Cell Tower Inspection

PSU researchers at the Robotics and Internet-of-Things lab, in collaboration with GM6, developed a system for inspection of telecommunication cell towers from beginning to end. The system helps mobile companies to ensure the better quality of telecommunication networks. It is based on recent advances in Artificial Intelligence and drone imagery. It is based on 3 main functionalities. First: Building an accurate 3D model of the cell tower. Second: Automatic detection of the antennas using A.I. Third: Automatic inspection of every antenna apart. Fourth: Generation of the inspection report of the cell tower.

The Robotics and Internet-of-Things Lab (RIOTU Lab) is the first research lab to be established in Prince Sultan University, under the College of Computer and Information Sciences back in 2017. Prof. Anis Koubaa currently leads the lab. The RIOTU Lab deals with multidisciplinary topics related to robotics and internet-of-things. In particular, it has four main overlapping research tracks: (i) deep learning and artificial intelligence, (ii) unmanned systems, (iii) internet-of-things, and (iv) emerging technologies.

One of the promising areas of research in the RIOTU lab is related to the Telecommunication industry. We founded a project to automate the inspection of cell towers using AI and drones. Cell towers are the main component of Telecommunication networks. Cell towers need to be inspected regularly to ensure the strength of the cellular signal. Usually, the inspector climbs the tower, and inspects it manually if there are any damages, which put them at a high risk as the operation is dangerous and may results in fatalities.

Our system solves this problem. It is an easy-to-use solution to inspect cell towers without any manual labor. We only need a drone to fly around a cell tower and take images. The rest is automated using A.I. to generate a detailed report about every antenna in the cell tower.

Our system is divided into 4 parts:

• Building an accurate 3D model of the cell tower
• Automatic detection of the antennas using A.I
• Automatic inspection of every antenna apart (anomaly detection, deviation, …)
• Generation of the inspection report of the cell tower

Part 1- BUILDING AN ACCURATE 3D MODEL OF THE CELL TOWER

In the RIOTU lab, we have the expertise to build an accurate 3D model of the cell tower.
A set of tools and methods are pipelined for this goal.

Part 2- AUTOMATIC DETECTION OF THE ANTENNAS USING A.I.

The A.I engine in the system is responsible for automatically detecting the antennas. For every detected antenna, the A.I. engine delineates, classifies, and extracts the associated 3d model.

Part 3- AUTOMATIC INSPECTION OF EVERY ANTENNA APART

The Inspection engine is responsible for measuring: the down tilt angle, the azimuth angle, and the plumb angle. Also, it generates the width, height, depth and altitude of the inspected antenna. These values are used to put the antenna in the best position.

Part 4- GENERATION OF THE INSPECTION REPORT OF THE CELL TOWER

Finally, a detailed report is generated about every antenna in the cell tower. This report heavily helps mobile companies to maintain the good status of their mobile networks.

The solution automates the inspection process for the cell towers. Inspection reports are saved to easily track the status of the cell tower from one visit to another. It limits the manual intervention on the cell tower site only to necessary cases