Real-Time Monitoring Ornstein–Uhlenbeck Process: Improving Fidelity Under Freshness Guarantees
Tianci Zhang, Zhengchuan Chen, Zhong Tian, Yunjian Jia, Min Wang, Tony Q. S. Quek
Abstract
This work studies improving the data fidelity for real-time monitoring of an Ornstein-Uhlenbeck process in a sensor-monitor status update system under stringent information freshness provision. Violation probability of age of information (VPA), i.e., the long-term proportion the age of information exceeds a given threshold is adopted to strictly assess freshness. Based on the automatic repeat-request transmission protocol, the VPA at the monitor is derived in closed-form. To evaluate the fidelity, the time-average mean squared error (MSE) between the real-time process value and the estimation at the monitor is also obtained, where the coupling between the time-average MSE and the VPA is characterized in detail. By analyzing the monotonicity of the time-average MSE and the VPA, the optimal quantization precision that achieves the minimal time-average MSE under a given VPA constraint is proposed, where an algorithm for quickly addressing the optimum is also given. Numerical results validate the effectiveness of the proposed optimal quantization design.