Reducing position instability of unaided inertial navigation systems in standstill

Abstract

Inertial navigation systems are frequently employed in robotic applications. It is well known that if they are used without the assistance of a reference system, position failures can increase rapidly in the course of time. This article presents various zero-velocity detection routines to reduce the position instability of robots equipped with unaided strapdown inertial navigation systems in standstill. To assess the detection performance under several environmental conditions such as varying temperature investigations are performed by various simulations as a function of noise level. The acceleration sensor signal was generated by a data synthesizer based on typical accelerometer noise characteristics. It can be shown that especially normality tests are independent from environmental fluctuations resulting in a high navigation accuracy.