GLS Integrity Estimation Based on Improved EM Parameter Estimation Method
Lunlong Zhong, Zhang Zhuoxuan, Fan Zhendong · 2020
Integrity is not only an important parameter to measure navigation performance, but also the basis of other performance parameters calculation. When the aircraft approaches, the pseudo range error is non Gaussian distribution or bimodal asymmetric distribution. However, the traditional integrity estimation algorithm is based on the assumption that the pseudo distance error follows the zero mean Gaussian distribution, which reduces the system availability. Based on the analysis of traditional integrity estimation algorithms, a pseudo range error distribution model based on Gaussian mixture distribution model is established, and an integrity estimation algorithm based on improved expectation maximization (EM) method is proposed. The algorithm introduces unsupervised learning method, which greatly improves the iterative efficiency and reduces the amount of calculation, and can be used in embedded software. The experimental results of real data show the effectiveness of the algorithm.