An adaptive Neuro-Fuzzy Rao-Blackwellized particle filter for SLAM

The Rao-Blackwellized particle filter SLAM (RBPF-SLAM) that is also known as FastSLAM is a framework for simultaneous localization using a Rao-Blackwellized particle filter. The performance and the quality of the estimation of the Rao-Blackwellized particle filter depends heavily on the correct a priori knowledge of the process and measurement noise covariance matrices (Q_t and R_t) that are in most applications unknown. On the other hand, an incorrect a priori knowledge of Q_t and R_t may seriously degrade their performance. To solve these problems, this paper presents an adaptive Neuro-Fuzzy Rao-Blackwellized particle filter. The free parameters of adaptive Neuro-Fuzzy inference systems are trained using the steepest gradient descent (GD) to minimize the differences of the actual value of the covariance of the residual with its theoretical value as much as possible.