Varun Gottumukkala, Delft University of Technology
Ritesh Balayan, GalaxEye Space
Rahul Singhal, GalaxEye Space
Kishan Thakkar, GalaxEye Space
Varun Gottumukkala, Mr, Delft University of Technology
The rapid growth in the number of Earth-observation satellite constellations presents a need to accurately determine a constellation’s imaging performance, which is a driving factor for the quality of data-products derived from the captured images. A primary objective for Earth-observation constellations is to maximize the coverage and revisit performance. Assessing these capabilities in a realistic manner for a defined satellite constellation requires the knowledge of several system, payload, and simulation parameters. The system parameters include the slewing range and rate, and power constraints; payload parameters include operational look-angles, field of view, and imaging duty cycles and durations; simulation parameters include the resolution of the defined area of interest, the simulation time step, propagation method, and accuracy of environment models, amongst others. Without these parameters, it is only possible to develop a preliminary understanding of a constellation’s imaging capability without detailed knowledge of how it will truly perform under operational constraints. The goal of this simulator is to assist mission planners in designing constellations for optimal mapping of their target regions and help mission operators to get the most out of their existing constellation by tasking their assets in an effective manner.
The conducted work involves the development of a simulator that allows a user to choose an area of interest and simulation duration and then propagates a constellation that is completely defined by the number of satellites and each of their orbital parameters. The area of interest is segmented into grids of a user-defined resolution, such that the real swath of the sensor is at least the size of one grid point. The simulator saves access data using the sensor’s virtual swath, which includes the entire operational range that the real swath can slew through. This eliminates the need to define an attitude slewing logic; instead, given knowledge about the slewing range and rate, constraints can be added to the final algorithm. The tool is therefore more straightforward as it only uses a static virtual swath to save access data instead of a computationally expensive slewing logic, which would require optimal solutions to be found at each time-step. This work distinguishes between the terms “coverage” and “mapping”. Coverage refers to areas which a constellation is capable of imaging if tasked accordingly. Mapping refers to the areas that can actually be imaged by the constellation. A mapping metric will provide insights into the areas which can be imaged under operational constraints.. This metric can be used within an optimization loop for constellation design purposes, where the objective is to perturb a constellation configuration to achieve a maximum mapping capability.
Once the constellation is propagated, the simulator saves data of every instant at which a sensor’s virtual swath accesses a grid-point in the area of interest. By analyzing this time-indexed data of accesses between the constellation and the area of interest, a methodology was developed to compute a maximum mapping capability. This was performed by formulating the problem as a bipartite graph and employing the Kuhn-Munkres algorithm to find an optimal mapping solution. The algorithm can be augmented by adding operational constraints to account for the aforementioned system and payload constraints. The formulation of the problem as a bipartite graph allows for the straightforward computational formulation as a two-dimensional matrix. Running the algorithm directly on this matrix provides the maximum mapping capability, while an augmented algorithm provides a more realistic mapping capability based on the constellation’s constraints.