The automobile industry is set to be changed drastically with the introduction of automated vehicles. We sat down with Mr Richard Jinks, Head of Portfolio Solutions – Insurance at XL Catlin, who has been working with leading robotics researcher Oxbotica on perfecting the formula for driverless cars.
By Ahmad Zaki
Insurers and governments around the world are conducting exhaustive tests on driverless cars to nail down all the possible risks that might emerge. XL Catlin recently formed a partnership with Oxbotica, one of the leading software creators for autonomous vehicles and a member of the
DRIVEN* consortium.
Currently, XL Catlin is working on a real-time risk assessment tool to work in conjunction with autonomous vehicles. This tool will be first used with the DRIVEN consortium’s London-to-Oxford trial, aiding the vehicles involved to spot possible dangers on the road and to avoid or navigate around it as appropriate.
Getting in on the ground floor
The question of how insurance might be bundled with autonomous vehicles is still up in the air – motor insurance is obvious, but will cyber insurance be included as well? That is one of the key questions XL Catlin’s partnership with Oxbotica will attempt to answer.
Mr Jinks explained that their partnership is just another facet in XL Catlin’s aim to use AI to make sense of the immense reservoirs of data available and to generate new insights on risk. “I worked with Oxbotica for 18 months, with the key objective being the assessment and management of risk, so we will be prepared to support our clients when they want to adopt this technology,” he said. “It’s an opportunity for us to understand this technology as it’s being invented, instead of trying to catch up to it when it has already entered the market.”
He also pointed out that Selenium – the operating system (OS) that Oxbotica has created – will work not just for private transportation, but also in industrial vehicles, such as forklifts and cranes. “For a lot of our clients, it will help minimise risk. Reducing the human element in warehouses will allow for greater efficiency and safety,” he said, adding that other industries, such as mining, have already adopted autonomy.
Figuring out the risks
He also spent a significant amount of time with Oxbotica attempting to understand the difference between good autonomy and bad autonomy – all controlled by the OS and –which is determined by how well a vehicle reacts to its sensory inputs. “The thing about Oxbotica is that they already have risk tolerances built into their systems, so safety concerns, such as product liability and cybersecurity are more easily addressed,”
Other risks involve data collection and sharing, especially in terms of the AI within the autonomous vehicle ”learning” how to navigate its environment. “An autonomous forklift truck on the first day is very different from the same truck six months later – it will have learned a lot more about the warehouse it is operating in,” Mr Jinks said.
And whatever that truck learns, it can share with other trucks in the same fleet, no matter the distance. “So then we start running into questions such as ‘who owns the data’, or ‘how do we secure the data’? And if the data gets corrupted or tampered with in a truck in the US, what are the chances of it ‘infecting’ the entire fleet – which might be scattered across the world – and how do we minimize that chance?”
How a robot navigates
A robot asks three key questions, said Mr Jinks. “Where am I, what’s around me, and what do I do?”
While the current models of autonomous vehicles are loaded with sensors and cameras, allowing it to achieve a full 360-degree view of its environment, there are translation and interpretation problems, which software designers must take into account. “A robot won’t be able to tell the difference between a shadow and a solid object with the ease that a human can. That’s where Selenium comes in, to help it interpret the world around it.”
*The DRIVEN consortium is a collection of companies who have received a grant from the UK government in order to cement the UK’s position as a leader in autonomous vehicle. The consortium will trial these vehicles in London and Oxford for 30 months beginning April 2017, culminating in a commute from London to Oxford. A