Integrated sensing and communication (ISAC) has emerged as a pivotal technology for enabling vehicle-to-everything (V2X) connectivity, mobility, and security. However, designing efficient beamforming schemes to achieve accurate sensing and enhance communication performance in the dynamic and uncertain environments of V2X networks presents significant challenges. While AI technologies offer promising solutions, the energy-intensive nature of neural networks (NNs) imposes substantial burdens on communication infrastructures. This work proposes an energy-efficient and intelligent ISAC system for V2X networks. Specifically, we first leverage a Markov Decision Process framework to model the dynamic and uncertain nature of V2X networks. This framework allows the roadside unit (RSU) to develop beamforming schemes relying solely on its current sensing state information, eliminating the need for numerous pilot signals and extensive channel state information acquisition. To endow the system with intelligence and enhance its performance, we then introduce an advanced deep reinforcement learning (DRL) algorithm based on the Actor-Critic framework with a policy-clipping technique, enabling the joint optimization of beamforming and power allocation strategies to guarantee both communication rate and sensing accuracy. Furthermore, to alleviate the energy demands of NNs, we integrate Spiking Neural Networks (SNNs) into the DRL algorithm. By leveraging discrete spikes and their temporal characteristics for information transmission, SNNs not only significantly reduce the energy consumption of deploying AI model in ISAC-assisted V2X networks but also further enhance algorithm performance. Extensive simulation results validate the effectiveness of the proposed scheme with lower energy consumption, superior communication performance, and improved sensing accuracy.

翻译：通感一体化（ISAC）已成为实现车联网（V2X）连接、移动性和安全性的关键技术。然而，在车联网动态且不确定的环境中，设计高效的波束赋形方案以实现精确感知并提升通信性能，面临着重大挑战。尽管人工智能技术提供了有前景的解决方案，但神经网络（NNs）的高能耗特性给通信基础设施带来了沉重负担。本文提出了一种面向车联网的节能智能通感一体化系统。具体而言，我们首先利用马尔可夫决策过程框架来建模车联网的动态和不确定性。该框架使得路侧单元（RSU）能够仅依赖其当前感知状态信息来制定波束赋形方案，从而无需大量导频信号和广泛的信道状态信息获取。为了赋予系统智能并提升其性能，我们随后引入了一种基于Actor-Critic框架并采用策略裁剪技术的高级深度强化学习（DRL）算法，实现了波束赋形与功率分配策略的联合优化，以同时保证通信速率和感知精度。此外，为了缓解神经网络的能耗需求，我们将脉冲神经网络（SNNs）集成到DRL算法中。通过利用离散脉冲及其时间特性进行信息传输，SNNs不仅显著降低了在ISAC辅助的车联网中部署人工智能模型的能耗，还进一步提升了算法性能。大量的仿真结果验证了所提方案的有效性，其具有更低的能耗、更优的通信性能和更高的感知精度。