Optimum Sensor Value Transmission Scheduling for Linear Wireless Networked Control Systems

In this work, an optimum communication resource allocation for sensor value transmission of wireless networked control systems consisting of independent linear subsystems is calculated. To use new service types of upcoming mobile radio standards in an optimal way, the communication resources used by each subsystem have to be adapted to the current individual subsystem requirements. In our case, the scheduling of the transmission of fresh sensor values to the controller of the subsystems is optimized. The quality of scheduling is directly related to the deviation of a subsystem from its control goal. A proof for the optimality of a regular updating scheme for minimizing this deviation is given. Based on this regularity, the optimum communication resource allocation for a limited amount of resources and a set of control systems with different given characteristics is derived. Using the resulting resource shares for the subsystems from this optimization, an algorithm to schedule the actual transmissions during runtime is given. This split approach saves computational effort during system operation. To simplify the evaluation of different scheduling policies, a deterministic calculation of the expected cost with mean absolute error cost function for each subsystem is presented, which removes the need of Monte-Carlo experiments for evaluation.