AI Technology Helps Drones Sense and Avoid Obstacles

Iris Automation, a Vancouver-based startup, raised $1.5M to bring their technology to industrial-type drones.

Using NVIDIA GPUs on the Amazon cloud to train their deep learning models and a Jetson TX1 onboard the UAV, Iris’ technology analyzes and draws insights from videos captured by the drone’s cameras in real-time. Its system has a detection range of over 500 meters, significantly farther than what others can currently do.

“We’re designing this to work like a human pilot’s vision and decision-making process,” says Iris Automation CEO and co-founder Alexander Harmsen.

Iris Automation’s artificial intelligence module combines 3D maps and real-time images to avoid mid-air collisions.

The company mentions that industrial drones need a special collision avoidance system versus consumer drones, since they need to cover long distances and capture video of expansive infrastructure that doesn’t show up on a street map, such as a change in weather or new construction.

An Iris investor believes their technology is applicable beyond drones and can be used for other situational awareness systems in cars or ships.

Read more >