An infrared earth horizon sensor for a LEO satellite
Thesis (MScEng (Electrical and Electronic Engineering))--University of Stellenbosch, 2008.
Horizon sensing is an effective way to determine the pitch and roll of a LEO satellite and Earth horizon sensors that operate in the visible range of the electromagnetic spectrum are commonly used. These sensors have the disadvantage that they cannot operate when the satellite is in eclipse. Earth horizon sensors that operate in the infrared spectral range are a solution to take attitude measurements when the satellite is in eclipse. Until recently infrared detectors could only operate at very low temperatures and needed to be cryogenically cooled. The result was that their power consumption and physical characteristics (like dimensions and mass) were such that they were not suitable for use in small and medium LEO satellites. As a result of technology expansion in the field of infrared imagers the past few years, infrared imagers were developed which do not require cooling. The scope of this project was to develop and implement an Earth horizon sensor by using a low-cost, uncooled infrared imager. The performance and physical characteristics of various imager were evaluated and it was decided to select a low resolution thermopile imager mainly as a result of the cost limitations of the project. Software algorithms were then evaluated and selected for horizon detection and attitude determination. The Earth horizon sensor that was developed did not comply with the accuracy requirement (3s < 0.1o) that was set for the project because of the low resolution of the sensor. Methods to improve the accuracy were investigated and finally a sub-pixel edge estimation algorithm was developed and implemented which resulted in an improvement of 69% in the pitch accuracy and 49% in roll accuracy. With the sub-pixel edge estimation algorithm implemented the horizon sensor almost met the accuracy requirements (s < 0.0811o for pitch and s < 0.2944o for roll). This project confirms that, with further improvement to the design and test facilities, developing a low-cost, uncooled infrared Earth horizon sensor that meets the accuracy requirements is feasible.