Print This Post

Global Infrastructure for Driverless Vehicles

Originally published on LinkedIn:

IoT and Driverless Cars

Driverless vehicles are fast becoming a reality. Several jurisdictions, such as the State of Nevada, have begun the process of developing the requirements for such vehicles to operate on our roadways and highways. So far, most of the discussion around such vehicles has focused on sensors, such as video and GPS, and on the artificial intelligence required to safely operate the vehicle using the inputs from those sensors.

Awareness of local issues, like sensing the boundaries of the roadway, determining the distance between vehicles, and the detection of obstacles, is critical; however, a truly successful driverless system will also need to consider global issues as well, including the position, velocity, etc., of other vehicles on the road, along with their intended routes, objectives (e.g. deliver a parcel), the state of vehicle resources such as parking spaces, and the status of the roadways on which the vehicles are travelling (things like surface conditions, obstructions, and road closures).

In terms of the Internet of Things, this means taking a broader, rather than narrower, view of IoT. Information infrastructures to support IoT are just as critical for success here as they are in other IoT domains.

Much of this is reminiscent of the System Wide Information Management (SWIM) systems being developed to support commercial aviation. As in SWIM, we can anticipate a global registry of vehicles tracking each vehicle’s position, velocity, and respective time history. Also like SWIM, we can anticipate the development and deployment of web services for things like road weather, road conditions (e.g. road closures, road surface degradation due to construction, weather), and conditions at parking facilities. The latter will clearly require the development of “parking procedures” which govern access to available parking spaces if chaos is to be avoided.

The new “rules of the road” need to go beyond the ability to sense that a space is available to include negotiating with other vehicles for that space. Since, for the foreseeable future, driverless vehicles will still need to operate with human drivers, these vehicles will, like their human counterparts, need to behave in a regionally-compliant manner. It’s clear that different “rules of the road” apply in New York City than in Lincoln, Nebraska, and driverless vehicles will need to be aware of these regional differences. The ability to adjust to local rules of the road will be especially true for vehicles like trucks, that operate across multiple jurisdictions.

To achieve such global objectives will require the use of open standards and, in particular, open standard registries. These will include registries of vehicles, registries of roadways and roadway conditions, and registries of “rules of the road” for vehicle global behaviour.

To create these open standards, we can build on existing Open Geospatial Consortium (OGC) standards such as the Web Registry Service Standard. This standard leverages the OGC/ISO Geography Markup Language (GML) standard for object geometry which, through Extension Packages, can easily be used to describe vehicle position and velocity, roadway geometry, points of interest, roadway surface conditions, as well as all the various regional rules of the road. The general approach to using this standard is described in “Towards a Universal Data Interchange Standards Factory”.

We can imagine a future driverless system infrastructure with vehicle registries, real-time roadway registries, web services for roadway weather, and registries of available services. But to conceive a driverless vehicle future without such a system infrastructure is simply not possible.