1 Research Work
In networked control systems, estimation of different process parameters/states is extremely important so that the controller is up to date with the processes, and hence, can control the system as desired, and maintain stability. Such estimation is done using sensors, which are usually placed at locations that are different from the controllers. These sensors take measurements, perform some local computations, and communicate the processed measurements to the remote controllers either through a wired or a wireless channel. For this purpose, there exists a need for a communication network protocol designed such that the controller is kept updated without network congestion. It would also be beneficial, if the protocol is such that it allows the network to be used for other applications and still maintains certain level of performance. Traditional protocols like TCP are not suited for applications where simple reliability is not of utmost importance and certain transmission requirements must be satisfied. Hence, we tackled this design issue by formulating a Network Utility Maximization (NUM) problem with the objective of minimizing estimation error at the controller and finding distributed solutions to this optimization problem.