Graham Jarvis investigates how Europe is finding standards to dispel consumer fears of autonomous vehicles. [Auto.Jarvis.2016.03.18]
A recent study has revealed 70% of UK motorists feel nervous about the prospect of the first wave of self-driving cars.
The data collected by car supermarket HPL Motors Research reveals that only 38% of men and 20% of women feel positive about the prospect of being a passenger in an autonomous vehicle.
Therefore, one of the future challenges of car safety testing organisation, The European New Car Assessment Programme (Euro NCAP) is to demonstrate that they are safer than in comparison to a driver-driven vehicle.
In fact, any new technology has to be tested rigorously before it can gain a 5-star Euro NCAP rating and that includes connected vehicles which aren’t quite yet self-driving.
“Euro NCAP is what manufacturers can offer in terms of the new safety equipment and how effective it is in the real world,” explains Michiel van Ratingen, secretary general of Euro NCAP. He says his organisation is about setting a standard that vehicle manufacturers can follow to permit consumers to benefit from the latest safety technology at the lowest price. Its aim is to encourage legislation to be made to ensure that the most effective solutions become installed in every car as a common standard.
He adds that Euro NCAP is leading up to the driverless car stage but there is more work to do before a star rating can be created specifically for autonomous vehicles. “Euro NCAP operates more closely to innovation than legislation does, which lags behind, whether this involves testing and rating a self-driving function, automatic braking or lane support systems.” In essence he explains that Euro NCAP is forward-looking; he claims that it is interested in what will benefit consumers in three to 10 years from now and it sets out to advise consumers about what they should demand in their vehicles today.
Anders Eugensson, director governmental affairs at Volvo adds: “EURO NCAP is pushing manufacturing to improve safety standards.” He adds that most vehicles these days are connected, but they aren’t used for self-driving. In his opinion direct short range communication (DSCR) is being discussed around the world but Volvo doesn’t support it for a number of reasons. For example, he claims that Volvo doesn’t see the need for it in autonomous vehicles “since the autonomous drive cars do not need DSRC and V2V, and because Volvo believes in systems and equipment that are built into the autonomous drive” without relying on other vehicles.
He adds: “Volvo knows that the roll-out of DSRC systems will take a long time and before the penetration will be high enough to the point, whereby you can secure a reliable signal from other vehicles.” He thinks this will take decades to achieve. Yet, if car manufacturers build in all of the systems needed to have autonomous drive in each vehicle, the market penetration of the systems becomes less critical. Volvo’s approaching is, therefore, to build in a number of sensors in its vehicles in addition to Cloud connectivity. He claims that having a detailed 3D map downloaded into the vehicle for the purposes of positioning and referencing “is not much use or needed for autonomy”.
“We have a system called Pilot Assist that helps us to stay in lane and to keep a distance to the car in front,” he explains. At the moment there is also a discussion within Euro NCAP which is about considering whether or not it is safe to have the driver as the fall back, to take over the control of the vehicle like an airline pilot would take over control of an aeroplane in the event of an incident occurring that you endanger the car and its passengers. “There are systems that give the impression that the car will deal with anything but Euro NCAP won’t encourage this because these systems will enhance safety so long as the driver is still in the loop,” he claims.
Euro NCAP Advanced
Meanwhile the complimentary nature of Euro NCAP Advanced “is almost like a gentle encouragement for auto manufacturers to install safety systems as standard,” explains Chris Davies, head of technical superiority at Autoglass. He finds that there are lot so systems that currently aren’t so widespread at the moment but Euro NCAP wants to see more of them. “So Euro NCAP Advanced is painting a picture of where they want the manufacturers to go and my view is that governments tend not to like to legislate unless they have to do so,” he says. So in the UK, for example where there is a relatively small vehicle manufacturing base, he believes that it’s better to shape it via EU legislation to continue to allow Euro NCAP to be a tool to change vehicle design and safety.
“Euro NCAP Advanced was set up in 2010 to demonstrate the safety benefit of some technologies, as a way to show the value of some innovations which we believe may have potential for the future,” adds Van Ratingen. In the past Euro NCAP Advanced has been used for things automatic emergency braking (AEB) but is a separate form of recognition to the Euro NCAP star rating system, which seeks to recognise innovation “based primarily on a presentation of a data provided by manufacturers to show how the system would prevent a real-world crash”.
To achieve this, automotive manufacturers ask Euro NCAP to assess their findings to see whether they actually offer any real safety performance benefits and then, if the manufacturers’ claims prove correct, Euro NCAP can publicly reward them. However, he explains that Euro NCAP Advanced isn’t a star rating but a price to show whether a technology is really innovative or not. This shows the way to developing further safety systems.
Cyclists and pedestrians
In terms of testing and rating systems for the detection of cyclists and pedestrians, Matthew Avery – head of research at Thatcham Research, describes the process. “Effectively the traditional way is that you look at the innovation, look at its real-world effect and encourage other manufacturers to apply it.” He claims that the technology is very good at avoiding crashes but this, as a result, makes it harder to measure its effect in a crash. For this reason a black box approach is taken.
“We can then design a test procedure to define how cars, pedestrians and cyclists react to estimate the likely effect by using a lot of science behind the tests,” he explains. This involves designing the tests to enable them to accurately complete real-world scenario tests to replicate people’s reactions. The results offer car manufacturers some guidelines to enable to them to design to develop effectiveness, and he says they offer a means against which the technologies can be assessed and compared.
Real pedestrians aren’t used in these tests and so they have to design a target to enable the testers to understand how the vehicles’ sensors are designed and what they are looking for in order to react to different scenarios. He adds: “Yet, because the frequency of car and pedestrian crashes is very low, it takes time before anyone can say that the technology is working because it’s difficult trying to measure crashes that haven’t occurred!”
Next generation ADAS
Avery claims that the next generation ADAS will occur in 2018: “We will have enhanced car-to-car AEB, introducing additional test configurations to cover a greater proportion of the real world crashes such as different crash overlaps, angles and manoeuvres.” He says manufacturers are working on this now and Thatcham is developing an ADAS test against a head-on crash with another vehicle where you leave your lane and have a head-on crash with something. “Research suggests that these crashes are common, so we are encouraging autonomous emergency steering based on a threat while next expecting any affect with lane support systems.”
Euro NCAP in fact finds that lane support systems can be very annoying. “They annoy the drivers by constantly nudging the vehicles back into a lane and we have studies about why these systems might not be working,” he explains. As a result of being annoying about 50% of lane support systems are turned off by drivers. They intervene whenever they don’t need to do so – such as when the vehicle is intentionally driven over a white line. To resolve this issue a new threat level is being added to the test procedure by providing the driver with a vibration through the steering wheel to remind the driver and whenever there is a real threat it will warn the driver.
“The challenge is for vehicle manufacturers to have a camera and radar sensors combined to see a car coming the other way on the road and identify the crash situation, and we have to design a test target to use as the opposing car to see and to make sure it does no damage to either car,” he says. For this to be accurate, he adds, the target has to have an appropriate radar signature and visual attributes to replicate the ongoing vehicle. There has to be 3D robotic car target to test potential accidents at different angles so that the radar and camera feels that it’s actually a real car.
So, for the moment, it’s the technology that offers driver assistance rather than autonomy that are driving the market today.
11 Jul 2016 - 12 Jul 2016, Tokyo, Japan
This event will be an information and networking platform aiming to bring together decision makers in the ADAS and autonomous arena from OEMs, Tier 1s, ADAS tech providers, map makers, IT tech providers, to discuss the challenges and establish consensus to enable the next phase of driving safety and autonomy.