Abstract:Fluid antenna systems (FAS) enable dynamic antenna positioning, offering new opportunities to enhance integrated sensing and communication (ISAC) performance. However, existing studies primarily focus on communication enhancement or single-target sensing, leaving multi-target scenarios underexplored. Additionally, the joint optimization of beamforming and antenna positions poses a highly non-convex problem, with traditional methods becoming impractical as the number of fluid antennas increases. To address these challenges, this letter proposes a block coordinate descent (BCD) framework integrated with a deep reinforcement learning (DRL)-based approach for intelligent antenna positioning. By leveraging the deep deterministic policy gradient (DDPG) algorithm, the proposed framework efficiently balances sensing and communication performance. Simulation results demonstrate the scalability and effectiveness of the proposed approach.

What problem does this paper attempt to address?

This paper attempts to solve the following problems: 1. **Integrated Sensing and Communication (ISAC) Performance Optimization in Multi - target Scenarios**: - Existing research mainly focuses on communication enhancement or single - target sensing, and the ISAC performance optimization in multi - target scenarios has not been fully explored. The paper aims to improve the ISAC performance in multi - target scenarios through intelligent antenna positioning. 2. **Joint Optimization of Beamforming and Antenna Position**: - The joint optimization of beamforming and antenna position is a highly non - convex problem. As the number of fluid antennas increases, traditional methods become impractical. The paper proposes a method based on Deep Reinforcement Learning (DRL) to solve this complex optimization problem. 3. **Application of Dynamic Antenna Positioning in ISAC Systems**: - The Fluid Antenna System (FAS) can achieve dynamic antenna positioning, providing new opportunities for ISAC systems. However, most of the existing work only focuses on single - target or pure communication - driven cases, and the application of multi - target sensing has not been fully explored. By introducing the DRL framework, the paper explores the potential of FAS in multi - target sensing. ### Specific Problem Description - **Balance between Communication and Sensing Performance**: Due to the limitation of fixed antenna positions, traditional ISAC systems have difficulty in simultaneously optimizing communication and sensing performance in the case of multiple sensing targets. By introducing the Fluid Antenna System (FAS) and using its dynamic positioning ability, the paper solves this problem. - **Solution to Non - convex Optimization Problems**: The joint optimization problem of beamforming and antenna position is highly non - convex, and traditional methods are difficult to deal with. The paper proposes a framework based on Block Coordinate Descent (BCD) and Deep Deterministic Policy Gradient (DDPG) algorithms, which effectively solves this problem. - **Support for Multi - target Sensing**: Most of the existing research is limited to single - target sensing, while the requirements for multi - target sensing are more complex. By introducing the DRL framework, the system can perform well in multi - target sensing scenarios. ### Solutions The paper proposes an intelligent antenna positioning framework based on Block Coordinate Descent (BCD) and Deep Reinforcement Learning (DRL), which specifically includes the following aspects: - **BCD Framework**: Alternately optimize the beamforming vector and antenna position to gradually approach the optimal solution. - **DDPG Algorithm**: Directly optimize the position of the fluid antenna without relying on traditional iterative schemes, improving the optimization efficiency. - **Simulation Results**: Verify the scalability and effectiveness of the proposed method and show its real - time optimization ability in a multi - user environment. Through these methods, the paper successfully solves the performance optimization problem of ISAC systems in multi - target scenarios and shows the advantages of FAS in dynamic antenna positioning.

Towards Intelligent Antenna Positioning: Leveraging DRL for FAS-Aided ISAC Systems