Star Scanner Research Articles

별추적기의 기술개요와 개발동향

전세계적으로 적용되는 개발개념인 보다 빨리, 더 좋게, 더 값싼 위성을 설계하는 일에 있어 향상된 자세결정시스템을 도입하는 것은 자명한 일이라고 할 수 있다. 오늘날 CCD나 APS 기반의 별추적기는 현존하는 위성의 자세결정용 센서중에서 가장 좋은 정밀도를 제공하고 있다. 본 논문에서는, 별추적기의 기본적인 동작원리와 관련 기술의 개요를 소개함과 동시에 별추적기 개발에서 주요 이슈가 되었던 기술들을 비교 분석한다. 또한 별스캐너로부터 MEMS기술을 적용한 별추적기에 이르기까지, 별추적기의 간추린 역사와 세계 각국에서의 개발현황을 소개한다. In order to accelerate the evolution of faster, better, cheaper spacecraft, it is evident that greatly enhanced general-purpose attitude determination methods are needed Currently, star tracker sensors based on charge coupled devices (CCD) or active pixel sensors(APS) enable one to obtain the best spacecraft attitude estimation among the existing sensors for attitude determination. In this paper, basic principles of star tracker technology are explained including major issues arising in design and development of star tracker. Also, an historical overview and worldwide survey associated with various star trackers from star scanner through microelectromechanical system(MEMS)-based star tracker is offered.

The Galileo star scanner observations at Amalthea

In November of 2002, the Galileo spacecraft passed within 250 km of Jupiter's moon Amalthea. An onboard telescope, the star scanner, observed a series of bright flashes near the moon. It is believed that these flashes represent sunlight reflected from 7 to 9 small moonlets located within about 3000 km of Amalthea. From star scanner geometry considerations and other arguments, we can constrain the diameter of the observed bodies to be between 0.5 m to several tens of kilometers. In September of 2003, while crossing Amalthea's orbit just prior to Galileo's destruction in the jovian atmosphere, a single additional body seems to have been observed. It is suspected that these bodies are part of a discrete rocky ring embedded within Jupiter's Gossamer ring system.

Pioneer Venus Orbiter search for Venusian lightning

During 1988 and 1990, the star sensor aboard the Pioneer Venus orbiter (PVO) was used to search for optical pulses from lightning on the nightside of Venus. In our previous effort to use the star sensor for this purpose [Borucki et al., 1981], we reported an upper limit to the lightning activity, but we were unable to show that lightning was being detected because the signal rate was indistinguishable from the false alarm rate. Because the periapsis altitude has increased by nearly a factor of ten since 1979, the star scanner views a much larger area of Venus than it did previously. This increased viewing area should provide an increased signal rate because the amplitude of optical pulses should still be above the detection threshold of the sensor if the flashes are as bright as terrestrial flashes. Because the false alarm rate did not increase, the increased viewing area allows a more sensitive search for lightning activity. Useful data were obtained for 53 orbits in 1988 and 55 orbits in 1990. During this period, approximately 83 s of search time plus 7749 s of control data were obtained. Our results again find no optical evidence for lightning activity. Within the region that was observed during 1988, the results imply that the upper bound to short‐duration flashes is 4×10−7 flashes/km²/s for flashes that are at least 50% as bright as typical terrestrial lightning. During 1990, when the 2‐Hz filter was used, the results imply an upper bound of 1×10−7 flashes/km²/s for long‐duration flashes at least 1.6% as bright as typical terrestrial lightning flashes or 33% as bright as the pulses observed by the Venera 9. The upper bounds to the flash rates for the 1988 and 1990 searches are twice and one half the global terrestrial rate, respectively. These two searches covered the region from 60°N latitude to 30°S latitude, 250° to 350° longitude, and the region from 45°N latitude to 55°S latitude, 155° to 300° longitude. Both searches sampled much of the nightside region from the dawn terminator to within 4 hours of the dusk terminator. These searches covered a much larger latitude range than any previous search. Our results show that the Beta and Phoebe Regio areas previously identified by Russell et al.[1988] as areas with high rates of lightning activity were not active during the two seasons of our observations. When we assume that our upper bounds to the nightside flash rate are representative of the entire planet, the results imply that the global flash rate and energy dissipation rate derived by Krasnopol'sky [1983] from his observation of a single storm are too high. Experimental measurements carried out to determine the amount of scattering occurring in the star scanner optics indicate that our previously determined upper bound to lightning activity is too low. The present results supplant the past results. The apparent conflict between the radio data that indicate the presence of a large amount of lightning activity and the optical data that indicate no lightning on the nightside, can be resolved by noting that theoretical considerations indicate lightning activity is to be expected only on the dayside.

Magellan in-flight gyro/star scanner misalignment calibration

Techniques are described for the in-flight calibration of gyro/star scanner misalignments for the Magellan spacecraft. The poor observability of one of the six components of misalignment is discussed in the context of a simple least-squares estimation model. The assumptions that lead to singularity in the information matrix are explicitly stated and it is shown that the singularity persists for all scanner slit configurations using only two slits, regardless of slit geometry or separation. A set of misalignment error state variables, a configuration of three stars, and a maneuver/scan sequence is described which yields a well-conditioned information matrix for least-squares estimation of five of the six misalignments. Finally, it is shown by convariance simulation that ground-based optimal estimation can satisfactorily resolve all six components of the misalignment error when more than two star scanner slits are used.

Inertial attitude determination for a dual-spin planetary spacecraft

Design and performance evaluations of an onboard inertial attitude determination system for a dual-spin planetary spacecraft are presented. A quaternion integration algorithm and a rate estimator sequentially determine the scan platform's inertial attitude and rate from the drift and misalignment compensated gyro outputs. A least-squares estimator algorithm processes the star transit data from a rotor-mounted star scanner, and provides periodic updates of the scan platform's attitude as a means of correcting the drift of the quaternion integration algorithm. Computer simulated algorithm performance in the presence of nutation and sensor noise are presented.

A radiation-hardened star scanner for spacecraft guidance and control

A star scanner for the Galileo Project is described. Because it must function in the intense Jupiter radiation field, which consists primarily of high-energy electrons and protons, this star scanner is designed with radiation-hardness as a prime objective. The optics and the multiplier phototube are specially designed to minimize fluorescence and Cerenkov effects, and the electronic components to minimize ionization and displacement damage. The star-scanner layout provides maximum shielding for the optics and phototube. Test results predict successful operation in the expected Jovian radiation environment.

Automated Test Set to Measure Star Scanner Accuracy with Arc-Second Resolution

Automated Test Set to Measure Star Scanner Accuracy with Arc-Second Resolution

Star Scanner Attitude Determination for the OSO-7 Spacecraft

Calculation of the inertial orientation of the OSO-7 spacecraft from the times at which known stars or planets transit planes fixed in the spacecraft. Both the reference planes and the timing information are provided by the star scanner instrument aboard the spacecraft, while the star identification and the statistical estimation of a set of parameters describing the spacecraft attitude are accomplished in a ground station computer facility. A recursive least-squares determination is made of a vector of first-order differential corrections to the attitude state vector. Preliminary analysis indicates the system accuracy to be 3 arc min in each attitude Euler angle.