The decision to repair the Hubble Space Telescope with robotics technology seemed a natural given the safety concerns raised by the crash of the space shuttle Columbia in 2003. That way the observatory prized by scientists and admired by the public could function for at least five more years with upgraded sensors. And all the experience gained by using teleoperated manipulators would form the basis of future semiautonomous servicing operations in space. Now, however, independent analyses have cast fresh doubts on whether a robotic mission can reliably save Hubble.
According to the original scheme, spacewalking astronauts would have replaced failing batteries, gyroscopes and fine-guidance sensors. In addition, new scientific instruments--including a wide-field camera and a spectrograph--would have boosted Hubble's observational ability by a factor of 10 or more.
But after the Columbia disaster, NASA scaled back its shuttle plans. Unlike flights to the International Space Station, a mission to Hubble provides no repair depot and safe haven for the shuttle and its crew in an emergency. In its stead, the space agency contemplated unmanned rescue efforts.
Various technical obstacles confront an unmanned mission, explains Preston Burch, program manager for the NASA Goddard Space Flight Center. One, the self-guided orbital repair module will have to "rendezvous and soft-dock" with Hubble without any assistance from the telescope or ground staff. (Previous automated dockings have been little more than "controlled crashes.") He notes that these advanced maneuvers will be demonstrated in the near future by a trio of NASA and Defense Department spacecraft. Two, the robot arms will have to unfasten several types of latches, connect cables, and exchange a variety of fragile devices, often in tight spaces. The dual robotic manipulator arms, fitted with tiny video cameras and precision positioning sensors, must accomplish the work mostly autonomously with only limited supervision from earthbound operators. (Because of the distance and other factors, communications face a delay of a few seconds.)
Nevertheless, detailed analysis and laboratory tests of key procedures with full-scale facsimile hardware give the Goddard staff and their contractors a good deal of confidence. Clever workarounds should ease matters considerably as well. Engineers, for example, would place the otherwise difficult-to-install gyroscopes inside the wide-field camera, which is to be attached to the internal telescope structure from the outside. Robotic repair "will be very slow--maybe 10 times slower than by the astronauts, but it's definitely doable," Burch says.
A National Research Council committee, led by Louis J. Lanzerotti of Bell Laboratories and the New Jersey Institute of Technology, does not seem so sure, however. In an interim study report, the panel stated: "The proposed Hubble robotic servicing mission involves a level of complexity, sophistication, and technology maturity that requires significant development, integration, and demonstration to reach flight readiness." The committee warned that the proposed unmanned mission is essentially an experiment that is expected to accomplish difficult real-world objectives and that relying on untested procedures is risky. It urged NASA to "take no actions that would preclude a space shuttle servicing mission." Not exactly a ringing endorsement.
The Aerospace Corporation, a nonprofit private research, development and advisory organization headquartered in El Segundo, Calif., is also skeptical. According to those who have seen its confidential assessment, the study concludes that the probability of success of a robotic mission attempting only minimal goals (extending Hubble's life) is 58 percent. Worse, the odds are one in three for a more ambitious automated effort to install a "deorbit module" (a rocket motor to control reentry) and to upgrade scientific instruments. These odds are less than those of a manned shuttle mission, which offers a 63 percent chance of success, say sources who asked to remain unidentified.