Ericsson and Optus have partnered to trial a new AI model that predicts whether a device on a frequency layer is also within coverage of another frequency in 5G Standalone networks. As the first collaboration of its kind between Ericsson and an operator, this solution enables the network to make decisions faster and with greater accuracy, supporting smoother connections and better performance for customers.
Accurately predicting frequency coverage is crucial because a device’s location within the cell determines whether it is within the coverage area of various frequencies. Relying on static methods to address this issue drains battery life, increases the risk of dropped connections, and requires complex manual configuration by the operator.
Drawing on months of live data from the Optus network, Ericsson evaluated a range of options before developing a specially designed model for the radio access network (RAN). This model can be retrained locally for each cell and delivers results in real time.
Recent tests show that the model can accurately predict – up to 95 per cent of the time – whether a device is within coverage of another frequency layer. With this reliable, cell-level coverage prediction, the network can reduce unnecessary checks and measurements from devices.
For Optus 5G Standalone customers on the move, this means faster handovers, fewer dropped connections, lower system load, and improved device battery life.
Mårten Lerner, Head of Networks Strategy & Product Management at Ericsson, says: “AI makes networks smarter and more responsive, ensuring uninterrupted connectivity for consumers and businesses. With enhanced real-time coverage prediction where it matters most, calls and data sessions move seamlessly from one area to another.”
Sri Amirthalingam, Chief Technology Officer at Optus, says: “With the help of this AI model in coverage prediction, Optus 5G Standalone customers on the move will enjoy enhanced connectivity with a lower risk of dropped calls thanks to the more accurate and smoother handover from cell site to cell site, improved energy efficiency due to reduced computing demands and better battery performance on their devices.”
