6G-LEADER

6G-LEADER PoC#1 will validate the 6G-LEADER solution for seamless real-time interaction between XR headsets and UAVs. The validation scenario involves an XR headset and two UAVs: the first is equipped with a neuromorphic camera offering low energy consumption at the cost of reduced resolution, while the second employs a high-fidelity camera that enables greater transmission quality but consumes more power. To optimise energy efficiency, the high-fidelity camera is activated only when the neuromorphic camera detects a relevant event.
Semantics-aware, real-time video transmission is performed on the XR headset. Semantic segmentation is applied to isolate and process only critical visual information, thereby minimising the volume of data that needs to be transmitted.
Furthermore, a split inference strategy is employed, wherein part of the processing is performed onboard the UAV and the remainder is offloaded to an edge node via compressed feature vectors. This approach supports responsive and energy-efficient reconstruction of the surrounding environment, ensuring a high quality of experience (QoE) for the end user. Given the UAVs’ mobility and the potential for ultra-low-latency requirements, the 6G-LEADER framework is applied to enhance the communication link using advanced AI/ML-driven physical-layer radio technologies. This enables high-throughput, bi-directional data exchange, even under peak traffic conditions.
Accurate localisation is also critical in this setup. In addition to video streams, the UAV's position must be transmitted to align spatial content with the user's visual perspective in the XR headset. Therefore, precise and timely localisation is essential to reduce transmission overhead. By leveraging preloaded environmental data and static maps stored on the user’s device, redundant data transmissions are avoided, enabling smooth and immersive real-time interaction without overloading the network.