R1-2600030
discussion
On remaining aspects of EE evaluations
From Nokia
Summary
This Nokia contribution presents 22 proposals and 8 observations on 6G Energy Efficiency for the FS_6G_Radio work item, covering base station and UE power consumption models, UE power saving methods, and idle mode evaluations.
Position
Nokia proposes adopting the 5G BS power consumption model (TR38.864 Category 2) as the 6G baseline while defining a new 'Category 2-plus' category with 10-20% relative power improvements across sleep and active states, plus faster transition times (light sleep 640ms→100ms, deep sleep 10s→5s) with common deep sleep normalization to enable meaningful 5G/6G comparison. They propose reusing the 5G scaling framework where only active-state dynamic power scales while static power remains fixed at P3 or 1.5*P3. For UE power saving, they propose supporting two adaptation delay values with T in {≥1ms, ≤5ms} and omit adaptation interruption in TDD, and present scaling factors for PDCCH+PDSCH, PDCCH-only, and micro-sleep slots across bandwidth utilization ratios from 5% to 400%. They require CFO inaccuracy of ~1ppm for WUS reception assuming SSS availability, arguing that 5ppm would impair coherent combining, and restrict WUS RX chains to 1 during ultra-deep/deep sleep states. For idle-mode WUS design, they argue the number of monitoring occasions (MOs) must be optimized as a function of subgroups to mitigate false wake-up probability.
Key proposals
- Proposal 1 (Sec 2): Adopt the 5G BS power consumption model as the reference for 6G and discuss necessary value adjustments based on factors like 6G numerology, parameters, and anticipated BS hardware advances.
- Proposal 2 (Sec 2): To reflect the anticipated BS hardware advances, define a new 6G BS power model category, “Category 2-plus”, for evaluations of deployments that leverage new and more power-efficient BS hardware.
- Proposal 3 (Sec 2): Improve the transition time and additional transition energy for light sleep (e.g. transition time from 640ms to ~ 100ms) and deep sleep (e.g. from 10sec to ~ 5sec) for Category 2-plus.
- Proposal 4 (Sec 2): Reuse the 5G BS power consumption scaling method as a starting point for 6G BS power model, where only active state power is scaled, the static part equals micro sleep (P3) or optionally 1.5*P3, and the dynamic part uses the same DL/UL formulas.
- Proposal (Sec 3.1): Assume that the adaptation delay T is sufficient to enable UE confirmation in UL, configurable in range of {≥1ms, ≤5ms}, omit adaptation interruption for TDD scenarios, and for FDD assume 1 slot switching interruption.
- Proposal (Sec 3.2): For XR traffic, assume that UE has a periodic UL activity at [5ms or 10ms] periodicity or aligned with C-DRX cycle.
- Proposal (Sec 3.3): Consider DCI-based UE receiver BW adaptation in study for UE power saving.
- Proposal (Sec 3.3): The maximum schedulable PDSCH throughput can be reduced to obtain UE power saving via reduced baseband processing by enabling network to configure the UE with maximum expected PDSCH allocation, TBS or rank in dynamic manner.
- Proposal (Sec 3.4.1): RAN1 should consider CFO inaccuracy of ~1ppm assuming the availability of SSS or other synchronization signals before WUS reception.
- Proposal (Sec 3.4.2): The number of RX chains shall be restricted to 1 when the sleep state is either ultra-deep or deep sleep as the primary objective is to reduce the power consumption.
- Proposal (Sec 4.1): RAN1 shall consider optimizing the number of MOs as a function of number of subgroups to avoid false wakeups, while addressing the WUS monitoring power.
- Proposal (Sec 3.3): C-DRX should continue to be considered as a baseline for further power saving enhancements in 6GRs.
- Proposal (Sec 3.3): Consider UE PDCCH monitoring adaptation in study for UE power saving.
- Proposal (Sec 3.1): Consider further the need of two adaptation delay values, and only support two if sufficient justification exists.
- Proposal (Sec 3.2): For VoIP traffic, assume activity based on VoIP model in both link directions, DL and UL.