Multisensor data fusion algorithms for estimation of a walking person position

This paper presents the problem of sensor fusion to estimate a walking person position. Simple walking person moving model is introduced. We propose two filtering algorithms to solve the present problem. The first algorithm represents an extended Kalman filter (EKF) model which is based on the principle of the state transition matrix and observation matrix linearization under Taylor series expansions. During walking, several various “walking modes” differ from each other in terms of moving model parameters. This fact is addressed to the second sensor fusion algorithm, namely, the interacting multiple model (IMM) algorithm, which also employes a single EKF for each considered mode, combines the state estimates and covariance matrices.