LAUSR

A source coding (quantization) scheme is studied for the feedback of discrete-time and continuous-state cyber-physical systems (CPSs). It is formulated as a sequential coding optimization problem. The goal is to find a deterministic but adaptive policy, as a series of mappings from the historical information to the quantization strategy. In particular, an optimization problem is formulated, and then solved by the Bellman equation in dynamic programming (DP). To overcome the challenge of continuous state space, a practical solution is proposed by leveraging the approximate DP (ADP). The performance of the proposed strategy is examined for both scalar and vector dynamical systems in two practical applications. It shows that the designed policy can significantly outperform the simple fixed quantization strategies in CPSs and can be applied to the mobile/vehicle communication.