Highly precise MEMS gyroscopes for fully automated driving

Future, fully automated vehicles pose strict requirements on the performance of inertial sensors. Achievable accuracies of micro-electromechanical (MEMS) gyroscopes are however challenged by a number of non-idealities and long-term drift effects like bias instability. This dissertation analyzes the effect of bias instability on purely inertial navigation in comparison with other noise effects. It is shown, that bias instability becomes the dominant error component during navigation periods as short as ten seconds. At the core of the dissertation, origins and mechanisms of bias instability in triaxial, mode-matched, force-feedback MEMS gyroscopes are examined. Results gain credibility through a combination of analytical investigation, model-based simulation and extensive experimental analysis on real-world, next-generation sensor prototypes. It is found, that a combination of flicker noise on the frequency tuning voltage ensuring mode-matching together with certain types of rate offsets form a dominant source of bias instability. Feedback control on the frequency tuning voltage using pilot tones leads to an improvement of up to a factor of ten for the z-axis. Bias instabilities of lower than 0.1 dph are reached, which is an unprecedented value for automotive-type MEMS gyroscopes. The out-of-plane sensing x- and y-axes are shown to experience an additional, yet-unknown contribution of bias instability and could not be improved by frequency tuning control. For the first time, scale-factor instability was described and analyzed in detail. This effect produces an increase in signal drift with higher measured angular rates. Lastly, a novel measure for mode-matching was devised, which, contrarily to the pilot tone scheme, only uses the existing noise in the gyroscope’s force-feedback structure to estimate the sense mode’s detuning and scale-factor.