AI-powered Ericsson Performance Optimizers make an impression

The new Ericsson Performance Optimizers, which is a suite of AI-powered applications, can analyse the CSPs’ Radio Access Network (RAN) to proactively provide mobile network optimisation recommendations and resolve specific network performance issues, according to a company release.

Ericsson Performance Optimizers can be deployed to enable a helicopter view of the network as a whole. The solution accounts for the invisible changes in the network caused by each addition to the environment, such as new applications, city growth, new sites or user behavior. It is equipped with digital twin technology and advanced AI techniques like deep reinforcement learning. 

Digital twin technology accurately mimics the network behaviour upon parameter changes, ensuring an approach that minimises risk and elevates the optimiser’s quality to telco grade from day one. Reinforcement learning is a machine-learning technique that learns from the network, where an agent (Performance Optimizer) interacts with the environment and takes actions towards a longterm goal.

Additionally, the solution can also be used to accurately predict network performance improvements and proactively provide one-shot optimisation recommendations for targeted cells.

Ericsson plans to periodically add more features to the existing solution to address the challenges brought on by different network issues, thereby increasing the coverage of automated optimisation. 

With 5G, CSPs are looking to transform essential parts of their operations to achieve optimal performance and return on investment. Vendor agnostic AI-native solutions such as Ericsson Performance Optimizers are key to automate and cope with growing complexity while keeping control of cost.  

In November 2021, Ericsson trialed the solution with Ooredoo Qatar for a major football tournament. As part of the partnership, Ericsson provided network optimisation and event management and 5G connectivity in three stadiums, airports and places of attraction. Massive efficiency gains were noted – where traditional worst cell analysis takes days to complete, the new automated approach with performance optimizers took only 15 minutes, producing hundreds of recommendations.

Juan Manuel Melero, Head of Network Design & Optimization, Ericsson, said: During the trial with Ooredoo Qatar, we noted that 89 percent of internal uplink interference cases were automatically resolved by the Performance Optimizers. This resulted in seven percent improvement in uplink capacity, along with significant gains in speed and traffic volume. We now look forward to leveraging our expertise and experience to support our customers around the world to achieve top network performance and deliver the best user experience.

Günther Ottendorfer, Chief Technology and Infrastructure Officer at Ooredoo Qatar, said: “For Ooredoo, customer experience and automation are top priority. We see AI-powered operations as key to deliver on our promise of world-class services to our subscribers. During the major football event held in Qatar in December 2021, Ericsson’s Cognitive Software provided near realtime automation of optimisation actions. The result was a superior user experience before, during and after the matches.

Ericsson Performance Optimizers suite is part of the Cognitive Software pack in Ericsson Operations Engine. It can be implemented through licensing, software as a service (SaaS), or as part of our services packs.