Temperature compensation method for low cost three-axis MEMS digital angular rate gyroscopes
Temperature compensation method for low cost three-axis MEMS digital angular rate gyroscopes

In recent years, a large number of small volume, low cost micro electro mechanical systems (MEMS) digital three-axis angular rate gyroscopes have been developed and widely used in civil and military fields. However, these kinds of gyroscopes have poor performances in initial zero-bias, temperature drift, In-Run bias stability, bias repeatability, etc., their output errors need to be compensated before being used. Based on a lot of experiments, the temperature drift and the initial zero-bias are the major error sources in the MEMS gyroscopes output data. Due to the poor repeatability of temperature drift, the temperature compensation coefficients need to be recalculated every time before using. In order to recalculate parameters of the temperature compensation model quickly, a 1st-order polynomial model of temperature is established, then a forgetting factor recursive least squares estimator will be adopted to identify the model parameters in real time. Static and dynamic experimental data shows that this method removed/compensated the temperature drift and initial zero-bias from the output of the gyroscopes effectively.
In recent years, a large number of small volume, low cost micro electro mechanical systems (MEMS) digital three-axis angular rate gyroscopes have been developed and widely used in civil and military fields. However, these kinds of gyroscopes have poor performances in initial zero-bias, temperature drift, In-Run bias stability, bias repeatability, etc., their output errors need to be compensated before being used. Based on a lot of experiments, the temperature drift and the initial zero-bias are the major error sources in the MEMS gyroscopes output data. Due to the poor repeatability of temperature drift, the temperature compensation coefficients need to be recalculated every time before using. In order to recalculate parameters of the temperature compensation model quickly, a 1st-order polynomial model of temperature is established, then a forgetting factor recursive least squares estimator will be adopted to identify the model parameters in real time. Static and dynamic experimental data shows that this method removed/compensated the temperature drift and initial zero-bias from the output of the gyroscopes effectively.