ARIS-RSMA Enhanced ISAC System: Joint Rate Splitting and Beamforming Design

1. Problem Statement

The paper addresses a critical bottleneck in Integrated Sensing and Communication (ISAC) systems operating in environments with obstructed Line-of-Sight (LoS) links. Specifically, it tackles the fairness issue in multi-target sensing.

• Context: Traditional ISAC systems often rely on Space-Division Multiple Access (SDMA), which uses interference nulling. In obstructed environments, this leads to sub-optimal spatial resource utilization and an inability to guarantee balanced sensing performance (Echo SINR) across multiple targets.
• Challenges:
  • Coverage Deficit: Physical blockages prevent direct LoS between the Base Station (BS) and targets.
  • Resource Competition: Balancing Quality of Service (QoS) for multiple users against the need to maximize sensing performance for multiple targets.
  • System Complexity: The integration of Active Reconfigurable Intelligent Surfaces (ARIS) (which amplify signals but introduce noise) and Rate-Splitting Multiple Access (RSMA) creates a highly coupled, non-convex optimization problem involving quartic terms in the objective function and constraints.

2. Methodology

The authors propose a novel framework combining ARIS and RSMA to overcome these limitations.

System Model

• Architecture: A monostatic ISAC BS with $M$ antennas communicates with $U$ single-antenna users and senses $Q$ blocked targets via an $L$-element ARIS.
• Signal Strategy:
  • RSMA: The transmit signal consists of a common stream ($s_c$) and private streams ($s_{p,u}$). Users decode the common stream first (treating private streams as interference) and then use Successive Interference Cancellation (SIC) to decode their private streams.
  • ARIS: Unlike passive RIS, ARIS elements have an amplification factor ($a_l$) and phase shift ($\theta_l$), allowing them to boost weak incident signals but also amplifying dynamic noise.
• Objective: Maximize the minimum Echo Signal-to-Interference-plus-Noise Ratio (SINR) among all targets, subject to user QoS constraints and power budgets for both the BS and the ARIS.

Algorithm Design

The authors formulate the joint optimization of BS transmit/receive beamformers ($F, w_q$), ARIS reflection coefficients ($\Phi$), and RSMA rate allocation ($c$) as a non-convex problem ($P_0$). They propose a Block Coordinate Descent (BCD) based iterative algorithm to solve it:

1. Receive Beamforming ($w_q$): Optimized using Generalized Rayleigh Quotient maximization (eigenvector decomposition).
2. Transmit Beamforming & Rate Splitting ($F, c$):
   • The fractional objective is handled using a Majorization-Minimization (MM) approach to approximate the SINR.
   • Sequential Rank-One Constraint Relaxation (SROCR) is employed to handle the non-convex rank-one constraints of the beamforming matrices, progressively tightening the Semidefinite Relaxation (SDR) solution.
   • First-order Taylor approximations are used to linearize the rate constraints.
3. ARIS Reflection Beamforming ($\Phi$):
   • The quartic terms arising from the round-trip BS-ARIS-target-ARIS-BS path are transformed into quadratic forms using vectorization ($\tilde{\phi} = \text{vec}(\phi\phi^H)$).
   • MM is applied to approximate higher-order terms, converting the problem into a Quadratically Constrained Quadratic Programming (QCQP) form, which is then solved via SROCR.

3. Key Contributions

1. Novel Framework: First integration of ARIS and RSMA specifically for multi-target ISAC under obstructed LoS conditions to address sensing fairness.
2. Fairness Mechanism: Demonstrates that RSMA's common/private stream structure is superior to SDMA and NOMA for balancing spatial resources between communication QoS and worst-case target sensing.
3. Algorithmic Innovation: Development of a low-complexity, convergent BCD algorithm utilizing SROCR and MM to solve the intricate non-convex problem involving coupled amplification and rate-splitting constraints.
4. Noise-Aware Design: Explicitly models the impact of ARIS-amplified noise on both communication reliability and echo quality, optimizing the trade-off.

4. Simulation Results

The proposed scheme (ARIS-RSMA) was compared against benchmarks: ARIS-NOMA, ARIS-SDMA, Passive RIS (PRIS)-RSMA, and Sensing-Only schemes.

• Performance Gain: ARIS-RSMA significantly outperforms all baselines. At a user rate threshold ($R_{th}$) of 5 bits/Hz, it achieves a 5.54 dB max-min echo SINR, compared to 3.02 dB for ARIS-NOMA and 1.91 dB for ARIS-SDMA.
• Robustness: The scheme approaches the "sensing-only" upper bound, indicating highly efficient resource utilization.
• Parameter Sensitivity:
  • Amplification ($a_{max}$): Increasing ARIS amplification improves SINR but with diminishing returns due to amplified noise and power constraints. A moderate amplification factor (e.g., 5) is often sufficient.
  • Deployment: Optimal ARIS placement is near the blocked sensing region but not necessarily the closest point, to balance the BS-ARIS-target round-trip gain with user link quality.
• Rate Allocation: As user rate requirements increase, the system dynamically shifts power from the common stream to private streams, maintaining QoS while sacrificing some sensing performance, yet still outperforming other schemes.

5. Significance

This work provides a critical solution for 6G ISAC systems operating in complex, non-LoS urban environments.

• Theoretical Impact: It proves that treating interference as a decodable stream (RSMA) combined with active signal amplification (ARIS) can resolve the "fairness bottleneck" in multi-target sensing, a problem where traditional nulling-based methods (SDMA) fail.
• Practical Application: The proposed algorithm offers a viable path for deploying ISAC in scenarios where targets are occluded by buildings or terrain, ensuring that no single target suffers from poor detection probability due to resource competition with communication users.
• Future Direction: It establishes a foundation for further research into active metasurfaces and advanced multiple access schemes for joint sensing and communication.