Formation and Attitude Control System (FACS) lies at the heart of the formation flying software. It is designed to provide 3-axis inertial attitude and inter-s/c range and bearing control of each of the spacecraft in the TPF-I formation. Additionally, FACS provides the capability of initializing/reacquiring the formation through the acquisition of inter-s/c range/bearing knowledge, using the available on-board formation sensing capability. FACS ensures collision free operation of the formation throughout all nominal phases of the TPF-I mission.

Two high-level scenarios have been demonstrated in FAST. The first consists of two spacecraft functioning as a distributed interferometer. The second is a two-robot simulation of the Formation Control Testbed (FCT).

Two-spacecraft distributed interferometer

Formation and Attitude Control System (FACS)

The purpose of this simulation was to demonstrate nominal, ground-based operation of a distributed interferometer formation with high-level commands and autonomous formation reconfigurations with collision avoidance. The spacecraft are in deep space and have several flexible modes due to a large sunshield with a fundamental mode at 0.5 Hz. An additional requirement for this FACS design was that all thruster firings for both attitude and relative position control across all spacecraft occur in the same 4 s window every 60 s. This requirement allows nanometer-level interferometer control loops to reacquire a fringe and make a science observation between impulsive disturbances due to thrusters. Since this simulation was demonstrated in 2004, ion thrusters have been adopted for TPF-I's baseline design that have changed the operating assumptions, and the 4-s window is no longer a requirement for operation.

The spacecraft go through individual modes to become power positive such as detumble, Sun-search, and Sun-point, and then begin to acquire the formation. In this scenario, the two-spacecraft are assumed to be in a stack and separate via springs. Propagation using accelerometers is used to determine an approximate position of the other spacecraft. With this knowledge, the spacecraft point their formation sensors at one another. The autonomous formation flying (AFF) sensor was used in this simulation, which is RF-based with a GPS-like signal structure. The field-of-view of the sensors is 70 deg. After acquiring relative sensor lock and establishing inter-spacecraft communications, a coarse control loop is closed to maintain a constant distance and formation rotates to the center the Sun on the formation solar panels. This formation is held until a command is received.

Two commands were sent in this demonstration, both of which specify stellar targets to observe. The stop-and-stare observations consist of holding a constant relative position with the thruster quiescence requirement. The commands include the baseline to hold while observing, and the time allotted for retargeting. In response to a retarget command, the formation guidance algorithm first plans collision-avoidance constrained, energy-optimal relative spacecraft trajectories to achieve the desired baseline. This process is non-trivial, since the unconstrained, energy-optimal trajectories for the second retarget lead to a collision (which is here successfully avoided).

Two-robot FCT

Visualization of FAST demonstration of two-robot FCT.

The purpose of this demonstration is to validate the FACS used on the FCT robots before implementation on the actual robotic hardware, including flight-like inter-spacecraft communications and control cycle synchronization. The formation software used the Real-Time protocol, wireless links, time offset estimation via echo packets, and control cycle synchronization. Additionally, the sensors models use measured sensor noise values. For example, each fiber optic gyroscope on the robots was calibrated and measured values of rate random walk, angle random walk, and angle white noise were used in the gyroscope model. The control cycles were initially 0.5 s out of synchronization and the clock models had a relative drift of 0.1 ms/s. In addition to autonomous reconfiguration, synchronized rotations were demonstrated in which the formation rotation rotates as a virtual rigid body. Synchronized rotations are used for observations "on-the-fly" and can also be used to retarget a formation. In the latter case, the rotation axis is not along the formation boresight.

These two FAST simulations demonstrated formation software for autonomous formation flying with realistic inter-spacecraft communication and asynchronous clocks. In particular, formation algorithms for actuation-constrained formation control, autonomous collision-free reconfiguration, and synchronized rotation were demonstrated. This formation software has been integrated with the Formation Control Testbed robots for a flight-like, hardware demonstration of precision formation flying.