Insurance in the age of drones - The technologist
26 Oct 2016
Robots are a technology developing at pace. We are living at a time when a confluence of technical advances – battery and power provision, computational ability and sensors – are allowing the rapid advance of robotics. Moreover, through continual self-optimisation, robots are able to become self-referential, and learn adaptively from their surroundings. Within a well-developed systems architecture, many robots can work collectively with each other on a common task. This collaboration is not centrally co-ordinated, it is the product of distributed autonomous intelligence.
Watch the interview with Raffaelo D'Andrea:
Raffaello D'Andrea has spent his career working with such robotics, highlights of which include the creation of Kiva Systems, which became Amazon Robotics, and regular participation in the Robot Soccer World Cup.
Drones, in this context, are robots that can move in three dimensions rather than just two. As such, their potential application is significant. They can be used to survey; to capture images; to aid reconnaissance in a war situation; to help within search-and-rescue; to survey crops, and to spray only those areas that require spraying; and, perhaps their most hyped quality, to deliver goods. Drones can be large fixed wing craft capable of carrying heavy payloads long distances; they could be micro, insect-sized cameras able to access tight confined spaces. The limits of drone technology are confined only by our imaginations.
From a technological perspective, there are two challenges of drone technology, which do not receive much media coverage:
- Battery efficiency: Getting power into a drone to make a journey is cheap. The journey becomes more expensive when the fixed cost of a battery is included. A typical drone battery can only be re-loaded 50-100 times, before it needs replacing; and replacing is not cheap. We are not currently on the verge of a breakthrough in lithium battery technology. For drone producers, it is therefore important to make their drones more efficient, reducing component weight or allowing for changes in structure. This fact also suggests it is currently only efficient to transport goods of high net worth. Drone pizza deliveries may have to wait.
- Safety and reliability: The potential for injury and damage from a drone is significant. One YouTube video shows a drone hitting a ski slope just seconds after the skier it was filming had passed. Drones have to target a mean-time-between-failure rate that is similar to that of commercial aircraft. To ensure this, drones built in Professor D'Andrea's lab are designed to be fail-safe – if one component fails, the drone has the ability to remain airborne. In terms of collision avoidance, drones are best placed to operate both through their own sensors, and through the order of an air traffic control system.
We do not yet see drones walking to work. In order for them to become part of our every day environment, they require public acceptance. There will be accidents. If a bad accident occurs before we are familiar with the ubiquity and the utility of drones, it could set the industry back several years in its development. Insurance will play a role in this process. It must evolve together with the industry, so that all parties can learn and develop in realising the many benefits that drones offer society.
See also:Drone event in October 2016. D'Andrea is Professor of Dynamic Systems and Control, Swiss Federal Institute of Technology (ETH) Zurich. Summary written by Simon Woodward.