Line of sight track guidance method for precise satellite imagery

For the need of the large scale images both in military and in civilian areas, the demand for the earth observation satellites increases because of they can acquire images from large areas at one time and they can cover globally all geographic regions in a short time interval depending on the orbit of the satellite. At the same time, with the developing technology, the mirror of the electro-optical payload equipment enlarging and the pixel size on the detector gets small which make it possible to achieve images having higher quality and higher resolution. Nowadays, the number of the electro-optical earth observation satellites being have below meter ground sampling distance increases. It is known that Turkey, Spain, Canada, China, England, Argentina, Russia, Kazakhstan, Australia and many other countries have tends to develop imagery satellites with sub-meter ground sampling distance levels. As an example, Spain launched their first very-high resolution satellite, DEIMOS-2 in 2014 and Turkey is working on Göktürk-1 satellite which has a launch date of 2015. Institutions and organizations have a trend of developing imagery satellites which have 0.5 meters of ground sampling distance values. For this reason, 0.5 meters of ground sampling distance requirement is taken into account for the analysis in this study. However, even if the capabilities of the electro-optical payload are enough to detect high resolution and high quality images, to perform the payload's tasks in terms of resolution, quality and targeting performance, the satellite shall have sufficiently correct position and orientation profile. To acquire an image, a satellite having an electro-optical payload in push-broom type shall scan the image on ground with a fixed rate (the velocity of the line of sight track on ground shall be constant). To allow this activity, the satellite shall be at the correct position and orientation (and at the correct attitude rate) at the starting point of the acquisition activity. At the same time, the satellite should track the reference attitude profile (which can be uploaded from ground station via mission plan) with high accuracy to maintain the line of sight axis in a fixed rate on ground during the image acquisition activity. In this study, a method named "line of sight track guidance" is developed which can be used to obtain reference attitude profile during an acquisition activity for a satellite equipped with high resolution push-broom electro-optical payload is presented. In the presented study, an algorithm called as "line of sight track driver" which uses this method is descripted with all related functions. The scope of this thesis is to describe the developed method, to identify its inputs (and to develop all algorithms necessary to obtain required inputs), to analyze the outputs (errors) and lastly to make analysis particular to applications which uses the presented method. The output (attitude profile) of the algorithm which uses the presented method is in the format of polynomial coefficients. The reasons for the usage of the polynomial coefficients are that polynomials are computationally simple to solve in satellites on board computer in any step size (such as in control frequency), it reduces the parameter size which is sent from the ground station via tele command and also for the simplicity of computation of analytical derivative of the polynomial. The all necessary inputs for the analysis of acquisition activities and errors are performed with the algorithms which are also developed in the scope of this thesis. At the top level, these algorithms cover the orbit propagation, rendezvous algorithm (for the calculation of data take opportunities), ground elevation values obtained from digital elevation model, incrementing a ground location in the given direction (direct geodetic problem), inverse geodetic problem, visualization tools, test & error analysis functions and functions & user interfaces for the implementation of the application analysis of the presented method. The line of sight track driver algorithm which is developed in the scope of this thesis uses the geometrical properties / requirements of the image and the fundamentals of the push-broom image acquisition technique. The output of this algorithm is the reference attitude profile (in the polynomial coefficients format) of the satellite for a given image acquisition activity. However, the method itself also an error source, primarily depending on the order of the polynomial. As an example, if the order of the polynomial is one, then the rate of change of the attitude will be constant which means a linear attitude profile, whereas the line of sight track on ground will be nonlinear due to the geometry of the Earth surface – satellite orbit relation. If the order of the polynomial is zero (it means constant attitude), then it is not possible to acquire images with different heading angles, only images which are oriented as parallel to satellites ground track can be acquired (this situation limits the operational usage of the satellite system). Higher than the order one of the polynomial makes it possible to define complex attitude behavior of the satellite during the image acquisition activity. In this study, the relative errors which can be occurred in the reference attitude profile and therefore in the geometrical properties of the image due to the order of the polynomial is analyzed. In this analysis, parameters such as integration time of different push-broom electro-optical payloads are also included to obtain a general conclusion. The errors in the inputs of the algorithm are also error sources effecting on the properties of the image which is going to be acquired. As an example, if the elevation data belonging to geographical place to be imaged is known with an error and/or the orbital position of the satellite is known with an error, the produced attitude profile will have errors and therefore the properties of the image will have errors, especially for the off-nadir acquisition activities. In the scope of this thesis, the analysis regarding to these errors also are performed and the results are investigated for the satellites having different altitudes and different push-broom electro-optical payload equipment. Even if there is no error in the produced reference attitude profile (ideal reference attitude profile), there could be errors occurred while the satellite is tracking the reference profile (realizing the attitude profile). These errors are investigated according to satellites pointing accuracy and stability (stability is analyzed assuming that the satellite is a rigid body and therefore high frequency jitter effects are not examined). The relationship between the geometrical properties of the image and the pointing accuracy and stability requirements is established and investigated for the errors which may occur due to the pointing accuracy and stability criteria by using Monte Carlo analysis. For this purpose, instead of modeling a complete disturbance model for a specific satellite configuration, an algorithm named as Attitude Dynamic Distribution is developed which produces perturbed attitude angles with respect to pointing accuracy and stability requirements. Monte Carlo analysis carried out by adding the attitude perturbation to the reference attitude profile and running the analysis code hundreds of time to observe the error behavior during the acquisition activities. This approach is selected to achieve generalized solutions. Furthermore, a simplified model which can be used to determine satellites stability requirement regarding to payload and orbit specifications is obtained. By using line of sight track driver algorithm, reference attitude profile can be obtained for the acquisition activities within geographically all regions of the world (constraints may come only by the orbits inclination and attitude limits of the satellite). However, even if the satellite performs the reference attitude profile perfectly, this may not be enough to acquire high quality image from all geographical regions. For the electro-optical observation missions, to receive enough light from the region to be acquired by the payloads detector, the region shall be sufficiently illuminated by the sun to ensure quality requirements such as signal to noise ratio. In this study, the regions which can be acquired with plenty amount of light are investigated associated to sun zenith angle and regarding to orbits local time of ascending node parameter, seasons and payload capabilities. Also illustrations of imageable zones for different sun zenith angle criteria and for different local time of ascending node parameters are given. Finally, the longest strip map activities are investigated regarding to payload capabilities and orbit altitude. The analysis under this scope is made by assuming that there are no constraints exist on satellites memory, power and thermal conditions. The only constraint is on the ground sampling distance related to range between the satellite and the image portion to be acquired. Furthermore, a method to be used to solve the problem of varying ground sampling distance throughout the strip for the long strip map images, is presented. In this method, it is shown that by giving an amount of concession on the ground sampling distance, it is possible and feasible to obtain constant ground sampling distance throughout the strip map image. Further, by using this method, there will be no constraints on the length of the strip map image excluding illumination factors, memory constraints, power and thermal effects. It should be noted that, it is known that companies which designs state of art earth observation satellite systems use these types of methods, however the way / approach is not known. Therefore, the presented method in this thesis is individual. As a conclusion, the method of line of sight track guidance which is developed in the scope of this thesis is investigated by the analysis of the errors in the inputs, the errors caused by the method itself and the errors related to the applications which use the outputs of the method. In especial, simplified models are presented which give approximately same results with the detailed analysis results, almost for all error and application analysis. By using the simplified models, results are investigated for the earth observation satellites having different altitudes and payload capabilities.