The underlying technology and initiatives powering autonomous vehicles (“big data”, cloud technology, machine-to-machine (“M2M”) technology and improved sensor technology - often collectively referred to as the Internet of Things (“IOT”)) is already part of today’s infrastructure. IOT presents some fascinating possibilities for autonomous vehicles, but manufacturers and partners need to be aware of real risks to their business and their customers.
Big data and sensor networks – creating the perfect driver
Data drives autonomous vehicles, and the act of driving produces even more. From the mainstream (GPS location data and cameras) to more left-field features (LiDAR/radar data), even mundane journeys collect crucial data to improve the autonomous vehicle experience. Human drivers are naturally limited to their own experiences – meaning that when an unexpected event occurs, it might not be clear what the best course of action may be. With a shared “consciousness”, driverless vehicles can, in theory, expect the unexpected and reduce the likelihood of accidents occurring. This data can also be used to improve infrastructure by allowing for a deeper analysis of driver behaviour than ever before.
However, the old mantra of “garbage in, garbage out” applies – poor quality sensors in a cheaper series of vehicles may lead to unreliable data models. Large-scale data collection can also raise some important questions regarding privacy, as we have previously highlighted as part of this series.
M2M – the glue that binds it all together
Autonomous vehicles require reliable, fast communications. Sensors may be able to collect thousands of data points per second and cloud computing may allow for complex calculations to be made almost-instantaneously, but without reliable, low latency (the time between data generation and delivery) wireless communications networks, driverless cars will struggle to achieve their true potential. Industry giants Nokia and Ericsson have both spoken of the importance of fifth generation (“5G”) wireless networks in driverless cars, reducing latency levels to under a millisecond and allowing for vehicles to be controlled remotely whilst on the roads. At the turn of the millennium, a broadcasting system allowed drivers to automatically tune into traffic announcements. Vehicle-to-vehicle communication could automatically relay traffic, weather information and road surface condition without human input, automatically correcting driving directions, speed limits or even suspension height. Google’s Waze software already provides community-based travel information to smartphone-equipped drivers.
Vehicle-to-infrastructure communication provides the greatest opportunity for M2M communications. Work has already started on the M6 “smart motorway”, automatically monitoring congestion levels and setting variable speed limits to ensure a smooth flow of traffic. Publicly-owned autonomous vehicle fleets could respond to reports of increased pedestrian footfall in certain areas or changing weather conditions. The key question is whether the UK’s mobile networks can realistically deliver upon manufacturers’ promises of “mission-critical” 5G connections in the Yorkshire Peaks, rather than in laboratory settings. Telefónica’s involvement in the Greenwich Automated Transport Environment project will certainly be one to watch.
The power of the cloud
Cloud computing allows for driverless cars to become a key part of the IOT. Valets could become a thing of the past as smart locks notify your car that you are leaving the building, music streaming could handoff straight to your vehicle and your smart thermostat could automatically adjust the heating in your apartment when you drive home from the office. The “smart car” is a natural extension of the “smart home”, but it brings with it a plethora of unique concerns.
Cloud resilience and service outages will provide particular challenges for vehicle manufacturers. Early adopters of IOT devices such as Nest have experienced significant service disruption issues, and drivers will be particularly aggrieved if their driverless car (for example) suddenly loses its navigation ability or ability to drive at higher speeds. Internet-connected vehicles are already experiencing significant security concerns - with the Nissan Leaf the latest car to be compromised remotely – and manufacturers will have to adopt a cautious balancing act between consumer desire for greater integration within the IOT versus the need for robust vehicle cyber-security.
Thanks to the IOT, driverless cars are far from a pipe-dream – but they are reliant on infrastructure and security improvements that may yet prove to be roadblocks.