Study On Spatial Information Analysis And Processing Of The Disicion-Making For Launching Safety Control

Space launching is a complex system engineering with risks. It will bring more damages under the condition of faults if measures are not proper. As the main measure which should be taken when the fault occurs, the safety controlling is the focus during launching which demands not only the correct analyzing and processing the launch vehicle's state, but also the fast processing the ground information, coverage analyzing for the measuring and controlling devices, the protecting regions analyzing of the range and fuzzy spatial decision making. And the research of the dissertation has significance and pratical utility.The dissertation utilizes artificial intelligent approaches,such as artificial neural network(ANN), clustering, fuzzy, particle swarm optimization and etc., to solve the problems of the safety controlling for space lunching, which are how to improve the precision of geographic data, how to arrange the measuring and controlling devices under the complex geographic condition, how to realize the real-time processing of spatial data and how to realize the spatial decision-making.For the geographic data is vital to the safety controlling system, the dissertation introduces an ANN correcting method for geographic data after analyzing the geographic data deformation and the deficiency of the polynomial correcting method. And based on the advantage of ANN's approximating function, the dissertation constructs the correcting model for the geographic data correction. The experiment results show that the ANN method performances far better than the polynomial fitting method does.In order to meet the realtime requirement of the geographic data processing for the safety controlling system, the dissertation introduces a method for the geographic data processing and accessing based on grid geographic data. By the gridding, the vector geographic data are transformed into grid data which are stored in a 2-D array. And with the relation of the geographic coordinates, grids coordinates and array subscript, the dissertation provides the transformation method for the vector data's spatial and attribute data into grid data. And both the algorithm analyzing and experiments show that the grid-based method is far better than the traditional method with vector data in accessing and processing the geographic data.Aimed at the site location of the measuring and controlling devices in complex geographic area such as mountain area, the dissertation presents a site selection method for the devices based on the modified maximum elevation angle. To utilize the former analyzing results, the method, along the helices whose center is the vehicle's projection point, analyzes vehicle's visibility on the helices at each moment during the flying period. And the visible area can be gotten by comparing the visibility at each moment during flying. So with the above visible area, the dissertation introduces the site selecting factors of roads, rivers, gradients and communication lines into the algorithm based on the particle swarm optimization to realize the conditional site selecting.The dissertation gives a flight range assessment method based on the spatial clustering analysis and fuzzy. With analyzing the influence of roads, rivers and the mountains to the distance between the entities, the dissertation introduces the factors that reflects the geography environment, and improves the spatial K-means clustering algorithm which is used to the analyzing the ground entities in the flight range. And the dissertation presents a method to divide the flight range into the groud protecting regions according to the clustering result. With the fuzzy approaches, the dissertation analyzes the importance of the above regions which can improve the online analyzing efficiency.At the end, the dissertation gives the spatial decision-making method, including the production rules with credible factors, the fuzzy representation of predicting fall-range and the uncertain spatial inference and analysis for safety controlling with fuzzy 9-intersection models.