Synesthesia of machine (SoM)-driven analog precoder optimization for enhanced ISAC performance in sub-THz systems

Area of Science:

• Electrical Engineering
• Wireless Communications
• Signal Processing

Background:

• Integrated Sensing and Communication (ISAC) is crucial for future sub-terahertz (sub-THz) systems.
• Line-of-sight (LoS) propagation dominates sub-THz channels, simplifying transmitter design but introducing challenges.
• The beam squint effect in sub-THz systems hinders the balancing of sensing and communication functionalities in ISAC.

Purpose of the Study:

• To explore the design flexibility of sub-THz analog hardware for ISAC.
• To enhance dual-functional gains by optimizing the analog precoder to adapt to sub-THz channel characteristics.
• To develop a computationally efficient solution for ISAC challenges.

Main Methods:

• Analog precoder optimization to adjust the equivalent channel for enhanced dual-functional gains.
• Development of a lightweight complex-valued squint-aware network (CSP-Net) inspired by the synesthesia of machine (SoM).
• Utilizing both communication and sensing channel data within CSP-Net, with an architecture tailored to specific data and task characteristics.

Main Results:

• Demonstrated that adjusting the equivalent channel through the analog precoder enhances dual-functional gains.
• Proposed a near-optimal benchmark for analog precoder optimization.
• Developed CSP-Net to effectively reduce algorithmic complexity while maintaining performance.
• Validated the effectiveness of the proposed schemes through simulations.

Conclusions:

• Analog precoder optimization is key to overcoming beam squint limitations in sub-THz ISAC.
• The proposed CSP-Net offers an efficient solution for ISAC by leveraging dual-channel data.
• The developed methods significantly enhance ISAC performance in sub-THz environments.