Predictive QoS
5G is a key enabler towards autonomous driving. To resolve QoS limitations in today's mobile network (e.g. latency, data rate), 5GAA is investigating mechanisms and solutions for in-advance QoS change notification in (e) NESQO WI, such as Predicitve QoS.
Predictive QoS, e.g. along a specific route and/or in a specific timeframe (~minutes) allows to adjust the application level behaviour:
- e.g. automation level, change vehicle speed, handover to driver, etc..
- This is not only valid for autonomous cars but also autonomous ships.
- Important to raise comfort level in the adoption of new autonomous technologies.
C-V2X supported by Predictive QoS involves collecting prediction-supporting data from multiple sources, making the prediction (with ML) and delivering the prediction, e.g. when entering patchy coverage area or cross-border area.
Prediction data spans vehicle data (location, planned route, speed), vehicle sensor data (hazardous road conditions, situational awareness), road infrastructure data (road topology, traffic signs), network data (coverage, load, resources), weather data (rainfall, visibility due to fog, wind speed), large events data (sports, concerts, festivals), analytics data (handover failure rates, statistics).
The mapping of common requirements at the 2nd 5G Vertical User workshop (July 2019) explored common requirements around predictive QoS, starting with automotive but also pinpointing other uses, such as:
- Industry automation and manufacturing may also benefit from predictive QoS (modern factories are changing/re-configuring dynamically) though more study is needed in 5G-ACIA.
- Public Safety also finds this important to better judge available QoS ahead of critical missions.
According to Steffen Schmitz, Volkswagen, Predictive QoS is a key enabler for complex use cases in the automotive sector.
Modest QoS is needed in the utilities and it must be reliable. Software-defined capability could significantly lower the cost of getting 5G.
Further standardisation work on this common requirement across several verticals would benefit from combining efforts, e.g. automotive, maritime, critical communications and public safety (e.g. PPDR), Industrial Internet of Things.
Pitches on common requirements for Predictive QoS: Maxime Flament, 5GAA, Steffen Schmitz, Volkswagen AG, Julian Stafford, EUTC, Philippe Agard, TCCA, Omar Erikkson, IALA.
Related verticals and associations: 5GAA, IALA, TCCA, PSCE, 5G-ACIA, EUTC.
Related 3GPP studies and reports for gap analysis:
- SA2 TR 23.786: Study on architecture enhancements for the Evolved Packet System (EPS) and the 5G system (5GS) to support V2X services.
- SA2 TS 23.287: Architecture enhancements for 5G system (5GS) to support vehicle-to-everything (V2X) services.
- SA2 TS 23.288: Architecture enhancements for 5G System (5GS) to support network data analytics services.
