ALBERT

Description: Multi-antenna global navigation satellite systems (GNSS) can provide attitude measurement information for GNSS/INS integrated system. However, once the wrong attitude measurement is used in the tight coupled model, it is easy to lead to filtering divergence. Therefore, fault detection and exclusion (FDE) is very important for multi-antenna attitude determination. Because the baseline length information is often used as prior information in attitude determination, it cannot be used as the statistic of FDE again. In this paper, we propose a new statistic for attitude FDE, differential inter-system biases (DISB), and jointly use DISB, baseline length, standard deviation, and ratio to perform FDE for multi-antenna GNSS attitudes. The obtained attitudes with high reliability is added to the GNSS/INS tight coupled positioning model for measurement update. A set of urban vehicular data is used to verify the method. The results show that the DISB has FDE capabilities comparable to the baseline length information, which will be a powerful supplement when the baseline length information cannot be used. Compared to the single-antenna GNSS/INS tight coupled model, The accuracy of yaw,pitch and roll of the RTK/INS tightly coupled model with multi-antenna attitude measurement is improved by 57-64%,26-29% and 26-32% respectively, and the accuracy of the PPP/INS tightly coupled model with multi-antenna attitude measurement is improved by 60-61%,51-53% and 41-44% respectively.

Description: The loss of earthquake disaster is mainly caused by the damage and collapse of buildings caused by ground motion, and most of civil buildings in our country are greatly affected by high-frequency ground motion. The real and accurate high-frequency ground motion is of great significance for building seismic design and earthquake damage simulation. Therefore, when evaluating the high frequency ground motion, it is particularly important to consider a variety of uncertain factors of source parameters to obtain reasonable and reliable near-field high frequency seismic motion which is more in line with the actual situation. Based on the abundant strong earthquake data and scientific research results of the Lushan earthquake, this project plans to take the Lushan earthquake affected area as the demonstration area, use the empirical Green's function method as the simulation tool. The logic tree method and statistical empirical relationship are used to comprehensively deal with various uncertainties of parameters. The non-uniform stress drop and dynamic corner frequency are introduced into the key source parameters. By establishing the sliding distribution characteristic source model considering the parameter uncertainty, the high frequency strong ground motions with higher probability of different set magnitudes are obtained. The quantifiable scheme to verify the reliability of ground motion evaluation results is tried to be given. The practical high frequency ground motion evaluation method based on real small earthquake records and parameter uncertainty is also developed. It can provide more real and reliable high frequency ground motion input for seismic design and damage simulation of buildings.