A start-up is deliberately trying to crash aircraft into drones to test a new collision-avoidance system.
US-based Iris Automation鈥檚 technology can detect, identify and react to airborne objects. The start-up says it can spot light aircraft 500 metres away, respond in a fifth of a second, and it works with the drone travelling at up to 210 kilometres per hour.
Traditionally, collision avoidance systems for drones are similar to those used for other aircraft, which rely on radar, but are large and expensive.
Advertisement
Instead, Iris Automation鈥檚 system relies on a single small camera and artificial intelligence to pick out objects that the drone could collide with. When it detects a potential collision, the drone drops by 50 metres, cuts its speed and starts circling.
To test the system, the start-up has set drones on a collision course with helicopters and other aircraft over seven thousand times, but so far had no collisions. 鈥淲e have flown aircraft from many different directions and trajectories,鈥� says Alex Harmsen at Iris Automation.
Companies like Amazon and Google are exploring launching drone delivery services, both of which have or are running limited trials. Harmsen says some of their early customers are in the final stages of obtaining licences to run trials in the US with approvals expected in the next two months.
However, for the services to go mainstream they will have to show there is little risk of a collision.
This is likely to be one of many such systems to enable operation in shared airspace, says Gary Mortimer, who runs an industry drone website.
They will open up the potential for widespread drone use, not just for delivering parcels and medical supplies but for further use by the emergency services, search and rescue, agriculture and industrial uses such as power line and pipeline inspection, he says.
