State-Driven Adaptive Deep-Unfolded PGA Algorithm for Hybrid Beamforming in MIMO-JCAS Systems

Area of Science:

• Wireless communication systems
• Signal processing
• Optimization algorithms

Background:

• Hybrid beamforming (HBF) offers a balance between performance and cost in massive MIMO joint communication and sensing (JCAS) systems.
• Designing HBF for MIMO-JCAS is challenging due to coupled precoders and high-dimensional search spaces.

Purpose of the Study:

• To propose a state-driven adaptive deep-unfolded HBF algorithm for MIMO-JCAS systems.
• To address the challenges of coupled optimization and dynamic tradeoffs in HBF design.

Main Methods:

• Redesigned analog precoder update using manifold-based form for constant-modulus constraints.
• Enhanced digital precoder update with a learnable gradient-balancing mechanism.
• Introduced a state-driven control network for adaptive hyperparameter scaling.

Main Results:

• The proposed algorithm demonstrates improved joint objective performance compared to conventional methods.
• Achieved faster convergence and enhanced robustness in simulations.
• Effectively manages the dynamic communication-sensing tradeoff.

Conclusions:

• The state-driven adaptive deep-unfolded HBF algorithm offers a more effective optimization strategy for MIMO-JCAS.
• The adaptive approach improves performance and efficiency over static methods.
• Validates the effectiveness of the proposed method through simulation results.