Simultaneous Calibration Method for Doppler Velocity Log Errors Based on a Genetic Algorithm

Abstract

Doppler velocity log (DVL) errors affect the accuracy of strapdown inertial-navigation system (SINS)/DVL integrated navigation. Most existing calibration methods do not consider the problem of DVL sampling frequencies that are not fixed, and they cannot cope with large misalignment angles. This study proposes a method to calibrate internal and external DVL errors simultaneously, based on a genetic algorithm. To solve the problem of asynchronous SINS and DVL sampling, a method for unifying the time reference using cubic-spline interpolation is proposed. An error model for the DVL is established; the scale factor and misalignment angles are used as optimization genes for a genetic algorithm. The method can achieve high accuracy without requiring the misalignment angles to be small, significantly reducing the practical difficulty of installation. The results of simulation and shipboard experiments indicate that the calibration accuracy of the proposed algorithm is significantly better than that of the traditional method. For example, when the misalignment angle reaches 15°, the proposed algorithm has only a small error. If the carrier pitch angle changes to 5°, the root-mean-square error is reduced by 31.15% in the x-direction and 69.15% in the z-direction under highly dynamic conditions. The method thus seems likely to be useful in real-world engineering settings.