This project proposes to develop time-windowed, efficient smoothing software for incremental estimation of the relative poses and velocities between multiple, small (potentially differently…

This federal contract award to the Massachusetts Institute of Technology (MIT) is for the development of software and hardware to enable precise relative navigation and positioning of small spacecraft in formation. The project aims to mature a smoothing algorithm from Technology Readiness Level (TRL) 3 to TRL 6, allowing for more accurate estimation of the relative poses and velocities between multiple spacecraft using available sensor data. This includes incorporating a long-range depth sensor based on coded aperture imaging to supplement other sensors like star trackers and GPS. The two-year, $199,796.12 cooperative agreement is funded by the Shared Services Center, a civilian government agency. The project supports research into advanced space technologies that could enhance the capabilities of small satellite formations and fractionated satellite systems. This contract does not have a set-aside designation.

This project proposes to develop time-windowed, efficient smoothing software for incremental estimation of the relative poses and velocities between multiple, small (potentially differently instrumented), spacecraft as well as a small, long range depth sensor based on coded aperture imaging capable of sensing other spacecraft in the formation. The smoothing algorithm will obtain the maximum a posteriori estimate of the relative poses between the spacecraft through the use of all available sensor information. IT will be inherently portable between different satellite platforms possessing different sensor suites and computational capabilities, and adaptable in the case that one or more satellites become inoperable. IT will obtain a solution which approaches an exact solution, as opposed to one with baked-in linearization errors typical of filtering algorithms. Thus, the algorithms developed and demonstrated as part of this program will enhance the applicability of small spacecraft to multi-platform operations, like precisely aligned constellations and fractionated satellite systems. Both near-field and far-field applications of the algorithm will be considered. In the near-field, IT will be assumed that relative pose measurements between satellites contain full 6-degrees-of-freedom (6-DOF) pose information. In the far-field, IT will be assumed that relative range measurements contain bearing (and possibly range) information. In near-field or far-field, the satellites will optionally be able to incorporate star-tracker and/or global positioning system (GPS) data. Depth sensing using stereo imaging relies on a large baseline, and is therefore not considered for in-space application where satellite separation is greater than 10-20 meters. Coded aperture cameras have been studied for close range applications, but not for the long range distances required for on-orbit formation flight. This proposal will develop coded aperture depth cameras for integration with the smoothing algorithms to perform long-range relative navigation. The recently developed 'depth from coded aperture' imaging technique provides the ability to compute a depth map from a compact (monocular) camera. By using typical passive cameras whose circular aperture is replaced by a patterned, or coded, aperture with high frequency components, and setting the scene out of focus, the size of the imaged aperture code can be used to deduce the depth map of the scene. The proposed two-year project aims to mature the smoothing algorithm from TRL3 to TRL6. This will be accomplished through maturation of the smoothing algorithms that have already been developed in the laboratory. Previously collected data from spheresvertigo (synchronized position hold engage and reorient experimental satellites with the visual estimation for relative tracking and inspection of generic objects linux-based extension) will be used to verify and validate the expected offline performance of the algorithms for two and three satellite scenarios, bringing the technology to TRL4. The smoothing algorithm will be demonstrated in 6 degrees of freedom (DOF) using spheres-vertigo aboard the international space station (ISS). Using the stereo cameras to obtain a depth map will imitate the coded aperture depth sensing cameras. Two initial test sessions will test the online performance of the smoothing algorithms for two and three satellite scenarios. Two final test sessions with the completed algorithm will demonstrate a variety of operational scenarios in the relevant environment of full 6- DOF microgravity, bringing the technology to TRL6. The coded aperture cameras will be integrated with the ground version of the spheres satellites at MIT, and tested at relevant distances at JPL's formation control testbed, to demonstrate the performance of the algorithm at long distances. At the end of the program, we intend to release the smoothing algorithm as an open source library.