CMStatistics 2019: Start Registration
View Submission - CMStatistics
B1086
Title: A SWL fixed-lag smoothing algorithm with intermittent observations in the presence of correlated noises Authors:  Rosa Maria Fernandez-Alcala - University of Jaen (Spain) [presenting]
Rahul Yadav - University of Jaen (Spain)
Abstract: A new fixed-lag smoothing algorithm in the quaternion domain is devised in order to handle intermittent observations in the presence of correlated noises at the same time. The problem is formulated in the state-space framework where a sequence of independent Bernoulli random variables is introduced to model random interruptions. Assuming the quaternion signal is C-proper, the proposed methodology is based on a semi-widely linear processing which reduces the dimension of the problem to a half. In this approach, the initial state space-model involving autocorrelated and cross-correlated noises is reformulated to make the noise sequence does not correlate with each. Subsequently, a similar reasoning that of the quaternion Kalman filtering equations is followed. The benefits of the proposed solution over the conventional fixed-lag smoothing algorithm are illustrated by means of a simulation example.