What happens when drones, machine learning and video analytics unite? Next-level automation.

At any point in time, we’re installing, repairing or inspecting one of our 65,000 cell towers. AT&T already has contracted with vendors who conduct drone inspections of towers. Now, the video analytics team at AT&T Labs has joined forces with our National Drone Team to bring automated cell tower inspections a step closer to reality. With automation in the mix, we can do the job faster, better and more efficiently. With regulatory changes and further research, we hope that automated inspections will be possible.

Imagine in the future a drone stationed by a cell tower. It fully charges itself. It then inspects and communicates the condition of the tower – and potentially repairs it.

What we’re investigating is similar to what’s been happening with driverless cars. The ultimate goal is full automation, but we’re not there yet. We created a deep learning-based algorithm that analyzes video footage and shows promise in detecting defects and anomalies. We’re investigating how drones can be used to inspect these towers and feed HD video to our technicians, who can view the video in real-time.

Watch the video below to learn more about our collaborative Drone AI project at AT&T. 

Mazin Gilbert - Vice President of Advanced Technology at AT&T Labs

Mazin Gilbert
Mazin Gilbert Vice President of Advanced Technology & Systems at AT&T Labs