Abstract: The fusion positioning of high-precision fiber-optic inertial navigation and high-precision position sensor is an effective method to realize the accurate positioning of coal mine roadway roadheader. However, the high-precision fiber-optic inertial navigation has high cost and the error accumulates with time. How to achieve high-precision fiber-optic inertial navigation performance and eliminate cumulative error through low-cost and low-precision fiber-optic inertial navigation adaptive zero-speed correction is an urgent problem to be solved. Therefore, an adaptive zero-speed correction method for fiber-optic inertial navigation of coal mine roadheader based on zero-speed detection and extended Kalman filter is proposed.Aiming at the problem of inaccurate zero-speed detection of traditional threshold method for roadheader fiber-optic inertial navigation, a zero-speed detection method based on PCA−SCSO−SVM ( Principal Component Analysis PCA, Sand Cat Swarm Optimization SCSO, Support Vector Machine SVM ) is proposed. This method uses roadheader vibration signal for zero-speed detection. Firstly, the vibration signal is decomposed by VMD and the IMF component is selected according to the correlation coefficient. Secondly, the time-frequency domain features of IMF components are extracted, and the principal component analysis method is used to reduce the dimension to reduce the complexity of the diagnostic model and the difficulty of data analysis. Finally, the accuracy of zero-speed detection is improved by introducing the sandcat swarm optimization algorithm to optimize the kernel function and penalty parameters.Aiming at the problem of high cost and error accumulation with time of high-precision fiber-optic inertial navigation, an adaptive zero-speed correction method is proposed. According to the correction interval time determined by the zero-speed detection results of the roadheader and the motion characteristics of the roadheader, the speed error and angular velocity error of the extended Kalman filter at the zero-speed moment are used as observations to perform adaptive zero-speed correction. In order to verify the effectiveness of the proposed method, the experimental verification of zero-speed detection and zero-speed correction is carried out. In the zero-speed detection experiment, this method, SVM method, GA−SVM method and PSO−SVM method are compared. The experimental results show that the zero-speed detection accuracy of this method is the highest, reaching 96.5%.The zero-speed correction experimental results show that the zero-speed correction method proposed in this paper can effectively reduce the attitude error of the fiber-optic inertial navigation and improve the attitude detection accuracy of the roadheader. The shorter the correction interval, the more accurate the error estimation and the higher the corrected attitude accuracy. When the correction interval is 10 minutes, the fiber-optic inertial navigation of 0.1（°）/ h can reach the attitude detection accuracy of 0.057（°）/h, and the low-precision fiber-optic inertial navigation can reach the high-precision positioning target.