AI-nativ integrerad avkänning och kommunikation för 6G-baserad smart mobilitet
| Diarienummer | |
| Koordinator | Kungliga Tekniska Högskolan - Avdelningen för teknisk informationsvetenskap |
| Bidrag från Vinnova | 493 080 kronor |
| Projektets löptid | juli 2026 - september 2026 |
| Status | Avslutat |
| Utlysning | 6G - Kompetensförsörjning |
| Ansökningsomgång | Individrörlighet inom 6G för samarbeten med USA och Singapore |
Viktiga resultat som projektet gav
The project studied multi-base-station ISAC with multiple sensing targets. We developed a multi-LLM agentic framework that coordinates specialised agents to allocate wireless resources, aiming to maximise users´ sum communication rate while meeting target-specific sensing requirements. In our evaluations, the method outperformed the conventional optimisation and heuristic baselines considered. The work deepened collaboration between KTH and NTU/COSMO.
Långsiktiga effekter som förväntas
The project is expected to strengthen KTH´s expertise in AI-assisted 6G optimisation and sustain collaboration with NTU/COSMO. Further development of the multi-LLM framework could support more efficient wireless resource allocation under sensing constraints. Broader validation is needed to assess robustness, scalability and practical benefits for connected smart mobility.
Upplägg och genomförande
The KTH staff exchange with NTU/COSMO ran from 1 July to 30 September 2026 and was completed on schedule, without delays or major changes. We formulated the multi-base-station, multi-target ISAC resource allocation problem, developed a coordinated multi-LLM framework and evaluated it against conventional optimisation and heuristic baselines. Technical discussions, seminars and internal reporting supported collaboration and knowledge transfer.