Hybrid Beamforming Design for Communication-Centric ISAC

Wireless communication and radio-sensing systems have evolved independently due to their distinct primary functions and design considerations. However, the increasing demand for spectrum, particularly in densely populated areas, has necessitated the development of more efficient coexistence approaches. Integrated Sensing and Communication (ISAC) is an emerging concept that combines sensing and communication functions into a single system. The wireless networks send information and "sense" or scan their surroundings using shared hardware, spectrum, and signals.

Beamforming is one of the fundamental techniques in ISAC systems. It focuses electromagnetic radiation in a specific direction by carefully adjusting the timing of signals transmitted or received by an antenna array. This helps improve signal quality, reduce interference, and increase capacity. The paper provides a detailed overview of existing Full Digital (FD) and Hybrid Analog Digital (HAD) architectures.

The earlier approaches mainly focused on radio sensing, prioritizing sensing metrics while ensuring communication performance. Some used dual function precoders to handle both simultaneously but often experienced interference in communication receivers. More recently, fully connected HAD beamforming designs have been explored, including zero-forcing approaches and weighted sum optimization methods, to balance spectral efficiency and sensing accuracy.

This paper presents a new approach for creating fully connected hybrid beamforming in multiuser, multibeam ISAC systems. In this system, a base station with multiple antennas transmits signals to users and detects targets. It uses a beamformer to focus signals and a bistatic receiver to avoid self-interference. The base station estimates channel conditions and determines the necessary power for target sensing. Unlike previous methods, this approach prioritizes improving communication quality while ensuring that radars continue to function effectively.

The problem was complex due to the non-convex nature of hybrid beamforming, power limitations, and the constant modulus constraint for analog components. To make the problem more manageable, the design reformulates the non-convex problem by exploiting the equivalence between weighted sum-rate maximization and weighted-sum mean-squared-error (WMMSE) minimization. However, the hybrid architecture and radio-sensing constraints in the problem meant that there was no closed-form solution. Therefore, the design uses a novel iterative algorithm that adjusts the precoder objective at each iteration, gradually improving performance until a solution is reached for the original problem. The method alternates between optimizing the digital and analog precoder matrices while ensuring that radar sensing constraints are always met.

The proposed WMMSE-based iterative alternate optimization, called ItAlHAD algorithm, was compared to FD precoding, TwoS-AltMin, and CADMM algorithms. Simulation results show that the ItAlHAD algorithm converges quickly and outperforms competing methods regarding weighted sum-rate and radar sensing accuracy. The algorithm achieves performance close to FD beamforming with far fewer RF chains, making it a practical solution for large-scale ISAC systems.

Hybrid beamforming is an effective method to balance the competing demands of communication and radar sensing, particularly in scenarios where spectrum efficiency and energy consumption are critical. The proposed algorithm offers significant benefits in terms of efficiency, cost-effectiveness, and performance, making it a promising solution for next-generation communication systems.