Autonomous Optical Navigation (AutoNav)

The first flight of NASA's New Millennium Program, DeepSpace 1, will include a new navigational technology: anautonomous optical navigation system. The D S 1Navigation system will be the first use of autonomousnavigation in deep space. The task for this system is to 1)perform interplanetary cruise orbit determination, usingimages of distant asteroids, 2) control and maintain the orbitof the spacecraft using the ion propulsion system (anothertechnology never before applied to deep space) andconventional thrusters, 3) perform approach orbitdetermination and control using images of the sciencetargets, 4) perform late knowledge updates of target positionduring close fast flybys in order to facilitate a high degree ofquality data return from 2 targets: asteroid McAuliffe andcomet West-Kohoutek-Ikemura. Additionally, an encounterwith Mars will probably be performed with possibly a closeflyby of one of the Martian moons, Phobos or Deimos.Several functional components are necessary to accomplishthese tasks. These include picture planning and imageprocessing, dynamical modeling and integration, planetaryephemeris and star catalog handling, orbit determinationdata filtering and estimation, maneuver estimation,spacecraft ephemeris updates and maintenance, and generalinteraction with the other onboard autonomous systems.These systems are described, as is the means of theiroperation onboard. Finally, performance statistics from trialruns of the system are given. INTRODUCTIONAutonomous onboard optical navigation will be a necessarycomponent of autonomous spacecraft operations for manyfuture planetary exploration missions. Because of light-travel times, there are experiments and even missions thatcannot be performed or have limited data potential unlessautonomous navigation systems are incorporated. Closeorbits or very fast flybys of small poorly characterizedobjects are examples of such missions. Reducingoperational complexity and costs is another goal ofautonomous navigation systems. In the not-too-distantfuture, many small robotic missions may be simultaneouslyexploring the solar system. To increase the efficiency ofthese missions, the spacecraft must take on more of theresponsibilities of their own maintenance, includingnavigation. Adapting many of the techniques proven foroptical navigation for Voyager and Galileo, the NewMil lenn ium DS1 onboard navigation system mustautonomously plan picture sequences, perform imageanalysis, estimate the trajectory and calculate trajectorycorrections using the low-thrust solar-powered ionpropulsion system (IPS). DS1 will be the first planetaryexploration mission to autonomously navigate all post-injection phases of its mission. The engineering of such anavigation system poses a number of very significantchallenges. An overview of Optical Navi-gation and how itwill be applied to DS1 is given in Ref. 1. This first experiment in deep space autonomous navigationwill be a closely monitored experiment. As a means ofvalidating the performance of the onboard navigationsystem, a conventional ground radio-navigation campaignwill be maintained. This ground effort offers the furtheradvantage of providing very high quality calibrations of IPSengine performance, something which the flight navigationsystem (The “Navigator”) would not be able to do. Thoughthe Navigator is designed to be capable of fully autonomousoperation, with many new technologies been tried on DS1,the capability has been maintained to quickly intervenewith, and modify the behavior of the system if missionemergencies require. DS1 MISSION ATTRIBUTESAn overview of the New Millennium Program and DS1 inparticular is given in Ref. 2. The DS1 mission includes avery ambitious and challenging set of mission objectivesand activities. Three targets are intended for flybyencounters: asteroid McAuliffe, Mars, with possibly a closeflyby of one of the Martian moons, and comet West-Kahoutek-Ikemura (WKI). Currently, it is anticipated thatlaunch will occur in July of 1998. The McAuliffe encounter Deep Space 1 Technology Validation Report—Autonomous Optical Navigation (AutoNav) 49will happen late January of 1999, the Mars flyby in late Mayof 2000, and the comet encounter about six weeks later.Figure 1 shows a heliocentric view of a likely missiontrajectory, with important mission events annotated. Theannotations are referenced to Table 1. Figure 1. DS1 Mission DesignFor the McAuliffe flyby, the DS1 spacecraft will performthe closest flyby encounter ever attempted in a deep spacemission: 10 or perhaps even 5 km from the surface of theasteroid. The encounter parameters of Mars have not yetbeen determined, but the flyby altitude of the comet willlikely be on the order of several hundred kilometers, due tothe dangerous environmental conditions near even arelatively inactive comet such as W-K-I. ID Time of EventDescription of Event A Jul. 1, 1998 DS1 Launch B Oct. 24, 1998 End of first principal thrust arc C Dec. 6, 1998 Beginning of second thrust arc D Dec. 27, 1998 End of second thrust arc E Jan 16, 1999 McAullife encounter F Jan 20, 1999 Beginning of third thrust arc G Feb. 8, 2000 End of third thrust arc H Apr. 26, 2000 Mars encounter I Jun. 4, 2000 WKI encounter Table 1. Principal DS1 Mission Events The ambitious nature of these encounters is enabled solelyby the presence of the autonomous navigation system.Performing navigation functions in a closed-loop senseonboard the spacecraft makes possible very late (beforeencounter) controls of the spacecraft encounter coordinates,and updates of knowledge about those coordinates. The objectives of the New Millennium Program (of whichDS1 is the first mission) is to develop and demonstrate newtechnologies which can enable future space explorationmissions. The Autonomous Navigation System is one ofthese technologies being demonstrated. Another such Figure 2. New Millennium DS1 SpacecraftEarth OrbitMars Orbit