Task-Aware Pilot Design for ISAC Systems via Mutual Information Optimization
Abdulmuneem Alselwi, Osamah Khaled, Magdi Dhaiban, Khaled Abdullah, Hussein Abdulhadi · 2025
This paper presents an environment-aware adaptive pilot design framework for integrated sensing and communication (ISAC) systems operating in dynamic wireless environments. Unlike conventional static pilot schemes, our approach leverages real-time updates based on evolving channel and scene statistics to jointly optimize sensing and communication performance. We formulate a scalarized optimization problem that jointly considers the communication and sensing MI, parameterized by a task-priority weight ρ. To solve this problem under orthogonality constraints, we develop a projected gradient descent (PGD) algorithm operating on the Stiefel manifold, enabling the generation of environment-adaptive and task-aware pilot matrices. The proposed approach exhibits superior performance across a wide range of operating conditions. Extensive simulations demonstrate that it achieves lower channel estimation error, improved mutual information trade-offs, and faster convergence compared to fixed or random orthogonal designs. Furthermore, the framework adapts to variations in user load, pilot length, SNR, and clutter power, while maintaining computational efficiency. The PGD method's ability to dynamically balance communication and sensing objectives makes it highly suitable for next-generation ISAC systems requiring flexible and efficient joint functionality.