Abstract:Fiber optic inertial navigation has the characteristics of high resolution, large dynamic range, high sensitivity, long lifespan, strong autonomy, and low cost, it is applied to driving behavior and vehicle attitude recognition systems. The error in the vehicle attitude recognition system for medium and low precision fiber optic inertial navigation system is analyzed, and a self-adaptive vehicle inertial navigation alignment revision method mode and a full parameter self-calibration error revision method mode are proposed to solve the problem of medium and low precision fiber optic inertial navigation system errors accumulating, and achieve autonomy in vehicle attitude recognition. Meanwhile， it provides strong support for driving attitude recognition and vehicle behavior recognition. The analysis and test results show that the vehicle attitude recognition system can guide driving safety.