R1-2508524
discussion
Discussion on evaluation assumptions for 6GR air interface
From ZTE
Summary
ZTE provides 27 proposals and 1 observation on 6GR evaluation assumptions covering spectrum, antenna models, multi-TRP operation, channel models, NTN, traffic models, and sensing scenarios. The document recommends specific antenna configurations for BS and UE across multiple frequency bands and introduces new evaluation frameworks for heterogeneous networks, AI/ML services, and ISAC deployment.
Position
ZTE recommends specific BS antenna configurations for 7GHz (16 horizontal elements limited by 0.5m installation constraint), 15GHz (both HBF and ABF candidate options), 700MHz, and 2GHz bands. They support co-channel HetNet deployment with macro and micro layers sharing the same carrier frequency for dense urban and urban macro scenarios, arguing this is more typical than inter-frequency deployment. They propose studying DL/UL codeword-to-layer mapping, UL high-resolution precoding/dynamic port-selection, and corresponding UL-power-control/calibration as dedicated enhancements due to extra performance degradation observed with the Rel-19 handheld UT antenna model over legacy ULA/UPA. They propose asymmetric DL sTRP/UL mTRP architecture for heterogeneous networks to improve UL performance with NW energy savings. They recommend near-field and SNS channel models in TR 38.901, citing 1.5dB additional performance loss for near-field UEs using legacy 2D-DFT codebook weights. For traffic models, they propose extended FTP model with three fixed packet sizes per service type, support token-based GenAI traffic using PHY-layer packet parameters rather than per-token simulation, and recommend reusing XR traffic model structure for immersive communication. They define new sensing scenarios including Nearshore waters with ship targets and Highway with vehicle and infrastructure collapse targets.
Key proposals
- Proposal 2-1-1 (Sec 2.1): 6GR evaluation should consider BS antenna model defined in Table 2-1-1 to Table 2-1-4, covering 7GHz, 15GHz, 700MHz, and 2GHz with specific (M,N,P,Mg,Ng; Mp,Np) configurations.
- Proposal 2-2-1 (Sec 2.2): For dense urban and urban macro scenarios, co-channel HetNet deployment should be evaluated, where macro and micro layers share the same frequency.
- Proposal 2-3-1 (Sec 2.3): 6GR evaluation should consider handheld UT antenna model defined in Table 2-3-1, with up to 4 antennas for FR1, up to 8 for 7-24GHz, and up to 32 elements for FR2.
- Proposal 2-4-1 (Sec 2.4): 6GR evaluation should consider multi-TRP operation with CJT targeting FR1/around-7GHz and NCJT transmission targeting FR2/around-30GHz.
- Proposal 2-4-2 (Sec 2.4): 6GR evaluation should consider asymmetric DL sTRP/UL mTRP architecture for reducing deployment cost, ensuring UL coverage, and saving NW energy.
- Proposal 2-5-1 (Sec 2.5): 6GR evaluation should consider near-field and SNS channel models in TR 38.901.
- Proposal 2-6-1 (Sec 2.6): 6GR evaluation for NTN satellite should consider fully-connected mapping approaches.
- Proposal 2-6-2 (Sec 2.6): In 6GR NTN scenario for non-GSO systems with phased antenna array, antenna modeling in TR 38.901 can be used with consideration of circular polarization (LHCP/RHCP).
- Proposal 2-7-1 (Sec 2.7): For 6G study, the evaluation assumptions listed in Table 2-7-1 are recommended for performance evaluation of multi-layer heterogeneous network with assisting nodes.
- Proposal 3-1-1 (Sec 3): For link budget template, Candidate 1 reuses TR38.830 template with parameter values based on agreed simulation assumptions, reuse MCL row definition, and delete maximum range row; Candidate 2 reuses TR38.913 Table 7.10.1-1 without update.
- Proposal 4-1-1 (Sec 4.1): Extended FTP traffic model assigns each UE one of three service types according to pre-defined probabilities, with fixed packet sizes per service type, and discards packets exceeding PDB while treating UPT as zero.
- Proposal 4-2-1 (Sec 4.2): When discussing token size, it is necessary to clarify input data type (e.g., text, image) and AI/ML task type (e.g., general-purpose tasks, specific-purpose tasks).
- Proposal 4-2-2 (Sec 4.2): Generative AI traffic model supports option 1b with modification consisting of image-based GenAI, video-based GenAI, and chatbot service types using FTP model 3 with fixed parameter values.
- Proposal 4-2-3 (Sec 4.2): Study traffic models for AI/ML model training/inference with three service types—training data collection, inference result transmission, and model download—each using FTP model 3.
- Proposal 4-3-1 (Sec 4.3): For 6G immersive communication services, reuse existing XR traffic model structure with updated parameter values; RAN1 should postpone further discussions until receiving service characteristics from SA4.
- Proposal 5-1 (Sec 5): For ISAC, support deployment scenarios including Indoor factory, Dense urban, Urban macro, Rural, Highway, and Nearshore waters, in accordance with MS excel attachment assumptions.