Optical Deep Space Navigation and Sensor Alignment with an Adaptive Cascade Filter for Small Satellites

As deep space missions for CubeSats are now possible, new autonomous technologies are required to navigate the Solar System. The low cost and low power consumption of off-the-shelf cameras make optical navigation an attractive technology for CubeSats. Therefore, this work proposes an adaptive cascade filter for the attitude and orbit determination of CubeSats. The attitude determination system (ADS) processes star tracker attitude measurements and gyroscopic rate measurements to estimate the spacecraft inertial attitude, gyroscope bias, gyroscope scale factor, and camera misalignments. The gyroscope Angle Random Walk (ARW) is additionally estimated through mutating Multiple Model Adaptive Estimation (MMAE). The ADS state estimate is then input to the Orbit Determination System (ODS) in a cascade filter. The ODS uses the attitude information and ephemeris data to filter the spacecraft position from bearing measurements to planets. Monte Carlo simulations show that position accuracy is attainable on the order of 10000 km with low-grade sensors, or 1000 km with mid-grade sensors. The MMAE scheme improves the ADS performance when the gyroscope ARW is unreliable.