Re: [Minutes] Auto WG 2018-11-20

The following are some of my thoughts on representing cars on IoT and
what has formed that thinking. With the holiday weekend I am only now
sending them.

Our spec work on signals is all about exposing a common data model and
means of accessing it. The intent to date has not been about making
vehicles directly accessible over the web and I think most would agree
that would be unwise from a privacy and security point of view.

As we embark on a second version we are being careful to give
consideration to the multitude of powerpul potential use cases that
would likely seek to leverage it. Having more capable apps that can
access data and interact with the user, their media, location etc on
the head unit has been our primary goal since we got started. What has
been drawing more attention of late is how our work can be used to get
data off the vehicle and into the cloud for things like preventative
maintenance, urban/traffic planning, fleet management, insurance etc.
This attention has warranted forming a task force in the Business Group
to see what else we should be working on in addition to the signal spec
work to facilitate these broader needs.

I wrote a blog previously on the forming of this data task force.


https://www.w3.org/community/autowebplatform/2018/06/18/w3c-automotive-big-data-task-force/

In this task force we have been looking into representing data sampling
methodology as metadata, adding value to what otherwise could be a
meaningless random sampling, capturing consent, privacy (GDPR)
considerations, starting to look more into policy language, an ontology
is being applied on VSS and we identified some areas we will defer on.

The different data needs for different audiences (eg fleet management
verus insurance - vehicle or usage based) will need to be represented
in policy language and part of the consent capture piece. Who is in a
position of authority to provide permission to access to what specific
information for collection and to send to which parties for what
specific, intended purpose or redistribution.

It is clear future vehicles need to interact with various things in
V2i/V2v/V2x, IoT and Smart Cities. Again a common data model and means
of accessing it on-board would be beneficial for being able to
communicate to the corresponding standards for V2x etc as too would our
data task force work. A car is a rather complicated thing in the
Internet of Things, so much so I would argue it is not necessarily a
single thing. A car can be different things based on the purpose
similar to how we are discussing carving up sets of signals in the data
task force.

To communicate with a traffic light via V2i/SmartCities it may only
need to provide location, heading and speed but not things like engine
temperature nor fuel/charge level. A car can measure road conditions
and be a roaving sensor for municiple governments to dispatch trucks to
lay down salt and sand or locations of potholes to be addressed.  It
may use engine temperature, outside temperature, oxygen intake,
location and other data points to be a roving air quality thing in IoT.
Additional after-market sensors can be added to measure actual number
of particulates as discussed on the IoT call. We have some
considerations from both VISS and ViWi approaches for after-market
addons providing additional or replacement signals information. I have
heard of replacement TPMS units that collect additional signals
information from the wheels that just the pressure.

As cars are often offline and one would want to access historical data,
we might want vehicles as things on IoT to be going through data
warehouses in the cloud. This model could also support close to
realtime for some times of interactions. Daniel expressed a possible
gateway solution. I would like to hear more from others on this topic
on this thread or subsequent calls.

On Tue, 2018-11-20 at 18:12 -0500, Ted Guild wrote:
> https://www.w3.org/2018/11/20-auto-minutes
> 
> I will be updating and including a link to Jean's slides later.
> 
-- 
Ted Guild <ted@w3.org>
W3C Automotive Lead
http://www.w3.org