Joint Detection, Channel Estimation and Interference Nulling for Terrestrial-Satellite Downlink Co-Existence in the Upper Mid-Band

TL;DR

The paper tackles downlink interference from terrestrial base stations to NTN receivers in the FR3 upper mid-band and proposes a beacon-based four-stage framework to detect victim NTN-UEs, estimate their channels, and apply interference nulling. It combines a simplified single-victim analysis with site-specific ray-tracing to study the tradeoffs between path loss, channel estimation quality, and nulling performance, and demonstrates substantial INR reductions with manageable TN-UE SNR loss. Key findings show that regularization and channel knowledge critically shape nulling effectiveness, and that large-scale (64+ element) MIMO at the TN-BS is essential to support high NTN densities. The work provides practical design guidance for TN-NTN coexistence in FR3, validated by realistic simulations, and points to calibration refinements and massive-MIMO scaling as important avenues for future networks.

Abstract

The upper mid-band FR3 spectrum (7-24 GHz) has garnered significant interest for future cellular services. However, utilizing a large portion of this band requires careful interference coordination with incumbent satellite systems. This paper investigates interference from high-power terrestrial base stations (TN-BSs) to satellite downlink receivers. A central challenge is that the victim receivers, i.e., ground-based non-terrestrial user equipment (NTN-UEs) such as satellite customer premises equipment, must first be detected and their channels estimated before the TN-BS can effectively place nulls in their directions. We explore a potential solution where NTN-UEs periodically transmit preambles or beacon signals that TN-BSs can use for detection and channel estimation. The performance of this nulling approach is analyzed in a simplified scenario with a single victim, revealing the interplay between path loss and estimation quality in determining nulling performance. To further validate the method, we conduct a detailed multi-user site-specific ray-tracing (RT) simulation in a rural environment. The results show that the proposed nulling approach is effective under realistic parameters, even with high densities of victim units, although TN-BS may require a substantial number of antennas.

Figures (9)

• Figure 1: Proposed interference mitigation framework for non-terrestrial and terrestrial co-existence networks: (1) NTN-UE transmits beacon signal, (2) TN-BS detects potential interference victim, (3) TN-BS estimates channel between TN-BS and NTN-UE, and (4) TN-BS implements spatial filtering with simultaneous null steering toward NTN-UE and optimized beamforming to TN-UE, maximizing spectrum sharing efficiency.
• Figure 2: INR after interference nulling versus uplink detection receiver per antenna SNR $\gamma^{\sf U}_i$ for different numbers of transmit antennas $N_{\mathrm{tx}}$.
• Figure 3: Simulation environment: 6km $\times$ 6km rural area near Boulder, Colorado.
• Figure 4: Ray tracing simulation in Sionna-RT showing line-of-sight (LOS), reflection, and penetration paths.
• Figure 5: Simulated INR radio maps for the rural scenario. (a) shows the result before nulling, and (b) shows the result after nulling.