The Research On Error Compensation Of Mems Gyroscope

Micro-Electro-Mechanical System(MEMS) gyroscopes have the advantages of small size, low cost and high volume production over the conventional mechanical gyroscope. Hence, they have been widely used in navigation systems, automotive electronics, mobile phones and commercial game products, etc. However, the present MEMS technology has its own limitations, such as large process variation, and the material used to implement the structures are sensitive to temperature change and abrupt impact.Large process variation in MEMS manufacturing process could causes the fabricated drive and sense vibration beams of the MEMS gyroscopes being not fully orthogonal to each other; this results in non-ideal effect of anisoelasticity in MEMS structure leading to serious problem called the quadrature error. The MEMS materials used are usually highly temperature dependent. It degrades the performance of gyroscopes and results in the output of gyroscopes varied with temperature. In addition, the MEMS structure is very fragile. Any abrupt impact on the MEMS gyroscope could create large outliers in the readout of gyroscopes leading to high linear acceleration error. However, the present commercial products rarely take care of such errors due to cost issues resulting in their specifications not able to meet the requirements of those applications that need high precision.To solve the above problems, the thesis proposes three corresponding methods to counteract each of the problems individually. First, a new improved PID controller with the addition of decision making algorithm to implement an adaptive feature and an improved integration function to prevent high overshoot in real-time is proposed to compensate the quadrature error. Second, to reduce the effect of temperature on gyroscopes, an improved polynomial curve fitting using segmentation fitting is proposed. Third, an improved adaptive Kalman filter with the function of updating the optimal estimation in real-time by a new estimation error equation is proposed to compensate the linear acceleration error.To verify the effectiveness of the above algorithms, the three algorithms are implemented in MATLAB. Using the improved adaptive PID controller, the quadrature error at the output of the gyroscope under zero bias condition has decreased from-0.350 V to-0.005 V. After going through the temperature compensation of the three fitting functions, an error variance has decreased from 1.827e-006 to 2.7545e-009. Applying the improved adaptive Kalman filter, the error variance has decreased from 0.0380 to 0.0014.