Optimal Beamforming for Full-Duplex Integrated Sensing, Communication and Computation Systems

This paper investigates optimal beamforming design for full-duplex (FD) Integrated Sensing, Communication, and Computation (ISCC) systems, which are envisioned as key enablers for next-generation wireless networks. To address the challenge of cross-domain interference from uplink transmissions, we employ a successive interference cancellation (SIC) approach to enhance the radar signal quality. However, due to the imperfection of interference cancellation algorithms, residual interference persists and is modeled as additive noise. Additionally, the portion of the radar signal embedded in the computation data is also modeled as additive noise. We derive closed-form expressions for the uplink receive combiners used in both sensing and computation, along with the corresponding power allocation strategy. We also propose an optimal beamforming algorithm to jointly maximize the aggregated downlink communication and radar information rate (RIR) under a total power constraint. Given the conflicting nature of these objectives, a multi-objective optimization (MOO) framework is adopted. To handle the non-convexities arising from inter-user and residual self-interference, we utilize semidefinite relaxation (SDR) in conjunction with sequential convex approximation (SCA). An alternating maximization approach is used to iteratively optimize the beamforming vectors and uplink power allocation.We further demonstrate that rank-one solutions can be reconstructed from the relaxed problem without loss of optimality. Extensive Monte Carlo simulation results confirm that the proposed algorithm significantly outperforms baseline schemes, yielding substantial improvements in communication rate, radar information rate (RIR), and computation mean squared error (MSE). These findings underscore the potential of full-duplex ISCC systems to deliver high-performance, interference-resilient solutions for future wireless networks.