Approximate Bayesian Smoothing with Unknown Process and Measurement Noise Covariances

Creative Commons License

Ardeshiri T., Ozkan E., ORGUNER U., Gustafsson F.

IEEE SIGNAL PROCESSING LETTERS, vol.22, no.12, pp.2450-2454, 2015 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 22 Issue: 12
  • Publication Date: 2015
  • Doi Number: 10.1109/lsp.2015.2490543
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Page Numbers: pp.2450-2454
  • Keywords: Adaptive smoothing, Kalman filtering, noise covariance, Rauch-Tung-Striebel smoother, sensor calibration, time-varying noise covariances, variational Bayes, MAXIMUM-LIKELIHOOD, MODELS
  • Middle East Technical University Affiliated: Yes


We present an adaptive smoother for linear state-space models with unknown process and measurement noise covariances. The proposed method utilizes the variational Bayes technique to perform approximate inference. The resulting smoother is computationally efficient, easy to implement, and can be applied to high dimensional linear systems. The performance of the algorithm is illustrated on a target tracking example.