Using artificial intelligence to shape 5G

​5G networks will be fast and able to support many more devices simultaneously than the networks that came before – this is particularly important considering that by 2025 there will be approximately 41.6 billion connected Internet of Things devices in use, according to analyst IDC. And many of them, such as autonomous vehicle systems and industrial sensors, will require reliable access to cloud services in order to function safely.

The wireless technology has been described by GSMA, the mobile network trade body, as delivering three pillars of next-generation connectivity: enhanced mobile broadband (eMBB), ultra-reliable and low-latency communications (URLLC) and massive machine-type communications (mMTC). 

The complexity of the infrastructure required will involve a huge number of devices drawing on high-speed (and low-speed) bandwidth, all demanding real-time, automated decision-making.

With the sheer amount of data being generated, 5G and artificial intelligence (AI) will need each other. 

Related content: 

Using AI to meet data demands

One difference 5G will bring is the requirement to co-exist with other technologies that operate in the same frequencies. In the past, spectrums have been licensed to network operators in blocks of frequencies, over which they have exclusive use. So when an operator isn't using that piece of the spectrum, it lies unused.

However, the demand for the spectrum – which is not only shared by commercial services, but supports government and military communication too – is growing. And with 5G, desire for spectrum access will be even greater.

Staying cyber-secure with 5G and AI

In this new era where an unprecedented number of devices and individuals are connected, there is an increasing demand for trusted communication environments. And one area where AI and 5G can successfully work together is in mitigating security risks. Developing machine learning in cyber security means machines can learn to detect and neutralize new threats.

According to David Yates, VP for product management at Guavus, a Thales company, AI will also play a vital role in improving the service offering of 5G networks. "5G will be all about services," he says. "That's how operators will make money. If they just supply a best effort broadband service they will struggle to achieve returns on investment."

While utilizing AI doesn't necessarily mean operators are going to start offering their own streaming video platforms, Yates points out there are also other untapped opportunities – particularly when it comes to analytics. "Better analytics can provide better insights into what consumers are doing as well as showing operators how to deliver improved services and monetize them," he says.

5G and the future

While 4G networks (that use Internet Protocol, or IP, broadband connectivity and are reaction-based) have dominated the connectivity landscape, AI and its various sub-categories, including machine learning and deep learning, have been evolving, ready to provide individuals with 5G networks that address the poor efficiency of the spectrums we saw with 4G.

One solution that can be offered through 5G is the use of dynamic spectrum allocation (DSA), a deployment used in countries that still broadcast analog television signals. DSA involves the optimization of spectrum use by radios that hunt for gaps through which they can transmit data, helping operators meet data demands.

In other words, 5G is predictive and proactive, and in itself will generate the need for real-time, automated decision making.

Related content: 5G will require tougher security protocols