NTN: from 5G NR to 6G

NTN: from 5G NR to 6G MOHAMAD SAYED HASSAN Qualcomm France, S.A.R.L. France [email protected] ALBERTO RICO ALVARINO Qualcomm Technologies, Inc United States of America [email protected] CHIRANJIB SAHA Qualcomm Technologies, Inc United States of America [email protected] JUN MA Qualcomm Technologies, Inc United States of America [email protected] QIANG WU Qualcomm Technologies, Inc United States of America [email protected] JI LIANGHAI Qualcomm Technologies, Inc United States of America [email protected] LE LIU Qualcomm Technologies, Inc United States of America [email protected] Abstract—5G New radio (NR) supports different vertical domain services (e.g., Vehicle to everything (V2X), Nonterrestrial networks (NTN), Virtual Reality (VR)), etc. In this paper, we provide an overview of 3rd generation partnership project (3GPP) NTN covering Release 17 and the current work items for Release 18. We also discuss the main research challenges to enhance NTN services. We then provide the innovations and challenges to be considered in 6G design. Keywords—5G NTN, 6G, satellite communications, 3GPP. I. INTRODUCTION This paper provides an overview of the evolution of 5G NR to support NTN communications. We focus on radio access network (RAN) covering from physical layer to upper layers. 3GPP has provided definition for the non-terrestrial network [1]. NTN refers to networks, or segments of networks, using an airborne or spaceborne vehicle for transmission: • Spaceborne vehicles: Satellites (including Low Earth Orbiting (LEO) satellites, Medium Earth Orbiting (MEO) satellites, Geostationary Earth Orbiting (GEO) satellites as well as Highly Elliptical Orbiting (HEO) satellites). • Airborne vehicles: High Altitude Platform Systems (HAPS) encompassing Unmanned Aircraft Systems (UAS) including Lighter than Air UAS (LTA), Heavier than Air UAS (HTA), all operating in altitudes typically between 8 and 50 km, quasi-stationary. NR NTN has experienced different phases including Release-15 Study Item (SI), Release-16 SI, Release-17 Work Item (WI) and release 18 WI for NTN evolution. The objective of Release-15 SI is to define the 3GPP channel model for the NTN, provide detailed description of deployment scenarios and related system parameters and identify key impact areas on the NR [2]. The objective of Release-16 SI is to identify necessary features to adapt NR protocol for NTN with a priority on satellite access [3]. The objective of Release-17 WI is to specify the solutions enabling NR to support LEO and GEO based NTN with basic support for HAPS and ATG based NTN [3], focusing on bands in frequency range 1 (e.g. S and L bands). Release 18 first meeting for NTN enhancements has started in May 2022. Considering the NTN characteristics such as large propagation delay and satellite movement, the WI objectives [4] are to specify enhancing features for coverage, to provide mobility and service continuity enhancements and to support scenarios for NR-NTN deployments in frequency bands above 10 GHz. In [5], an overview about NTN architecture has been presented. The physical layer features and higher-layer protocols are not fully covered. In [6], a survey on NTN systems is presented. The survey focusses on the NTN use cases and architectures, the satellite network roadmap and role of NTN in cellular communications, 3GPP research activities. This contribution focuses on NTN physical layer and upper layers from 3GPP standard point of view. Further, it provides the challenges and innovation directions to be considered in 6G based on the issues encountered during release 17 and 18. The paper is organized as following: in section II, a summary about the NTN system architecture supported in release 15 to release 18 is described. In section III, we focus on physical layer aspects. The time and frequency synchronization and the enhancements agreed in release 17 are presented. In addition, the enhancements on Hybrid Automatic Repeat reQuest (HARQ) to support NR for NTN systems are described. We provide an overview about the work item in release 18. Section IV provides the upper layer design. In section V, we provide our view for NTN in 6G and the aspects that require enhancements or new design with focus to capacity solutions. Section VI concludes this contribution. II. NTN SYSTEM ARCHITECTURES 3GPP discussed two options for system architecture [8-9]: • A bent pipe payload: The satellite payload corresponds to a Radio Frequency (RF) filtering, frequency conversion and amplification. This is referred as transparent satellite. • A regenerative payload: The satellite payload corresponds to a RF filtering, frequency conversion and amplification as well as demodulation/decoding, switch and/or routing, coding/modulation. This is effectively equivalent to having base station functions (e.g., gNB) on board the space/airborne vehicle. There is no need to modify the NG-RAN architecture to support transparent satellite access. The timing enhancement is needed to cope with the long delay of the feeder link, between 173979-8-3503-3870-6/23/$31.00 ©2023 IEEE WISEE 2023 2023 IEEE International Conference on Wireless for Space and Extreme Environments (WiSEE) | 979-8-3503-3870-6/23/$31.00 ©2023 IEEE | DOI: 10.1109/WiSEE58383.2023.10289427 Authorized licensed use limited to: Politecnico di Milano. Downloaded on December 08,2025 at 17:44:59 UTC from IEEE Xplore. Restrictions apply. gateway and the satellite, and service link, between the user equipment and the satellite (see Fig. 1). Release 17 considers the system architecture with transparent satellite. FIGURE 1. NTN system architecture. Table I summarizes the main frequency bands and deployment scenarios with Frequency-division duplexing (FDD) for: 1) Moving beam case where the beam footprint on the ground is moving with the satellite’s movement. 2) Quasi earth fixed beam case where NTN network (NW) provides beams/cells fixed with respect to a certain location on the earth during a certain time duration (e.g., with steerable beams at the satellite). TABLE I. Deployment scenarios with FDD duplexing Deployment GEO LEO LEO Frequency range of Service Link Around 2 or 20 GHz for DL Around 2 or 30 GHz for UL Around 20 GHz for DL & 30 GHz for UL Around 2 GHz for DL & UL Beam pattern Quasi earth fixed beam Quasi earth fixed beam Moving beam III. PHYSICAL LAYER ASPECTS FOR NTN Unlike in terrestrial network (TN) systems, communications in NTN systems need to overcome large and time-varying delay and Doppler frequency offset. This requires special time and frequency synchronization, HARQ enhancements and redesign of some other physical layer aspects. A. Time synchronization Release-17 and 18 design assumes that the NTN UE is equipped with Global Navigation Satellite System (GNSS) capability which can be used by the NTN UE to know its position. This information, combined with satellite ephemeris (either broadcasted by the network or provisioned via other means) information, allows the UE to carry out frequency precompensation and timing adjustment (see Fig. 2). The ephemeris information can be represented by either position-velocity vector format or orbital element format. In either case, UE applies orbit propagation to determine the delay and Doppler frequency of the service link at any time. Some NTN systems architecture requires the UE to estimate the feeder link delay and delay variation. For this purpose, the network provides the UE with another information called TA_common, which is in the form of coefficients of a quadratic polynomial and indicates the feeder link variation over time. Combined with location knowledge, the UE can apply a timing advance described by the following equation. 𝑇𝑇𝐴 =(𝑁𝑇𝐴 + 𝑁𝑇𝐴,𝑈𝐸−𝑠𝑝𝑒𝑐𝑖𝑓𝑖𝑐 + 𝑁𝑇𝐴,𝑐𝑜𝑚𝑚𝑜𝑛 + 𝑁𝑇𝐴,𝑜𝑓𝑓𝑠𝑒𝑡)× 𝑇𝑐 (1) • NTA is the closed-loop timing advance element based on feedback received from the NW. • NTA,UE−specific is UE self-estimated TA based on ephemeris and UE location information to precompensate for the service link delay. • NTA,common is network-controlled common TA that can be computed by the UE based on received TA_common parameters. • NTA,offset is a fixed offset reserved by the NW for additional timing advance control. • Tc is the smallest timing unit equal to ~0.508ns. FIGURE 2. Timing pre-compensation where the GW is used as reference point. Besides the newly designed timing advance formula, the NTN system design also introduces two parameters, K_offset and k_mac. Parameter K_offset is used to account for the extended round trip delay between the UE and system timing reference point in NTN systems. Parameter k_mac is used to account for the round-trip delay between gNB and system timing reference point. Note that system timing reference point refers to the point in the communications link where downlink (DL) and uplink (UL) slots are aligned. B. Frequency synchronization According to Release 17 NTN design, an NTN UE is only required to compensate for service link Doppler frequency shift. The Doppler frequency shift associated with the feeder link should be compensated either by the ground NW or the satellite payload. Just like timing synchronization, the ephemeris information, combined with UE location knowledge, can be used to estimate the service link Doppler frequency. When pre-compensating for the UL Doppler frequency shift, UE also needs to account for the fact that its own reference clock is not accurate. C. HARQ enhancement In TN, the NW may schedule a transmission from/towards the UE based on the reception status of a previous transmission associated to a considered HARQ process. In NTN release 17 [19-20], an increase of the number of HARQ processes has been specified. The maximum number of supported HARQ process is extended from 16 in TN to 32 in NTN for UL and DL. Increasing the number of HARQ processes, results in increased complexity/cost at the UE. In addition to increasing the number of HARQ processes, release 17 supports the possibility of enabling/disabling HARQ feedback: NW can configure an NTN UE to enable/disable HARQ feedback for downlink transmission per HARQ process via UE specific RRC signaling (dynamic signaling is discussed in release 18 for IOT over NTN). In this case, NW may reuse 174 Authorized licensed use limited to: Politecnico di Milano. Downloaded on December 08,2025 at 17:44:59 UTC from IEEE Xplore. Restrictions apply. the HARQ process without the need to wait for the HARQ feedback from the UE, which solves the issue of HARQ stalling. Enabling/Disabling HARQ has an impact to HARQ codebook. Release 17 supports solutions to treat these issues. For type-1 HARQ (i.e., semi-static codebook), it has been agreed that the UE will report NACK only for the feedbackdisabled HARQ process regardless of the decoding results of corresponding PDSCH. For Type-2 HARQ codebook (i.e., dynamic codebook) in NTN, the codebook size with HARQACK codebook is reduced by only including HARQ-ACK of PDSCH with feedback-enabled HARQ processes. For Type-3 HARQ codebook in NTN, the UE should skip the codebook feedback for a feedback-disabled HARQ processes. The Type3 codebook size is reduced by excluding the bit positions of disabled HARQ processes. D. Release 18 enhancements In Rel-18, a new work item [13] has been approved with the objective of defining enhancements for NTN communications, including: • Support deployments in frequency bands above 10 GHz. Very Small Aperture Terminal (VSAT) devices are supported in above 10 GHz. • Enhance the coverage enhancement when addressing smartphones with more realistic assumption on antenna gains. LEO-1200 operating at LoS and considering -5dBi UE antenna gain is agreed for performance evaluation. The targeted services are Voice over IP (VoIP) using lowrate codec Adaptive Multi-Rate (AMR) 4.75 kbps and data transmission services with low data rate of 3 kbps. The objectives to determine the bottleneck channels or signals. • Provide mobility and service continuity enhancements considering the NTN characteristics such as large propagation delay and satellite movement. • With NTN it is possible to deploy very large cells with coverage over different countries. NW operators must know reliably the location information of a UE attached to the NW to select the appropriate core NW. This enables services like emergency call, Public Warning System, Lawful interception. The GNSS measurements send by UE to the NW can be faked (select another PLMN which leads to another core NW). In addition, the GNSS measurements can be send before Access Stratum (AS) security is set up which raises privacy issues. To solve this issue, 3GPP defined RAT dependent positioning methods is needed to cross check the UE location reported [14][15]. IV. UPPER LAYER ASPECTS A. Mobility support Comparing to TN, a NW node such a satellite in NTN can move with a high speed, e.g., in case of LEO and MEO satellites. Thus, though a UE in NTN may be static and not move, it may still need to be served by different satellites at different times. Meanwhile, the movements of satellites are predictable, which makes it possible for NW to make certain predictions, such as when the UE’s current connected satellite will stop serving the UE and which satellite can next serve the UE. These special characteristics are considered in Rel-17 for NTN mobility support. To support handover for a UE in RRC_CONNECTED state, the legacy handover and conditional handover procedures are leveraged. In addition, the following enhancements have been introduced in the specifications: • Event A4-based conditional handover (CHO) is introduced, i.e., the UE is triggered to apply conditional handover to a configured candidate target cell, if the radio conditions of the candidate target cell become better than a configured threshold. • A timing-based CHO triggering event and a locationbased CHO triggering event, in combination with the measurement-based event, are introduced. It is noted, since the radio condition is still the most dominant factor impacting the handover decision and its performance, measurement-based event should be configured to the UE for UE to determine if it should trigger the execution of the timing-based or location-based CHO. In other words, besides the timing-based event or the location-based event, the configured measurement event for the candidate target cell, e.g. A3, A4, or A5 event, needs to be fulfilled, to trigger a timing-based or location-based CHO execution. For a UE in RRC_IDLE or RRC_INACTIVE state, the UE performs cell-reselection to camp on a proper cell, based on its neighbor cell measurement result. Like handover, a timingbased and a location-based mechanisms have been developed to enhance neighbor cell measurements in the quasi-earth fixed cell scenario. For example, the serving cell will broadcast when it stops serving a UE’s located area, and the UE should start performing neighbor cell measurements before that. In another example, the serving cell broadcasts a reference location, e.g., its cell center, and the UE should perform neighbor cell measurements if its distance to the serving cell’s center is above a configured threshold. B. NTN-specific SIB A new NTN-specific SIB, i.e., SIB19 [19], has been introduced to carry the NTN-specific configurations, such as the common TA parameters and the ephemeris information for calculating NTA,common and NTA,UE−specific in Equation (1), respectively. The ephemeris information in SIB19 indicates the satellite’s location at an epoch time, which is also indicated by SIB19 to the NTN UEs. Accordingly, a UE can utilize the broadcasted ephemeris information to predict the satellite’s position, which is further used by the UE to derive its signal propagation delay to the satellite and calculate its NTA,UE−specific. It is noted that, to ensure a good accuracy of NTA,UE−specific calculated by the UE, it is important to ensure a good accuracy of the ephemeris information at the UE. For that purpose, the ephemeris information needs to be updated to the UE by the NW from time to time. Thus, in SIB19, the NW indicates a validity duration associated with the epoch time and the ephemeris information provided in that SIB19. Accordingly, the UE can use the ephemeris information to derive its uplink timing advance within the validity duration starting from the epoch time. The UE may consider assistance information valid as soon as it is received. In other words, after the validity duration expires, the ephemeris information is considered as invalid by the UE for deriving its uplink timing advance. Thus, to avoid any uplink transmission interruption caused by an expired ephemeris, the UE should try to obtain a new 175 Authorized licensed use limited to: Politecnico di Milano. Downloaded on December 08,2025 at 17:44:59 UTC from IEEE Xplore. Restrictions apply. ephemeris information from a new SIB19 before the current ephemeris expires. In addition, other NTN-specific information may also be broadcasted in SIB19. For example, neighbor cell’s NTN configuration can also be transmitted there, which can help the UE to derive the right timing and frequency offset for measuring a neighbor cell, based on its estimated neighbor satellite’s position. C. NTN-TN and NTN-NTN mobility and service continuity enhancements in Rel 18 In Rel-18, it is proposed to consider further enhancements to improve the mobility support and serving continuity in NTN. For example: • The timing-based and location-based cell reselection mechanisms in Rel-17 have only considered the quasiearth fixed cell scenario and, thus, Rel-18 will consider how to enhance the cell reselection mechanisms for the earth moving cell scenario. • To improve the NTN efficiency, Rel-18 will consider how to reduce the signaling overhead related with the handover procedure. • Due to the large coverage of an NTN cell, a first UE served by the NTN cell may locate on the boundary between NTN and TN, while a second UE served by the same NTN cell may locate far away from TN. Thus, if both UEs in RRC_IDLE or RRC_INACTIVE state is configured by a broadcasted SIB to perform neighbor measurement for TN in their cell reselection procedure, it would introduce unnecessary power consumption for the second UE. To solve this issue, in Rel-18, enhancements to support the cell reselection will be considered, with a higher priority to consider the mobility scenario between NTN and TN, as described. • Due to the movement of a satellite, it is possible that a satellite must switch its feeder link between two different gateways on the ground. In this case, Rel-18 will study the need and specify enhancement(s) for the signaling over Xn and NG interface to support the feeder link switch-over and CHO, e.g., by exchanging necessary information between gNBs. V. NTN IN 6G The functionality of NTN access has been added to NR after its initial Release. It is possible that 6G will be designed and introduced with NTN support from the beginning. The learnings and challenges from NR NTN can be used as input for further improvement on the 6G design. One important aspect for 6G design to consider is the dense constellation. We categorize our view for 6G design as following: A. Coverage NTN enhancements is a topic in RAN1, Release 18 to study NTN-specific solution to enhance the uplink and downlink coverage. Voice over IP and low data rate are assumed as main use cases for coverage study under the assumption of handset terminal (maximum power is limited based on UE category). The NTN-specific solutions in release 18 will not address fundamental change in physical layer design. 6G waveform for NTN should consider: • Line Of Sight (LOS) communication. • Resiliency to Doppler. • Low Peak to Average Power Ratio (PAPR). Lower PAPR can enhance the coverage. • Low Out of Band (OOB) is required mainly for fullduplex operation, integrated satellite-terrestrial communications. • Mechanism for fast switching between waveforms as it is expected that 6G to support multiple waveforms to cover different use cases/scenarios. B. Capacity/reliability The first deployment of NR NTN will be coverage-centering, ensuring global coverage supporting low-data rate applications. With emerging dense constellations, especially at LEOs, 6G NTN may consider further coverage and capacity enhancements through multi-satellite connectivity. FIGURE 3. (a) Iridium Constellation (orbit height 780 km, inclination 87°, min elevation angle 8.6°. (b) Box plot of number of visible satellites in latitude range. Note that at 30°-50° latitude range, where most of the continental landmass is located, number of visible satellites is more than 1 with high probability. The motivation of NTN multi-connectivity can be derived from the visibility of more than one LEO satellites. In Fig. 3, we demonstrate how a LEO constellation targeted for global coverage leads to the visibility of more than one satellites closer to the pole. Users in such regions can potentially be served by multiple satellites by leveraging the techniques like multiple transmission-reception point (mTRP), carrier aggregation (CA), or dual connectivity (DC). Multiconnectivity can potentially solve the problem of satellite blockage by improving the LOS probability. Also, because of non-ideal UE antenna characteristics, the UE needs to be pointed at a particular direction in the sky to lock a particular satellite and close the link. Multi-satellite connectivity may significantly reduce the time and effort to find a satellite connection. Furthermore, mTRP may bring the traditional advantages and degrees of freedom of MIMO, such as: (a) space and multiuser diversity gain, (b) spatial multiplexing gain, (c) array and coding gain, and (d) interference reduction. A. NTN MIMO Design The main challenge of designing mTRP over NTN is the large difference between the delays and Dopplers of multiple TRPs (especially at LEOs). Current mTRP design can handle delay difference within cyclic prefix (CP). This means, in the DL, signals from two (or more) satellites need to arrive at the UE within CP. Also, the Doppler frequencies need to be compensated so that the resources from two (or more) satellites align at the UE within a tolerable range. To this end, NR NTN 176 Authorized licensed use limited to: Politecnico di Milano. Downloaded on December 08,2025 at 17:44:59 UTC from IEEE Xplore. Restrictions apply. needs to consider some fundamental designs in RAN such as follows: 1) Delay-Doppler Beams: Since NTN cells are larger than the typical TN cells, different regions in a cell may have significantly different delay and Doppler differences from the TRPs. An NTN cell may be divided into several regions (or beams) where the delay and Doppler differences from TRPs are within certain tolerable range. For example, consider a 2 TRP scenario. The target coverage that can be provided mTRP service can be split into regions ℛ(𝑛, 𝑚) where delay difference between satellites 1 and 2 is in the interval [(𝑛 − 1)𝐶𝑃,𝑛𝐶𝑃] and the Doppler difference is in the interval [(𝑚 − 1)𝛥𝑓, 𝑚𝛥𝑓], 𝑚,𝑛 ∈ 𝕫, 𝐶𝑃 is the length of cyclic prefix, and 𝛥𝑓 is the tolerable frequency offset. The network can configure DL reference signal (e.g., Tracking reference signal (TRS)) over a single frequency network (SFN) for each beam, where the TRS from satellite 2 is applied a given time and frequency offset with respect to the TRS from satellite 1. Note that unlike spatial beamforming by applying spatial filters on signals in FR2, this operation is equivalent to beamforming in the delay-Doppler domain. 2) New Scheduler Design: Once the UE identifies a particular beam, the scheduler needs to consider the relative time and frequency offset during the joint transmission of data. If the UE is in ℛ(𝑛,𝑚), then the joint scheduler needs to move the PDSCH grants for TRP 2 by 𝑛𝐶𝑃 in time and 𝑚𝛥𝑓 in frequency with respect to TRP 1. Thus, additional time-frequency gaps need to be introduced while co-scheduling multiple UEs in different beams. 3) Receiver Design Problems: The scheduler complexity of scheduling UEs in different beams can be reduced by increasing the receiver complexity at the UE. Since the UE receivers are equipped with single FFT, the signals from multiple satellites need to overlap in time and frequency at the receiver. Furthermore, if the UE receivers have multiple FFTs, where each processing chain can individually process the streams from multiple TRPs, then it removes the constraint on the joint scheduler as described before. However, the receiver needs to cancel the inter-stream interference. Furthermore, the MIMO spatial multiplexing or diversity gains can no longer be exploited. 4) Uplink Transmission: Like the DL, the UL has its own set of design problems. To keep the UE transmitter design simple, the UE can transmit a single PUSCH, but the network applies different time frequency compensation to receive the PUSCH at multiple TRPs. Naturally, similar constrains may apply for the uplink scheduler design. B. Simulation results To demonstrate potential benefits of MIMO over NTN, we provide preliminary simulation results for 2 TRP downlink (see Fig 4(a)). We consider a 19-cell layout with three regions 𝒮𝑖, 𝑖 = 1, 2,3. UEs in 𝒮1 are served by satellite 1, UEs in 𝒮2 are served by satellite 2, and UEs in 𝒮3 are provided mTRP connectivity from both satellites. Satellite 1 and 2 are equipped with circularly polarized directional antennas. We focus on the UEs in S3, which are dual-polarized and observes 2 Tx-2 Rx effective channel in DL. The simulation parameters are presented in Table III. We assume that satellites apply noncoherent joint transmission since it is not possible to control the phases of two satellites coherently. In this simulation, we neglect the effect of delay and doppler differences of two TRPs (assuming that these differences have been resolved perfectly) and assume full-buffer traffic. In Fig. 4(b), we plot the rate distributions of different MIMO schemes. We observe that 2 satellite transmission have at least 5 Mbps throughput gain over single satellite. This gain is more prominent for frequency reuse factor (FRF) of 3 because of the reduction in intercell interference (see Fig. 4(c)). Furthermore, in FRF 3, 2-layer transmission provides an additional 10 Mbps median gain over single layer transmission. Note that, in real system, these gains will be limited once we consider the scheduler constraints, and inter-stream interference because of the loss of orthogonality during joint transmission with large time and frequency compensations. Although the initial results look quite promising for the mTRP NTN design in 6G. Figure 4. (a) Cell/beam layout for System-level simulation with FRF 1. Users are dropped in all beams. Only the users in the middle region (served by satellites 1 and 2) are considered for statistics, (b) Throughput distribution for FRF 1, (c) Throughput distribution for FRF 3. Table III. SIMULATION PARAMETERS, SEE TABLE 6.1.1.1.-1, SET 1, LEO-600 IN [23] FOR DETAILS. Variable Value Thermal Noise power -174 dBm/Hz Noise Figure 7 dB Sat EIRP density 34 dBW/MHz Bandwidth 10 MHz UE antenna pattern Quasi-isotropic Sat antenna pattern Bessel function Max antenna gain of UE 0 dB Sat antenna aperture 2m C. Scalability: 3D and multi-layer network 6G NTN expects an integrated space-air-ground NW (Fig. 5): • Multi-segment NW with massive NTN nodes. In table IV, we summarize the altitude of different types of NTN nodes. Table IV. Altitude of NTN nodes Satellite LEO/MEO /GEO Air-to-ground NW HAPS UAV Altitude 600Km & above ~10 Km ~20 Km ~100m • Inter-node link supports connections among different nodes. Higher frequencies (> 50GHz and sub-THz) are expected to be applicable as inter-satellite link. The link between satellite and unmanned aerial vehicle (UAV) is 177 Authorized licensed use limited to: Politecnico di Milano. Downloaded on December 08,2025 at 17:44:59 UTC from IEEE Xplore. Restrictions apply. expected to use similar frequencies as the serving link between satellite and UE. The UAV NTN node is seen as a NW node and/or a user device. Service anywhere and anytime, with LoS or NLoS condition • Backhaul support: non-terrestrial nodes can serve as wireless backhaul for the terrestrial nodes [9]. The NTN backhauling can be used when no wired backhaul solutions are available. • Coordination between NTN and NTN, as well as between NTN and TN. The NTN-TN and NTN-NTN mobility and service continuity enhancements is an item in RAN2 Release 18. In 6G, it is expected to support inter satellite link with denser satellite constellation and with different system architecture (transparent or regenerative payload), thus NTN-TN or NTN-NTN mobility should be revisited for 6G design. Figure 5. 3D and multi-layer network. D. Spectrum sharing Spectrum is essential for operators. Spectrum sharing between TN and NTN is essential to mitigate interference between TN and NTN. Dynamic spectrum sharing between TN and NTN is a good candidate for 6G NTN design. A dynamic spectrum sharing between TN distributed unit (DU) / central unit (CU) and NTN distributed unit (DU) / central unit (CU) requires TNDU/CU and NTN-DU/CU to interact. New requirements for interface between TN-DU/CU and NTN-DU/CU are needed. E. Regenerative payload Having more processing at satellite aims to reduce the latency. Regenerative payload can allow the satellite to support a distributed Unit (DU). This is requiring new NG-RAN architecture and new interfaces to be defined between DU and CU. With regenerative payload architecture, more processing is expected at satellite side which allows from one side to reduce the energy consumption as the satellite is empowered by solar system. From other side, more processing at satellite requires intelligent cooling system as the traditional cooling systems (e.g., using fans, liquid cooling, …) applied at TN cannot be applied. Regenerative architecture requires to enhance or to re-consider the interface between DU and CU as due to satellite movement, frequent switch of DU-CU interface is expected. VI. CONCLUSION Non-terrestrial NWs over NR have been standardized in release 17 and enhancements are currently discussed in release 18. NTN enhancements is considered as one of the key components for 6G. In this paper we overviewed release 17 and release 18 from physical layer to upper layers. We discuss the challenges that need to be addressed in 6G design. NTN coverage is one of the main challenges to consider for NTN physical layer design. The two architectures discussed in NR can be supported in 6G. Integrated TN-NTN is one of the fundamental objectives for 6G. The integration of TN-NTN can allow to support mobility and spectrum sharing between TN and NTN. REFERENCES [1] Solutions for NR to support non-terrestrial networks (NTN), 3GPP RAN#91-e, RP-210908, Thales. [2] Study on NR to support Non-Terrestrial Networks, 3GPP RAN1 # 88bis, RP-171450, Thales, Dish network, HUGHES Network Systems Ltd, ESA. [3] RP-190710, Study on solutions for NR to support non-terrestrial networks (NTN), RAN #83, Thales. [4] NR NTN (Non-Terrestrial Networks) enhancements, 3GPP RAN #95e, RP-220953, Thales. [5] X. Lin et al., “5G from Space: An Overview of 3GPP,” IEEE Communications Standards Magazine, vol. 5, no 4, p. 147 – 153, December 2021. [6] F. Rinaldi et al., “Non-terrestrial networks in 5G & beyond: A survey,” IEEE Access, vol. 8, pp. 165178-165200, 2020. [7] Study on New Radio (NR) to support non-terrestrial networks, 3GPP TR 38.811 V15.4.0. [8] Solutions for NR to support non-terrestrial networks (NTN), 3GPP TR 38.821 V16.1.0. [9] M. Giordani and M. Zorzi, “Non-Terrestrial Networks in the 6G Era: Challenges and Opportunities,” IEEE Network, vol. 35, no. 2, pp. 244251, Mar. 2021. [10] I. P. Nasarrei, T. Levanen, K. Pajukoski, A. Lehti, E. Tiirola, and M. Valkama, “Enhanced Uplink Coverage for 5G NR:Frequency-Domain Spectral Shaping With Spectral Extension”, IEEE open journal of the communication society, vol 2, pp. 1188-1204, June 2021. [11] N. Michailow, M. Matthé, I.S. Gaspar, A.N. Caldevilla, L.L. Mendes,A. Festag and G. Fettweis, “Generalized Frequency Division Multiplexing for 5th Generation Cellular Net-works,”, IEEE Transaction on Communications, Vol. 62, No. 9, Sep. 2014. [12] F. Schaich and T.Wild “Waveform contenders for 5G - OFDM vs. FBMC vs. UFMC”, 6th International Symposium on Communications, Control and Signal Processing (ISCCSP), May 2014. [13] Revised WID: NR NTN (Non-Terrestrial Networks) enhancements, 3GPP TSG RAN Meeting #97-e, RP-222654. New SID: Study on requirements and use cases for network verified UE location for Non-Terrestrial-Networks (NTN) in NR, 3GPP TSG RAN Meeting #96, RP-221820. [14] Study on requirements and use cases for network verified UE location for Non-Terrestrial-Networks (NTN) in NR, 3GPP TR 38.882, V18.0.0. [15] Radio Resource Control (RRC) protocol specification, 3GPP TS 38.331, V17.2.0. [16] RP-221348, Study on Artificial Intelligence (AI)/Machine Learning (ML) for NR Air Interface, 3GPP TSG RAN Meeting #96, Qualcomm Technologies, Inc. [17] Study on enhancement for Data Collection for NR and EN-DC, 3GPP TR 37.817, V17.0.0. [18] Physical layer procedures for control, 3GPP TS 38.213 V17.0.0.0. [19] Physical layer procedures for data, 3GPP TS 38.214 V17.0.0.0. [20] M. Polese, L. Bonati, S. D'Oro, S. Basagni, T. Melodia, “Understanding O-RAN: Architecture, Interfaces, Algorithms, Security, and Research Challenges”. arXiv:2202.01032. [21] R. De Gaudenzi, M. Luise and L. Sanguinetti, "The Open Challenge of Integrating Satellites into (Beyond-) 5G Cellular Networks," in IEEE Network, vol. 36, no. 2, pp. 168-174, March/April 2022, doi: 10.1109/MNET.011.2100116. [22] “Technical Specification Group Radio Access Network; Solutions for NR to support non-terrestrial networks (NTN) (Release 16),” 3GPP TR38.821, version 16.1.0. 178 Authorized licensed use limited to: Politecnico di Milano. Downloaded on December 08,2025 at 17:44:59 UTC from IEEE Xplore. Restrictions apply.