By Paolo Franchi
Adaptive Bayesian tracking under time-varying target detection capability of sensor networks with measurement origin uncertainty. Papa, Giuseppe; Braca, Paolo; Horn, Steven A. CMRE-FR-2014-026. December 2014.
In practical target tracking applications the target detection performance of the sensors may be unknown and also change rapidly in time. This work considers a network of sensors and develops a target tracking procedure able to adapt and react to the time-varying changes of the network detection capability. These types of changes can seriously degrade the overall performance of the tracking system in terms of both the tracking of the target state and the detection of the presence or absence of the target. In this work, the above problem is solved proposing a tracking strategy based on a Bayesian framework, in which the dynamic target state is augmented with the sensor detection probabilities. The proposed method, referred to as adaptive tracker, is validated using extensive computer simulations and real-world experiments, conducted by the North Atlantic Treaty Organization (NATO) Science and Technology Organization (STO) - Centre for Maritime Research and Experimentation (CMRE). The adaptive tracker is studied using a dataset collected during the CMRE high frequency (HF)-radar experiment, which took place between May and December 2009 on the Ligurian coast of the Mediterranean Sea. Also studied is a dataset collected during Proud Manta 2012 using the CMRE underwater tracking system, composed of an underwater wireless sensor network of autonomous underwater vehicles for anti-submarine warfare applications.