Short Blocks, Fast Sensing: Finite Blocklength Tradeoffs in RIS-Assisted ISAC

Integrated sensing and communication (ISAC) is a cornerstone for future sixth-generation (6G) networks, enabling simultaneous connectivity and environmental awareness. However, practical realization faces significant challenges, including residual self-interference (SI) in full-duplex systems and performance degradation of short-packet transmissions under finite blocklength (FBL) constraints. This work studies a reconfigurable intelligent surface (RIS)-assisted full-duplex ISAC system serving multiple downlink users while tracking a moving target, explicitly accounting for SI and FBL effects in both communication and sensing. We formulate an optimization framework to minimize service adaptation gaps while ensuring sensing reliability, solved via alternating optimization and successive convex approximation. Numerical results show that short blocklengths enable fast adaptation but raise radar outage from fewer pulses and motion sensitivity. Longer blocklengths improve signal-to-interference-plus-noise ratio (SINR) and reduce outages but allow motion to degrade sensing. A "sweet spot" arises where blocklength and beamformer allocation optimize throughput and sensing, seen as a local minimum in radar SINR variance. RIS-assisted optimization identifies this balance, achieving reliable communication and radar sensing jointly.