Ericsson · 5

Ericsson questions whether capability 58-2-11 (Support of more than one activated PRS processing windows across all active DL BWPs) should be known only to the gNB or both the gNB and LMF, citing that legacy capability 27-23 is sent only to the gNB. They further question whether capabilities 58-2-9 and 58-2-10 should be known to the gNB, observing that these are currently set to N/A for the 'Need for the gNB to know' field. Ericsson argues that in legacy PPW configuration, details are known to the gNB so that the gNB can configure the PPW configuration as per UE capability, implying a similar need for AIML-related capabilities. They present this as a request for RAN1 feedback to resolve the visibility of these specific NR_AIML_air-Core capabilities.

Ericsson proposes that capability 58-2-11 should be known only to the gNB, arguing that this mirrors the legacy Rel-17 feature which is sent only to the gNB since it impacts the enabling of PPWs via RRC signaling. They further propose that capabilities 58-2-9 and 58-2-10 should be known to both the gNB and the LMF. Ericsson argues that since AIML-case1 positioning mirrors DL-TDOA configuration, these capabilities should follow the same principles as legacy features for gapless measurements (capabilities 27-3-2 and 27-3-3), which are reported to both nodes to inform of the types of priority handling the UE supports. They emphasize that in legacy PPW configuration, details are known to the gNB so that the gNB can configure the PPW configuration as per UE capability.

Ericsson clarifies the visibility of specific UE capabilities for PRS Processing Window (PPW) in the context of NR AIML air interface. They state that capability 58-2-11, which relates to support for more than one activated PRS processing window across all active DL BWPs, should be known only to the gNB, aligning with legacy non-AIML positioning behavior. Conversely, they confirm that capabilities 58-2-9 and 58-2-10 should be known to both the gNB and the LMF, ensuring the LMF has the necessary information for positioning configuration. This distinction ensures that the gNB can configure PPW details based on UE capability while the LMF retains awareness of relevant positioning features.

RAN2 is advocating for the inclusion of UE-side data collection capabilities for UE-side model training in AI/ML applications for the NR air interface, as evidenced by their endorsement of TP R2-2407807 and their request for RAN1 to incorporate this work into the technical report TR 38.843.