visited in the survey actively provided information on eSafety technologies and only one in seven salespeople focused on safety as an important vehicle feature.
“The sales people we saw simply failed to explain and sell the technology to us. The emphasis tended to be on styling and closing the deal swiftly, which isn’t helped by having to explain complex concepts such as advanced braking systems at the point of sale.
“Some manufacturers have said that education is a dealership responsibility and that the public doesn’t want to pay for technologies such as Electronic Stability Control. That’s patently untrue; our own studies and the European Commission’s Eurobarometer study on eSafety have shown that safety is a preoccupation of car buyers and that they are willing to pay for a safe car.”
An issue is the wide variation of systems development within different auto manufacturers. Top-end marques producing highly featured vehicles tend to want to promote their technical prowess for commercial advantage, whereas budget vehicle brands prefer to divert attention away from their offerings’ relative paucity.
“The car manufacturers simply don’t have a common interest,” Bangsgaard continues. “Another problem in an already fragmented market is branding, creating confusion of the functionalities behind the different ADAS names. As a consumer organisation, the FIA is left with the task of transferring that knowledge to the buying public because consumer involvement sometimes is neglected at the development stage.
“Admittedly, it’s very difficult to explain ADAS to users in a static environment but the more technologies you have, the more important it becomes to have the right HMIs. I’ve been in vehicles with eight different ADAS on board, and even though I knew what to expect I still got confused by the proliferation of warnings I got. I could manage because I know the technology but what chance does a 70 year-old person have? You need to get only the most essential warnings. I’m not convinced it’s always that safe to be distracted by numerous visual and acoustic warnings at the same time. Throw in navigation and infotainment, smartphones and so on and driver distraction becomes a key issue – perhaps even the main cause of accidents. We might avoid a greater number of less serious accidents with applications such as low-speed emergency braking systems but bombarding drivers with too much information is a concern.”
The solution, he feels, is maintaining tight control of HMIs and training drivers.
Some FIA clubs, in conjunction with Tier One suppliers such as Bosch, Continental and TRW, already operate training centres for professional drivers. This includes training police drivers as well as truck and camper van drivers. Bangsgaard also notes that countries like Austria, Finland and Luxembourg have mandatory second-phase driver training which individuals have to undertake shortly after passing their initial tests. The emphasis is less on making the individuals more skilled drivers and more on making them aware of the technology and the cars’ capabilities and limitations.
“It’s about balance. There’s little point in giving people training which only makes them better drivers. That has the potential to make them over-confident,” Bangsgaard says. “Similarly, training must also concentrate on the ADAS likely to be encountered in the near term – there’s no point in training people to use technology that they may next see in a decade’s time.”
Second-phase driver legislation is something that the FIA has been working on with the European Commission for some time now and Bangsgaard hopes to see tangible results within five years.
Work is also going on for longer-term systems: “Ultimately, ADAS developments are taking us towards cooperative driving systems. The FIA is now promoting the concept of the ‘five-star driver in a five-star car on five-star roads’, thus drawing together all elements of road transport.
“That’s going to need more involvement from road operators and organisations like Ertico have been doing a great job here. For instance, there’s no point in having a top-flight vehicle which can read information from the infrastructure if that information isn’t available. Advanced in-vehicle systems can’t just be mandated – we need the infrastructure peoples’ input, too. This is another area where the EC is working with us on securing the needed conditions.”