IEEE Consumer Electronics Magazine - October 2015 - 105

based on available data from all the entities in the vicinity,
and precalculate/weigh reactions of taking an evasive action,
or better, preventing critical situations from arising altogether. In this context, it will be necessary that a vehicle in an
imminently dangerous position communicate in some way
with those outside, in particular, with the non- or lesser technologically equipped VRUs.
In addition, the ethics of who owns vehicles or which
parts of vehicles and the software code within them is an
ongoing debate. Did an algorithm cause the accident, or is the
physical hardware it controlled faulty? John Deere wants to
own its tractors due to software licensing issues [25]. This
may (pardon the pun) backfire. Manufacturers who insist on
owning vehicles purchased by others may change their mind
if liability becomes aligned with ownership. If software is the
main controller of automobile behavior and the software
algorithms are owned by manufacturers to protect their intellectual property, then liability may also be in their custody as
well, and they may not want that.

SOCIAL CONTRACTS
In current manual contexts, social contracts are what enable people to cooperate in the traffic space, and crucial traffic situations
may mostly be avoided by adherence to these. Each driver roughly
agrees to obey traffic rules and infrastructure measures (unless
they decide to take agency in certain situations). These rules and
measures are, for the most part, mitigated through implicit communication between the traffic participants through their bodies
(e.g., eye contact, waving gestures, and learned behavior) and
through their operation of their vehicles (e.g., turn signals, horn,
brake lights, and emergency flashers). Traffic flow is maintained
through the exchange of behavioral intentions as humans agree to
cooperate through the contract of participating in the driving system. However, in a future with automated cars at different levels
[15], human-driven vehicles, and VRUs [26, p. 6] existing side by
side, the abilities of humans to de-escalate conflicts and critical situations are limited. The limitations are due to interacting human
agencies, situation-specific complexity, and automation effects on
humans (like being out of the loop) along with the properties of
how automation itself functions in relation to other circumstances.
As a consequence, automated vehicles and future road users must
possess the skills of cooperation and negotiation within the car
(i.e., between automation and its driver), with other road users,
and with traffic/ITS infrastructure [3], [6], [22]. Social participation within a cooperation negotiation requires agency, for all parties have to agree that they will cooperate for successful
cooperation to take place. On a road with mixed levels of automation, sociability, whether real or constructed, is vital.

CONCLUSIONS
The environment for automated vehicles for the next few decades
will exist within a mixed vehicle and pedestrian/VRU context. In
addition, there will be an added layer of third-party apps running
on mobile devices that may or may not be integrated into vehicle
interaction controls. This will add a layer of heterogeneous complexity to the interaction model that will not be easy to resolve.

However, initially, the concern will be broader, applying to concerns with how automated vehicles operate on roads as a small
minority alongside otherwise human-operated transportation.
Gradually, this will change to include a larger proportion of automated vehicles. Because the context of the proportion of automated vehicles will be in flux, it will be difficult to base design
decisions on historical data and methods. What has transpired in
the past with regard to errors and performance will simply not be
useful data for forecasting future parameters, and new design
methods and benchmarks will need to be developed alongside
new vehicle paradigms.
The transition complexity from mixed autonomy is compounded by the legal need to establish clear responsibility and
liability when a mishap occurs. As people are unlikely to assume
this responsibility in the long term as passengers in self-driving
vehicles, and corporations are likely to resist absorbing this

Whatever solution evolves, it is
likely to involve much more
information flow between vehicles
as well as outside to monitors.

responsibility in full, more collective solutions will need to
emerge, where responsibility and liability are distributed beyond
the immediate vehicle and manufacturer of the vehicle. There
are several means to achieve this, including the following:
▼ Centers could be established that use people and assistive technology to control vehicles in traffic. This will be challenging,
given the heterogeneity of the traffic composition, although it
would not be unlikely that legislation could be applied to force
nonautomated vehicles to incorporate communications equipment and conform to the requests of controllers.
▼ Another prospect would be the establishment of floating
dynamic autonomous controllers for groups of vehicles. Individual cars might be under the partial control of several such
controllers as the vehicles progress and break off from a
group. Again, this might require some kind of accommodation for manually driven vehicles.
▼ Though a privacy issue that will require established trust,
smart sensing networks that can anticipate driver needs [19]
might be used to improve quality aspects on the local driver-
vehicle interaction level, such as driver convenience, driving
experience, ergonomics, user friendliness, and other factors.
Whatever solution evolves, it is likely to involve much more
information flow between vehicles as well as outside to monitors. This means that automated vehicles will not be autonomous, in the sense of behaving on their own, but rather elements
of a dynamic group whose composition is determined by an
authority outside the domain of the vehicle. These may, in part,
be determined on a commercial basis, as a function of the state
october 2015

IEEE ConsumEr ElECtronICs magazInE

105

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - October 2015