Collision Warning Capabilities
The DOD SSN warns the space shuttle program of possible close conjunctions with cataloged orbiting objects. The SSN is composed of ground-based radars, electro-optical sensors, and command and control systems that detect, track, and catalog man-made objects in Earth orbit. The SSN uses ground-based sensors to collect data on orbiting objects. The data are used to estimate the positions and trajectories (called orbital element sets) of objects. The element sets can then be used to predict whether a close approach or collision may occur.
The SSN can only warn about collisions with objects identified in their satellite catalog. The current SSN catalog of tracked objects includes essentially all objects larger than 100 cm in diameter in low Earth orbit, and about 95 percent of objects larger than 30 cm in diameter. Although some objects as small as 10 cm are also included, from 15 and 50 percent of objects between 10 cm and 20 cm may be missing. Virtually no objects smaller than 10 cm in diameter are included in the catalog (Kessler, 1996; Lord, 1996).
The U.S. Space Command’s Space Control Center (SCC) notifies NASA flight controllers when it predicts close conjunctions between the orbiter and cataloged objects. The information is used to determine if a launch should be delayed or an orbiting vehicle should be maneuvered to avoid a possible collision.
Two different prelaunch warnings are used to protect the orbiter from collisions with objects shortly after launch. The SCC (using data from the SSN)
determines whether any cataloged objects are predicted to enter an area around the orbiter (called an alert box) with dimensions approximately 5 km radially, 25 km along the track of the orbiter (either leading or trailing), and 5 km out of the orbital plane during the first two hours of the mission (Flight Rule A4.1.1–3). The Eastern Range is responsible for notifying NASA if the orbiter will enter a 50 km×200 km×50 km region around another crewed vehicle during the first orbit after a launch (Flight Rule A2.1.1–1.).
If either of these warnings indicates a possible collision, the launch is usually delayed until the next even minute. Additional analyses are requested using the new launch time, and further holds may be ordered. To date, two shuttle launches have been delayed to avoid potential collisions with orbiting objects (Reeves, 1997).
When the shuttle orbiter is in orbit, the SCC screens the entire satellite catalog for objects that could approach the orbiter within a 5 km×25 km×5 km alert box at any time during the mission. The SSN is tasked with providing more intensive tracking of the approximately one to two objects per day that penetrate this box. The objects are then reassessed using a more accurate and computationally intense “special perturbations” algorithm to determine if any will come within a “maneuver box” of 2 km radially, 5 km along the orbiter’s track, and 2 km out of plane. (The alert and maneuver boxes are shown in Figure 5–1.) The large box size relative to the size of the orbiter is necessary because the current and future positions of tracked objects are not known precisely. The accuracy of the special perturbations algorithm is obviously dependent on the availability of accurate sensor data.
Information about potential close conjunctions is passed to NASA flight controllers who apply Flight Rule A4.1.3–6 (see Box 5-1), which stipulates that a maneuver be performed “if the maneuver does not compromise either primary payload or mission objectives.” Like all flight rules, this one can be superseded by real-time decisions.
A collision avoidance maneuver will be performed for a conjunction predicted by the United States Space Command if the predicted miss distance is less than 2 km radially, 5 km down track, and 2 km out-of-plane and if the maneuver does not compromise either primary payload or mission objectives. Propellant redlines will not be budgeted for any potential maneuvers.
If NASA flight controllers decide a maneuver is necessary, the orbiter uses its on-board propulsion system to execute the maneuver. On average, the orbiter changes its velocity by about 30 cm/s to avoid a collision, which requires the expenditure of about 11 to 14 kg of propellant (Loftus, 1997). Between shuttle missions STS-26 and STS-82, the orbiter logged approximately 527 days of on-orbit operations. During that time, nine cataloged objects penetrated the 2 km× 5 km×2 km box. In five of these nine cases, avoidance maneuvers were not performed because they would have interfered with primary mission objectives. In the other four cases, evasive maneuvers were performed by the shuttle. On two other occasions, maneuvers were performed when penetrations of the larger 5 km×25 km×5 km alert box, but not the 2 km×5 km×2 km maneuver box, were predicted (NASA, 1997).
Proposed New Collision Warning Technique
The current technique for determining the threat of collision between the orbiter and a tracked object does not directly take into consideration the geometry of the conjunction or uncertainties about the position of either the orbiter or the other object. In the future, NASA plans to switch from this deterministic approach to avoiding collisions to a probability-based approach, which will be used for ISS operations.
The new method is based on the probability of collision (Pc), which is defined as the probability that an object will penetrate a sphere around the spacecraft. The calculation of Pc is based on the uncertainties of the positions of the spacecraft and the other object at conjunction and the geometry of the predicted conjunction.
In this probability-based approach, the U.S. Space Command’s computation of misses between orbits (COMBO) program will be run with current data for all cataloged objects for 72 hours into the future. The SSN will increase the
frequency of tracking observations for each object COMBO indicates will penetrate the 5 km×25 km×5 km alert box, thus increasing the accuracy of position estimates. After processing the updated tracking data, COMBO will be run again in the “special perturbations” mode. If the conjunction is still within the alert box, an alert warning will be sent to NASA, along with the updated vectors and covariances for each object and the ISS at conjunction. (Covariances represent the uncertainty in the element sets of the ISS and conjuncting object.) The SCC will continue tracking these objects more intensively and send the updated vectors and covariances to NASA.
NASA will use these data to calculate Pc. If Pc exceeds a predetermined threshold, called the yellow threshold, the flight director will consider ordering a collision avoidance maneuver. If Pc exceeds a higher threshold, called the red threshold, the flight director will order a collision avoidance maneuver.
ANALYSIS AND FINDINGS
The SSN can only observe objects with radar cross-sections on the order of tens of square centimeters. As a result, the Satellite Catalog does not include many objects that could seriously damage the orbiter. The full extent and quantity of the uncataloged population is not known. As shown in Table 2–1, however, NASA’s ORDEM96 model predicts that the probability that the shuttle will collide with a trackable object (i.e., greater than 10 to 30 cm in diameter) if no collision avoidance maneuvers are performed is less than 0.5 percent over the design life of the shuttle fleet (.002 impacts per 400 10-day missions). The probability of impact with an object that is untrackable with current sensors but still able to cause critical damage to parts of the orbiter (i.e., objects 5 mm to 10 cm in diameter) is closer to 20 percent (0.2 impacts over 400 10-day missions). Thus, according to NASA estimates, more than 95 percent of the objects that can cause critical damage to the orbiter are not being tracked or cataloged.
Finding. Warnings of collision are provided only for cataloged objects. NASA estimates that more than 95 percent of the objects that could cause critical damage to the orbiter are not being tracked or cataloged.
The current methodology for providing collision warnings to the orbiter requires coordination between NASA and the DOD SSN. Although the Memorandum of Agreement between United States Space Command and Johnson Space Center for Space Control Operations Relationship, Space Shuttle Program Support, and International Space Station Program Support (USSPACECOM and JSC, 1996) addresses requirements in a general fashion, it does not state requirements
for the timeliness of warnings, minimum object size, or the accuracy or uncertainty of predictions. In August 1997, NASA levied more detailed requirements on the SSN, but the SSN may be unable to meet these requirements.
In the absence of specific requirements, budget limitations have forced the DOD to reduce the number of ground-based sensors that supply most of the information about debris at orbiter operational altitudes. Since 1989, the number of radar sensors in the SSN has been reduced from 19 to 13, and no new radar sensors have been added, although upgrades have been made at a few sites. Plans for new or upgraded SSN sensors do not include requirements that would improve debris tracking. Unless action is taken, the SSN’s ability to provide collision warnings to the shuttle will probably diminish.
Finding. The capabilities of the SSN to provide collision warnings to NASA are eroding. Until recently, NASA had not issued requirements that might have helped to halt this erosion.
New Approach to Collision Warnings
The planned use of covariance data should help NASA make better decisions about collision avoidance maneuvers and reduce the number of unnecessary maneuvers. NASA is now waiting for the U.S. Air Force to complete development of covariance matrices. The covariance data will also be used for the ISS, for which it will be needed by late 1998.
NASA has recently specified the level of accuracy it requires for covariance or state vector uncertainties at conjunction. The method the SSN currently uses to compute a covariance does not represent uncertainty in the state vectors accurately enough to calculate Pc accurately at the orbiter’s operational altitude, where the major uncertainty is atmospheric drag. Inaccurate calculations of Pc could result in the orbiter performing unnecessary maneuvers or not performing necessary maneuvers. Current work (Barker, 1996, 1997) for the Air Force will greatly improve the computation of covariance but will not incorporate the uncertainty in atmospheric drag.
Finding. NASA plans to use a new probability-based approach to determine when a collision avoidance maneuver is necessary, but the collision avoidance data currently provided by the SSN is not accurate enough for the new approach to be effective
NASA Flight Rules for Collision Avoidance
In deciding whether to make a collision avoidance maneuver, NASA flight controllers assess whether the maneuver would compromise primary payload or mission objectives. The current wording of Flight Rule A4.1.3–6, suggests that a very close conjunction of the orbiter and a large, well tracked object could occur
without requiring a maneuver. Unlike the vast majority of flight rules, which place safety first and then allow exceptions in limited cases, this flight rule places mission needs first and requires those who provide the collision warning to prove that action needs to be taken.
Future shuttle missions that support the ISS will have a limited ability to maneuver. NASA Flight Rule C4.3.2–1, Space Station Translation Maneuvers During Joint Shuttle Operations, states that debris avoidance maneuvers will not be performed during docked operations. In addition, NASA reports that maneuvers will probably not be performed when the orbiter is undocked but is involved in assembly operations (Reeves, 1997).
Finding. NASA Flight Rules A4.1.3–6 and C4.3.2–1 appear to place mission success ahead of flight safety. The mechanism for making trade-offs between success and safety is not explicit.
Recommendation 6. NASA has recently documented and provided collision avoidance requirements to the DOD. NASA and the DOD should work to satisfy these requirements, to identify impending changes to the SSN that will affect debris tracking, and to identify changes that would improve the SSN’s ability to track smaller objects that pose a hazard for crewed spacecraft.
Recommendation 7. NASA should re-examine Flight Rules A4.1.3–6 and C4.3.2–1 and consider restating them to establish when a maneuver is mandatory for safety reasons.
Barker, W.N. 1996. Space Station Debris Avoidance Study. Initial Report. Kaman Sciences Report KSPACE 96–114. Colorado Springs, Colo.: Kaman Sciences Corporation.
Barker, W.N. 1997. Space Station Debris Avoidance Study. Final Report. Kaman Sciences Report KSPACE 97–47. Colorado Springs, Colo.: Kaman Sciences Corporation.
Kessler, D.J. 1996. Private communication to Robert Culp, member of the National Research Council Committee on Space Station Meteoroid/Debris Risk Management, April 17, 1996.
Loftus, J.P. 1997. Letter to Committee on Space Shuttle Meteoroid/Debris Risk Management from Joseph Loftus, Jr., Assistant Director (Plans), NASA Johnson Space Center. SA-97–178. July 15, 1997.
Lord, L.W. 1996. Memorandum to the Committee on International Space Station Meteoroid/Debris Risk Management from U.S. Air Force Major General Lance Lord, April 3, 1996.
NASA. 1997. Conjunction Summary for STS-26 through STS-82. Letter DM3–97–7. Houston: NASA.
Reeves, W. 1997. Presentation to the Committee on Space Shuttle Meteoroid/Debris Risk Management, Houston, Texas, June 16, 1997.
USSPACECOM and JSC (United States Space Command and Johnson Space Center). 1996. Memorandum of Agreement for Space Control Operations Relationship, Space Shuttle Program Support, and International Space Station Program Support. April 10, 1996.