Inertial measurement units are widely used in pedestrian indoor positioning because of their advantages such as simple structure

small size and independence.However

the inertial positioning suffers from accumulative error

outlier and so on.The paper first preprocesses the original signal

then fusion filter and zero-speed update are taken to suppress the accumulative error.The strengthened adaptively outlier-restrained Kalman filter is proposed based on adaptive Kalman

which weakens the interference of outliers and suppresses the filter divergence.The experiments signify the positioning system could satisfy pedestrian’s accurate positioning in closed environment.