|
|
Formation Control
This task develops the formation control architectures,
algorithms and dynamic models needed to plan optimal formation
trajectories for next-generation science missions and to
ensure that these trajectories are precisely tracked while
avoiding collisions, minimizing and balancing fuel consumption,
and minimizing sensing and communication requirements.
Task Objective
The objective of this WA is to develop and demonstrate formation
architectures, integrated guidance, control, and estimation
algorithms, and formation dynamic models for precision, autonomously-supervised,
small-to-moderately sized formations in both Earth-orbit
and deep space.
Statement of Work
- Develop 3 and 6 DOF formation guidance algorithms
- Develop scaleable formation estimation algorithms
- Develop formation flying control algorithms
- Develop modeling architectures of formation dynamics
Task Description
Develop and demonstrate formation architectures, integrated
guidance, control, and estimation algorithms, and formation
dynamic models for precision, autonomously-supervised, small-to-moderately
sized formations in both Earth-orbit and deep space
Formation flying spacecraft refers to a set of spatially
distributed spacecraft whose dynamic states are coupled through
a common control law. This coupling enables the set of spacecraft
to interact and function collaboratively. See Figure 1. Several
of NASA’s future Earth and Space science missions involve
formation flying of multiple coordinated spacecraft.
For Earth science applications, formation flying (FF) spacecraft
can provide distributed sensing for gravitational field mapping,
contemporaneous spatial sampling of atmospheric data, co-observations
(i.e., near-simultaneous observations of the same science
target by instruments on multiple platforms), and synthetic
radio-frequency and radar apertures. In particular, it will
become possible to expand our science capabilities by deploying large
numbers of low cost, Earth-orbiting miniaturized spacecraft.
An added advantage of this paradigm is that new formation
members can be introduced expand or upgrade the formation,
or to replace a failed member.
|
Figure
1. Spacecraft Formations in Deep Space and Earth Orbit |
Similarly, several space science missions include either
distributed instruments, large phased arrays of lightweight
reflectors and antennas, or variable baseline space interferometers.
For example, synthetic apertures composed of multiple collector
and combiner spacecraft will image the event horizon of a
black hole, study stellar evolution, and search for Earth-like
planets circling other stars and probe their atmospheres
for indications of life. Figure 2 shows the proposed black
hole-imaging formation MAXIM. Included in Figure 2 is a possible
predicted image of the black hole suspected to be at the
center of Galaxy M87.
|
Figure
2. Micro-Arcsecond X-Ray Imaging Mission (MAXIM) Formation
Concept and Example Science Return1 |
A new set of critical requirements is emerging out of the
distributed spacecraft mission architecture that goes beyond
traditional single spacecraft attitude guidance and control.
By its very nature distributed spacecraft missions (especially
the deep space science missions) require:
- Collision free operation, even under fault scenarios
- Robust lost-in-space initialization, i.e., the ability
to establish relative sensor measurements and inter-spacecraft
communication after initial formation deployment or after
reset/recovery events
- Optimal path planning in both attitude and translation
under multiple constraints (e.g. solar, thermal, glint,
hardware capability)
- Consumable resource balancing across all spacecraft
- Precision formation control coordinated across multiple
spacecraft, i.e., relative positions are controlled to
the centimeter level or better, and attitudes are controlled
to the arc-minute level or better
- Observation-on-the-fly capability to enable continuous
science measurements during formation maneuvers, i.e.,
synchronized actuation across a formation to ensure self-generated
disturbances do not corrupt science measurements
- Autonomous operation: the formation can be commanded
as a single unit, with lower level, individual spacecraft
actions derived from higher-level commands
These new, distributed spacecraft requirements must be met
under traditional single spacecraft-level constraints and
new, formation-level constraints. For example, optimal
formation guidance algorithms must still account for individual
spacecraft pointing constraints (e.g. star tracker field-of-view
must exclude the Sun), but now must also account for spacecraft-to-spacecraft
glint and thermal constraints (i.e., the hot sun-shield of
one formation member cannot thermally irradiate the cryogenic
payload of another). In these examples, spacecraft attitudes
and relative positions are coupled: formation guidance is
naturally a six degree-of-freedom problem. This rotational/translational
coupling is a common aspect of formation guidance and estimation
problems.
The requirements and constraints which apply to a given
formation depend upon the operational phase of that formation.
A formation’s operations can be divided into three
high-level phases: 1) formation initialization, 2) science
observation, and 3) formation reconfiguration. Formation
initialization is the process of using on-board hardware
to autonomously establish inter-spacecraft communications
and measure relative spacecraft positions for control. The
latter part, measuring relative spacecraft positions, is
particularly important for deep space formations, which are
beyond the reach of the Global Positioning System. This relative
position measurement phase is illustrated in Figure 3.
|
Figure 3. Initialization
(Relative Position Measurement): on-board, limited field-of-view
relative sensors must simultaneously “lock” to obtain
a measurement. While spacecraft 3 is in the field-of-view
of spacecraft 1, the reverse is not true, and so there
is no measurement (grey cone). However, spacecraft 1
and 2 see each other simultaneously, and so there is
a measurement (green cones). |
As the name suggests, the science observation phase consists
of formation maneuvers that are necessary to carry out a
formation’s primary science objective. For example, a tetrahedron
formation can be used to measure the gradients of the Earth’ magnetic
field or a formation can be rotated as a virtual rigid body,
which requires synchronized relative positions and attitudes.
See Figure 4. Another example, shown in Figure 5, is planning
a “u,v-Coverage” for a synthetic aperture image.
Figure 4. Formation Rotations:
rotating the formation as a virtual rigid body |
|
|
Figure 5. u,v-Coverage:
a path through a set of relative spacecraft positions
(u,v-points) that leads to the desired synthetic aperture
performance |
Formation reconfiguration is the least structured and most
complex operational phase. It consists of formation maneuvers
that are needed to move a formation from its current configuration
to the configuration needed for the next science observation.
For example, a synthetic aperture formation can be rotated
as a virtual rigid body to observe a new star, or an Earth-observing
sparse antenna array can reconfigure its constituent spacecraft
to form two sub-apertures with different wavelengths and
baselines. See Figure 6.
This task develops integrated precision formation estimation,
guidance and control architectures and algorithms for each
of these three operational phases subject to the requirements
and constraints discussed previously. Formation dynamic models
that are the basis for analytic developments and precision
performance validation are also developed. Note that integrated
algorithms are needed since, similar to the rotational/translational
coupling mentioned previously, estimation, guidance and control
functions can be coupled. For example, a formation control
algorithm’s ability to tolerate delay affects the latency
requirements for communicated measurements in a formation
estimator.
|
Figure
7. Leader/Follower Formation Control Architecture |
Due to the distributed nature of a formation, FF architectures
are being developed to support flexible, scalable, and robust
sensing, communication and control capabilities (e.g. communication
latencies, intermittent loss of lock of relative sensors,
and asynchronous spacecraft). In terms of robustness, analytical
redundancy within the control and estimation architectures
is necessary to retain functionality even in the face of
hardware failures. Such analytical redundancy can be utilized
to, for example, autonomously implement translation maneuvers
under a limited set of thruster failures or maintain formation
control in the presence of intermittent loss of relative
sensing during formation reconfigurations.
The precision formation architectures and algorithm developed
in this task will be (and have been) demonstrated in appropriate
test environments. A selection of the developed algorithms
will be (and have been) demonstrated in a distributed, real-time
simulation environment. Select algorithms will also be demonstrated
in the 6 degree of freedom, robotic Formation Control Testbed
(FCT). The FCT, which is nearing completion, uses flight-like
hardware and will allow validation of algorithms to the centimeter/arcminute
level required by precision formations.
1 Images courtesy of Keith Gendreau, Webster Cash,
Ann Shipley, and Nick White at http://maxim.gsfc.nasa.gov/documents/MAXIM-Documents.html and
NASA/STScI at http://hubblesite.org/newscenter/newsdesk/archive/releases/1994/23/
|