Vehicle Automation: Proceed with care

This is a guest post by Dr Scott Le Vine of Imperial College, London

Professor John Polak and I are very pleased to share our contribution to the Independent Transport Commission’s Occasional Paper series, Paper Number Five, which addresses the broad range of opportunities and tough challenges posed by Road Vehicle Automation.

Figure 1: The ‘Navia’ driverless vehicle designed for low-speed public transport application. It is being tested as part of European-Commission-sponsored demonstration studies: http://induct-technology.com/en/480.

There can be no doubt that we live in interesting times for the transport sector.  The technical capabilities of cars to operate increasingly-autonomously from direct human control are nothing short of mind-boggling.  Nearly two years have passed since Google posted video online purportedly showing a blind man being chauffeured – “no hands no feet”.  And the company has publicly stated its commitment to introducing fully-automated driving to the general public by 2018, thugh the precise form this could take has not been made clear.

Figure 2: Tyler Folsom’s ‘minimum autonomous vehicle’ – a kitted-out bespoke tricycle

We will certainly see automated-driving technologies advance incrementally in coming years.  We can be equally certain that this unstoppable force will increasingly bump up against the immovable object that is the legal and regulatory system.  Quite simply, there is a major disconnect between the pace at which events move in these two worlds.  As my colleague Bryant Walker Smith has revealed, there is not even a straightforward answer to the question of whether autonomous vehicles are legal – and one advantage GB has is a relatively flexible set of regulations for testing emerging automotive technology.

Clearly new ways of thinking will be needed, if we are to meet the aspiration that Business Secretary Vince Cable has expressed – to make effective use of automation technology to cut congestion, improve road safety, and position GB plc as a leading innovator.  Some of this new thinking will involve policy and will of necessity come from Government and civil servants, but other bits can only originate from industry – such as ways to take advantage of new forms of infotainment and retailing.

Figure 3: View of a vehicle’s surroundings as sensed by AutonomouStuff’s aftermarket products. By processing the raw data feed, the system has categorised the items in the surroundings into different types (represented by colour shades)

Through this Occasional Paper, we make a plea for a sustained and thought-through approach to the challenges of vehicle automation.  No outcome would be worse than a high-profile setback that sours the public, so great care must be exercised, and we must continually err on the side of caution.

In preparing this Occasional Paper, I took part in Summer 2013’s Road Vehicle Automation Workshop at Stanford University.  I hope you find the pictures below from the Workshop thought-provoking.  As my colleague Stan Young (University of Maryland) remarked, what is immediately noticeable upon stepping into Google’s self-driving demonstrator vehicles (besides the human operator sitting in the driver seat, waiting to take control if required) is how much the rough edges of the system have been smoothed off – quite literally.  The controls are designed into the vehicle’s steering wheel and dashboard with the panache of an expert product designer – this is a system nearing consumer-readiness.