Search

Aiming at the problem of unknown dynamic parameters of the new assembly during the on-orbit service, a parameter identification algorithm based on convolution neural network was proposed with the help of deep learning in multi-parameter optimization. The algorithm realizes the identification of the combined spacecraft′s multi-parameter under the condition of external force and non- conservation of linear momentum and angular momentum. A 4-layer convolution neural networks was designed by using the characteristic of the weight sharing of the convolution neural network. The identification of inertial parameters with high precision was achieved by plenty of training of state data in a specific form of storage in a short time. The feasibility of the convolution neural network algorithm was proved by simulation calculation. The results show that the proposed method can accurately and quickly identify the mass, centroid position and inertia matrix of combined spacecraft under the influence of external random force and moment, the identification accuracy is within 3%. [ABSTRACT FROM AUTHOR]

The method of real-time positioning for LEO satellites was studied. GPS positioning is usually used in the real-time positioning of low orbit satellites. Due to the use of broadcast ephemeris parameters and ordinary crystal vibration, the precision is generally within 10m, which cannot meet the needs of high-precision real-time positioning. IGS organization generates the precise ephemeris and clock error products through the global GPS tracking analysis centers, and releases the SSR ephemeris and clock error correction products on the Internet. These IGS products were investigated, and it was put forward that under the condition of existing observing and controlling support, high-precision location of SSR could be achieved by high-density injection of information flow. The background of augmentation load for a low orbit micro satellite was applied. The ephemeris and clock difference of a dual frequency GPS receiver were corrected with 1057 and 1058 data in IGS03 products, using the recursive least square estimation and LAMDA fuzziness to deal with the carrier phase and pseudo distance information. [ABSTRACT FROM AUTHOR]

The technology of estimating position and pose of spacecraft based on laser lidar is a research hotspot of onorbit services. For the position and pose estimation of failure spacecraft, the technology of general graphbased optimization SLAM was applied to the research of noncooperative target in space. In order to solve the problem of accumulative error generated in the dynamic scene, an improved SLAM algorithm based on detection of prior submap according to the characteristics of failure spacecraft was proposed. In this algorithm, laser lidar and inertia measurement unit were used to sample the point cloud data of failure spacecraft as well as the environment and the movement information of service spacecraft, to construct a relative pose graph of service scenario. Then the method of detecting prior submap was used to generate the constraints of discontinuous poses. At last, the constraints information was used to optimize the result of pose graph. The simulation result shows that the method decreases the cumulative deviation and improves precision of relative position and pose estimation compared to general SLAM algorithm, providing information for onorbit tasks such as navigation and controlling. [ABSTRACT FROM AUTHOR]