TL;DR
This paper tackles the challenge of integrating radar sensing and communication (ISAC) in a full-duplex (FD) system, where the base station (BS) simultaneously transmits and receives signals for both functions. The authors propose novel joint optimization algorithms for beamforming and power allocation to maximize spectral efficiency and minimize power consumption, demonstrating significant improvements over existing half-duplex ISAC systems.
What Problem Does This Paper Address?
The increasing demand for advanced wireless networks, like those in the Internet-of-Vehicles (IoV) and smart factories, necessitates efficient communication and reliable sensing capabilities. Integrated sensing and communication (ISAC) systems offer a promising solution by sharing spectrum and hardware. While previous research has explored ISAC, most works have focused on half-duplex (HD) communication or sensing, or have not fully considered the complexities of full-duplex (FD) operation for both functions simultaneously. This paper addresses the challenge of designing an FD ISAC system where the BS performs both radar sensing and communication with multiple uplink and downlink users, all while reusing the same time and frequency resources. The core problem is the tight coupling between radar sensing and communication, and the inherent interference in FD systems, which makes joint optimization of transmit and receive signals, as well as power allocation, a complex task.
Methods
The paper proposes a framework for joint optimization of the BS's transmit beamforming, uplink user transmit power, and receive beamforming for both radar sensing and communication. Key methodological contributions include:
- FD ISAC System Model: A comprehensive model is developed for an FD ISAC system with a BS equipped with uniform linear arrays (ULAs), multiple uplink users, and multiple downlink users, capable of detecting a point radar target.
- Closed-Form Receive Beamformers: The authors derive optimal linear receive beamformers for both radar sensing and uplink communication in closed-form expressions. These are optimized to maximize the respective Signal-to-Interference-plus-Noise Ratios (SINRs).
- Iterative Optimization Algorithms: Two main optimization problems are formulated: power minimization and sum rate maximization. For both, the derived closed-form receivers are substituted into the original problems, leading to equivalent, albeit still nonconvex, optimization problems. These are then solved using an iterative approach based on the Successive Convex Approximation (SCA) technique.
- Special Case Optimization: For the power minimization problem, a special case of uplink-only communication is considered. Here, a more computationally efficient alternating optimization (AO)-based algorithm is proposed, leveraging semidefinite programming (SDP) and second-order cone programming (SOCP) for solving subproblems.
- Rank-One Solution Recovery: A theorem is presented to show that the rank-relaxed solutions obtained from the optimization can be deterministically converted into rank-one solutions without performance loss, which is crucial for practical beamformer implementation.
Datasets
Not discussed in the paper. The evaluation relies on simulated channel models and parameters.
Key Insights and Ideas
- FD ISAC Offers Significant Gains: The paper strongly suggests that embracing full-duplex capabilities for both sensing and communication in ISAC systems leads to substantial improvements in spectral and power efficiency compared to conventional HD approaches.
- Tight Coupling is the Core Challenge: The inherent interdependence between radar echo signals and communication signals at the BS receiver is identified as a primary difficulty, necessitating sophisticated joint optimization techniques.
- Receive Beamforming as a Key Enabler: Deriving closed-form solutions for receive beamformers significantly simplifies the overall optimization problem, allowing for efficient iterative solutions.
- SCA and AO for Nonconvexity: The paper effectively demonstrates the application of SCA for general FD ISAC optimization and a specialized AO approach for a simplified uplink-only scenario, showcasing their power in tackling complex nonconvex problems.
- Trade-offs are Inherent: The results highlight the fundamental trade-off between sensing performance (radar SINR) and communication performance (sum rate), which must be carefully managed through joint optimization.
Key Results
- Power Minimization: The proposed algorithms achieve significant reductions in transmit power consumption compared to HD ISAC systems, especially when the radar SINR requirement is less stringent. For instance, Algorithm 1 yields lower power consumption than the HD mode.
- Sum Rate Maximization: The FD ISAC scheme can achieve nearly doubled sum rates compared to HD modes, particularly when residual self-interference (SI) power is low.
- Convergence: The proposed iterative algorithms (Algorithm 1 and Algorithm 3) demonstrate fast convergence, typically within 6 iterations, even with varying numbers of users.
- Special Case Efficiency: Algorithm 2, designed for the uplink-only scenario, offers a lower computational complexity per iteration compared to Algorithm 1 while achieving similar performance.
- Beampattern Control: Simulation results show that the proposed algorithms can effectively steer transmit beams towards desired targets and users while simultaneously creating deep nulls towards interferers, demonstrating the dual-functional capability.
Limitations and Future Work
- Ideal SI Cancellation: The paper assumes that self-interference (SI) at the BS is suppressed to an acceptable level using advanced SI cancellation techniques. The performance impact of imperfect SI cancellation is not explicitly studied.
- Perfect CSI: The algorithms assume perfect channel state information (CSI) is available at the BS. Future work could investigate the impact of imperfect CSI and explore robust beamforming designs.
- Flat Channels: The analysis assumes flat fading channels. Extending the framework to frequency-selective channels would be a valuable future direction.
- Environmental Dynamic Database: The paper assumes that target and interferer angles can be pre-estimated using an environmental dynamic database. The robustness of the proposed methods to estimation errors in these parameters could be explored.
- Non-Flat Channels and Imperfect CSI: The authors suggest that future extensions could investigate the impact of non-flat channels and imperfect CSI on the FD ISAC design.
Summary updated