Accuracy Evaluation Of Trajectory Measurement System And Analysis For Trajectory Reconstruction

This dissertation researches some questions about accuracy evaluation of trajectory measurement system and trajectory reconstruction.Firstly, the mathematical method, wavelet transform of random error processing and Kalman filtering algorithm of system error estimation, is designated; The process of development of wavelet transform and its present research is recommended; And, the definition and development situation of real-time data processing and post-flight data processing of trajectory measurement data is summarized briefly.Statistics characteristics of random error are analyzed by wavelet transform method. Basic physical implication of wavelet transform and its time-frequency characteristic is studied, and the theorem of denoising with wavelet transform is analyzed; Multiresolution analysis of wavelet transform is introduced, random error is processed with Mallat algorithms, and thresholds is discussed; Modified MAD algorithms is introduced which is used to calculate variance based on coefficient.System error is estimated with Kalman filtering algorithm in real-time data processing and post-flight data processing. Dynamic models and measurement models of real-time data processing and post-flight data processing are given respectively. Based on Kalman filtering, two improved Kalman filtering algorithms are obtained, which correspond to different demand of real-time data processing and post-flight data processing. It also has comparison between Kalman filtering algorithm and Least Square method in post-flight data processing.System error models are analyzed and diagnosed. Some basic system error models are introduced, and theorem and method of diagnoses of error models is researched. How to diagnose system error models and how to estimate parameter of dynamic system error is discussed through analyzing practical example. The process of data processing of trajectory reconstruction is studied using the knowledge above-mentioned. Firstly, random error is processed, then system error is compensated, and trajectory is measured with several radars switched, finally a new integrated trajectory is accomplished after data processing. Additionally, a practical example will emulate the process.