Uni­ver­si­ty of Oulu launch­es 6G-ROCK­ET to build AI-native radio net­works

A new €1.4 mil­lion research and devel­op­ment project called 6G-ROCK­ET has start­ed to accel­er­ate AI and machine learn­ing solu­tions for future 6G radio access net­works. The Uni­ver­si­ty of Oulu’s 6G Flag­ship leads the project, sup­port­ed by Busi­ness Finland’s 6G Bridge pro­gramme. The team includes Nokia, NVIDIA, DNA, Keysight, and Medi­aTek, along with North­east­ern Uni­ver­si­ty (Boston) and the Uni­ver­si­ty of Tokyo as inter­na­tion­al part­ners.

The project builds on a GPU-accel­er­at­ed AI-RAN test­bed con­nect­ed to the Uni­ver­si­ty of Oulu’s 5G/6G Test Net­work. The project aims to advance glob­al progress in AI-RAN by com­bin­ing 6G Flagship’s wire­less research with its part­ners’ indus­try expe­ri­ence.

“6G-ROCK­ET is about build­ing the foun­da­tions for AI-native mobile net­works where com­mu­ni­ca­tion, sens­ing and intel­li­gence work togeth­er seam­less­ly. Our goal is to move beyond sim­u­la­tions and demon­strate prac­ti­cal AI-RAN tech­nolo­gies in GPU-accel­er­at­ed test­beds. Through this project, Fin­land con­tin­ues to strength­en its lead­er­ship in 6G research, inno­va­tion and future wire­less ecosys­tems,” says Dileepa Maras­inghe, AI-RAN Research Man­ag­er at 6G Flag­ship, Uni­ver­si­ty of Oulu Fin­land.

“By inte­grat­ing intel­li­gence direct­ly into the radio access net­work with AI-RAN, we are turn­ing the base sta­tion from a sig­nal proces­sor into a high-per­for­mance com­put­ing node that can also inter­pret changes in the phys­i­cal envi­ron­ment. It allows the net­work to see, rea­son and react in mil­lisec­onds, which is essen­tial for robot­ics. The 6G-ROCK­ET project in col­lab­o­ra­tion with NVIDIA is advanc­ing 6G research for phys­i­cal intel­li­gence, help­ing devel­op­ers build, test and opti­mize their next break­through on NVIDIA AI-RAN plat­forms,” says Soma Velayutham, Vice Pres­i­dent, Tele­coms, NVIDIA.

From reac­tive to proac­tive

Most mobile net­works today are reac­tive, rely­ing on man­u­al adjust­ments and sep­a­rate AI tools. 6G-ROCK­ET aims to change this by cre­at­ing AI-native RAN tech­nolo­gies that proac­tive­ly antic­i­pate needs. The project cov­ers AI-native RAN designs, real-time dig­i­tal twins for net­work opti­miza­tion and con­trol, GPU-accel­er­at­ed test­ing plat­forms, mul­ti-sen­sor per­cep­tion for bet­ter aware­ness, and autonomous net­work man­age­ment based on intent.

The project puts AI direct­ly into the radio access net­work, mak­ing intel­li­gence a native part of the RAN archi­tec­ture rather than an add-on capa­bil­i­ty. This approach builds proac­tive capa­bil­i­ties into the net­work from the start. This makes the net­work flex­i­ble and scal­able, rather than sim­ply adding a per­cep­tion sys­tem to exist­ing net­works.

Part­ners across the full AI-RAN stack

Each part­ner in the con­sor­tium brings in their com­ple­men­tary exper­tise. 6G Flag­ship brings with it its wire­less research and the Uni­ver­si­ty of Oulu’s 5G/6G test net­work infra­struc­ture. Nokia, in turn, offers exper­tise in mobile net­work sys­tems and AI-native RAN evo­lu­tion. NVIDIA pro­vides AI-RAN infra­struc­ture and AI Aer­i­al open-source soft­ware libraries and tools. Keysight con­tributes exper­tise in dig­i­tal twin tech­nolo­gies, includ­ing sim­u­la­tion, test­ing, and val­i­da­tion capa­bil­i­ties. Medi­aTek pro­vides chipsets and user-equip­ment exper­tise. DNA offers oper­a­tor insight and real-world net­work expe­ri­ence. North­east­ern Uni­ver­si­ty and the Uni­ver­si­ty of Tokyo add inter­na­tion­al research col­lab­o­ra­tion to align the work with glob­al AI-RAN and 6G efforts.

The project also sup­ports Finland’s broad­er 6G strat­e­gy by strength­en­ing con­nec­tions between uni­ver­si­ties and indus­try and by help­ing devel­op the stan­dards, tech­nolo­gies, and skills the coun­try will need.

“As an oper­a­tor, we want to help ensure that future net­work tech­nolo­gies are devel­oped around real cus­tomer needs and can be test­ed in prac­ti­cal envi­ron­ments from an ear­ly stage. 6G-ROCK­ET brings togeth­er world-class research, tech­nol­o­gy com­pa­nies and oper­a­tor expe­ri­ence in a high­ly valu­able way. For DNA, the project is an oppor­tu­ni­ty to influ­ence how AI-native net­works devel­op, while gain­ing prac­ti­cal insight into solu­tions that can improve net­work effi­cien­cy, reli­a­bil­i­ty and future ser­vices,” says Sari Lep­pä­nen, CTO at DNA.

“Val­i­dat­ing AI-dri­ven net­works requires pre­cise, repeat­able mea­sure­ments under real­is­tic con­di­tions. Keysight is pleased to con­tribute its sim­u­la­tion, mea­sure­ment, and chan­nel val­i­da­tion exper­tise to 6G-ROCK­ET, help­ing the con­sor­tium advance from con­cept to demon­strat­ed per­for­mance,” says Michael Dieudonne, 6G R&D Man­ag­er at Keysight Tech­nolo­gies.

“Device-side intel­li­gence is essen­tial to clos­ing the loop between net­work and ter­mi­nal. Medi­aTek is glad to bring its chipset and user-equip­ment exper­tise to 6G-ROCK­ET, help­ing ensure these AI-RAN inno­va­tions work end to end,” says Chih-Ming Hung Asso­ciate Vice Pres­i­dent, Tech­nol­o­gy, Medi­aTek.

What will the next two years deliv­er

In the next two years, 6G-ROCK­ET will cre­ate tech­nol­o­gy blue­prints and pro­to­type demos to pre­pare for smart, sens­ing-enabled AI-RAN and future 6G net­works. The project also plans to estab­lish an AI-RAN Per­cep­tion Lab at the Uni­ver­si­ty of Oulu. This lab will sup­port ongo­ing col­lab­o­ra­tion between uni­ver­si­ties and indus­try and help Fin­land remain a leader in 6G and AI-native wire­less research.

Besides tech­nol­o­gy, the project aims to strength­en Finland’s 6G plans and glob­al rep­u­ta­tion, build local exper­tise in AI-native RAN and wire­less inno­va­tion, encour­age new inter­na­tion­al part­ner­ships, and grow Finland’s work­force and skills in AI and wire­less com­mu­ni­ca­tions.

By bring­ing togeth­er advanced AI com­put­ing, wire­less sys­tems, and real-world test­ing, 6G-ROCK­ET is mov­ing toward the next gen­er­a­tion of smart, adapt­able 6G net­works.

Source (text and image): Uni­ver­si­ty of Oulu