Application of the stVBRD5-CKF Algorithm for Joint Calibration of a Combined Onboard Navigation System in Mining Operations

The objective of this study is to address the challenges associated with the online joint calibration of external parameters for vehicle-mounted combined navigation systems used in special mining operations. These systems are susceptible to interference from the complex dynamics of the working environment and the inherent variability of sensors. To this end, a variational Bayesian (VB) adaptive dimensionality reduction fifth-order cubature Kalman filter (CKF) algorithm based on the student’s t-distribution (stVBRD5-CKF) is proposed. First, a joint calibration error model of the onboard combined navigation system is derived and established for the case of large Eulerian platform misalignment angle and azimuthal mounting error angle while considering the pole arm error. Based on the analysis of the statistical properties of engineering noise, it is modeled as a t-distribution that shares the same mean and variance as Gaussian white noise. Subsequently, an adaptive algorithm based on variational Bayes is employed to facilitate the real-time estimation of the covariance array of the measurement noise. Finally, the estimation performance and robustness of the stVBRD5-CKF algorithm proposed in this article are further verified by simulation and experiment.