1 Introduction 2 FS Receiver Modelling and Protection

Technical Appendix

Low Power Indoor (LPI) and Very Low Power (VLP) operation of Unlicensed National Information Infrastructure (U-NII) Devices and coexistence with Fixed Links at 6 GHz

Authors: Fabiano Chaves, Lauri Sormunen, Antti Piipponen, Prakash Moorut

Nokia Bell Labs & CTO

1 Introduction

This document presents analysis of coexistence between U-NII Devices and Fixed Service (FS) Links for the 6 GHz band addressing two items of the Commission’s FNPRM [1]:

• Very Low Power (VLP) Operation. • Power Spectral Density Increase for Low Power Indoor (LPI) Operation.

For VLP operation, numerical analyses are presented with three interference scenarios, i.e. same building, neighboring building and outdoor interference, where a transmit power is assumed for the U-NII devices as per the FNPRM, and propagation models and the FS receiver antenna are considered.

For LPI operation, numerical analyses are presented for two interference scenarios, i.e. same building and neighboring building interference, where the Power Spectral Density (PSD) assumed for the U-NII devices are as per [1], and propagation models and the FS receiver antenna are considered. These numerical analyses are followed by system simulations where interference caused by U-NII networks to FS receivers is simulated using a dynamic system simulator, which models the U-NII radio access at a detailed level with realistic traffic loads, medium access protocols, transmit output power levels, and transceiver impairments. The deployment models, propagation models and device antenna patterns are described. Both co-channel and adjacent channel interference scenarios are considered, and the study focuses on two different deployment scenarios of neighboring building interference.

2 FS Receiver Modelling and Protection

Rec. ITU-R F.758 [2] provides representative system parameter values to be used in sharing and compatibility studies involving FS in various frequency bands. From Table 7 in [2], the long-term interference power density (in dBW/MHz) for Point-to-Point (PP) FS systems allocated in bands 5925-6425 MHz and 6425-7125 MHz can be considered as −139 + 퐼/푁 (dBW/MHz). Considering the interference protection criterion specified by the Commission, i.e. 퐼/푁 = −6 dB, the long-term interference power density limit is -115 dBm/MHz.

The antenna gain for FS at the 6 GHz band ranges from 33 dBi to 47 dBi, also according to Table 7 in [2]. For the numerical analysis and simulations performed in the next sections, it is assumed a FS antenna gain of 40 dBi. Different antenna pattern models can be considered in the analysis. Figure 1 illustrates the antenna pattern generated with the model in Rec. ITU-R F.699 [3], as well as FCC CFR 47 Part 101.115 Directional Antennas, i.e. the FCC antenna patterns of type A, B1 and B2, always for antenna gain of 40 dBi.

Figure 1: FS antenna patterns (40 dBi antenna gain): Rec. ITU-R F.699 [3] model and FCC types A, B1 and B2.

FCC types A and B2 are the most common microwave link antennas in the 6 GHz band in the US. Considering that commercial antennas usually have performance at least slightly better than the reference patterns, the FS antenna pattern generated with the model in Rec. ITU-R F.699 and showed in Figure 1 is adopted for the numerical analyses presented in the next sections. This antenna model provides better off-axis discrimination than FCC types A and B2 in most angular ranges relevant for the analyses. It should be noted that the adopted FS antenna modeling for the numerical analyses represents an intermediary choice in terms of antenna gain and off-axis discrimination. Modelled and commercial FS antennas with gain lower than 40 dBi, for example, could present worse off-axis discrimination than the one considered in the analysis, due to lower directivity of lower gain antennas.

For system simulations, the FCC antenna type A shown in Figure 1 is adopted as FS antenna pattern. This antenna model provides better off-axis discrimination than FCC type B2 model and type B2 commercial antennas in most angular ranges relevant for the simulations. It should be also noted that the adopted FS antenna modeling for system simulations represents an intermediary choice in terms of antenna gain and off-axis discrimination, with modelled and commercial FS antennas with gain lower than 40 dBi, for example, potentially presenting worse off- axis discrimination than the one considered for simulations, due to lower directivity.

3 Very Low Power (VLP) Operation 3.1 Interference scenario VLP – 1: Same building interference

The interference scenario studied in this section refers to a single VLP U-NII device transmitting indoors at the highest floor of a building where a FS receiver antenna is placed. Figure 2 illustrates the scenario.

Co-channel interference PSD, 퐼푃푆퐷, is estimated as follows:

퐼푃푆퐷(푑) = 퐸퐼푅푃푃푆퐷 − 푃퐿(푓, 푑) − 퐿표푠푠퐵표푑푦 − 퐵퐸퐿 − 퐿표푠푠푃표푙 + 퐺퐹푆−퐴푃(휃, 휙) [dBm/MHz],

2 where:

1 • 퐸퐼푅푃푃푆퐷 is the VLP U-NII device PSD EIRP, assumed -8 dBm/MHz, with isotropic 0 dBi antenna gain . • 푃퐿(푓, 푑) is the path loss for frequency 푓 and distance 푑, calculated as free space loss at 6.5 GHz for short distances.

• 퐿표푠푠퐵표푑푦 is the body loss. • 퐵퐸퐿 is the building entry loss.

• 퐿표푠푠푃표푙 is the polarization loss, assumed 3 dB. • 퐺퐹푆−퐴푃(휃, 휙) is the FS receiver antenna gain in the direction of the VLP U-NII device, which according to the model in Rec. ITU-R F.699 [3] gives -6.15 dBi for the considered interference geometry.

Figure 2: Interference scenario VLP – 1: Single indoor VLP U-NII device transmitting at the highest floor of a building where a FS receiver antenna is placed.

For distances between the VLP U-NII device and the FS receiver antenna ranging from 9.5 m to 21.5 m, free space loss at 6.5 GHz gives 68.26 dB and 75.35 dB loss, respectively. Then, co-channel interference PSD for the considered distances are given by:

퐼푃푆퐷(푑 = 9.5 푚) = −85.4 − (퐿표푠푠퐵표푑푦 + 퐵퐸퐿) [dBm/MHz]

퐼푃푆퐷(푑 = 21.5 푚) = −92.5 − (퐿표푠푠퐵표푑푦 + 퐵퐸퐿) [dBm/MHz]

In order to have co-channel interference PSD below the threshold for protection of FS receiver, i.e. -115 dBm/MHz, one must rely on the following conditions of body loss and building entry loss:

퐼푃푆퐷(푑 = 9.5 푚) ≤ −115 ⇒ (퐿표푠푠퐵표푑푦 + 퐵퐸퐿) ≥ 29.6 [dB]

퐼푃푆퐷(푑 = 21.5 푚) ≤ −115 ⇒ (퐿표푠푠퐵표푑푦 + 퐵퐸퐿) ≥ 22.5 [dB]

Rec. ITU-R P.2109 [4] provides a statistical method for estimating building entry loss for use in sharing and compatibility studies. Two building classes are considered in [4], with “thermally-efficient” building class presenting higher building entry loss than the “traditional” building class. The Commission has considered appropriate a 70% traditional construction/30% energy efficient construction mix of building types for determining building entry loss, which results in a 50th percentile building entry loss of 20.62 dB [1]. For body loss, a wide range

1 Considering the desirable and foreseen characteristics of VLP devices, including small size and simple operation, the VLP U-NII device is assumed not to employ any power control mechanism, and therefore transmits with fixed power. It is assumed that the VLP U-NII devices transmit with 14 dBm power, i.e. power spectral density of -8 dBm/MHz (for a 160MHz channel).

3 of values may apply according to different assumptions. For reference, a standard value of 4 dB has been indicated for interference analysis involving cellphones or smartphones as representative of all different user cases (i.e. speech position, browsing position, etc.) [5].

Therefore, taking the values of body loss and building entry loss mentioned above as appropriate estimates, there is some potential for co-channel interference to a FS receiver due to transmissions of a single VLP U-NII device operating in the highest floor of the same building that the FS receiver antenna is placed. It is noted again that commercial FS antennas may present off-axis discrimination worse than the one assumed in the analysis.

3.2 Interference scenario VLP – 2: Neighbouring building interference

Analysis similar to previous section is presented now for a different interference scenario: a single VLP U-NII device transmitting at the highest floor of a neighboring building to the one where a FS receiver antenna is placed. Figure 3 illustrates the scenario.

Co-channel interference PSD, 퐼푃푆퐷, is estimated as follows:

퐼푃푆퐷(푑) = 퐸퐼푅푃푃푆퐷 − 푃퐿(푓, 푑) − 퐿표푠푠퐵표푑푦 − 퐵퐸퐿 − 퐿표푠푠푃표푙 + 퐺퐹푆−퐴푃(휃, 휙) [dBm/MHz], where:

• 퐸퐼푅푃푃푆퐷 is the VLP U-NII device PSD EIRP, assumed -8 dBm/MHz, with isotropic 0 dBi antenna gain. • 푃퐿(푓, 푑) is the path loss for frequency 푓 and distance 푑, calculated as free space loss at 6.5 GHz for short distances.

• 퐿표푠푠퐵표푑푦 is the body loss. • 퐵퐸퐿 is the building entry loss.

• 퐿표푠푠푃표푙 is the polarization loss, assumed 3 dB. • 퐺퐹푆−퐴푃(휃, 휙) is the FS receiver antenna gain in the direction of the VLP U-NII device, calculated according to antenna pattern generated with the model in [3].

Figure 3: Interference scenario VLP – 2: Single VLP U-NII device transmitting at the highest floor of a neighboring building to the one where a FS receiver antenna is placed.

4 receiver antenna at 6.5 GHz is 79.39 dB, and the FS receiver antenna gain in the direction of the U-NII AP is -0.64 dBi. For d’ = 40 m, these values become 83.13 dB loss and 4.35 dBi, respectively. Then, co-channel interference PSD for the considered distances between buildings are given by

퐼푃푆퐷(푑′ = 20 푚) = −91.0 − (퐿표푠푠퐵표푑푦 + 퐵퐸퐿) [dBm/MHz]

퐼푃푆퐷(푑′ = 40 푚) = −89.8 − (퐿표푠푠퐵표푑푦 + 퐵퐸퐿) [dBm/MHz]

In order to have co-channel interference PSD below the threshold for protection of FS receiver in the considered scenario, i.e. -115 dBm/MHz, one must rely on the following conditions of body loss and building entry loss:

퐼푃푆퐷(푑′ = 20 푚) ≤ −115 ⇒ (퐿표푠푠퐵표푑푦 + 퐵퐸퐿) ≥ 24.0 [dB]

퐼푃푆퐷(푑′ = 40 푚) ≤ −115 ⇒ (퐿표푠푠퐵표푑푦 + 퐵퐸퐿) ≥ 25.2 [dB]

As previously mentioned, the Commission has considered appropriate for building entry loss a value of 20.62 dB [FCC-DOC363490A1], and body loss of 4 dB has been indicated for interference analysis involving cellphones or smartphones as representative of all different user cases (i.e. speech position, browsing position, etc.) [5].

Therefore, taking the values of body loss and building entry loss mentioned above as appropriate estimates, there is very low potential for co-channel interference to a FS receiver due to transmissions of a single VLP U-NII device operating in the highest floor of a neighboring building to the one where the FS receiver antenna is placed, if the VLP U-NII device is indoors. However, as VLP devices are foreseen for usage as portable devices or wearables, it is reasonable to consider events where the radio propagation path between VLP U-NII devices and a FS receiver antenna are not subject to building entry loss, e.g. U-NII device in use at an apartment balcony or in the terrace of a neighboring building to the one where the FS receiver antenna is placed. It is noted again that commercial FS antennas may present off-axis discrimination worse than the one assumed in the analysis.

3.3 Interference scenario VLP – 3: Outdoor street level interference

Analysis similar to previous section is presented now for an interference scenario where a single VLP U-NII device is transmitting outdoors, at the street level, and at a short distance from the building where a FS receiver antenna is placed. Figure 4 illustrates the scenario.

Co-channel interference PSD, 퐼푃푆퐷, is estimated as follows:

퐼푃푆퐷(푑) = 퐸퐼푅푃푃푆퐷 − 푃퐿(푓, 푑) − 퐿표푠푠퐵표푑푦 − 퐿표푠푠푃표푙 + 퐺퐹푆−퐴푃(휃, 휙) [dBm/MHz], where:

• 퐿표푠푠퐵표푑푦 is the body loss.

Figure 4: Interference scenario VLP – 3: Single VLP U-NII device transmitting outdoors, at the street level, and at a short distance from the building where a FS receiver antenna is placed.

With no loss of generality, the distance between the VLP U-NII device and the building is considered the same as the height of the building, 퐿, and two values are assumed to 퐿: 30 m and 50 m. Considering the geometry of the interference scenario, for L = 30 m and 50 m, the FS receiver antenna gain in the direction of the U-NII AP is -6.0 dBi and -5.9 dBi, respectively. Then, co-channel interference PSD for the considered distances between buildings are given by

퐼푃푆퐷(퐿 = 30 푚) = −17 − (푃퐿 + 퐿표푠푠퐵표푑푦) [dBm/MHz]

퐼푃푆퐷(퐿 = 50 푚) = −16.9 − (푃퐿 + 퐿표푠푠퐵표푑푦) [dBm/MHz]

퐼푃푆퐷(퐿 = 30 푚) ≤ −115 ⇒ (푃퐿 + 퐿표푠푠퐵표푑푦) ≥ 98.0 [dB]

퐼푃푆퐷(퐿 = 50 푚) ≤ −115 ⇒ (푃퐿 + 퐿표푠푠퐵표푑푦) ≥ 98.1 [dB]

As previously mentioned, body loss of 4 dB has been indicated for interference analysis involving cellphones or smartphones as representative of all different user cases (i.e. speech position, browsing position, etc.) [5]. Considering that free space loss between the VLP U-NII device and the FS receiver antenna at 6.5 GHz in this interference scenario is 84.13 dB and 87.53 dB for L = 30 m and 50 m, there is potential for co-channel interference to a FS receiver due to transmissions of a single VLP U-NII device operating outdoors, at street level and short distances from the building where a FS receiver antenna is placed. It is noted again that commercial FS antennas may present off-axis discrimination worse than the one assumed in the analysis.

4 Power Spectral Density Increase for Low Power Indoor (LPI) Operation 4.1 Numerical analysis 4.1.1 Interference scenario LPI – 1: Same building interference

The interference scenario studied in this section refers to a single indoor U-NII Access Point (AP) operating at the highest floor of a building where a FS receiver antenna is placed. Figure 5 illustrates the scenario.

Co-channel interference PSD, 퐼푃푆퐷, is estimated as follows:

퐼푃푆퐷(푑) = 퐸퐼푅푃푃푆퐷 − 푃퐿(푓, 푑) − 퐵퐸퐿 − 퐿표푠푠푃표푙 + 퐺퐹푆−퐴푃(휃, 휙) [dBm/MHz], where:

• 퐸퐼푅푃푃푆퐷 is the U-NII AP PSD Equivalent Isotropically Radiated Power (EIRP), assumed 5 dBm/MHz, with isotropic 0 dBi antenna gain. • 푃퐿(푓, 푑) is the path loss for frequency 푓 and distance 푑, calculated as free space loss at 6.5 GHz for short distances. • 퐵퐸퐿 is the building entry loss.

• 퐿표푠푠푃표푙 is the polarization loss, assumed 3 dB. • 퐺퐹푆−퐴푃(휃, 휙) is the FS receiver antenna gain in the direction of the U-NII AP, which according to the model in Rec. ITU-R F.699 [3] gives -6.15 dBi for the considered interference geometry.

Figure 5: Interference scenario LPI – 1: Single indoor U-NII AP operating at the highest floor of a building where a FS receiver antenna is placed.

For distances between the U-NII AP and the FS receiver antenna ranging from 8 m to 20 m, free space loss at 6.5 GHz gives 66.76 dB and 74.72 dB loss, respectively. Then, co-channel interference PSD for the considered distances are given by:

퐼푃푆퐷(푑 = 8 푚) = −70.9 − 퐵퐸퐿 [dBm/MHz]

퐼푃푆퐷(푑 = 20 푚) = −78.9 − 퐵퐸퐿 [dBm/MHz]

In order to have co-channel interference PSD below the threshold for protection of FS receiver, i.e. -115 dBm/MHz, one must rely on the following conditions of building entry loss:

퐼푃푆퐷(푑 = 8 푚) ≤ −115 ⇒ 퐵퐸퐿 ≥ 44.1 [dB]

퐼푃푆퐷(푑 = 20 푚) ≤ −115 ⇒ 퐵퐸퐿 ≥ 36.1 [dB]

As previously mentioned, the Commission has considered appropriate for building entry loss a value of 20.62 dB [1]. Therefore, there is a discrepancy of 15.5 dB to 23.5 dB between the appropriate estimate of building entry loss and the levels that would be required to prevent co-channel interference to a FS receiver due to transmissions of a single U-NII AP operating in the highest floor of the same building that the FS receiver antenna is placed. This interference assessment assumes U-NII AP transmitting with PSD EIRP of 5 dBm/MHz. A PSD increase to 8 dBm/MHz would increase the mentioned discrepancy proportionally. Some variation on the assumptions of this interference scenario (e.g. lower U-NII AP antenna gain in the direction of the FS receiver antenna) could lead to variations on the gap for protection of FS receiver, but considering that commercial FS antennas may present off- axis discrimination worse than the one assumed in the analysis, it becomes evident that U-NII APs are likely to cause co-channel interference to the FS receiver in such a scenario even transmitting with PSD EIRP of 5 dBm/MHz.

4.1.2 Interference scenario LPI – 2: Neighbouring building interference

Analysis similar to previous section is presented now for a different interference scenario: a single indoor U-NII AP operating at the highest floor of a neighboring building to the one where a FS receiver antenna is placed. Figure 6 illustrates the scenario.

Co-channel interference PSD, 퐼푃푆퐷, is estimated as follows:

퐼푃푆퐷(푑) = 퐸퐼푅푃푃푆퐷 − 푃퐿(푓, 푑) − 퐵퐸퐿 − 퐿표푠푠푃표푙 + 퐺퐹푆−퐴푃(휃, 휙) [dBm/MHz], where:

• 퐸퐼푅푃푃푆퐷 is the U-NII AP PSD EIRP, assumed 5 dBm/MHz, with isotropic 0 dBi antenna gain. • 푃퐿(푓, 푑) is the path loss for frequency 푓 and distance 푑, calculated as free space loss at 6.5 GHz for short distances. • 퐵퐸퐿 is the building entry loss.

• 퐿표푠푠푃표푙 is the polarization loss, assumed 3 dB. • 퐺퐹푆−퐴푃(휃, 휙) is the FS receiver antenna gain in the direction of the U-NII AP, calculated according to antenna pattern generated with the model in [3].

Figure 6: Interference scenario LPI – 2: Single indoor U-NII AP operating at the highest floor of a neighboring building to the one where a FS receiver antenna is placed.

For the distance between the neighboring buildings, 푑′, two values are considered: 20 m and 40 m. Considering the geometry of the interference scenario, for d’ = 20 m, free space loss between the U-NII AP and the FS receiver antenna at 6.5 GHz is 79.21 dB, and the FS receiver antenna gain in the direction of the U-NII AP is 0.27 dBi. For d’ = 40 m, these values become 83.05 dB loss and 5.28 dBi, respectively. Then, co-channel interference PSD for the considered distances between buildings are given by:

퐼푃푆퐷(푑′ = 20 푚) = −76.9 − 퐵퐸퐿 [dBm/MHz]

퐼푃푆퐷(푑′ = 40 푚) = −75.8 − 퐵퐸퐿 [dBm/MHz]

퐼푃푆퐷(푑′ = 20 푚) ≤ −115 ⇒ 퐵퐸퐿 ≥ 38.1 [dB]

퐼푃푆퐷(푑′ = 40 푚) ≤ −115 ⇒ 퐵퐸퐿 ≥ 39.2 [dB]

As previously mentioned, the Commission has considered appropriate for building entry loss a value of 20.62 dB [1]. Therefore, there is a discrepancy of 17.5 dB to 18.6 dB between the appropriate estimate of building entry loss and the levels that would be required to prevent co-channel interference to a FS receiver due to transmissions of a single U-NII AP operating in the considered scenario, i.e. in the highest floor of a neighboring building to the one where the FS receiver antenna is placed. This interference assessment assumes U-NII AP transmitting with PSD EIRP of 5 dBm/MHz. A PSD increase to 8 dBm/MHz would increase the mentioned discrepancy proportionally. Some variation on the assumptions of this interference scenario (e.g. lower U-NII AP antenna gain in the direction of the FS receiver antenna) could lead to variations on the gap for protection of FS receiver, but considering that commercial FS antennas may present off-axis discrimination worse than the one assumed in the analysis, it becomes evident that U-NII APs are likely to cause co-channel interference to the FS receiver in such a scenario even transmitting with PSD EIRP of 5 dBm/MHz.

4.2 System simulations 4.2.1 Simulator description

The system simulator models indoor U-NII devices, Access Points (APs) with configurable antenna patterns, and client devices at randomly dropped positions within the simulation area. The client devices are connected to the APs. Propagation models are used to compute the path loss from transmitting nodes to the receiving ones (including nodes which consider interference). Scheduling and power control algorithms are used and at the receiving nodes the simulator calculates the signal to interference and noise ratio (SINR) and uses statistical methods to determine whether data was successfully received or not.

In the context of this study, the interfering system is parameterized according to the proposed U-NII device limits. The interference to the FS receivers is calculated at each time instance, and cumulative distribution functions of the interference probability are presented. In all scenarios, the U-NII devices form a network which is simulated using the dynamic system simulator, and the aggregate interference to the FS receiver is calculated. Both Co- Channel Interference (CCI) and Adjacent Channel Interference (ACI) cases are calculated.

4.2.2 Simulation scenarios and parameters

Two interference scenarios are considered in this study. They are described below, along with the relevant parameters for modeling the FS receiver and U-NII devices.

Scenario 1 – FS receiver antenna oriented towards building with indoor U-NII devices:

The first scenario has an FS receiver antenna placed on the roof of a single-story or multi-story building, and U-NII devices operating at the highest floor of a neighboring building placed within the FS receiver antenna boresight range in azimuth. Different distances between the buildings are considered. Only one floor of U-NII devices is modeled. In a multi-story building, the lower floors will have significantly larger penetration loss to the rooftop FS antenna, compared to the top floor, hence the contribution to interference can be considered negligible. Figure 7 illustrates the interference scenario.

Figure 7: FS receiver antenna oriented towards building with indoor U-NII devices.

Six ceiling-mounted U-NII APs are placed in a 80 m x 120 m area floor with 3 m height. Five U-NII client devices are associated to each AP, at 1.5 m height related to the floor. U-NII devices are simulated with PSD EIRP are as per [1], i.e. 5 dBm/MHz for AP and -1 dBm/MHz for client device. Only one user transmits at a time, with Listen-Before- Talk (LBT) implemented and no transmit power control considered, i.e. transmissions follow the mentioned PSD EIRP. Two traffic models are assumed in the simulations: full buffer, and FTP3 with 0.5 packets of 500 kB per second per user.

Two different antenna models are considered for the U-NII AP: an omni-directional antenna [6], described in Table 1, and a 3GPP directional antenna, described in Table 2. Isotropic antenna with 0 dBi gain is assumed for client devices.

Table 1: Omni-directional antenna [6] considered for U-NII AP.

Elevation angle 휽 (degrees) Antenna gain (dBi) ퟒퟓ < 휽 ≤ ퟗퟎ -4 ퟑퟓ < 휽 ≤ ퟒퟓ 0 ퟎ < 휽 ≤ ퟑퟓ 3 −ퟏퟓ < 휽 ≤ ퟎ -1 −ퟑퟎ < 휽 ≤ −ퟏퟓ -4 −ퟔퟎ < 휽 ≤ −ퟑퟎ -9 −ퟗퟎ < 휽 ≤ −ퟔퟎ -8

Table 2: 3GPP directional antenna considered for U-NII AP.

Parameter Value Antenna gain 6 dBi Horizontal 3dB beam width 60 degrees Horizontal front-to-back 15 dB ratio Vertical 3dB beam width 60 degrees Vertical front-to-back ratio 15

The distance between the buildings is varied from 20 m to 500 m. As mentioned in Section 2, the FCC antenna type A shown in Figure 1 is adopted as FS antenna pattern for simulations. Channel bandwidth is assumed to be 10 MHz for FS receiver and 20 MHz (TDD) for U-NII devices. Combined propagation models for urban macro environment

11 and indoor environment from 3GPP TR 38.901 [7] are used. For building entry loss, the model in [4] was adopted with the configuration considered appropriate by the Commission, i.e. 70% traditional construction / 30% energy efficient construction mix of building. Adjacent channel interference is calculated assuming that the U-NII device Adjacent Channel Leakage Ratio (ACLR) is 35 dB for the AP and 30 dB for the client device2.

Scenario 2 – Indoor U-NII device providing video streaming in the neighborhood of FS receiver antenna:

The second interference scenario for simulations uses similar topology as the first scenario, but only a single U-NII AP is considered to provide video streaming in the highest floor of a building which is in the neighborhood of the building where the FS receiver antenna is placed. Different distances between the buildings are considered, as well as varying azimuth offset from the FS receiver antenna boresight. Figure 8 illustrates the interference scenario.

Figure 8: Indoor U-NII device providing video streaming in the neighborhood of FS receiver antenna.

Both U-NII AP and client are considered at 1.5 m height within the 10 m x 10 m apartment area. The AP is placed in the center of the apartment, while the client device position is randomized for each simulation realization. Transmit power settings for the U-NII devices are the same as in Scenario 1. U-NII device antennas are assumed isotropic in this scenario.

The video streaming traffic model consists in the following assumptions:

• U-NII device channel bandwidth of 80 MHz. • Streaming traffic rate for HD quality: 3GB/hour, i.e. 900 kB/s split into 5 DL packets/s (from Poisson distribution).

2 The ACLR is based on 3GPP specification for unlicensed 5 GHz band operations.

• Small amount of UL traffic, i.e. 1250 B/s (1 packet/s).

This traffic model results in approximately 6% usage of channel capacity.

Distance between the buildings is varied from 50 m to 5 km. The FS receiver antenna is placed on the same level as the apartment, and the azimuth offset between the building and the FS receiver antenna boresight is varied with the following granularity:

• Between 0 and 5 degrees by 0.1-degree steps. • Between 5 and 30 degrees by 0.5-degree steps. • Between 30 and 60 degrees by 1-degree steps. • Between 60 and 90 degrees by 2-degree steps.

As mentioned in Section 2, the FCC antenna type A shown in Figure 1 is adopted as FS antenna pattern for simulations. Channel bandwidth is assumed to be 10 MHz for FS receiver. Combined propagation models for urban macro environment and indoor environment from 3GPP TR 38.901 [7] are used. In order to allow easy extrapolation of results, a fixed 20 dB building entry loss is considered in simulations for this scenario.

4.2.3 Simulation results

Scenario 1 – FS receiver antenna oriented towards building with indoor U-NII devices:

Figure 9 presents the Cumulative Distribution Function (CDF) of the I/N at the FS receiver for Scenario 1, with omni-directional antennas assumed for the U-NII devices. Results for full buffer traffic and FTP3 traffic (described in Section 4.2.2) are shown in Figure 9(a) and Figure 9(b), respectively.

(a) Full buffer traffic. (b) FTP3 traffic.

Figure 9: I/N distribution for Scenario 1, with omni-directional antennas assumed for the U-NII devices.

Figure 10 presents the same CDF of the I/N at the FS receiver for Scenario 1, but with 6 dBi directional antennas assumed for the U-NII devices. Results for full buffer traffic and FTP3 traffic are shown in Figure 10(a) and Figure 10(b), respectively.

(a) Full buffer traffic. (b) FTP3 traffic.

Figure 10: I/N distribution for Scenario 1, with 6 dBi directional antennas assumed for the U-NII devices.

The most important observation from Figures 9 and 10 is the high likelihood of co-channel interference for all considered distances between the buildings, and for both types of antennas assumed for the U-NII AP, i.e. omni- directional and 6 dBi directional. For the most favorable configuration studied, i.e. U-NII AP with directive antenna and FTP3 traffic (Figure 10(b)), approximately 8% of I/N samples cross the -6 dB limit for FS protection at 500 m distance. For distances between the buildings as short as 100 m and 20 m, the I/N is above the protection limit for approximately 17% and 28% of samples. Some probability of adjacent channel interference is also observed.

This study assumes 5 dBm/MHz as PSD EIRP for the U-NII AP. A PSD increase to 8 dBm/MHz would increase the mentioned discrepancy proportionally by roughly right shifting the curves plotted in Figures 9 and 10 by 3 dB.

Scenario 2 – Indoor U-NII device providing video streaming in the neighborhood of FS receiver antenna:

Figure 11 presents the percentage of I/N samples over -6 dB and -10 dB protection thresholds for Scenario 2 as a function of the distance between the buildings and the azimuth offset with respect to the FS receiver antenna boresight. Curves in different colors refer to different distances between the buildings.

It is observed that for buildings at 50 m distance in this interference scenario, a single U-NII AP providing video streaming, e.g. during 1-2 hours, has at least 1% of its transmissions crossing the -6 dB I/N protection threshold for azimuth offsets from the FS receiver antenna boresight as large as approximately 70 degree. For buildings at 250 m of distance, more than 1% of transmissions of such U-NII AP will cross the I/N protection threshold if it is placed within 5 degree offset of the FS receiver antenna boresight.

Figure 12 presents the scatter plot of all non-zero I/N samples observed for Scenario 2 as a function of the distance between the buildings and the azimuth offset with respect to the FS receiver antenna boresight. Scatter points in different colors refer to different distances between the buildings.

Figure 11: Percentage of I/N samples over the -6 dB protection threshold for Scenario 2.

Figure 12: Scatter plot of all non-zero I/N samples observed for Scenario 2.

From the I/N data presented in Figures 11 and 12, a distance of 2.5 km between buildings in this interference scenario seems to be required to provide enough isolation for all angles between the single U-NII AP and the FS receiver. This study assumes 5 dBm/MHz as PSD EIRP for the U-NII AP. A PSD increase to 8 dBm/MHz would increase the interference levels proportionally by up shifting the scatter plots in Figure 12 by 3 dB and presenting some increase on the percentage of I/N samples over the protection threshold.

5 Conclusions

This technical appendix presents analyses and studies on the coexistence between U-NII Devices and Fixed Service (FS) Links for the 6 GHz band regarding two items of the Commission’s FNPRM [1]:

• Very Low Power (VLP) Operation. • Power Spectral Density Increase for Low Power Indoor (LPI) Operation.

The numerical analyses on VLP operation considered three interference scenarios, including indoor and outdoor VLP transmissions. These analyses indicate potential for co-channel interference to a FS receiver due to transmissions of a single VLP U-NII device at power spectral density level of -8 dBm/MHz. Nokia understands that the adoption of a power limit for VLP U-NII device operation on the lower side of the power range considered by the Commission, e.g. 4 dBm for a 160MHz channel (-18 dBm/MHz PSD EIRP) would minimize the potential for co- channel interference to a FS receiver due to such devices.

For LPI operation, numerical analyses are presented for two interference scenarios, with FS receiver antenna and U-NII devices placed at the same building and in neighboring buildings. System simulations with more realistic modeling and assumptions were also used to study interference scenarios involving LPI operation. All interference scenarios analyzed showed high potential for co-channel interference to a FS receiver due to LPI transmissions. U- NII AP PSD EIRP was assumed 5 dBm/MHz. The 3 dB increase of PSD to 8 dBm/MHz considered by the Commission would increase the potential for interference in the studies scenarios. Therefore, Nokia understands that an AFC an AFC would be needed if the Commission decides to adopt this increase in PSD EIRP from 5 dBm/MHz to 8 dBm/MHz.

6 References

[1] Unlicensed Use of the 6 GHz Band, Order and Notice of Proposed Rulemaking, GN docket No. 18-295, et al., FCC 20-51 (rel. Apr. 24, 2020).

[2] Rec. ITU-R F.758-7, “System parameters and considerations in the development of criteria for sharing or compatibility between digital fixed wireless systems in the fixed service and systems in other services and other sources of interference”, Nov. 2019.

[3] Rec. ITU-R F.699-8, “Reference radiation patterns for fixed wireless system antennas for use in coordination studies and interference assessment in the frequency range from 100 MHz to 86 GHz”, Jan. 2018.

[4] Rec. ITU-R P.2109-1, “Prediction of building entry loss”, Aug. 2019.

[5] Rep. ITU-R M.2292-0, “Characteristics of terrestrial IMT-Advanced systems for frequency sharing/interference analyses”, Dec. 2013.

[6] ECC Rep. 244, “Compatibility studies related to RLANs in the 5725-5925 MHz band”, Jan. 2016.

[7] 3GPP TR 38.901, “5G; Study on channel model for frequencies from 0.5 to 100 GHz (3GPP TR 38.901 version 15.0.0 Release 15)”, Jul. 2018.