Coexistence Study Between GEO and LEO Satellites in the Q/V-Band

Coexistence Study Between GEO and LEO Satellites in the Q/V-Band Xavier Leturc, Dorin Panaitopol, Sorya Tong, Christophe J. Le Martret Thales SIX GTS SAS, France [email protected] Abstract—This paper studies the coexistence between a geosynchronous Equatorial orbit (GEO) satellite at nadir and a low Earth orbit (LEO) satellite. Both satellites operate in the Q/V frequency band, identified as potential candidate for the service link of 6G non-terrestrial communications. The two satellites on different orbits use either the same communication channel or adjacent channels. As a consequence, communications between the satellites and their associated user equipment on Earth are subject to interference in both downlink (DL) and uplink (UL), thus reducing the throughput. The level of experienced interference depends on the LEO satellite elevation, and on the adjacent channel interference ratio (ACIR) in case of adjacent channel interference. The coexistence results obtained for both DL and UL of the two considered orbits provide guidelines regarding the maximum elevation for the LEO satellite at which the 3GPP requirements are fulfilled. For instance, the required value of ACIR with respect to the maximum acceptable throughput loss is evaluated. Index Terms—Coexistence, GEO, LEO, 6G, Q/V-band. I. INTRODUCTION The Q/V frequency band corresponds to a carrier frequency between 37.5 GHz and 42.5 GHz for the downlink (DL) (Qband), and between 47.2 GHz and 50.2 GHz for the uplink (UL) (V-band), as identified in [1] for the seemless integration of non-terrestrial network (NTN) within the 6G. As a matter of fact, in [1], Task 4.3 is devoted to study the spectrum sharing in the Q/V-band between terrestrial network (TN) and NTN, and between several satellites with different orbits, see [2]. This paper studies the possible coexistence between a geosynchronous equatorial orbit (GEO) satellite located at Nadir and a low Earth orbit (LEO) satellite with elevation α assuming that the two satellites operate on the Q/V-band, and considering either co-channel or adjacent channel coexistence. A conventional technique to enable spectrum sharing between satellites with different layers is to define an exclusion angle, corresponding to defining a minimum angle between the GEO and the LEO boresight, minimum angle below which the interference caused by each satellite on the other one is considered as unacceptable, i.e. it causes an unacceptable performance loss. Defining proper exclusion angle is thus crucial before deploying new LEO satellites constellations. It is worth mentioning that other coexistence techniques such as beam steering [3] or coordinated beam hopping [4] techniques can be used to improve coexistence, but they are beyond the scope of this paper. Spectrum sharing between different satellites with possibly different orbits has attracted a lot of interest during the last decade, especially for coexistence in either the Ka or in the Ku-bands [5]. In [6], the authors develop a power control algorithm keeping the level of interference caused by a LEO satellite on a GEO satellite operating in the Ka-band is below a threshold. In [7], the authors study spectrum sharing between non-geostationary orbit (NGSO) constellations from different operators involving LEO and medium Earth orbit (MEO) satellites operating in the Ka and in the V-band. The authors in [8] study coexistence between several LEO and MEO satellites, by quantifying the level of interference as a function of the exclusion angles. In [9], the authors provide a comprehensive coexistence study between several constellations involving GEO and NGSO satellites in the Ku-band. The authors study several coexistence techniques including exclusion angle strategy, referred to as ”GSO protection technique”. The authors in [10] minimize the exclusion angle for the coexistence of the DL of a GEO and a LEO satellite operating at 10.7 GHz. Finally, [11] and [12] study the exclusion angle strategy for the coexistence between LEO and GEO satellites in the Ku and the Ka-band respectively. All these studies consider cochannel interference. Moreover, there is no paper studying the coexistence between GEO and LEO satellites in the Q/V-band to the best of our knowledge. The main contribution of this paper is a coexistence study between a GEO satellite and a LEO satellite in the Q/Vband assuming both co-channel interference and interference between adjacent channels. The obtained results provide guidelines regarding the exclusion angle that should be ensured between both satellites, and regarding the level of protection that must be ensured between adjacent channels that is driven by the adjacent channel interference ratio (ACIR) [13] (more details are provided in Section III-A). The rest of this paper is structured as follows. Section II describes the system model. Section III explains the coexistence simulation methodology. Section IV is devoted to numerical results. Finally, Section V draws concluding remarks. II. SYSTEM MODEL We study the coexistence between a GEO satellite located at Nadir and a LEO satellite with elevation α. We assume that the two satellites are pointing towards the same point on the Earth surface, i.e. the center of the beam of both satellites is NTN UE GEO satellite LEO satellite Fig. 1: Considered setup with a GEO satellite at Nadir and a LEO satellite with αelevation angle value. Both satellites points towards the same point on Earth, where the UE is located. located at the same place, and that they are serving co-located user equipment (UE) also located at the center of the beam. Such a setup is depicted in Fig. 1. The exclusive angle strategy consists in defining a minimal separation angle φex between the boresight of both satellites in order to keep their respective level of interference below a threshold. Since the GEO satellite is at Nadir, defining φex is equivalent to defining a maximum elevation αmax at which the LEO satellite can operate, and then deduce φex =π/2−αmax. Our objective is to study the coexistence between the two satellites when they use either the same communication channel or adjacent channel, interfering thus with each other. Such a study requires the computation of the link budget, i.e. expressing the received power as a function of the transmitted one, that is detailed in the next sections. A. Attenuation model We consider the attenuation model from the 3rd Generation Partnership Project (3GPP) technical report [14, Section 6.6], in which the loss of power in dB due to the propagation, assuming that the UE is outdoor, is: L:= LF S +LS+Lg+Lit,(1) where LF S is the free space path loss, LSis the shadowing, Lgis the attenuation due to atmospheric gasses, and Lit is the scintillation loss that is due to either ionospheric or tropospheric scintillations. Notice that (1) is obtained by combining [14, Eq. (6.6-1) and (6.6-4)] and by removing the terms that are related to indoor UE. The expression for the free space path loss Lbis provided in [14, Eq. (6.6-2)] and can be written in dB as: Lb:= 32.45 + 20 log10(fc) + 20 log10(d),(2) where fcis the carrier frequency in GHz, and dis the distance between the UE and the satellite in meters. The shadowing LSis modeled as a Gaussian random variable with zero mean and variance σ2 S(fc, α)that depends on the carrier frequency fc, the elevation α, and the type of scenario, i.e. dense urban, urban or rural. The values of σ2 S(fc, α)for the S-band and the Ka-band and for these three types of scenarios are provided in [14, Tables 6.6.2-1 to 6.6.23]. We use in Section IV the value of the Ka-band in suburban scenario, i.e., σ2 S(fc, α) = 4 regardless of the elevations. The loss due to atmospheric gasses Lgcan be computed using [14, Eq. (6.6-8)] as: Lg=Az(fc) sin(α)(3) where Az(fc)is the attenuation when the satellite is at Nadir, that is provided in [15, Fig. 6]. The values used in this paper are provided in Table III. The scintillation loss Lit depends on fcas discussed in [14, Section 6.6.6]. Since we assume in this paper that fc≥ 6GHz, [14, Section 6.6.6-2] states that Lit is given by the tropospheric scintillation. Table III provides the values used in the simulations (values are from [14, Table 6.6.6.2.1-1]). B. Received power The received power at the UE or at satellite side can be written in dB as: PRx =PT x +GU(θS U) + GS(θU S)− L (4) where PT x is the transmit power, GU(θS U)is the antenna gain of the UE that depends on θS U, the angle between the direction in which the antenna of the UE is pointing at and the line between the UE and the satellite, GS(θU S)is the satellite antenna gain that depends on θU S, the angle between the direction in which the satellite is pointing (i.e. the center of the beam) and the line between the satellite and the UE. The antenna gain in dB of both the satellite and the UE can be written as: Gx(θ) = Gmax,y x+ 20 log10 fx(θ),(5) where x∈ {U, S},y∈ {Tx, Rx}indicates whether the node is transmitting or receiving, Gmax,y xis the maximum antenna gain (that might be different when transmitting or receiving), and fx(θ)depends on the antenna diagram, with typically fx(0) = 1. The maximum antenna gain can be expressed in dB as: Gmax,y x= 10 log10 π2D λ2,(6) where is the antenna efficiency, Dis the antenna diameter, and λis the wavelength. The considered antenna diagram for the UE and the satellite are modeled as follows in the linear domain: fU(u) = 2J1(u) u(7) fS(u) = 2 32J1(u) u+4J2(u) u2(8) where u:= πD/λ sin(θ), and Ji(x)is the Bessel function of the first type and order i. One can check from (6), (7) and (8) that a higher value of D(i.e. a larger antenna) leads to a more directive antenna, and to a higher value for the maximum antenna gain. We assume in this paper that i) the UE is located at the center of the beam and thus θU S= 0 for both the useful signal and the interfering one, and ii) the UE is pointing towards the satellite he wishes to communicate with, meaning that θS U= 0for the useful signal, whereas for the interfering one, θS U corresponds to the angle between the line UE-GEO satellite and the line UE-LEO satellite, which is equal to θS U=π/2−α. III. COEXISTENCE SIMULATION METHODOLOGY The figures of merit used in this study are the 3GPP coexistence requirements from [5, Section 6.2.8] that are related to the throughput degradation induced by the coexistence as compared with a baseline without coexistence. This section first provides the expressions of both signal-to-noise ratio (SNR) and signal-to-interference-plus-noise ratio (SINR) that are required for the throughput calculation, and then details the coexistence requirements. A. SNR and SINR computation The SNR is computed as: SN:= P` Rx 100.1NfN0 (9) where P` Rx is the linear useful received power that is computed using (4) and converting the result into the linear domain, Nf is the noise figure in dB, and N0is the noise power in the considered bandwidth which is computed as: N0= 100.1(Nd 0−30)B, (10) where Nd 0=−174 dBm/Hz is the noise level in the power spectral density, and Bis the bandwidth computed as: B= 12RB∆µ(11) where RBis the number of resource block and ∆µis the subcarrier spacing. The values used for these parameters are provided in Table I. The SINR is computed as: SI:= P` Rx 100.1NfN0+cP` I (12) where P` Iis the linear interfering received power that is computed using (4) and converting the result into the linear domain, and c:= 1CC +F1AC , with 1CC (resp. 1AC ) is the indicator function whose value is 1 in case of co channel interference (resp. adjacent channel interference) and 0 otherwise, and Fis the so-called ACIR. The ACIR is defined as the ratio between the power transmitted by the interferer on its channel and the power received by the interfered system on its channel, due to both transmitter adjacent channel leakage ratio (ACLR) and receiver adjacent channel selectivity (ACS) imperfections. The equation linking the ACIR F, the transmitter ACLR ACand receiver ACS ALin the linear domain is [13, Section 5.2.6]: F:= 1 1 AC+1 AL .(13) We see from (12) that the greater the ACIR, the lower the contribution of the interference, and the closer the SINR from the SNR. As a consequence, a conventional procedure is to find an ACIR value enabling acceptable performance degradation (more details regarding the acceptable degradations are provided in the next section), which then enables to define acceptable range for the ACLR and ACS of the transmitter and receiver, respectively. B. Performance metrics The requirements provided in [5, Section 6.2.8] involve comparing the throughput with and without coexistence. The throughput is computed according to the attenuated and truncated Shannon capacity [13, Section 5.2.7] expressed as: T(x) =      0,if x < Smin, λS(x),if x∈[Smin,Smax], λS(xmax),if x > Smax (14) where S(x) := log2(1+x)is the Shannon capacity, λ∈(0,1) is an attenuation factor representing implementation losses, and Smin and Smax are the minimum and maximum SINR of the codeset. The values of λ,Smin and Smax used in this study are provided in Table II. The throughput loss due to the coexistence is defined as: TD:= 100 1−T(SI) T(SN).(15) Finally, [5, Section 6.2.8] states the following two requirements: 1) The average throughput loss should be less than 5%. 2) The 5th percentile throughput loss value should be less than 5%, meaning that the throughput loss should not exceed 5% for the the worst 5% user throughput. Remark: the second above requirement is more strict than the first one and thus fulfilling 2) implies that the 1) is fulfilled as well. However, this second requirement might lead to very high ACIR value that cannot be achieved in practice and thus it is also of interest to consider the first one. IV. NUMERICAL RESULTS A. Simulation methodology and setup 1) Simulation methodology: we vary the LEO satellite elevation αfrom 75 to 90° by step 0.1°, and perform 5×104 Monte Carlo simulations for each value of α. The random component in the link budget (1) for a given value of αis the shadowing LS. We collect for each value of αthe SNR and SINR of the DL and the UL of both orbits, and we compute the throughput loss (15) when subject to interference. Once these metrics are collected, we plot: (i) the average and 5th percentile throughput loss vs. αin case of co-channel interference, and (ii) the maximum value of αat which the requirements are fulfilled as a function of ACIR value in case of adjacent channel interference. The curves (i) enable us to define a maximum elevation αCC max at which LEO communications can occur on the same channel as GEO communications and still respect the 3GPP requirements. The graphs (ii) enable choosing jointly the ACIR, that might be driven by hardware capacity of the satellite or the UE, and, for a given ACIR, find a value αAC max at which LEO transmissions can occur on an adjacent channel with respect to GEO transmissions. As discussed in Section II, αCC max and αAC max are linked to the exclusion angle. 2) Communications parameters: Table I provides the considered carrier frequency for the DL and the UL as well as the values considered to compute the bandwidth in (11). TABLE I: Carrier frequency and bandwidth in DL and UL. Parameters DL UL fc(GHz) 37 47 ∆µ(kHz) 120 120 RB(#) 132 13 B(MHz) 190.08 18.72 The parameters for the truncated Shannon capacity (14) are provided in Table II, that corresponds to [13, Table 5.2.7-1]. TABLE II: Truncated Shannon capacity parameters [13, Section 5.2.7]. Parameters DL UL λ0.6 0.4 Smin (dB) -10 -10 Smax (dB) 30 22 3) Attenuation parameters: Table III provides the values of Azinvolved in the computation of Lgin (3), and Lit. TABLE III: Azand Lit values. Parameters DL UL Az(dB) 0.3 0.8 Lit (dB) 0.12 0.12 4) UE and satellite parameters: The UE and satellite parameters are provided in Table IV and Table V, respectively. We simulate two values for the UE antenna diameter, D1 and D2, yielding different antenna directivity and maximum antenna gain. Remark: Table V provides both the satellite effective isotropic radiated power (EIRP) density and the corresponding transmit power value PT x that are linked through the following relation: PT x =EIRP + 10 log10(B)−Gmax,T x S.(16) B. Co-channel interference scenario Fig. 2 and Fig. 3 depict the throughput loss for the DL of the GEO and LEO satellites, respectively, in case of co-channel interference, as a function of α. First, one can observe that the TABLE IV: User Equipment parameters. Parameters Values D1(cm) 15 D2(cm) 40 Gmax,T x Ufor D1(dBi) 35.2 Gmax,Rx Ufor D1(dBi) 32.9 Gmax,T x Ufor D2(dBi) 43.7 Gmax,Rx Ufor D2(dBi) 41.4 PT x (dBW) 1.5 Nf(dB) 2 TABLE V: Satellite parameters. Parameters GEO LEO Altitude (km) 35786 600 Din DL (m) 2.7 0.27 Din UL (m) 2.1 0.21 Gmax,T x Sand Gmax,Rx S(dBi) 58.5 38.5 EIRP density (dBW/MHz) 45 15 PT x (dBW) 9.29 -0.71 Nf(dB) 4 4 throughput loss is not a monotonic function of α, which is due to the fluctuating behavior of the Bessel antenna diagram of the UE provided in (7). Second, we can see that larger antenna yields lower throughput loss, except for elevation close to the Nadir. This can be explained because larger antenna are more directive and thus the throughput loss is generaly lower when the angle between the two satellites is large enough. On the other hand, when the elevation of the two satellites is very close, the antenna gain is close to its maximum value, which is higher for larger antenna, yielding in the considered setup a higher throughput loss. We can further observe that the curves cross the 5% line several times, meaning that several value for αCC max could be chosen. We propose in this paper to define αCC max as the angle at which the throughput loss exceeds 5% for the first time. Table VI provides the value of αCC max for the two orbits (GEO and LEO) and two 3GPP requirements (average and 5th percentile throughput loss of maximum 5%) for both DL and UL (the curves related to the UL are omitted to save space), and for the two antenna diameters. One can observe that GEO suffers more than LEO from the interference in DL, whereas the converse also holds for UL. This last observation can be explained because the UE transmits with the same power regardless of the orbits of its associated satellite, and thus its transmit power is set sufficiently high to reach the GEO satellite. As a consequence, the level of interference experienced by the UL of LEO is high. Finally, one can observe that larger antennas enable to chose much higher value for αCC max, especially for LEO UL. The values provided in this table could be used as informative for regulatory purposes. C. Adjacent channel interference scenario 1) DownLink (DL): Fig. 4 and Fig. 5 represent the maximum value of αat which the constraints are fulfilled as a function of the ACIR for GEO DL and LEO DL. Assuming for example that the satellite hardware imposes a maximum Fig. 2: DL throughput loss of the GEO satellite in case of co-channel interference vs. αelevation angle value. Fig. 3: DL throughput loss of the LEO satellite in case of co-channel interference vs. αelevation angle value. TABLE VI: Values of αCC max for the DL and UL of the GEO and the LEO satellites (in degrees [°]) for the two considered antenna apertures. DL UL Average GEO VSAT D183.9° 87.4° VSAT D286.5° 88.2° LEO VSAT D186.9° 82.8° VSAT D287.8° 87.2° 5th percentile GEO VSAT D180.8° 85.2° VSAT D284.3° 87.2° LEO VSAT D184.1° 78° VSAT D286.6° 84.7° value of ACIR of 5 dB for the GEO, then one can observe that the maximum value of αis about 84° for D1, and 86.9° for D2. These curves could be of interest with regards to the satellite filtering capability as currently defined for instance in 3GPP TS 38.108. To explain the stepwise behavior of the previous curves, we plot as an example in Fig. 6 the required ACIR vs. αfor GEO DL with VSAT antenna diameter D1. One can infer that the maximum elevation increases gradually from 80.8 to 81.6° for ACIR between 0 and 3.1 dB for the 5th percentile metric with D1. Then, the maximum elevation jumps to 83.7° as soon as the ACIR is greater than 3.2 dB, explaining the behavior observed in Fig. 4 and Fig. 5. Fig. 4: LEO maximum elevation value at which 3GPP requirements are fulfilled as a function of the ACIR for GEO DL. Fig. 5: LEO maximum elevation value at which 3GPP requirements are fulfilled as a function of the ACIR for LEO DL. 2) UpLink (UL): For the sake of completeness, Fig. 7 and Fig. 8 represent the maximum value of αat which the constraints are fulfilled as a function of the ACIR for GEO DL and LEO DL, which could be of interest for future 6G regulatory purpose. Fig. 6: ACIR value ensuring that the 3GPP requirements are fulfilled for GEO DL vs. αelevation angle value. Fig. 7: LEO maximum elevation value at which 3GPP requirements are fulfilled as a function of the ACIR for GEO UL. Fig. 8: LEO maximum elevation value at which 3GPP requirements are fulfilled as a function of the ACIR for LEO UL. V. CONCLUSION This paper presented a coexistence study between a GEO satellite located at Nadir and a LEO satellite with arbitrary elevation, both operating in the Q/V-band. The coexistence study was performed in the context of co-channel and adjacent channel interference, and 3GPP requirements were used as a figure of merit. The provided results could be used for future 6G systems to ensure coexistence between GEO and LEO satellites in the Q/V-band in order to maximize the spectrum use in a context of multi-layer satellite deployment. ACKNOWLEDGEMENTS This work has been funded by the 6G-NTN project, which received funding from the SNS JU under the European Union’s Horizon Europe research and innovation program under Grant Agreement No 101096479. The views expressed are those of the authors and do not necessarily represent the project. The Commission is not liable for any use that may be made of any of the information contained therein. REFERENCES [1] 6G-NTN project, “6G-NTN website.” [Online]. Available: https://6gntn.eu/ [2] ——, “Spectrum regulation analysis for 6G NTN scenarios D4.7 V1.0.” [Online]. Available: https://6g-ntn.eu/public-deliverables/ [3] A. Hills, J. M. Peha, and J. Munk, “Feasibility of Using Beam Steering to Mitigate Ku-Band LEO-to-GEO Interference,” IEEE Access, vol. 10, pp. 74 023–74 032, 2022. [4] C. Wang, D. Bian, S. Shi, J. Xu, and G. Zhang, “A Novel Cognitive Satellite Network With GEO and LEO Broadband Systems in the Downlink Case,” IEEE Access, vol. 6, pp. 25 987–26 000, 2018. [5] 3GPP, “TR 38.863 Technical Specification Group Radio Access Network; Solutions for NR to support non-terrestrial networks (NTN): Nonterrestrial networks (NTN) related RF and co-existence aspects (Release 18),” September 2024. [6] S. K. Sharma, S. Chatzinotas, and B. Ottersten, “In-line interference mitigation techniques for spectral coexistence of GEO and NGEO satellites,” International Journal of Satellite Communications and Networking, vol. 34, no. 1, pp. 11–39, 2016. [7] S. Tonkin and J. P. de Vries, “NewSpace spectrum sharing: Assessing interference risk and mitigations for new satellite constellations,” in Proc. TPRC, Sept 2018. [8] J. M. P. Fortes and R. Sampaio-Neto, “Impact of avoidance angle mitigation techniques on the interference produced by non-GSO systems in a multiple non-GSO interference environment,” International journal of satellite communications and networking, vol. 21, no. 6, pp. 575–593, 2003. [9] C. Braun, A. M. Voicu, L. Simi´ c, and P. M¨ ah¨ onen, “Should We Worry About Interference in Emerging Dense NGSO Satellite Constellations?” in 2019 IEEE DySPAN, 2019. [10] F. ¨ Ozt¨ urk and A. Kara, “Exclusion zone minimization and optimal operational mode selection for co-existent geostationary and nongeostationary satellites,” International Journal of Satellite Communications and Networking, vol. 40, no. 3, pp. 191–203, 2022. [11] A. Susanto and Iskandar, “Interference Analysis between LEO and GSO Satellites at Ku Band Frequency: Case Study on Starlink and Telkom3S,” in 2022 16th International Conference on Telecommunication Systems, Services, and Applications (TSSA), 2022, pp. 1–4. [12] H. Wang, C. Wang, J. Yuan, Y. Zhao, R. Ding, and W. Wang, “Coexistence Downlink Interference Analysis Between LEO System and GEO System in Ka Band,” in 2018 IEEE/CIC ICCC, 2018, pp. 465–469. [13] 3GPP, “TR 38.803 Technical Specification Group Radio Access Network; Study on new radio access technology: Radio Frequency (RF) and co-existence aspects (Release 14),” June 2024. [14] ——, “TR 38.811 Study on New Radio (NR) to support non-terrestrial networks (Release 15),” September 2020. [15] ITU, “Recommendation ITU-R P.676-13,” September 2013.