Event-triggered filtering with application to target tracking in binary sensor networks

This paper presents an event-triggered filtering algorithm with application to target tracking in binary sensor networks. The target is modeled by a stochastic differential equation and the binary sensors provide one-bit information about the target´s presence or absence within the sensing range. The sensors are triggered when the target enters or leaves the sensor´s sensing range. Based on the sensor model, target tracking problem is formulated as a filtering problem where the state of the stochastic dynamic system is estimated using only two types of measurement data: the times when the sensors are triggered and the positions of the triggered sensors. The event-triggered filtering problem is then solved by a proposed algorithm based on a Markov chain approximation method.