ESTIMATING FUTURE NETWORK CELL LOAD WHILE PRESERVING USER EQUIPMENT PRIVACY

Smart overview of the Invention

The focus is on estimating future network cell load while maintaining the privacy of user equipment (UE). This is crucial for self-organizing networks (SON) in upcoming 6G mobile networks, which aim to enhance user experience, network efficiency, and reduce costs.

Background

Current SON functions in 4G and 5G networks are generally reactive, responding to network issues after they occur. This approach often results in delayed responses that are inadequate for the low-latency demands of 6G networks. The transformation to proactive SON, enabled by machine learning (ML), could allow networks to predict and adapt to future states, optimizing coverage and capacity before issues arise.

Challenges

Estimating future network states based solely on historical data is prone to inaccuracies, as it may not account for new UEs connecting to the network. Moreover, privacy concerns arise when predicting future network usage, as it involves sensitive data like UE locations. Thus, a method is needed to estimate future states without compromising UE privacy.

Proposed Method

A method is proposed where a UE obtains an ML model to predict network usage and receives network configuration information from a base station. The UE generates trajectory data about its movement, uses the ML model to predict network usage, and transmits this data to the network node. This approach allows for accurate load prediction while preserving privacy.

Network Node Method

The network node receives predicted network usage data and encoded trajectory data from UEs. It combines this information to generate a comprehensive prediction of network usage for multiple UEs. This enables proactive network optimization without compromising user privacy, aligning with the goals of 6G networks.