Holographic type communication (HTC) is an emerging use-case of 5G and beyond networks. Holographic multi-source teleportation is an example of HTC, which places stringent requirements on the communication networks in terms of data rate and latency, along with frame synchronisation which is critical because there being multiple participants involved and communicating in real-time.

We first design a synchronisation algorithm for two cases, with local and global synchronisation windows. We decouple the synchronisation functions in order to early-drop the late frames, which saves the network resources and bandwidth along the routing path, and improves the user’s QoE. We implement efficient synchronisation units using a concept of on-time service guarantee, which is a part of time-engineered communication services from FG-NET-2030. We design three scenarios with increasing number of synchronisation units, and implement them on an emulated platform based on a local synchronisation window. Through extensive experimentation, we demonstrate that for each of these scenarios, we observe an improved performance in terms of specific QoE metrics in various network conditions.

We then quantify the motion misalignment between remote sources within different network contexts in order to justify the necessity of frame synchronisation on a real-life teleportation platform. Motivated by this, we propose HoloSync, a novel edge-computing based scheme capable of achieving controllable frame synchronisation performance for multi-source holographic teleportation based on a frame production timestamp. We carry out systematic experiments on a real system using the proposed HoloSync scheme, and evaluate its frame synchronisation performance in various network scenarios, and its sensitivity to different control parameters.

Finally, we present an AI-based adaptive mechanism for selecting the synchronisation control parameters in order to achieve and maintain the QoE close to user-defined target QoE parameters. The proposed AI-based adaptive algorithm dynamically selects the synchronisation control parameters based on the instantaneous context, that is, the current measured QoE parameters. We demonstrate, by carrying out extensive experiments, that the AI-based adaptive scheme provides distinct advantages, in terms of the user QoE, i.e., frame rate, frame pairing error, and average playback latency, over the scheme with manually fixed synchronisation parameters.