BBRv2-TI: A Utility-Guided Congestion Control Algorithm for Tactile Internet over QUIC

Tactile Internet applications such as remote surgery and real-time haptic control require ultra-low delay, high reliability, and steady throughput. Existing congestion control algorithms (CUBIC, Reno, BBRv2) degrade under jitter and loss. This work presents BBRv2-TI for QUIC-based TI, modeling congestion control as a multi-objective optimization over throughput, jitter, and loss with a gradient-free utility. Eleven BBRv2 parameters in QUICHE are tuned to improve responsiveness without kernel or cross-layer changes. In emulation, BBRv2TI reduces jitter by up to 98%, lowers loss by about 30%, and increases throughput by as much as 1.5× over baselines, indicating suitability for delay sensitive TI systems.