Distributed State Estimation with Random Switching Topology Under Sensor Saturation Constraints

This paper presents a distributed robust filtering technique for sensor networks that incorporates randomly switching communication topologies and sensor saturation nonlinearities. In the sensor network, numerous sensor nodes are placed to detect and measure the target plant, transmitting data to a remote filter via the network. Among the filter network, each local estimator receives measurement data from sensor nodes and estimation data from neighboring nodes through a randomly varying topology. The sensor network employs a Bernoulli binary distribution to model random saturation nonlinearity, while random switching topologies are characterized using a non-homogeneous Markov chain. Next, a sufficient condition for the existence of a distributed filter is presented using the linear matrix inequality (LMI). Lastly, a numerical example illustrates the viability of the designed strategy.