02) Machine-Learning Driven Coverage and Capacity Optimization

Companies Participating
in this Scenario



Device Under Test

  • Mavenir Non-Real-Time RAN Intelligent Controller

    Mavenir O-CU-CP, O-CU-UP [NR]

Test scenarios introduction

The Mavenir RAN Intelligent Controller (RIC) makes use of advanced machine learning algorithms to optimize network performance.

The target objective function is configurable by the operator, guiding the optimization process to meet potentially dynamically changing business requirements at runtime. Bayesian Optimization is used to build a model of the relationship between configuration parameters and performance, and the algorithm makes optimal use of the prior data samples to refine the model and choose the next proposed configuration update, making a tradeoff between exploitation, which attempts to maximize the performance immediately, and exploration, which aims to reduce the uncertainty of the model.

Testing Results/Summary

The demonstration shown the RIC in action, ingesting configuration data and performance metrics from a demonstration RAN, and iteratively optimizing the configuration parameters to achieve improvement in key performance indicators.

Benefits for the Industry

  • The resulting closed-loop automation provides a low cost way to continuously monitor and tune the system parameters to achieve operator business objectives.