1. Introduction
Hayabusa spacecraft was launched targeting on Itokawa, a near Earth asteroid in May 2003, arrived to
Itokawa in September 2005, and left for the Earth in April 2007.
Hayabusa has been operated by several ways different from the ones, supposed to be used at the beginning.
In July 2005, the control mode with ion-engines was switched to Dual Reaction Wheel Mode because of the
failure of one of three reaction wheels. In September, Hayabusa arrived to Itokawa. Three weeks after the
science missions started, the attitude control mode was switched to RCS mode due to the failure of the
second RW. The rehearsal for touch-down had been repeated three times and then, repeating taking-on/off
two times has been successfully done by RCS. Furthermore, the attitude control based on bias-momentum
stabilization with the third RW and no thrust of RCS was established and succeeded in estimating the
mass of Itokawa.
After the second take-off, however, Hayabusa spacecraft had lost the attitude due to the leakage of RCS
and had been out of communication for two months. Instead of the regular actuators out-of-order, the
xenon-gas from neutralizers of the ion-engines has been used for spin-down and attitude control along the
spin-axis.
These neutralizers (Figure 1) are the devices to neutralize ion-beams by xenon-gas. The torque arms with
one meter of the distance from the center of mass are fortunately enough to generate torques required for
the maneuvers, even though their thrust is only 20ƒŹN.
1 NEC Aerospace Systems Ltd
2 NEC Corporation
Earth return of Hayabusa has been postponed
to June 2010 and the generation of ĢV for the
return has been started from March 2006.
Figure 2 shows the Hayabusa return orbit. To
reduce the consumption of xenon-gas required
for ion-engine drive, the new technique for
passive attitude control by means of solar
radiation pressure for sun-tracking with no
thrust of xenon-gas has been established [1].
In February 2007, Hayabusa spacecraft was
switched from spin-stabilization to three-axis
attitude control based on bias-momentum and
in March, three-axis control with ion-engine
drive has been established. This paper reports
the results of spin-stabilization and three-axis
attitude control based on bias-momentum.
2. Sun Tracking while spin stabilized
In the passive sun-tracking mentioned in the previous chapter, the solar radiation pressure torque acts on
the angular-momentum vector of the spinning spacecraft and then, changes the attitude being based on
gyro-effect. Considering the directions of the sun and spin-axis in ecliptic coordinate system, the motion of
the sun from Hayabusafs point of view is almost parallel with the longitudinal direction. At this time,
pointing the spin-axis to the north/south of the sun, the solar radiation pressure torque is parallel to the
daily motion of the sun and proportion to the sun angle.
Because the angular momentum vector of Hayabusa in spin-stabilization is pointing to anti-solar
direction (i.e. ?z axis direction), it is possible to track the sun by pointing to the north side of the sun in
ecliptic coordinate system. The mechanism is as follows: the greater the angle from the optimum increases,
the more the solar radiation pressure torque increases and then, it causes the increase of the angular
velocity along the spin-axis in proportion to the torque. As a result, the spacecraft moves faster than the
sun (figure 3 (a)). When the spin-axis goes in front of the direction of the sun (figure 3 (b)), the solar
radiation pressure torque starts pointing to the side which the sun approaches to. It causes the reduction of
the angle and then, the angular velocity reduces in proportion to the angle (figure 3 (c)) By repeating this
sequence, the sun-tracking attitude control which keeps the sun angle constant has been successfully done
without fuel.
Figure 1 Ion Engines and Neutralizer
Figure 2 Hayabusa Return Orbit
Figure 3 Relation of Sun direction and Spin axis.
The history of the direction of the spin-axis until ion-engine drive restarted is shown on figure 4. The
history of z-axis of the spacecraft in ecliptic coordinate is plotted on the figure. The sun is located around 0
degrees in the ecliptic coordinate system. In March, the spin-axis had pointed to the Earth because the
direction is almost the same as the one to the Sun.
To start sun-target-pointing from June, the spin-axis was pointed to +2 degrees of the north side of the sun.
In April, to reduce the sun angle, the direction along z-axis had been lowered. After July, the error of the
sun angle had been successfully kept within less than 3 degrees and the difference of the longitudinal
direction with the one of the sun within }1 degree.
Figure 4 History of Spin Axis
3. Momentum-Biased Attitude Stabilization
To utilize the third/last RW parallel to z-axis, three-axis attitude control based on bias-momentum had
been established. In the case when the ion-engines were not in use, Hayabusa tracks the sun by utilizing
solar radiation pressure torque and pointing z-axis to the desired direction by thrusting xenon gas from the
neutralizers (as same as operation in spin stabilization).
On the other hand, the actuators used in accelerating ion-engines are not only RW but two-axis gimbals
to control the thrust axis of ion engines to pass through the center of mass. These gimbals are able to shift
the axis from the center of mass to generate torques. One is to keep unloading all the time to hold the rpm
of z-axis constant, and another one is to control y-axis directly.
X-axis has been controlled indirectly by two types of torques; (1) the very small, swirl torque along the
-2
0
2
4
6
8
10
12
14
140 160 180 200 220 240 260 280 300
Latitude (J2000.0 EC)
Longitude (J2000.0 EC)
March June Sep. Dec.
Pointing to the Earth Pointing to the Sun
(a) Large Sun angle (b) Middle Sun angle (c) Small Sun angle
Spin axis
Solar radiation pressure torque
thrust axis and (2) the solar radiation pressure torque pointing to the opposite direction. The mechanism is
shown in figure 5. The swirl torque mentioned above can be generated due to the structure of ion-engines,
which mechanism had been already clarified on the way to Itokawa. The second torque can be controlled by
adjusting the sun angle with the gimbals of ion-engines. The sun angles balancing with the swirl torque of
ion-engines are around 15 degrees when two ion-engines drive and around 3 degrees when only one engine
drives.
Figure 5 Solar Radiation Torque and Swirl Torque of Ion Engin
4. Flight Results
The history of the solar direction and z-axis of the spacecraft in ecliptic coordinate system from the
beginning of March 2007 to the end of July is shown in figure 6. At the beginning of March (around 50
degrees in ecliptic longitude), z-axis had pointed to the north side of the sun to determine the orbit prior to
driving ion-engines, based on the same idea as spin-stabilization. From March 12, the ion-engines have
been restarted. To balance with the torque generated by the engines, z-axis was pointed to the south side of
the sun and then, the test for attitude stabilization with solar radiation pressure torque was performed.
After the test, the ion-engines have been driven keeping the sun angle constant from the middle of April
(ecliptic longitude of around -15 degrees). Then, at the end of April (ecliptic longitude of 0 degree), the
control mode was switched to one ion-engine drive and the sun angle was reduced. To determine the orbit
-20
-10
0
10
20
-50 -40 -30 -20 -10 0 10 20 30 40 50 60 70 80 90 100 110 120
Ecliptic Latitude [deg]
Ecliptic Longitude [deg]
Sun
Z-axis
March
April May Jun July
Figure 6 Sun direction and Z-axis history
+Z-axis
Swirl Torque of Ion Engine
(fixed)
Solar Radiation Torque
(proportional to sun angle)
Sun Angle
Angular Momentum
(RW-Z)
at the beginning of May, the spin-axis had been pointed to the north side of the sun again without
ion-engine drive. After restarting ion-engine drive, the attitude had temporally drifted by reducing rpm of
RW from 3,200rpm to 2,000rpm in the middle of May (ecliptic longitude of 30 degrees approximately). After
that, it recovered back to the stable attitude pointing to the south side of the sun. To determine the orbit at
the end of July, the spin-axis pointed back to the north of the sun, which was stable even with no thrust of
ion-engines. The details of the remarkable ĢV operations are descrived in the followings.
4.1. Momentum-Biased Attitude Stabilization without
Thrusting
In the middle of February, spin-down and nutation-dump
had been performed and then, three-axis attitude control
started. Before starting ion-engine drive, the orbit was
determined while keeping the control with no thrust. At
this time, the passive attitude control for sun-tracking with
pointing z-axis to the north of the sun without xenon-gas
thrust had been successfully performed, based on the same
idea as spin-stabilization. The history of z-axis at this time
is shown on figure 7. This figure shows the history of z-axis
with respect to the sun in ecliptic coordinate system. In
about a week from 2nd to 9th of Marth, it goes around the
direction of 1 degree of the north side of the sun. It satisfies
the requirements of the attitude accuracy to guarantee the
communication with the Earth.
4.2. Momentum-Biased Attitude Stabilization with Thrusting
After determining the orbit, flight plan was updated and ion-engine drive re-started. The torque generated
by ion-engine is on the ecliptic plane because the direction of ĢV is almost on the same plane. On 12th of
March, the axis was pointed to 10 degrees apart from the south side of the sun by thrusting ion-engines and
operating gimbals. Then, the torque of ion-engine thrust axis could be balanced with solar radiation
pressure torque by increasing the elongation with the sun. The change of drift is shown in figure 8 and
table 1. After that, the optimum sun angle was determined and in April, stabilization with ion-engine drive
around 17 degrees from the south side of the sun had been successfully done. It had been required for
attitude control to keep the direction of ĢV within 5 degrees and successfully satisfied.
Figure 8 +Z axis history with thrust
Figure 7 +Z axis history without thrust
(Sun Tracking)
Table 1 Change of drift
Date IES and Spin axis direction Drift
3/12 One engineASouth, 10deg Left, 3deg
3/13 Two enginesASouth, 12deg Right, 4deg
3/14 Two enginesASouth, 14deg Right, 3deg
3/15 Two enginesASouth, 17deg Right, 1deg
4.3. Precise Attitude Control
Afterwards, the Earth return journey has been started by changing two-engine drive to one-engine drive.
During the operation, it was clarified that the balancing point of thrust axis torque and solar radiation
pressure torque could be changed by the influence of the small change of ion-engine thrust to the
thrust-axis torque.
Approaching to the perihelion, it became apprehensible that the temperature at the baked point might
exceed the baking temperature. In the worst scenario, the attitude might be disturbed by out-gas. In this
case, out-gas could be a very large disturbance for Hayabusa spacecraft, which controls the attitude by
adjusting very small sun angle.
Because it becomes clear that the temperature increase is proportion to the solar incidence angle on +X
side panel, the decision to expose solar heat flux on ?X side panel was made on 16th of May. Figure 8 shows
the history of TSAS-X corresponding to the solar incidence angles from 5th of May to 30th of June. +side
means that the plane of incidence is +X side panel and ?side for -X side panel. Approaching to perihelion at
the solar distance 0.95AU on Jun 9 (The lower side of figure 9), this constraint for attitude control had
become more severe. In the most cases, the operation with the precision of 0.3 deg/day has been achieved.
In the cases with the stable thrust of ion-engines, the precision of 0.1 deg/day could be achieved. Afterwards,
Hayabusa has been operated with the precision of 1 degree of the relative attitude to the sun.
5. Concluding Remarks
Three axis attitude control strategy using solar radiation torque was developed and has been used to control the spacecraft
for delta-V operation. The two axes were directly controlled using single reaction wheel and two gimbals on the ion
engines table, and another axis was indirectly controlled applying a balance between the swirl torque of the ion engine
thrust and the solar radiation torque proportional to the sun angle. It was realized two factors. First, the swirl torque was as
small as solar radiation torque with small sun angle. Secondly, the difference torque was effective in controlling the
angular momentum generated by Z-axis reaction wheel.
It was successful that the pointing accuracy was achieved within 0.3deg/day. The attitude has been kept the sun tracking
since the delta-V operation. On the other hand, the strategy was established by operational experience and was still not
enough study. These techniques and experience will be useful for future solar powered sail mission, so that we will
continue to be conducted the attitude control strategy.
Reference
[1] Kenfichi Shirakawa et. al., gHayabusa Attitude Control by Xe Gasjet for Rescue Operationsh, 17th. Workshop on
Astrodynamics and Flight Mechanics, 2006
[2] Jun'ichiro Kawaguchi and Kenfichi Shirakawa, A Fuel-Free Sun-Tracking Attitude Control Strategy And the Flight
Results in Hayabusa (MUSES-C), AAS 07-176, AAS Flight Mechanics Conference, Sedona, AZ, Jan. 30 ? Feb. 1,
2007
[3] Junfichiro Kawaguchi et. al., gAttitude Control Flight Experience: Coping with Solar Radiation and Ion Engines Leak
Thrust in Hayabusa (MUSES-C)h, 20th ISSFD Conference, Annapolis, Maryland, 24-28 September, 2007