3G: The Next Generation

The first 3G network in the UK was launched back in 2003. However, in the coming weeks, Virgin Media O2 (VMO2) will cease to offer 3G service across much of the UK (though isolated pockets remain for slightly longer). This follows the lead of both BT/EE and Vodafone, who withdrew their 3G services in 2024. We will therefore shortly witness the end of 3G service in the UK for the vast majority of users [1]. Whilst this may represent a source of frustration for users of older handsets, the withdrawal of 3G services should enable spectrum reallocation to 4G and 5G services, thereby improving the coverage and performance of these later generations.

In contrast to the decline of 3G, the most recent issue of the 3GPP newsletter [2] highlights some of the interesting directions that the both 5G and 6G services are developing in. Of particular interest are the applications of AI and machine learning (ML) to 5G-Advanced, with the long-term goal of facilitating future generations of mobile networks that are AI-native.

Specifically, the 3GPP newsletter indicates three key enhancements to 5G-Advanced which should offer significant benefits to users:

(1) AI/ML-based UE positioning, including deriving UE location based on ML Models in the Location Management Function (LMF) rather than on signaling measurements exchanged with the network, thereby improving the accuracy of UE positioning;

(2) Using ML Models to analyse quality of service (QoS) input parameters and predicting the quality of experience (QoE) for a list of QoS parameter sets, enabling an application’s quality requirements to be met more efficiently;

(3)  Mitigating unexpected surges in signalling which may reduce the QoS provided by identifying the UEs and/or RAN nodes responsible for the increased load.

For example, the benefits offered by these developments could be experienced by users while trying to navigate congested urban pathways, or when streaming music on the train during rush hour.

From a legal perspective, it will also be interesting to see how EPO practice will evolve with a greater number of applications for inventions spanning AI/ML and telecommunications. The EPO guidelines set out various examples of technical contributions, some of which may be applicable to these inventions (e.g., signal enhancement and analysis, efficient data encoding techniques, encryption/decryption), but additional examples may provide useful guidance for applicants. For now, to improve the chances of successful prosecution at the EPO, it is advantageous for European patent specifications to indicate a parallel with at least one of these currently accepted examples of technical contribution.