Regulating Autonomous Vehicles: Land, Sea and Air…

Whilst the phrase “autonomous vehicle” is quite possibly meaningless to many people, the notion of driverless cars seems to be rapidly leaving the world of science fiction (was it ever even a thing in science fiction?!), even to the extent of a RoboRace driverless support series being announced for the Formula-E (electric single seater race cars) championship in 2016-17.

Whilst putting together a general interest talk on The Current Future of Robotics for an OU residential school last year, it struck me that tracking the development of regulation around a technology might be a useful way of signalling those technologies likely to make some sort of impact over the next 3-10 years: if a technology is going to become widespread in the physical world, there are likely to be policy considerations as well as Health and Safety guidance notes, if not regulations, associated with it. (As an aside, it might be worth considering what sorts of regulatory protections are associated with the roll out of widespread digital technologies…)

So for example, towards the end of last year, the California Department of Motor Vehicles released draft regulations that “require a third-party testing organization to conduct a vehicle test to provide an independent performance verification of the vehicle. That is in addition to the manufacturer being required to certify the robot car meets safety and performance standards” (New DMV Robot Car Rules Prioritize Safety; Follow Consumer Watchdog’s Call To Require Steering Wheel And Pedals; Privacy & Cybersecurity Also Addressed.). You can see the regulations here: Autonomous Vehicles in California: Deployment of Autonomous Vehicles for Public Operation. Interestingly, accidents involving autonomous vehicles must be reported, and are listed here: Report of Traffic Accident Involving an Autonomous Vehicle (OL 316).

In the UK, a light touch is being applied to regulation around autonomous vehicles in the hope of driving (?!) related research activity. See, for example, the guidance that appeared last year (July, 2015) in the form of the DfT Automated vehicle technologies testing: code of practice. (A historical review is also available in the form of the Driverless cars in the UK: a regulatory review policy paper from February 2015.)

Associated work also reviewed the development of autonomous vehicles in a more general sense – that is, not just focussing on “robot cars”. For example, the Innovate UK network’s Robotics and Autonomous Systems Special Interest Group’s RAS 2020 Robotics and Autonomous Systems Roadmap [PDF] (July, 2014) identifies autonomous aerospace and marine robots, as well as autonomous road transportation systems, as areas worthy of development.

So when it comes to thinking about autonomous vehicles, what else is going on apart from robot cars and where might we be able to pick up regulatory or legislative signals that might enable – or announce – their widespread arrival?

In the aerospace domain, drones are being hooked up to ever more autonomous control systems, but when it comes to thinking of “autonomous vehicles” I think we should separate the levels of autonomy associated with how the system operates as a transportation system (e.g. in terms of navigation and control of the vehicle purely as a transportation system) and the operation of other systems mounted on the transportation system (for example, weapons systems). There are undoubtedly considerable concerns associated with the use of military drones, (for example, as reviewed in the University of Birmingham’s Policy Commission report of October, 2014, on The Security Impact of Drones [PDF] or the House of Commons Library briefing paper of Sept/Oct 2015 providing an [o]verview of military drones used by the UK Armed Forces [PDF]; see also the POST Briefing on Civilian Drones from October, 2014) but many of these relate to concerns about the uses to which drones might be put as platforms (for example, mounting surveillance systems or lethal weapons systems).

More generally, then, when it comes to considering the aerospace potential of autonomous vehicles, we might look to reviews produced by industrial programmes such as ASTRAEA (Autonomous Systems Technology Related Airborne Evaluation & Assessment), representing as they do a broader set of concerns, such as this review of the challenges facing the development of unmanned aerial vehicles, or this ASTRAEA progress report presentation from December 2014 reviewing the process steps associated with “[e]nabling the routine use of Unmanned Aircraft Systems (UAS) in all classes of airspace without the need for restrictive or special conditions of operation”.

Cars and drones, then… so far, so familiar. But in the marine world, too, there is much interest in the development of Maritime Autonomous Systems (MAS) (“smart ships”) (for an introduction, see this blog post from a policy advisor at the UK Chamber of Shipping). In terms of regulation, the UK Marine Industries Alliance: Maritime Autonomous Systems Group was set up to explore how a regulatory framework may extend to the marine environment; for example, The Development of a UK Regulatory Framework for Marine Autonomous Systems [PDF] and this Maritime Autonomous Systems (Surface) MAS(S) Industry Code of Practice.

As well as tracking regulatory activity, and any lobbying associated with it (for example, from folk like KPMG and their March 2015 report on Connected and Autonomous Vehicles – The UK Economic Opportunity [PDF], or this RAND Autonomous Vehicle Technology – Guide for policy makers from 2014), it probably also makes sense to see how interested parties might also be seeking to protect their own interests in the form of any insurance they may seek to take out, and how the insurance industry responds to such approaches. So for example, this 2014 report from Lloyd’s on Autonomous vehicles – Handing Over Control: opportunities and risks for insurance [PDF] is perhaps a useful place to start?