Abstract
Cooperative-integrated sensing and communication (C-ISAC) networks have emerged as promising solutions for communication and target sensing. However, imperfect channel state information (CSI) estimation and time synchronization (TS) errors degrade performance, affecting communication and sensing accuracy. This paper addresses these challenges by employing movable antennas (MAs) to enhance C-ISAC robustness. We analyze the impact of CSI errors on achievable rates and introduce a hybrid Cramer-Rao lower bound (HCRLB) to evaluate the effect of TS errors on target localization accuracy. Based on these models, we derive the worst-case achievable rate and sensing precision under such errors. We optimize cooperative beamforming, base station (BS) selection factor and MA position to minimize power consumption while ensuring accuracy. We then propose a constrained deep reinforcement learning (C-DRL) approach to solve this non-convex optimization problem, using a modified deep deterministic policy gradient (DDPG) algorithm with a Wolpertinger architecture for efficient training under complex constraints. Simulation results show that the proposed method significantly improves system robustness against CSI and TS errors, where robustness mean reliable data transmission under poor channel conditions. These findings demonstrate the potential of MA technology to reduce power consumption in imperfect CSI and TS environments.
Introduction
The evolution toward 6G wireless networks demands the seamless integration of communication and sensing functionalities. Integrated Sensing and Communication (ISAC) utilizes the same hardware and spectrum for both tasks, offering significant gains in efficiency. However, in cooperative ISAC (C-ISAC) systems, where multiple base stations (BSs) coordinate to serve users and track targets, performance is highly sensitive to real-world impairments. Specifically, imperfect Channel State Information (CSI) and Time Synchronization (TS) errors between BSs can severely degrade the signal-to-interference-plus-noise ratio (SINR) for communication and the localization accuracy for sensing.
To mitigate these issues, this research explores the use of Movable Antennas (MAs). Unlike traditional fixed-position antennas (FPAs), MAs can be dynamically repositioned within a local region to optimize the wireless channel's spatial characteristics. This added degree of freedom allows the system to "steer" around deep fades or timing misalignments, potentially reducing the transmit power required to meet strict performance constraints.
System Model and Robustness Analysis
The considered C-ISAC network consists of multiple MA-aided BSs, several single-antenna communication users (UEs), and a target to be sensed. Each BS is equipped with a set of MAs that can move within a designated transmit or receive region. The system must simultaneously satisfy the communication requirements of UEs and the sensing accuracy requirements for the target.
The paper introduces a robust modeling framework to handle uncertainties. For communication, the worst-case achievable rate is derived to ensure reliability even under the maximum expected CSI error. For sensing, the Hybrid Cramer-Rao Lower Bound (HCRLB) is utilized to characterize the fundamental limit of target localization accuracy in the presence of stochastic TS errors. This dual-constraint approach ensures that the system remains functional in non-ideal environments.
Proposed C-DRL Framework
The joint optimization of beamforming weights, BS selection, and MA positions is a highly non-convex problem with complex constraints. To address this, a Constrained Deep Reinforcement Learning (C-DRL) approach is proposed. The framework utilizes a modified Deep Deterministic Policy Gradient (DDPG) algorithm.
To handle the discrete nature of BS selection alongside the continuous variables of beamforming and antenna positioning, the Wolpertinger architecture is integrated. This allows the DRL agent to map continuous actions to the nearest valid discrete actions efficiently. The agent is trained to minimize total transmit power while being penalized for violating communication SINR or sensing HCRLB thresholds.
Performance Evaluation
Simulation results demonstrate the significant advantages of the MA-enabled system. By exploiting spatial diversity through antenna movement, the system can achieve the same performance as FPA systems with substantially lower transmit power.
As shown in the figure above, increasing the number of MAs leads to a dramatic reduction in the required average transmit power. For instance, with 32 MAs, the system can operate at power levels significantly lower than those required by FPA systems, even when CSI errors are present. Furthermore, the results indicate that MAs can compensate for infrastructure limitations; a system with fewer BSs but equipped with MAs can outperform a larger FPA-based network in terms of both power efficiency and sensing reliability.
Conclusion
This work highlights the transformative potential of Movable Antennas in 6G C-ISAC networks. By providing a new spatial dimension for optimization, MAs effectively counter the performance degradation caused by imperfect CSI and time synchronization errors. The proposed C-DRL framework offers a robust and efficient method for managing these complex systems, paving the way for more energy-efficient and reliable integrated sensing and communication deployments.
Summary updated