IRS-Aided Massive MIMO ISAC System

System model

Problem context & motivation

• 6G will require joint communication and sensing capabilities to support applications such as autonomous driving, industrial automation, and environmental monitoring.

• Intelligent Reflecting Surfaces (IRSs) and massive MIMO are promising enablers for ISAC, but most IRS-aided ISAC designs rely on computationally complex non-linear precoding methods.

• There is a lack of fundamental performance analysis for IRS-aided massive MIMO ISAC systems with computationally efficient linear precoders.

Main contribution

• Proposes and analyzes an IRS-aided massive MIMO ISAC system using maximal ratio transmission (MRT)-based linear precoders for both communication and sensing beams.

• Incorporates spatially correlated Ricean fading and imperfect CSI obtained via LMMSE uplink channel estimation.

• Derives closed-form achievable user rates for communication and models sensing via a 2D MUSIC-based target localization algorithm.

• Optimizes IRS phase-shifts using statistical CSI to maximize the weakest user's average composite channel gain while meeting a minimum sensing channel power constraint.

System model

• BS with large uniform planar arrays (UPAs) for transmit and receive.

• IRS with UPA of passive elements aiding both communication and sensing.

• Single target for sensing, \(K\) single-antenna users for communication.

• Composite channels include direct BS-user links and cascaded BS-IRS-user links.

IRS phase-shift optimization

• Ojective funtion:

  • Maximize the minimum average power gain among communication users.

• Constraints:

  • Maintain an average sensing channel power threshold.

  • Strict modulus constraint

• Reformulates the IRS phase-shift design into a convex problem using first-order approximations and solves iteratively via CVX/Gurobi.

• Ensures constant-modulus IRS reflection constraints after each iteration.

Key numerical results

• Optimized IRS phase-shifts yield sum-rate gain over random IRS phase-shifts

• Increasing IRS elements provides additional sum-rate gains .

• Higher sensing channel power constraints reduce achievable sum rate, demonstrating the sensing-communication trade-off.

• MUSIC algorithm accurately recovers target location parameters in all tested scenarios, confirming the sensing capability.

Conclusion

• IRS-aided massive MIMO ISAC with simple linear MRT-based precoding can effectively balance sensing and communication trade-offs.

• Statistical CSI-based IRS phase-shift optimization offers substantial performance improvements with random counterparts.
  
  

References

1. R. Kulathunga, J. K. Dassanayake, and G. Amarasuriya, “IRS-Aided Massive MIMO ISAC Systems,” in Proc. IEEE Int. Conf. Commun. (ICC), Montreal, Canada, Jun. 2025.