Afraid of heights? Drones and AI to the rescue!

As 5G roll-outs continue to gather momentum worldwide, network complexity and site numbers will grow. Here’s a look at how drones, machine learning, and AI-powered solutions are helping to accelerate site deployment and roll-out while reducing health and safety hazards.

When I started my career at Ericsson many years ago, one of the first responsibilities I had as an engineer was to perform site surveys for our customers. I quickly learned that this task is an important one when it comes to the roll-out of telecommunications networks. During the survey stage, it is the site engineer’s responsibility to identify the best location to deploy network equipment, which is critical. For example, batteries that weigh several tonnes need to be aligned with or very close to the strongest part of a building’s structure. Nobody wants batteries or equipment cabinets falling through the ceiling!

For me, the most exciting aspect of a site survey was perhaps the outdoor part. I would have to evaluate potential new antenna locations and optimum cable runways. This would often mean working on rooftops, which had the added value of some amazing views. In certain cases though, a survey would involve climbing up towers. That was pushing it for me, and was when my career as a site engineer ended: I’m afraid of heights.

5G roll-out aims high

As 5G deployments continue to gather momentum, we already know network site numbers will increase. We need to deploy more small cells in more places to bring mobile services closer to users, while existing macro sites will also be boosted with 5G to add much needed network capacity. According to the Ericsson Mobility Report (June 2020 edition) , by 2025 in North America alone, monthly data consumption will reach 45GB per smartphone.

Rolling out more sites also means finding innovative ways to deploy 5G equipment. In crowded city centers, one option is to use lampposts, traffic lights and other street furniture, which is being done in busy cities such as New York. Macro sites are also making way for 5G equipment, which is already translating into an increased demand for experienced tower climbers to carry out site surveys, equipment installation and maintenance. With 5G deployments increasing, thousands of brave climbers might be needed to cater for the demand, as we see in this recent report from CBS News.

 How are drones used in engineering?

Here is where technology can augment the work our people are already doing. For over two years now, Ericsson has been flying drones with several service providers around the world to assist with the site survey process. Traditionally, several people would be required to carry out a comprehensive site survey, including tower climbers who can verify and document a site in its entirety. Depending on variables such as weather and visibility, occupational health and safety requirements can prevent field personnel from going up the towers unless minimum safety conditions are met. Such circumstances can delay a survey and subsequent activation of a network site. Moreover, in some countries including Denmark, cranes and bulky heavy machinery can sometimes be required by law to lift technicians up onto the towers safely. This can eat into precious roll-out time, be expensive, troublesome and further complicate the activation of a site.

Enter the assisting drone… Once on site, a drone pilot can quickly set up and fly a camera-equipped drone on a predetermined path around a tower, snapping digital photos of the whole structure. Experience has shown that a complete set of images of a typical telecommunications tower can be captured in just 30 minutes!

Flying drones can record all exterior details of a site, including not only the antennas and radio equipment, but also cable runs, access roads and so on. A short flight around a tower allows the capture of enough photos to recreate the site digitally at a later stage. The interior of a network site (the equipment room, cabinets and so on) can also be documented with the help of technology, albeit of a more “terrestrial” kind: using high-definition 360-degree cameras, field personnel can scan an entire equipment room, enabling the documentation of its layout and wide range of equipment installed there with sub-centimeter accuracy.

What is even more impressive is that site surveys can now be completed by one engineer in about one hour as opposed to four, enabling more surveys to be carried out in a given period while allocating more time to other site engineering activities.

Intelligent Site Engineering

Indeed, as service providers all over the world continue to evolve their networks to 5G, increased network complexity calls for more efficient ways to deploy and manage these networks. As a result, traditional network roll-out site engineering services require not only the latest technological aids, but also constant innovation, often including artificial intelligence (AI) and machine learning (ML) capabilities to cope with the predicted demand of 5G roll-outs quickly and effectively.

Ericsson’s Intelligent Site Engineering approach now includes various capabilities that enable engineers to create site documentation based on digital representations of sites rather than relying on traditional survey methods such as tape measurements and tower climbing. Like this, the time to market for network delivery is being shortened, benefiting every user. We have crafted a very interesting summary of our work with TDC NET in Denmark about the use of this innovation.

Once photos (both aerial and those take on the ground) have been gathered, an AI-based desktop application is used to reconstruct the site digitally based on the pictures taken. A 3D model or digital twin of the site is then created, enabling engineers to make engineering decisions remotely based on the model and avoiding unnecessary site visits. Fewer site visits mean less risk to people, reduced carbon emissions and a more efficient engineering process altogether.

 Faster, accurate and sustainable 5G roll-out

Several benefits can be gained from the digitalization of site engineering activities. With Intelligent Site Engineering, data is always accurate and consistent between all sites, enabling a faster turnaround of site-related documentation and increased productivity with gaps between processes now minimized. In this way, remote centers of excellence can potentially perform site engineering activities such as calculating site materials remotely. The reduced health and safety hazards involved bring much greater flexibility and reduced long-term risk to projects. Through the use of drones and the digitalization of over 20,000 sites in more than a dozen countries, we have observed this methodology is facilitating the acceleration of site-build planning and decision-making processes by enabling the secure sharing of site data in near-real time with relevant stakeholders such as customers and interdependent teams, including Supply, Sourcing, Network Build, Radio Network Design and more. 

The future looks exciting where these kinds of initiatives are concerned. With the support of AI -based innovation, additional functionality will be added to further improve Intelligent Site Engineering services, eventually allowing full site inventories and lists of materials to be generated automatically, minimizing costly mistakes and waste. Digitalization will also facilitate the comparison of before and after scenarios, enabling the generation of “as-built” documentation in a more accurate way. Furthermore, preventive and corrective maintenance will continue to be enhanced, building on techniques previously developed by Ericsson, which have allowed us to demonstrate how drone inspections can be used for detecting and diagnosing cabling problems. AI and ML applications can assist not only with the identification of issues with cables and connectors, but also other problems affecting performance such as corrosion. As network roll-out keeps on getting smarter, the best thing of all for people (like me) who’re afraid of heights is there’s far less need for any climbing!

This post was first published on the Ericsson Blog.