Content for  TR 22.856  Word version:  19.2.0

The mobile metaverse server receives the data from various sensors, performs data processing, rendering and provide feedback to the vehicles and users.

Examples of typical data volume including 1) camera: 10 Mbit/s per sensor (unstructured), 2) LiDAR: 90 Mbit/s per sensor (unstructured), 3) radar: 10 Mbit/s per sensor (unstructured), and 4) real-time Status information including Telemetry data: [< 50 kbit/s] per sensor/vehicle/VRU (structured). This is to support at least 80 vehicles and 1600 users present at the same location (e.g. in an area of 40m*250m) to actively enjoy immersive metaverse services for traffic simulation and traffic awareness, the area traffic capacity is calculated considering 2 cameras, 2 Radars, 2 LiDARs on road side, 1600 user's smart phones and 80 vehicles with 7 cameras, 4 radar and 2 LiDAR for each vehicle.

The frequency considers different sensor types such as Radar/LiDAR (10Hz) and camera (10~50Hz).

The service area for traffic flow simulation and situational awareness depends on the actual deployment, for example, it can be deployed for a city or a district within a city or even countrywide. In some cases a local approach (e.g. the application servers are hosted at the network edge) is preferred in order to satisfy the requirements of low latency and high reliability.

The calculation is this table is done per one 5G network, in case of N 5G networks to be involved for such use case in the same area, this value can be divided by N.

User experienced data rate refers to the data rate needed for the vehicle or human, the value is observed from industrial practice.

The network based conference focus is assumed, which receives data from all the participants, performs rendering (image synthesis), and then distributes the results to all participants. As rendering and hardware introduce some delay, the communication delay for haptic feedback is typically less than 5ms.

To support at least 15 users present at the same location (e.g. in an area of 20m*20m) to actively enjoy immersive Metaverse service concurrently, the area traffic capacity is calculated considering per user consuming non-haptic XR media (e.g. for video per stream up to 40000 kbit/s) and concurrently 60 haptic sensors (per haptic sensor generates data up to 1024 kbit/s).

In practice, the service area depends on the actual deployment. In some cases a local approach (e.g. the application servers are hosted at the network edge) is preferred in order to satisfy the requirements of low latency and high reliability.

The arrival interval of compressed haptic data usually follow some statistical distributions, such as generalized Pareto distribution, and Exponential distribution [26].

The end-to-end latency does not include sensor acquisition or actuator control on the vehicle side, processing, and rendering on the user side (estimated additional 100ms total). Target e2e user experienced max delay depends on reaction time of the remote driver (e.g. at 50km/h, 20ms means 27cm of remote vehicle movement).

UL data transfer interval around 20ms (video) to 100ms (sensor), DL data transfer interval (commands) around 20ms.

The service area for teleoperation depends on the actual deployment; for example, it can be deployed for a warehouse, a factory, a transportation hub (seaport, airport etc.), or even a city district or city. In some cases, a local approach (e.g., the application servers are hosted at the network edge) is preferred to satisfy low latency and high-reliability requirements.

The area traffic capacity is calculated for one 5G network, considering 4 cameras + sensors on each vehicle. Density is estimated to 10 vehicles/km2, each of the vehicles with one user controlling them. [25]

KPI comes from [5] clause 7.11 "remote control robot" use case.

Examples of typical data volume including 1) ~8Mbps video stream. Four cameras per vehicle (one for each side): 4*8=32Mbps. 2) sensor data (interpreted objects), assuming 1 kB/object/100 ms and 50 objects: 4 Mbps [25].

These KPIs are only valid for cases where the viewport rendering is done in the tethered device and streamed down to the AR glasses. In the case of rendering capable AR glasses, these KPIs are not valid.

These values are aligned with the tactile and multi-modal communication KPI table in clause 7.11 of TS 22.261.

These values are aligned with "high-speed train" DL KPI from clause 7.1 of TS 22.261.

To leverage existing streaming assets and delivery ecosystem, it is assumed that the legacy streaming data are delivered to the rendering device, which incrusts this in the virtual screen prior to rendering. For a live streaming event, the user-experience end-to-end latency is expected to be competitive with traditional live TV services, typically [1-5] seconds.

For example, the glTF format [60] can be used to deliver avatar representation and animation metadata in a standardized manner. Based on this format, the required bitrate for transmitting such data is highly dependent on avatar's complexity (e.g., basic model versus photorealistic).

These values are aligned with "immersive multi-modal VR" KPIs in clause 7.11 of TS 22.261.