Experimental Transmission Power Control Policy for Remote State Estimation in Industrial WSN

Industrial Wireless Sensor Networks (IWSN) are used as a backbone for Networked Control Systems (NCS). In a wireless NCS, a closed-loop control system periodically delivers the instantaneous state of a physical plant under control to a remote state estimator (RSE) via a wireless channel. The RSE evaluates the state of the plant in order to compute control commands for the actuators. Loss of control packets due to channel interference leads to instability of the system under control. The maximum transmission power can be selected to circumvent this problem, however, it further exacerbates the problem of limited energy resources of the hardware-constrained IWSN transmitters. To overcome this issue, adaptive transmission power control (TPC) policies have to be applied to minimize the power consumption while maintaining the desired RSE distortion. The usage of TPC policies for RSE has been intensively investigated in theory, however, their effectiveness has not been confirmed in a realistic scenario within real IWSN. In this work, an IWSN testbed is deployed and a measurement-based TPC policy is calculated. The policy is evaluated via measurements of a simulated NCS and its performance is compared to fixed power and model-based TPC from the literature. With respect to fixed power, the measurement-based policy leads to an improvement of RSE performance up to 20%, while decreasing the energy consumption. The comparison with the model-based policy shows comparable RSE performance with more than 50% reduction in transmission power.