Gyro-free accelerometer-based SINS: Algorithms and structures

Abstract

Recently researchers all over the world direct their efforts to the creation of the low-cost gyro-free SINS using linear accelerometers only for navigation of various unmanned vehicles, robots etc. The basic problem of such kind of SINS is the conversion of the redundant linear accelerometers readouts in the estimation of angular accelerations, angular rates and the attitude determination. The relation between amount of redundant sensors and performance (accuracy) of signal processing software is investigated. It is shown that increasing amount of sensors from minimal value 6 to 12 gives additional possibilities to correct estimation results, thus increasing the estimation accuracy. Simulation of such systems shows their efficiency especially in case of the moving vehicle, spinning with high angular rate, when traditional use of gyros might be very problematic.