Rayleigh Quotient Iteration for a Total Least Squares Filter in Robot Navigation

Abstract

Noisy sensor data must be filtered to obtain the best estimate of the robot position in robot navigation. The discrete Kalman filter, usually used for predicting and detecting signals in communication and control problems has become a common method for reducing the effect of uncertainty from the sensor data. However, due to the special domain of robot navigation, the Kalman approach is very limited. Here we propose the use of a Total Least Squares Filter which is solved efficiently by the Rayleigh quotient iteration method. This filter is very promising for very large amounts of data and from our experiments we can obtain more precise accuracy faster with cubic convergence than with the Kalman filter.