Nine Tons of Hardware-in-the-Loop
SwRI engineers use simulation and autopilot designs to help NASA develop low-cost space vehicle
Hardware-in-the-loop simulation is a cost-effective method for testing complex, flight-critical hardware before it is used in the real world. When the hardware is attached to a nine-ton test vehicle falling to Earth beneath the world's largest steerable parachute, considerable amounts of time and money--not to mention human lives--ride on the ability of the engineering team to get the test vehicle's guidance and navigation system right.
The vehicle in question is the X-38 lifting body, NASA's prototype of the crew return vehicle, a fully automated craft designed to carry up to seven astronauts safely to Earth in case of emergency aboard the new International Space Station.
NASA is developing the X-38 vehicle at a fraction of the cost of a traditional space vehicle such as the space shuttle orbiter. Southwest Research Institute (SwRI) is providing NASA with the simulation tools and technology to help NASA validate the guidance and navigation system both in the air, using subscale and full-scale flying models, and on the ground, linking the guidance hardware and software to a flight simulator.
The X-38, which resembles a miniature shuttle orbiter, is attached to the orbiting space station, from which it can be jettisoned quickly in an emergency like a lifeboat from a troubled ship. Like the shuttle, it re-enters the atmosphere shielded by ceramic tiles on heat-exposed surfaces and then maneuvers like a high-speed aircraft, using its "lifting body" shape and its fins and flaps.
Unlike the space shuttle, however, the X-38 requires no long runway for a landing. Fitted with skids instead of wheels, it is designed to land slowly and softly on unpaved and relatively short surfaces, thanks to the automated deployment of its 7,500 square-foot parafoil (so named because it combines the functions of a parachute and a wing or airfoil). The desired landing point can be preprogram-med into the vehicle's GPS-based navigation system, which steers the giant parachute by manipulating control lines connected to the parafoil's flaps.
NASA is using a multi-phase process for testing the guidance, navigation and control (GNC) software for the X-38. The first test phase is flying the GNC software on a powered, unmanned aerial vehicle (UAV) that also uses a parafoil for lift but is much smaller than the X-38. SwRI engineers integrated critical parts of the X-38 GNC software, provided by the European Space Agency, into the UAV autopilot. Like the X-38, it is steered through movement of the parafoil's two control lines. SwRI developed the UAV originally as a demonstration and test vehicle for military applications by outfitting a recreational ultralight aircraft with an autopilot and autonomous control system that navigates using a GPS receiver.
The UAV's parafoil wing is only one fifteenth of the size of the full-scale X-38 parafoil.The second test phase uses an 18,000-pound cargo pallet that can test the X-38's full-scale parafoil. It is dropped from a cargo aircraft such as a C-130 or C-141. The same SwRI autopilot that controls the small UAV also controls the pallet so that the GNC design can continue to gain maturity without the risks associated with rewriting the software for a different control.
For the third test phase, the X-38 lifting body is dropped from the wing of a B-52 bomber from 40,000 feet. The same GNC software tested in the previous phases is integrated into a comprehensive flight control system. By the time the lifting body is tested, much of the risk has been eliminated in the earlier and less costly test phases. The result is a high degree of confidence in the lifting body test--NASA's excellent track record on these tests proves that this process works.
A critical part of the preparation for UAV flights and pallet drops is a series of tests using SwRI's hardware-in-the-loop simulator (HILS). An HILS fools the embedded system into operating as it would in the real world. In other words, it would make an autopilot system think the aircraft is flying.
Sensor inputs, such as those from a GPS and magnetic compass, are synthesized from the simulator. Each simulation run tests a different function of the GNC software or the SwRI autopilot. All of the backup systems are exercised in this way to guarantee that a component failure during the flight will not result in an unsafe condition. The HILS can provide simulated failures of sensors and subsystems that cannot be mimicked by simple means such as disconnection from the system. This level of preflight preparation could not be accomplished without the simulator.
As a vehicle, the smaller SwRI-developed UAV has several advantages as a test bed for the control system: It costs much less to operate than either the pallet or the actual lifting body, and it can fly multiple missions in a single day without need of a cargo aircraft and crew to carry it aloft. The UAV completed 28 descents in the first week of testing. In contrast, several weeks of preparation are required for a single pallet drop or an X-38 drop.
SwRI will continue to support NASA in testing the X-38 design. Future developments under consideration include enhanced ground control station functions and navigational systems. These enhancements will be tested on the UAV and the simulator to provide maturity at a low cost.
Comments about this article? Contact Ogden at (210) 522-6928 or firstname.lastname@example.org.
Published in the Spring 2002 issue of Technology Today®, published by Southwest Research Institute. For more information, contact Joe Fohn.