6G-TWIN

This use case focuses on teleoperated driving supported by a predictive Network Digital Twin (NDT). Here, teleoperated driving means moving the driver of a road vehicle out of their car/truck/etc. And into a (remote) control centre – either partly or fully. This requires vehicle sensors capable of sensing the environment and transmitting this real-time information (e.g., detected objects, Light Detection and Ranging -LIDAR- point clouds, video streams) to the remote operator over the mobile network. It also requires data fusion and pre-processing on edge servers to overcome the limited computing capabilities of the vehicles while limiting the amount of information transmitted (i.e., reducing network load). Such a system is particularly interesting for public transport, remote hailing, or remote parking – even more so when combined with some autonomous vehicle functions.

The teleoperation of a vehicle has high demands on the network, such as extreme reliability, low latency, and high throughput. Often-quoted values are 99.999% reliability and latencies below 20 milliseconds. This is in contrast to our assumption that the full capacity of the network is not always deployed throughout. In particular, we assume that edge servers at different locations can be turned off and virtual network functions might be undeployed during periods of inactivity, e.g., in rural areas. Before the journey of a teleoperated vehicle can begin, the vehicle is therefore first driven virtually, in a simulation, to ensure adequate resources are available – to investigate changes that would ensure they are – and to then realise these changes before the trip of the vehicle commences.

6G-TWIN will realise this use case through lab-scaled experiments involving a mix of physical nodes (e.g., robots) and simulated environments with real data coming from the project partners.