Optical Sensor Uncertainties and Variable Repositioning Times in the Single and Multi-Sensor Tasking Problem
thesisposted on 14.12.2020, 22:34 by Michael James Rose
As the number of Resident Space Objects around Earth continues to increase, the need for an optimal sensor tasking strategy, specifically with Ground-Based Optical sensors, continues to be of great importance. This thesis focuses on the single and multi-sensor tasking problem with realistic optical sensor modeling for the observation of objects in the Geosynchronous Earth Orbit regime. In this work, sensor tasking refers to assigning the specific?c observation times and viewing directions of a single or multi sensor framework to either survey for or track new or existing objects. For this work specifically, the sensor tasking problem will seek to maximize the total number of Geosynchronous Earth Orbiting objects to be observed from a catalog of existing objects with a single and multi optical sensor tasking framework. This research focuses on the physical assumptions and limitations on an optical sensor, and how these assumptions affect the single and multi sensor tasking scenario. First, the concept of the probability of detection of a resident space object is calculated based on the viewing geometry of the resident space object. Then, this probability of detection is compared to the system that avoids the computational process by implementing a classical heuristic minimum elevation constraint to an electro-optical charged coupled optical sensor. It is shown that in the single and multi-sensor tasking scenario if the probability of detection is not considered in the sensor tasking framework, then a rigid elevation constraint of around 25o-35o is recommended for tasking Geosynchronous objects. Secondly, the topic of complete geo-coverage within a single night is explored. A sensor network proposed by Ackermann et al. (2018) is studied with and without the probability of detection considerations, and with and without uncertainties in the resident space objects' states. (then what you have). For the multi-sensor system, it is shown that with the assumed covariance model for this work, the framework developed by Ackermann et al. (2018) does not meet the design requirements for the cataloged Geosynchronous objects from March 19th, 2019. Finally, the concept of a variable repositioning time for the slewing of the ground-based sensors is introduced and compared to a constant repositioning time model. A model for the variable repositioning time is derived from data retrieved from the Purdue Optical Ground Station. This model is applied to a single sensor scenario. Optimizers are developed using the two repositioning time functions derived in this work. It is shown that the constant repositioning models that are greater than the maximum repositioning time produce results close to the variable repositioning solution. When the optimizers are tested, it is shown that there is a small increase in performance only when the maximum repositioning time is significant.