摘要

This paper develops a novel approach to obtain the fixed ambiguity solution with orthonormal constraints for Beidou attitude determination. The approach extends the traditional attitude model by taking the attitude matrix elements as unknown parameters directly. The orthonormal property of the attitude matrix is used as constraints to assist in calculating the float ambiguity solution and its corresponding variance-covariance matrix. An iteration algorithm with the linearization of orthonormal constraints is developed to solve the nonlinear conditional extremum problem and a sequential filtering multi-epoch ambiguity resolution method is derived for real-time applications. Long-time static experiments with three antennas tracing Beidou signals are employed to demonstrate the reliability and effectiveness of the proposed approach. The results indicate that, compared to unconstrained ambiguity resolution, the proposed approach can improve the reliability of the integer ambiguity vectors for single epochs and can also decrease the time to fix for multiple epochs.