Study On The Initial Orbit Determination Of Space Targets With Space-based Surveillance

The space-based surveillance is valued highly on civil and military applications, which would mainly use the SBV, space-based visible, currently and for a fairly long future period. Today, USA has its on-orbit space-based surveillance system(SBSS). However, China is weak on this research. So it is urgent to develop the key technologies including space-based orbit determination(OD) for our own system in the future.Under the above background, this dissertation studies the space-based initial orbit determination(SBIOD) by combining theoretical analysis and numerical verification, stresses on relative issues including methods, models, algorithms, etc, especially on improvements of tradition methods for the space-based purpose. At the end of the dissertation, the SBIOD of the cislunar transfer trajectory is researched, incidentally designs of the trajectory and the formation flying(FF) configuration.The main content and conclusions are drawn as follows:First, visibilities of space targets to satellite and constellation platforms and the data pretreatment are researched. Results show that the constellation platform can restrain adverse impacts of visibility constraints, the visibilities near the earth are worse than the ones of high targets, and smoothing rand error is a tradeoff among fitting polynomial degree, sampling rate and amount.Second, applications and improvements of the double r iteration method, the Gauss method and the Laplace method for space-based purposes are studied, including OD accuracy, algorithms and initial values. The impact of measure errors on OD is also analyzed and models under error are presented as a result.Conclusions of this part consist of four points as follows:1. The traditional methods are fit for low platforms to measure high targets and would bring on phenomena of multivalue and self-solution. The phenomena, which derive from the absence of range measurement, couldn't be eliminated and only could be decreased by improvements of iteration algorithms and initial values.2. The space-based doubleρiteration method together with iteration algorithms and initial values brought forward in the dissertation could basically solve the above phenomena with sparse data. The numerical result shows an over 90% success rate.3. The model error is negligible and measurement errors play an important role.4. The systemic measurement error model is vital to error analysis and elimination when the data pretreatment is unconsidered or failed. Under the sparse data, the rand errors could result in the self-solution.Third, using the constellation to determine target orbits is discussed. Based on the solution process, the self-solution and the model singularity are analyzed.The following conclusions could be achieved :1. The constellation has a preferable visible geometry and can eliminate the self-solution, whose synchronized model is excellent. Also has the constellation some merits, such as effectively weakening or eliminating impacts of rand errors, lowering or partly eliminating influences of systemic measurement errors, and greatly shortening required measurement arcs. However, it still needs a fair measurement arc for high ODs.2. The effect of sampling rate on required measurement arcs is limited. The visibility influences the single platform more than the constellation. Increasing arc length and sampling rate is favorable to overcome visibility influences and to gain a high success rate.3. The self-solution can be reduced at a certain extent by improving measure conditions and by avoiding some particular times. The model singularity is intrinsic and cannot be eliminated by outside resorts. So it should avoid to determinate orbits at that time.Finally, as the further verification and expanded applications, the SBIOD of the cislunar transfer trajectory is investigated. The platforms consist of a single satellite, a constellation and especially a FF. The FF configuration design is also researched at the last chapter.According to the studies of the last chapter, model errors make the cislunar OD error far more than common space targets. The long measurement arc is not always preferable for constellation platforms. The FF synchronized model is obviously more favorable than the unsynchronized one which is nearly the same as the single satellite platform on impacts of measurement arcs and visibilities. When the FF is small, the size of formation has tiny impact on the results.The dissertation could support the foundation of our SBSS theoretically and technically and could be used as a reference of space-based orbit measurement and determination in our future lunar and deep space exploration.