Innovating 5G/AI – Where are we now?

We are now in week 4 of the process of combining concepts from mobile executives, technical leaders, and students.  Special thanks to our most recent guest on Friday, Mostafa Essa, Vodafone’s Distinguished Engineer, who spoke on the intersection of Radio Access Networks, AI, and Data Analytics

After just one month, here are a few things I have observed.  Consider this to be a map of what we now know and what we still need to figure out:

1. Many mobile players have deployed at least some 5G, but most are waiting for the business case before larger scale investments.
2. Everyone still is looking for the first signals of 5G deployments. Will it be stadiums for high speeds? Or will it be the application that requires both a) mobile and b) delay sensitive?  (Stay tuned, this is the study we all need to work on)! 
3. No one seems to know the 5G application space beyond the common concepts of a) improved gaming and b) AR/VR for remote work/training. 
4. The biggest challenge to 5G is the fact that 4G works fairly well.
5. Customers: There are two consumer segments. A higher-end segment will likely adopt 5G even if they are not sure what new services will be available – simply as an evolution. When Apple and others release 5G handsets, desire from this top segment will further increase. Then there is a second segment that is increasingly cost sensitive. They will look for value alternatives, which may pave the way for disruptive solutions.
6. Enterprise: We don’t know enough about this segment and its needs just yet. 
7. We do believe “Edge Computing” is a real trend. Old-world Internet providers should beware. Amazon, Microsoft, and others need to deploy applications and servers closer to their customers for higher availability and reduced delay. Carriers have some of the best real estate, and wiring/access options, whether they are related to mobility or not. Architecture is currently being developed and the case studies still need to be written!
8. IoT: It may be massive, but deployment models are still incremental.  Why? Because WiFi chipsets are $2 and 5G chipsets are $30. So, your IoT microwave or IoT smart pencil are still most likely to be tethered to a WiFi router or Bluetooth phone for quite some time.
9. 5G home broadband might be real.  This means that a segment of people will actually throw out their DSL or cable modems and replace them with 5G modems and WiFi, as per the standard’s intention. For some carriers, the cost of spectrum compared to fiber will make this prohibitive. In regions where cables/fibers have not been laid, it can work assuming deployment is possible. It will definitely be on a case by case basis, but it can work better if the 5G home router devices use peer to peer 5G services to solve part of the deployment and backhaul problems. 
10. Crediting AT&T & Vodafone, among others, seem to know a wide range of deployment-related challenges including: 
    1. • Issues with cell size and larger deployment 
       • Complexities with beam formation and the fact that it’s harder to switch back and forth between two cells
       • Global warming has been causing mobile signal ducting that randomly carries signals 20 to 100 miles away.
       • Increased need for self-healing during power outages at radio towers 
       • The 5G power requirements is still a big issue for operators and AI might play a role in the solution.

On the topic of deployments, it has been interesting to note that higher frequency spectrums may carry only a short range but they will penetrate brick, glass and windows. On the other hand, more traditional lower frequency signals may carry farther, but they will not penetrate walls and windows.

   11. And finally, there is a large opportunity to increase the customer engagement for phone carriers. Today, customers engage very positively with companies like Netflix even though they rarely speak with any of their agents. In contrast, mobile customers often expect to speak with mobile phone company agents and yet engagement is less positive and less sticky.  More can be learned and done in this area.

Background: The Innovating 5G with AI Course uses the Innovation Engineering Framework developed at UC Berkeley to apply technology and create transformation by aligning human talent in an efficient, effective, and positive manner. This framework offers practical guidance for how large firms, research labs, new ventures, and even student projects can execute their innovation projects and make their ideas a reality. Interested in learning more? Check out innovation-engineering.net and read Professor Sidhu‘s new book: Innovation Engineering.