|
NASA autopilot test suffers crippling flaw BY JUSTIN RAY SPACEFLIGHT NOW Posted: April 16, 2005 NASA's $110 million high-risk, high-tech demonstration of a space autopilot for future human spaceships and robotic cargo craft closed within 300 feet of its target Friday night, then inexplicably ran out of fuel to the dismay of helpless engineers on the ground. The Demonstration of Autonomous Rendezvous Technology was propelled into polar orbit aboard an air-launched Pegasus rocket off the coast of California at 1727 GMT (10:27 a.m. local time; 1:27 p.m. EDT) Friday, beginning a day-long mission entirely without human assistance. Dubbed DART for short, the sophisticated 800-pound, 6-foot spacecraft was equipped with a scifi-like "brain" and "eye" to smartly chase down a target satellite and maneuver within feet of the object devoid of any input from mission control. A half-hour into the flight, DART flew over the McMurdo tracking station in Antarctica and relayed initial data. Although engineers could not issue commands to the satellite, they were anxiously awaiting any and all information from the craft to assess the mission's progress. The satellite completed a checkout of its systems and instruments, which appeared normal. In hindsight, officials said there were some navigation errors beyond what was expected. "About 30 minutes in the mission, we went into our on-orbit checkout. We had all of the instruments powered up. Then we got a pass over McMurdo," Jim Snoddy, DART project manager from NASA's Marshall Space Flight Center, told reporters in a teleconference today. "The first event that we consider anomalous was we had higher than anticipated navigation errors that we observed after the first McMurdo pass over upon completion of the on-orbit checkout." But DART flew on successfully over the next 7 hours or so, executing the rendezvous phase of the mission to reach the U.S. military's Multiple Paths, Beyond-Line-of-Sight Communications (MUBLCOM) satellite. Launched in May 1999, the tiny craft features optical targets -- like bicycle reflectors -- for DART to see. "We had high accuracy, our autopilot was exercised, performed flawlessly as designed," Snoddy said, recalling that DART had completed an orbit transfer and aligned with MUBLCOM's orbital plane as it moved to a point 25 miles behind and 5 miles below the target. "We completed our free drift below the target and prepared for final approach. Again, that was all as planned precisely." As DART moved to 1.8 miles behind MUBLCOM, the onboard closed-loop guidance kicked in. "This is when you get into the GPS state vector, get into a very tightly precise, controlled approach. We noticed at that time we had some excessive propellant consumption. That operation goes on for approximately a couple of hours." The key instrument on DART, the Advanced Video Guidance Sensor, successfully spotted the reflectors mounted on MUBLCOM as the two flew inside two miles of each other. "That means we were locked onto MUBLCOM, we were doing things and being able to stay with MUBLCOM," Snoddy said. Mission engineers were in the blind during a large part of the DART flight, having to wait 30 to 60 minutes for the next spurt of data from the spacecraft as it passed over designated ground tracking stations scattered around the globe. Eleven hours after launch, DART approached MUBLCOM within a distance of 302 feet. But the two spacecraft would not get any closer as a mission-ending failure struck DART. A guidance system command commenced a safe retreat due to propellant depletion. DART has sensed the 70 pounds of nitrogen for its 16 cold-gas reaction control system thrusters needed for minute control and fine-tuning during close approach was gone. The safety instructions built into the satellite told DART it could no longer continue with the maneuvers and needed to move into the next phase of the mission script -- departing MUBLCOM's vicinity and preparing for a concluding retirement engine burn. And DART did just that, coasting away for over 12 hours before a scheduled Saturday afternoon firing of its larger engines to deplete the remaining hydrazine fuel supply to enter an orbit that should lead to re-entry in the atmosphere for burnup in about 10 years. The mission was over. "The automated rendezvous and proximity to go do all of this, to create the brain to do all of that was the key, which I think we demonstrated a lot of. So that was one key technology," Snoddy said. "The other one was to have this eye, the Advanced Video Guidance Sensor, actually look at (MUBLCOM), pull itself in, pull itself back. We got the eye open, it seen it, it knew how to go get to it, but we didn't get any further than that." DART was supposed to close within 16 feet of MUBLCOM during a series of maneuvers flying in, backing away and circling around the target, then repeating the sequence a second time. Designers believed there would be enough thruster propellant for the entire mission plan, plus 30 percent margin. But DART ran out even before flying down the rendezvous corridor to closest approach the first time. "We really were excited and hoping to go show we could drive the DART spacecraft, with its AVGS, down to five meters, back out, keep doing all the transitions, re-validate your models and keep repeating these steps. We actually programmed in to go do all the steps twice. So really what we didn't do is really show the robustness of the Advanced Video Guidance Sensor. We showed that it would power on and work (but) we didn't let it really go do the workhorse job we designed it to go do," Snoddy said. Telemetry did not show any signs of leakage or pressure loss from the thruster tanks. Why the spacecraft expended the fuel supply so quickly is a mystery that must be unraveled. "It was doing what the guidance told it to go do," Snoddy said, adding the thruster system was guzzling the propellant based on "exactly what the guidance told it to consume." NASA is forming an investigation to determine what went wrong. "The real truth is you have to be perfect in space. There was obviously something that we didn't model...That's what the investigation will go help drive out. They'll tell us what we didn't know, assign the root cause so that all of the future mission can learn from this. "This was the first time anybody has ever tried to hand over something totally autonomous in orbit and let it happen. It took a lot of faith to go do that." Although it didn't get a chance to demonstrate challenging maneuvers next to MUBLCOM, DART did make the first autonomous rendezvous conducted without any human interaction in U.S. space program history. "We've done what nobody has ever tried to do before,"Snoddy said, adding that "we met some key objectives" and called the mission "a partial success." "But again, we left out a key area of science," he said. "We left a 12-hour gap of data to go show how we could maneuver this vehicle precisely in orbit." DART was designed from the outset to be a relatively cheap flight with no ground commanding capability, built with batteries lasting just 24 hours, and limited ground station coverage, versus much broader, perhaps near-continuous communications via NASA's network of space-based Tracking and Data Relay Satellites. The 24-hour mission life in and of itself meant there was little chance for engineers finding and patching the fault even if they had the opportunity to wrestle control away from DART's onboard systems in a desperate attempt to save the day. "It was a technology demonstration, a flight experiment within a budget." NASA developed the DART mission to test autopilot capabilities for the now-cancelled Orbital Space Plane project. But the technology is considered valuable for future human spacecraft, cargo-carrying resupply ships and satellite servicing concepts. The space agency views such autonomous features critical to the Vision for Space Exploration, which calls for the U.S. to return astronauts to the moon and eventually put humans on Mars. "We had a rigorous test program, we went through verification, we went through all of the steps to dot all of the i's and cross the t's. That is what the failure investigation will go find out -- what i we didn't dot and what t we didn't cross." |
|
||||||||||||||||||||||||||||||||||||||||||||