Collaborating For a Safe, Autonomous Future Article written by Dr Sam Chapman
Every press article or news report you read about ‘driverless’ cars or autonomous mobility seems to be accompanied by an illustration of people relaxing in the back of a ‘pod’, which looks nothing like one of today’s passenger vehicles, proceeding down a motorway or through the busy streets of a city. The passengers are sat facing each other, relaxing, chatting, or playing games. Disengaged from the driving process, they are completely confident in the safety of their vehicle. When I see these illustrations I am often struck by the enormity of the task to transition the mobility options we have today into a future where this sort of driverless vehicle could be the norm. But before we run away with how hard it’s going to be it’s important to remember that semiautonomous features have been present in the cars that we’ve all been driving for decades. Automatic braking assistance and cruise control are types of semi-autonomous features that we use without thinking. However there’s a big difference between a feature that can help the driver and those that allow them to relinquish control of the vehicle entirely. Importantly, the defining element in moving to full autonomy is human behaviour.
Many proponents of autonomous mobility say that if every vehicle on our road was fully programmable then this could make our roads safer. After all, human error, human negligence or poor decision making is considered to be the most frequent cause of accidents and collision. However, what about the fact that our roads are also populated by pedestrians crossing the street, or cyclists weaving in and out of traffic? They are still just as likely to make errors in their decisions which could lead to accidents. Some of the most notorious incidents and fatalities concerning autonomous vehicles have happened not because the vehicle’s system being used has failed, but because a human road user has made a mistake, for example, stepping out into the road without leaving enough time for the vehicle to spot them and take evasive action. It is these variables of human behaviour that makes the introduction of full or part-autonomy such a complicated thing to get right. And it’s this very complexity that we have been addressing as part of the UK Government-funded MOVE_UK initiative which has just concluded after three years of work. MOVE_UK’s objectives have been to complete the ‘connected validation’ required to properly understand the parameters required for a safe, autonomous future and to fuel helpful thinking around how the nature of risk will be considered when this is a reality. The project has gathered an enormous set of frequent and wide ranging vehicle, sensor and vehicle control usage data across thousands