Substantial altimetry datasets collected by different satellites have only become available during the past five years, but the future will bring a variety of new altimetry missions, both parallel and consecutive in time. The characteristics of each produced dataset vary with the different orbital heights and inclinations of the spacecraft, as well as with the technical properties of the radar instrument. An integral analysis of datasets with different properties offers advantages both in terms of data quantity and data quality. This thesis is concerned with the development of the means for such integral analysis, in particular for dynamic solutions in which precise orbits for the satellites are computed simultaneously. The first half of the thesis discusses the theory and numerical implementation of dynamic multi-satellite altimetry analysis. The most important aspect of this analysis is the application of dual satellite altimetry crossover points as a bi-directional tracking data type in simultaneous orbit solutions. The central problem is that the spatial and temporal distributions of the crossovers are in conflict with the time-organised nature of traditional solution methods. Their application to the adjustment of the orbits of both satellites involved in a dual crossover therefore requires several fundamental changes of the classical least-squares prediction/correction methods. The second part of the thesis applies the developed numerical techniques to the problems of precise orbit computation and gravity field adjustment, using the altimetry datasets of ERS-1 and TOPEX/Poseidon. Although the two datasets can be considered less compatible that those of planned future satellite missions, the obtained results adequately illustrate the merits of a simultaneous solution technique. In particular, the geographically correlated orbit error is partially observable from a dataset consisting of crossover differences between two sufficiently different altimetry datasets, while being unobservable from the analysis of altimetry data of both satellites individually. This error signal, which has a substantial gravity-induced component, can be employed advantageously in simultaneous solutions for the two satellites in which also the harmonic coefficients of the gravity field model are estimated.

The objectives of the subject-matter of this thesis were to appraise some methods of solving non-singular optimal control problems by their degree of success in tackling four chosen problems and then to try the most promising methods on chosen singular problems. In Part I of this thesis, the chosen problems are attempted by quasilinearisation, two versions of shooting, Miels's method, differential dynamic programming and two versions of parameterisation . Conclusions on the various methods are given. NOC shooting, developed by the Numerical Optimisation Centre of The Hatfield Polytechnic, and constrained optimisation were found to be very useful for non-singular problems. In Part 11, the properties and calculation of possible singular controls are investigated, then the two chosen methods used. It was found that NOC shooting was again very useful, provided the solution structure is known and that constrained parameterisation was invaluable for determining the solution structure and when shooting is impossible. Contributions to knowledge as as follows. In Part I, the relative merits of various methods are displayed, additions are made to the theory of parameterisation, shooting and quasilinearisation, the best solutions known of the chosen problems are produced and choices of optimisation parameters for one chosen problem, the satellite problem, are compared. The satellite problem has dependent state variables and the Maximum Principle is extended in Appendix III to cover this case . In Part II, a thorough survey of the properties of singular controls is given, the calculation of possible singular controls clarified and extended, the utility of the two chosen methods is displayed, the best solutions known of the Goddard problem obtained with improved understanding of transitions in soluti on structures , Cl problem studied with control dependent on the costate variables and singular solution structures found.

In this research study, several new algorithms are developed to achieve spacecraft attitude determination from carrier phase information of GPS (Global Positioning System) signals. The first focus is on resolving integer ambiguity in carrier phase difference measurements. A newly developed algorithm based on Gram-Schmidt Orthonormalisation (GSO) is proposed for medium length baseline observations. Using this newly developed attitude algorithm from vector observations, an instantaneous estimated attitude solution is obtained, which we call 'coarse attitude', from only four phase measurements collected from only two baseline observations. Then a 'fine' attitude solution from all phase measurements is estimated, using a sophisticated Kalman filtering estimator, once integer ambiguity has been resolved. The second focus is on estimating the relative phase offset error (line bias) in carrier phase difference measurements. A newly developed block bias search is proposed which finds an initially plausible solution of line bias for each individual baseline. The line bias from all phase measurements collected from each individual baseline is then re-estimated using a developed recursive least squares (RLS) estimator. A newly developed parallel architecture GPS receiver is being flown on the UoSat-12 minisatellite, with the capability for simultaneous measurements from 24 channels for attitude sensing. The final goal of this research study was to apply the developed algorithms to real GPS data, and a number of data files of phase differences of GPS signals logged on UoSat-12 were tested. Independent ADCS (Attitude Determination and Control System) data was used for the reference attitude determination. The results show that an instantaneous attitude error less than 4 degrees is achieved during coarse attitude acquisition, relative to the reference ADCS system. When all measurements are processed during fine attitude tracking, the error in attitude estimation is reduced to one degree error (1 sigma RMS), without any error mitigation for multipath, relative to the reference ADCS system.

In global navigation satellite systems (GNSS) a fundamental operational component is the calculation of the orbits of the system spacecraft. This requires understanding and modelling the forces that act on the spacecraft. Solar radiation pressure (SRP) is the force caused by the impact of solar photons on the spacecraft surface. For GNSS spacecraft this is a significant force. If SRP is not included in the force model, then the calculated position of the spacecraft can be in error by between one and two hundred metres after one 12-hour orbit. SRP can be modelled using either analytical or empirical methods, or by some combination of the two. Historically, analytical SRP modelling has been somewhat neglected and high precision orbit estimation has relied upon empirical methods to account for SRP. Even so, most of these empirical methods start the estimation process with an a priori analytical model. The success of this empirical approach relies upon having many observations of the range between the system spacecraft and ground-based tracking stations, and works well within the context of the International Global Positioning System Service (IGS) network, which provides the necessary data volume. However, empirical methods do not work as well in operational GNSS, as these typically have a relatively small number of tracking stations. Moreover, empirical methods cannot be applied at the GNSS design stage, where knowledge of the system dynamics plays a key role. Existing methods for calculating analytical SRP models can only be used with relatively simple spacecraft structures, and lack flexibility as tools for analysis. In this study a new method is developed for calculating analytical SRP models that can cope with a high level of complexity in the spacecraft structure. The method is based upon simulating the solar photon flux with a pixel array. Using the method, models are calculated and tested for the Russian GLONASS IIv spacecraft. This particular spacecraft was used as the testbed because, at the time the study was being conducted, an international scientific campaign - called IGEX-98, the International GLONASS Experiment - was being carried out to analyse the Russian system. Developing force models for the spacecraft was one of the campaign goals, and the IGEX-98 steering committee accepted a proposal to use SRP models for GLONASS from this study. A detailed description is given of all the mathematics and physics that was used to develop the modelling technique. The method by which the models can be calculated and applied in practical orbit determination is also provided. In order to test the performance of the SRP models computed for the GLONASS spacecraft using the new method, comparisons were made between two kinds of trajectory. The first kind was calculated by numerical integration of the spacecraft's second order differential equation of motion, where this force model included the custom SRP models developed in the thesis. The second kind of trajectory, which is used as a 'truth' model in the study, was a precise orbit computed by the University of Berne using IGS range data and an empirical SRP model. Such precise orbits are the best estimates available of the true trajectories, as they are derived from the simultaneous estimation of multiple receiver tracking station network positions and spacecraft force model parameters. The repeatability of the Berne orbit is circa 0.75m. The RMS differences between the two trajectories over one twelve-hour orbit (an arc length of circa 160,000km) were 0.7m in height, 1.3m across track and 3.5m along track. This shows that the trajectory derived from the force model alone is very close to the precise orbit. The time-varying pattern of the differences between the two trajectories strongly indicates that the residual mismodelling of the forces acting on the spacecraft is due to thermal re-radiation effects. Further tests of the method were also conducted using satellite laser ranging (SLR) data to calculate arc lengths of 400 days, again using SRP models from the study. This enabled the calculation of model scale factors and additional empirical terms. The average SRP model scale factor was circa 1.01, which implies that the average error in the a priori SRP models calculated for the GLONASS IIv spacecraft is at the 1% level. This is consistent with an error budget based on an assessment of the accuracy of the source data supplied by the Russian authorities. The magnitude and parameterisation of the SLR empirical terms again strongly suggest that most of the remaining mis-modelling is caused by thermal effects. An analysis is given of the effect on the a priori SRP model of unmodelled, SRP-related forces acting along the spacecraft Y-axis. This is the so-called Y-bias. It is shown that whilst Y-bias effects are important in orbit determination, they are less critical in the process of calculating the a priori SRP model. A discussion is provided on how the new method can be adapted to improve the modelling and understanding of thermal re-radiation and Y-bias effects, and also on what benefits might accrue from such studies. The new method is an improvement over existing techniques as it enables the calculation of high precision SRP models that can be applied in the design, operation and scientific analysis of GNSS. A UK patent application has been made in respect of the new method.

In this thesis, we investigate the orbit control strategies of small satellites in Low Earth Orbits (LEO) where the disturbance effects are significant, in particular the nonspherical Earth and atmospheric drag effects. These orbits are not suitable to be controlled by using traditional ground-based control strategies which generally require high-thrust propulsion systems and other expensive resources, both onboard and in the ground segment. In order to react to those disturbances spontaneously and keep a small satellite at a pre-defined station using its limited resources, autonomous orbit control technology needs to be enabled. With the current advances in navigation and propulsion technology, as well as onboard computation systems, the only key issue that needs further investigations for practical implementation of an autonomous orbit operation system is the control algorithm. The orbit control strategies we investigate here are treated separately for each of the orbital control phases, i.e. orbit deployment and acquisition, orbit transfer and orbit maintenance. We present various forms of the solutions of the epicycle motion which allow us to treat each control problem according to the control requirements, nature of perturbations, control time scales and available resources. Although applied in different manners, the optimal low-thrust control scheme is a common aim for all control problems investigated here, as we mainly focus upon applications for low cost small satellites in LEO. The verifications of the strategies proposed in this thesis have been demonstrated not only via computer simulations, but also successfully demonstrated on in-orbit small satellite platforms thanks to an active small satellite programme at Surrey Space Centre. The success of this study is hoped to provide a valuable basis for satellite orbit operations which will involve larger number of satellites with more complex configurations in the future.

In this research study, several new in-orbit algorithms are proposed to improve the performance of Attitude Determination and Control System (ADCS) by estimating the inertia matrix and calibrating the cold gas thruster system of the UoSAT-12 spacecraft. Computer-based simulation models will be constructed using MATLAB and SIMULINK in order to evaluate the expected performance. The first focus is on the identification of the satellite inertia matrix. A new algorithm based on a Recursive Least Square (RLS) estimation technique is proposed for in-orbit use to estimate the inertia matrix (moments and products of inertia parameters) of a satellite. To facilitate this, one attitude axis is disturbed using a reaction wheel whilst the other two axes are controlled to keep their respective angular" rates small. Within a fraction of an orbit three components of the inertia matrix can be accurately determined. This procedure is then repeated for the other two axes to obtain all nine elements of the inertia matrix. The procedure is designed to prevent the build up of momentum in the reaction wheels, whilst keeping the attitude disturbance to the satellite within acceptable limits. It can also overcome potential errors introduced by unmodeled external disturbance torques and attitude sensor noise. The second focus is on a new algorithm for in-orbit use to calibrate thruster coefficients for thrust level and alignment, using three reaction wheel actuators. These algorithms will ensure robustness against modeling errors. The algorithms assume no prior knowledge of the thruster parameters and only an initial guess of the inertia matrix. It is proposed that this calibration can be used during normal mission conditions when the satellite is stabilised. The final goal of this research study was to apply the proposed algorithms in real-time. Firstly, the thruster calibration algorithm was tested on an air-bearing table. Finally, both thruster calibration and moment of inertia algorithms were tested using data generated by UoSAT-12 while in orbit. The practical estimation results proved the feasibility of proposed algorithms.

No description available.

The research project describes the key mission and systems engineering trade-offs involved in the end-to-end design of an orbiting mission to the Moon, using a "Smaller, Faster, Cheaper" mission approach. This approach is extended to enable the design of a new payload - within the management, cost, schedule, and physical constraints - of the low cost lunar orbiter mission. The payload is designed to image the Moon's permanently dark regions that are believed to contain water ice. To determine the best cost reduction management and engineering approach, the principles for reducing space mission cost are examined and planetary missions are assessed in terms of cost and risk drivers. 'Interplanetary' trajectories and attaining orbit around another body are shown to be the major risk areas encountered by traditional planetary missions. In addition to this, programme management is highlighted as an emerging high risk area for smaller, faster cheaper planetary missions. The preliminary mission design, covering lunar transfer, spacecraft and ground station is described. A 400 kg, three-axis stabilised, lunar orbiter, capable of delivering 20 kg of payload into a low lunar polar orbit is designed. The ground segment comprises one (possibly two) low cost ground stations, linked via the Internet. Images, raw data and telemetry can also be accessed via the Internet. The design-to-launch timeframe spans three years and the total mission cost target of $20 Million is met. The spacecraft is compatible with a range of existing lunar payloads, but the prime mission requirement will be to return images of the Moon's permanently dark craters for the first time. In order to design a low cost payload for imaging the Moon's permanently dark regions, the areas likely to contain the water ice are first characterised. The best and worst case lighting conditions for imaging are then calculated for these regions. The amount of light reaching a crater floor is a function of the crater depth-diameter ratio, solar irradiance incidence angle and local topography. The limiting case is shown to be imaging under starlight illumination only, which is modelled and estimated between 5 to 10µW/m2 over the 350 to 900 nm spectral band. These ultra-low light level conditions have led to identification and evaluation of several solutions in terms of both signal-tonoise ratio performance and development within the constraints of the smaller, faster, cheaper programme. This is achieved using a charge coupled device (CCD) camera employing a commercial sensor and optics. Large format Charge Injection Devices and Complimentary Metal Oxide Semiconductors (Active Pixel Devices) were identified as promising emerging technologies. The baseline low light level imager solution is a CCD array operated in Time Delay Integration mode in order to provide optical images from areas within permanent shadow. An intensified CCD offers a back up solution. The research demonstrates that a low cost lunar mission is technically feasible and additionally, that it is possible to meet a specific (if modest) application target through `smaller, faster, cheaper' payload design. It provides an approach that meets key challenges of planetary exploration at very low cost that can potentially be applied to other near Earth targets.

No description available.

The NAVSTAR Global Positioning System (GPS) has been under development by the US Department of Defense since 1973. Although GPS was developed for precise instantaneous position and velocity determination, it can be used for high precision relative positioning, with numerous applications for both surveyors and geodesists. The high resolution of the satellite's carrier phase has enabled relative positioning accuracies of the order of one part per million to be routinely obtained, from only one or two hours of data. These accuracies are obtained using the broadcast ephemeris, which is the orbit data that is broadcast in the satellite's radio transmission. However, the broadcast ephemeris is estimated to be in error by up to twenty five metres and this error is one of the principle limitations for precise relative positioning with GPS. An alternative to the broadcast ephemeris, is to determine the satellite orbits using the carrier phase measurements, obtained from a network of GPS tracking stations. This thesis describes the algorithms and processing techniques used for the determination of GPS satellite orbits using double differenced carrier phase measurements. The data from three different GPS campaigns have been analysed, which demonstrate a GPS orbital accuracy of between two and four metres, giving baseline accuracies of the order of one or two parts in ten million.