Much has been reported in the media about autonomous vehicles. Indeed, some cars makers are even marketing their new models with names that imply the vehicles can be driven autonomously. Although the technology being integrated into vehicles is fantastic and highly sophisticated, we are really taking about driver aides here – not fully ‘driverless’ cars’. Active cruise control, lane guidance systems, emergency braking and even auto steering systems are all there to assist the driver, but ultimately the vehicle still needs a human driver to stay alert and in control.
So, when can we expect truly autonomous vehicles (AVs) in car showrooms? There are a huge number of challenges to overcome. Apart from the legal and liability issues associated with driverless cars, the technology itself has some way to go.
AVs will be fitted with a bewildering array of technology to help them find their way around, which in turn requires staggeringly complex data and sensor fusion. The AV will have LIDAR – Light Detection and Ranging – radar, proximity sensors, cameras, GPS and perhaps a geo-map of its environment. All these different sensors and systems generate a huge amount of data on which the AV has to make a split-second driving decision. For example, if it ‘sees’ something ahead in the road it has to calculate what the object is, its direction of travel and its speed before making a decision.
This is no easy feet as this decisions includes if the AV should ignore the object or input instructions to the steering, brakes or accelerator. And if it decides to deploy any counter measures, it has to be aware of if there is anything close behind the AV that might collide with it if the car slams on the brakes.Also,do all sensors and systems agree regarding the object in the road and the action to take?
AVs could even be charged with making a decision between the least damaging of two crash scenarios when faced with an unavoidable obstruction in the road. To match human decision-making the AV has about 0.1 seconds to assimilate, evaluate and decide how to react on the road. And I haven’t even mentioned bad weather, poor light, snow, fog or blinding sunlight which can all affect AV sensor systems.
A key expectation is that AVs will be safer on the roads but it is accepted that they will still be involved in fatal accidents. Some hope that AVs will reduce fatal accidents but gain the statistical confidence to achieve UK levels of road safety would take billions of miles of road testing per vehicle type.
Human drivers make mistakes and people do die in road accidents. While this is of course tragic, it’s an accepted fact in society and we do our best to reduce such accidents. But what would the public acceptance be of machines killing people in traffic collisions? Personally, I suspect the public would be highly intolerant.
This issue of liability when accidents occur is a huge dilemma for carmakers, who would have liability for the safety of their designs for the lifetime of the vehicle. Saying it is out of warranty after a few years just wouldn’t cut it when people are entrusting their lives to the vehicle. Regardless of the legal consequences of a fatal AV accident, the car manufacturer and the brand will come under scrutiny. A lifetime of liability is very different from selling cars with a three or even five-year warranty covering defects.
Fully driverless cars will need constant software and potential computer hardware upgrades, as well as constantly updated cyber security systems.
Even assuming all the technological, legal and liability issues are resolved, and AVs achieve a fatal accident rate that is somehow acceptable, do you want a fully autonomous vehicle requiring zero driver input?
Many people suffer from motion sickness so using the time in an AV to read or surf the net may be impossible, so nothing is really gained by being a passove passenger in an AV. Many people like driving and the sense of freedom it brings. Experts suggest that only 15% of cars sold globally in the year 2030 will be operating as level 4 autonomous cars and level 4 still requires a driver to be present, concentrating and ready to take over.
I don’t think we will buying fully autonomous cars anytime soon although I am sure complex technology will continue to be developed in the field of driving aides – but not driver replacements. We should embrace technology and develop driver training to meet the needs of future drivers, but I think we will be teaching the fundamentals of safe driving – for life – for many decades to come.