Sensor-Enabled Automation Fuels the Future of Connectivity DX Week Panelists Predict AI, ML will Solidify the Human-Digital Connection

TDK Ventures
TDK Ventures
Published in
5 min readJul 13, 2022


The progress of the world’s digital transformation (DX) hinges in large part on how well personal and business devices will be able to connect to each other and the vast network of communications media, computing power, and smart equipment. Machine learning and artificial intelligence has the potential to predict network demand and allocate capacity and to help systems recover more quickly from outages.

“AI drives a lot of use cases for connectivity but AI itself is going to make a significant impact on the connectivity solution,” noted Investment Associate Henry Huang, as he kicked off a portion of TDK Ventures’ DX Week discussion during day 3’s panel on connectivity. “How much network automation will we see in the next five to 10 years because of AI?”

“I surmise that operators are already using AI to do a lot of things, from load balancing to resource allocation, because they have the data and have been using it,” said Nada Golmie, wireless networks chief at the National Institute of Standards and Technology (NIST). “So, who doesn’t have the data? It’s the researchers that would like to have it in order to develop algorithms.”

Sanyogita Shamsunder, head of Google’s global edge networking and a former executive at Verizon, provided a state of the industry

“There have been a lot of improvements. The standards themselves have provided a vehicle to add smarts to the network,” she said. “How the industry optimizes 3G networks and how we build 4G involved a lot of automation. There is a lot more intelligence in the networks. The self-organizing network — the SON concept in 4g — came in early, but it wasn’t implemented to the full degree by industry because of how the network was built.

As 5G becomes the standard, however, Shamsunder said the open radio access network (Open RAN) creates a wider ecosystem for developing cellular functionality and architecture.

“Open RAN And all these capabilities that provide further incentive to build those networks and run them using tools that do not involve manual intervention. If we’re talking about building large networks with the capabilities that we want, taking out manual optimization is going to be an important part, and there’s still a lot more opportunity.”

Boingo Chief Technology Officer Derek Peterson reiterated that operators are taking advantage of AI to automate their networks.

“Ten years ago, when I started running Wi-Fi networks, it was a pain because I just didn’t have the data insights to manage it like I had when I was working in the license realm,” he said. “So, we ended up developing a lot of automation, a lot of tools that created great improvements.”

He noted that As 400 people get off a plane, the first thing that happens is their phones connect to a single node at the arrival gate.

“How are you going to manage that when I only had 20 people there before, watching Netflix or sending email. All these devices are built so that as soon as they connect to Wi-Fi they start updating all their apps in the background. That’s going to destroy the bandwidth for everybody else,” he explained. “What do you do? You put in network smarts — thank goodness we’re allowed to do that now.”

That allows smart networks to automatically throttling down trivial updates so the movies can keep streaming and slow the film to allow it to stream without sapping additional resources, so everyone can use the service.

“Those are the kind of smarts we’ve build that dynamically recognize what’s going on at a single access point and makes those shifts in the network.

Audience members weighed in on what they find most challenging and exciting in the connectivity field, and Shamsunder agreed with viewers’ assessment.

“Data and data-driven intelligence is permeating everything we do,” she said. “Wireless networks are as complex as they can get The wireless channel is not an easy one (to manage). Traditionally, there has been a lot of modeling used in understanding it, so AI is definitely a tool. Automation is making planning easier for wireless networks, deriving analytics from various parts of the network. Converging all those different tools that engineers use on the wireless side and convert them into smart, autonomous networks is a big open area.”

She expressed that the standards and network architecture continue to evolve through open-RAN and other protocols allows advances developments made by all providers, operators, and other contributors to be more quickly adopted throughout the industry and adapted to diverse use cases.

Healthcare is one use case Peterson sees as taking off as a result of more sophisticated sensing and connectivity, especially in the wake of COVID-19. He said the pandemic demonstrated that “the data we had available wasn’t as good as it could have been. A lot of us have watches that keep track of our heart rates, temperatures, oxygen levels, and everything else, Those are the kinds of things that, might have prevented some of the missteps we made in addressing the last two years.”

While audience members placed metaverse development in the 10-year timeframe, many panelists remain optimistic that this digital realm will be available much sooner. Matt Grob, XCOM Labs cofounder and chief technology officer, acknowledged that “some aspects of it are very challenging. The device itself, the optics in particular, and how you do augmented reality in an outdoor setting, make it look good, and be registered to the real world, that’s a touch problem. Being able to upgrade the network — not just at the demo center — but for all society in outlying areas — to have the kind of throughput and performance you need to get these experiences. I think it will happen faster in some areas but providing it ubiquitously with devices [that are more akin to reading glasses than bulking AR headsets] will take a while.”

Like many, he sees gaming and entertainment as primary use cases for the metaverse, but he also mentioned education, training, and simulation, along with machine diagnostics, teleoperation of semi-autonomous vehicles as applications that can take advantage of the technology.

Like all the panels presented during TDK Ventures’ DX Week, this final segment of the discussion on connectivity inspired the current and future generations of inventors and entrepreneurs. The breakthroughs, use cases and scalability present reasons for optimism as well as challenges to those on the front lines. We believe the information presented during DX Week will help startups and investors identify opportunities where they can contribute to a more connected and sensitive future.