Gayan Amarasuriya
Home
Biography
Awards
Experience
Research
Research Interests
Research Group
Publications
News
Projects
ISAC Projects
Teaching
Teaching at SIU
Service
Service at SIU
Links
Google Scholar

Near-Field Performance of ELAA-Based ISAC

alt text 

Context & motivation

  • 6G's newly allocated mid-band (FR3: 7-15 GHz) offers better propagation propagation

  • With smaller bandwidths than mmWaves, 6G base stations are needed to be equipped with extremely large aperture arrays (ELAAs) to boost spatial multiplexing gains at these frequencies.

  • Then, the large apertures push the near-field boundary to hundreds of meters, making spherical-wave channel models necessary.

  • Existing near-field ISAC works often ignore effects such as near-field scatterers, spatial correlation, partial visibility, non-wide sense stationarity (non-WSS), and imperfect CSI.

Main contributions

  • Develops a comprehensive near-field performance analysis for ELAA-based ISAC that includes:

    • Spatially correlated Rician fading due to near-field scatterers.

    • Generalized spatial correlation model for non-uniform antenna spacing.

    • Partial visibility regions to model non-WSS.

    • Imperfect CSI estimation.

    • Clutter and extended targets in sensing.

  • Proposes a computationally-efficient conjugate precoding-based superimposed ISAC waveform.

  • Formulates and solves a transmit power allocation problem to maximize the minimum user rate while satisfying a sensing power threshold.

System setup

  • ELAA base station with \(M_T\) transmit and \(M_R\) receive antennas, serving \(K\) single-antenna users and sensing one extended target.

  • Communication channels modeled with near-field spatially correlated Rician fading.

  • Sensing model includes extended targets and multiple clutter sources.

  • CSI acquired via uplink pilots; precoders built from estimated CSI.

Performance metrics

  • Communication: Achievable user rates derived under imperfect CSI using statistical downlink CSI.

  • Sensing: Optimal Neyman-Pearson (NP) detector derived, with performance measured via ROC (probability of detection vs false alarm rate).

Optimization

  • Jointly optimizes user and sensing power allocation.

  • Objective: maximize weakest-user SINR while meeting a minimum sensing power constraint.

  • Solved as a convex problem via geometric programming.

Key numerical findings

  • Partial visibility and spatial correlation significantly degrade user rates.

  • Near-/far-field channel model mismatch in precoder design can cause notable performance loss.

  • Increasing sensing threshold improves detection probability but reduces achievable communication rate, showing a clear sensing-communication trade-off.

  • Larger receive arrays improve sensing performance even with lower sensing power.

Conclusion

  • Accurate near-field modeling is essential for ELAA-based ISAC in FR3 bands.

  • Conjugate precoding with optimized power allocation can balance communication and sensing.

  • ELAAs can leverage spatial multiplexing gains to improve the fundamental ISAC trade-off when near-field effects are properly considered.

References

1. J. K. Dassanayake and G. Amarasuriya, “Near-Field Performance of ELAA-Based ISAC,” in Proc. IEEE Int. Conf. Commun. (ICC), Montreal, Canada, Jun. 2025.