Ran Slicing Design Policy for Integrated Sensing, Communication, and Computation

In the emerging Sixth-Generation (6G) telecommunications era, achieving ubiquitous intelligence requires perception technology to sense and collect information from the surrounding environment. This technology enables real-time, accurate environmental awareness, supporting informed decision-making and adaptive responses. While offering substantial opportunities for intelligence delivery, it also presents significant challenges, such as interoperability and network complexity. Integrated sensing, communication, and computation (ISCC) has emerged as a key enabler of 6G intelligence. However, traditional one-size-fits-all network architectures are inadequate for the diverse service demands of ISCC, as its performance spans multiple dimensions. In response, network slicing offers a transformative architecture, supporting multiple slices over a shared physical infrastructure to meet various service requirements of different use case, such as enhance mobile broadband and massive machine type communications. We propose a RAN slicing framework for ISCC that supports sensing, communication, and computation (SCC) services through tailored slices. A slice design objective is also defined to address the distinct SCC needs. We introduce a Deep Deterministic Policy Gradient (DDPG)-based algorithm of RAN slicing design policy. This algorithm optimizes sensing accuracy, transmission latency, and processing delay. Numerical simulations demonstrate that the proposed algorithm achieves stable convergence and outperforms baseline Deep Reinforcement Learning (DRL) methods in latency and sensing accuracy.