In order to detect the attitude of the vehicle in real-time

a MEMS accelerometer based on single chip microcomputer（ MSC1214） is designed. Front-end sensitive component which senses accelaration is extracted and output by the digital circuit. The Kalman filtering and hardware average filtering are used to effectively filter out the random noise and improve the smoothness and response speed of signal curve. A software elimination zero error algorithm is designed

aiming at the zero offset problem

which effectively reduces the zero offset and shortens the digital output convergence time by about 16 times. Aiming at the zero drift problem of the system

a software algorithm is designed to track signal

which effectively suppresses the drift of zero position with time. After testing

the accelerometer can measure the linear acceleration of the carrier in the range of-60 ～ + 60 gn. The digital output stability after filtering is improved by 8. 77 times

and the linearity is not more than 1% FS.