The EU’s document proposing guidelines on the future approval of automated vehicles (version 4.1), although still in its early stages, should be of considerable
interest to all OEMs working in the field of automated vehicles.
While there are a significant number of detailed
design essentials, unfortunately many are of a far more generalized nature. In this week’s
insight, we select some of the more interesting requirements and consider how they may
impact the design strategies of the industry’s key players.
Determining what represents a sufficient lead time between
issuing a request and expecting the driver to take control is one of the most
challenging tasks in the field of automation. A sufficient or ‘safe’ lead time
depends not only on the vehicle’s immediate situation concerning traffic and
road situation, but also on the driver’s mental state at the time it is issued.
Most OEMs are developing Driver Monitoring (DM) technology that can analyze a
person’s biometric parameters and so build up a picture of their cognitive
and/or alertness state, but understanding the true effect of these on
determining a safe ‘take-over’ interval is still some way from being achieved.
Although this appears straightforward, there may need to be
a range of approaches for different people and situations. An audio message
cannot be the only means to inform as hearing-impaired people, who can currently
drive, would no longer be able to do so. Visual messages may be preferable, but
in order to be seen will need to be delivered to wherever the driver’s
attention is focused, or the driver's attention may need to gained with haptics (or by other means).
To meet this requirement, the automated vehicle will not
only need to follow basic traffic flow rules, but also interpret road signage
and other infrastructure-based instructions. This has already proved to be a
difficult area since there is a large variety of road signs, they frequently
appear with others, and their meaning can vary with context. While a human can
often interpret a combination of instructions that lack clarity, the process
they use to do this is complex and currently beyond most automatic recognition
technology. Signs which combine graphics and text further increase the
difficulty of securing an accurate interpretation. Combinations of signs can create a confusing picture even
within a single country. The complexity increases greatly when
vehicles need to be capable of cross-border travel and must, therefore, deal with
additional groups of local signage. A final point to consider is the use of
human-based instructions, such those delivered by a traffic officer using hand
signals to adjust traffic flow temporarily. Again, different countries and
conventions will need to be considered.
While the expectation placed on manufacturers from these
technical requirements is mostly clear, there are also some of a more open nature
that may cause additional debate. For example, the requirement that the
automated vehicle ‘shall not cause any traffic accidents that are
rationally foreseeable and preventable’ will give manufacturers much to
consider. The cybersecurity obligation that ‘Vehicle manufacturers shall take
measures such as those related to the updating of software, etc. …. to ensure
in-use cybersecurity over its lifetime ‘ may also cause some anxiety.
Overall, these guidelines will provoke many discussions
which, it is hoped, will lead to the effective and safe deployment of automated
vehicles which meet the expectations of their customers and other road users, but it is clear that OEMs and their suppliers need to pay close attention to the constantly moving legislative bar that is being set for them.