We’ve all watched TV programmes and movies that depicted funky new modes of transport for the future, from the Jetson’s flying car to Luke Skywalker’s Landspeeder. But in reality things seems like they’ve not radically evolved in transport. We still drive cars powered (mostly) by internal combustion and fly in planes with wings and engines. However, a revolution in transport is on the horizon, and it’s powered by AI, in the form of autonomous transport. It seems everything will be free of human intervention soon, from cars, to planes to trains and ships. Will this be the new reality or is it just pie in the sky?
Despite their high profile teething problems, including fatalities in testing, driverless cars are firmly on the agenda and already high on the wishlist for consumers. According to chip manufacturing CEO Jensen Huang, fully autonomous vehicles could be on the road as early as 2022, though Uber boss Dara Khosrowshahi said we’re still 10 to 15 years from “full autonomy”. For many autonomous vehicles will be an inflection point for AI in our lives, a gateway AI, where humans begin to trust AI, not just to complete a task, but to do so in a way that maintains human safety. We’ll need some convincing, with the first truly driverless vehicles likely to be ‘robo-taxis’ confined to specific campus areas that prove the concept beyond doubt.
For some the only thing scarier than a driverless car is a pilotless plane, but that’s exactly what’s on the agenda for major manufacturers, Boeing and Airbus. Boeing announced its intention to move to fewer and no pilots in 2017, partly to reduce the risk of human error and drive up efficiencies, but also to address the fact there may not be enough pilots to go round as the aviation industry grows in the next decade. Boeing sees its AI initiative as a logical next step beyond autopilot, with cargo journeys likely to see the first effects. It will be some time before we see no pilots in the cockpit of passenger flights.
The shipping industry has been squeezed by tighter margins and oversupply of vessels for quite some time now, so it’s no surprise that this industry is in the vanguard when it comes to exploring the savings available from autonomy and AI. Perhaps surprisingly, those leading the field are not traditional ship brands, with Rolls Royce and Google now partnering to bring autonomous shipping a step closer. Rolls Royce is using Google’s Cloud Machine Learning Engine across numerous applications to improve ship safety and develop better ships for the future. The challenges for AI here are different, as for a crewless vessel the AI will need to replace many functions, from piloting the ship to maintaining shipboard systems and managing cargo handling. The Engine will be used initially to improve existing image recognition systems used to identify objects and hazards during the voyage. Initial machine learning will be augmented by huge datasets captured from onboard cameras, sensors and scanners, with the insights gleaned then shared globally to develop best practice and processes.
Rail has been slow on the uptake in terms of driverless trains, but it has begun the journey to AI by improving other processes using AI tools. In the UK a joint project by the Universities of Huddersfield and Leeds is trialing AI software designed to make railways safer. Part of this study is scrutinizing rolling stock maintenance and the human visual inspections currently used to check the condition of components. This SMaRTE project unites human experience with AI software, with safety analysts interacting with the IT, teaching it the words and phrases relevant to safety management. By using machine learning, the program aims to reduce costs, improve safety and improve vehicle availability.
All of these AI initiatives in transport share one thing in common: they’ll all generate and rely upon vast amounts of data, from traditional and new sources. It’s fitting then that the datacentres and IT storage that will power this revolution are themselves at the forefront of AI adoption. Already HPE’s Nimble storage and InfoSight management software are implementing machine learning, minimising downtime and leveraging global datasets for stronger insights.