Dr. Greg Chavers, test lead at the Marshall Space Flight Center in Huntsville, Alabama, helped to design the “Mighty Eagle” robotic prototype lander. The vehicle, which can guide itself to a specified target, flew “open loop” to an altitude of 100 feet in late August.
NASA Tech Briefs: What is the Mighty Eagle?
Dr. Greg Chavers: The Mighty Eagle is a test vehicle, and it was built originally to demonstrate that we can control a small vehicle that is dynamically similar to a small robotic lander that could land on the moon or other airless body. We started with a flight design concept and built this vehicle with a propulsion system that uses pulse-width modulated thrust, with very fast-acting valves so they’re either on or off. They’re not throttled to control the altitude and the attitude of the vehicle.
It was built as a demonstrator for the control algorithms. However, to leverage more of the design team, we actually included landing legs, and operate the vehicles to be very similar to what the spaceflight implementation would be.
Over the past six years we’ve been teaming with Johns Hopkins University Applied Physics Lab to develop the actual spaceflight concepts and start design on what we would actually do. We designed this vehicle about three years ago, and 18 months after we started the design, we started flying it.
We’ve done almost 30 flights now. In the latest series, since we demonstrated the controllability, we have an onboard camera. We can take the optical images that are onboard the vehicle and process them through the onboard computer. It can update its guidance algorithms based on what it sees in the images so that it can fly to a target and land. The Mighty Eagle is just a testbed or test vehicle which is similar in size and characteristics to what the real spaceflight implementation would be.
NTB: What is the most unique aspect of this lander design?
Dr. Chavers: The most unique aspect is the pulsed thrusters. A lot of landers, if you look at the Masten Zombie or Morpheus, have throttled engines, and they’re designed more for the bigger landers that would go and land on, say, the moon. This one specifically was designed to be a small, simple, robotic lander. It has a very high thrust-to-weight ratio, and very small thrusters are used. This one targets more of the science and the precursor-type missions to landing payloads on perhaps the lunar surface.
NTB: You mentioned the onboard camera. How is the lander able to navigate autonomously?
Dr. Chavers: We have accelerometers. We fly inertially, using an inertial measurement unit. Inside that are the accelerometers. It knows that the motion of the vehicle is changing by the acceleration that the accelerometers detect. From that, we can calculate what the velocity is and what the position is; that’s flying inertially.
We put in a profile to fly, but because of inaccuracies in the inertial measurement unit, we also have an onboard sensor. It’s a radar altimeter. After several seconds, errors build up when you fly inertially. The radar altimeter updates our navigation solution such that we can accurately determine the altitude, and the onboard camera provides the guidance.
For the previous test series, we had a pre-defined target that was painted on the ground. It knows what the target is supposed to be. For example, if we were doing orbital debris, or capturing a dead satellite, the camera would know the image it would be looking for, and could update a vehicle’s navigation to track to that vehicle. And that’s what we had shown. It’ll take an image, and look for a specific target, and then update your navigation solution to get you to that target, and provide those commands to the thrusters completely autonomously.
NTB: Can you take us through your most recent test run and how that went?
Dr. Chavers: We finished up the controllability demonstration in November of 2011. We started building for this latest test run in the spring. We actually changed locations. We went from the Redstone Test Center on Redstone Arsenal to the Marshall Space Flight Center test area, which is still on Redstone Arsenal but it’s on NASA property.
So in this series, using the same vehicle, we had an area that gave us enough room to do the vertical and lateral translations. We added the targets to the ground again, and we also added in several young engineers. The vehicle was already in operation. We took this opportunity to bring in several engineers that had been out of school for two or three years. They worked at Marshall, we trained them, and we put them in lead positions for the test operations.
We did some checkouts because the landing hadn’t flown in eight months. We went right into the test series. We picked it up with a crane to about 100 feet, so we could see what the camera would see without actually having to fly it, and verify that our images were crisp and clear, that the camera focus was correct and so forth. We loaded some propellant, did two low hover tests to make sure all systems were good, and then we went up to 30 feet, and flew open loop. The guidance didn’t update, but it was running in the background and got optical solutions. After that flight, we ran it back thorough a simulation to verify that the optical solutions would provide the correct change in path, so we didn’t go outside of our test area.
And on our next flight, we closed the loop so that the lander autonomously flew to the target. Then we came back and flew open loop again at 100 feet on August the 28th. Last week, Sept 5th, we flew to 1000 feet closed loop again. We completed that part of the test series.
NTB: Through that whole process, what was your biggest technical challenge?
Dr. Chavers: There are two challenges. This optical guidance is very similar to an automated rendezvous and capture requirement, in order to provide autonomous updates to a vehicle on orbit. We were developing techniques and algorithms using a very inexpensive camera that we have onboard. We had challenges to go through with the image processing because our onboard computer has limited processing capability. We’re actually using a RAD750, which is a spaceflight-qualified processor. It is very sturdy, but limited in processing capability for such a small package. One challenge was to process the images with that computer, while still proving the controllability. The computer does operate all the onboard systems.
The other challenge is to make sure all the subsystems work correctly, because this is a vehicle system. Any subsystem can malfunction and cause the whole system to have a bad day. Having flown almost 30 times, one of the challenges is to make sure our younger, new engineers pay attention to the trends of the performance system and treat it almost like a launch vehicle.
NTB: How many engineers worked on this project, and how did these teams work together?
Dr. Chavers: When we first started, we were building this as a demonstrator for our controllability, but we also had assembled a design team to design the real spaceflight implementation of this. We were put on hold because of budget shortfalls within the Lunar Quest Program. We had the team assembled from Marshall Space Flight Center and Johns Hopkins University Applied Physics Lab, and we had not built hardware directly with each other before. We had to go through the processes of how each organization designs hardware, integrates hardware, and operates hardware. That’s how we intentionally went about with the teaming.
And we actually included as much of the spaceflight implementation we could afford at the team. We did landing legs with energy absorption, which was very similar to what we would do on the moon. We used the same software and ground system implementation we would use when we were going to the moon. We did not use the exact same propulsion system, because the propulsion system for the moon is very high performing and a little more costly than what we use for the test vehicle. For the test vehicle, we’re using 90 percent hydrogen peroxide -- very simple system, very green, and not expensive at all, so we can do many test flights with a low budget. It does simulate what the high performance thrusters would do, in that they’re pulsed operation. It’s just that the specific impulse is not good. In other woods, the gas mileage is not as good for that propellant.
We also included the Von Braun Center for Science & Innovation. The Teledyne Brown Engineering here in Huntsville also participated in the testbed, to actually go fabricate these non-flight like thrusters that we were using. It was a diverse team, and organizationally challenging to do a small quick project like this. The people who were on the team were all behind it, so we were successful just because the people on the team wanted it to be successful, and we all worked together. The organizational diversity was challenging, but we clearly stated who the lead on the team was, and who had the technical authority. Everyone on the team recognized that, and we were able to move forward quickly.
NTB: How do you see these being used in the future?
Dr. Chavers: For this testbed, we’ve already demonstrated controllability. We’ve already demonstrated that we can update the guidance autonomously in real time. One of the challenges is to demonstrate that we can do hazard avoidance with the testbed. Our friends that are working Morpheus are planning to demonstrate the Autonomous Landing and Hazard Avoidance (ALHAT) technology.
This technology is built for small robotic landers, so for hazard avoidance, it would have to be a very low-mass, low-power system to do that technique. We’re thinking about ways to do that, and working with our friends at Jet Propulsion Laboratory and the Applied Physics Lab as well.
The future for this testbed is allowing engineers to get some hands-on hardware [experience] using the full six-degree of freedom vehicle. It would be a demonstration, not a verification or validation unit for demonstrating other sensors. For the robotic lander project at Marshall, this has been a very low percentage of our investment: About five percent of the total investment on small robotic landers went to the testbed. We have other subsystems that we are maturing, and that are ready for infusion into a real spaceflight implementation.
NTB: You mentioned debris capturing. Are there other applications for a lander that has the autonomous rendezvous and capture capabilities?
Dr. Chavers: Maintenance on satellites. Refueling the satellites. Actually getting to assets on orbit if there are communication issues with those satellites, to retrieve them and get them back to a space station to be repaired and redeployed.
Landing on the moon or rendezvousing with an asteroid is one application. Landing on the moon, trying to navigate to a specific target on the moon, is an application. There are both landed and free-flight orbiting applications for this technology.
NTB: What are you working on now? What are your day-to-day responsibilities?
Dr. Chavers: The Science Mission Directorate at headquarters is what has funded this effort for the past several years. Due to the budget challenges that we currently have, my day-to-day activities now include informing the leaders at NASA of the capabilities that we’ve acquired, and the knowledge that we’ve learned through this testbed, as well as the testing of other subsystems, to infuse these to accomplish NASA’s near-term mission. In particular, we understand that we have enough technology now to build a small lander to land on the moon within the next three years, and the challenge is basically funding. There aren’t technical challenges involved in that right now.
NTB: What is your favorite part of the job?
Dr. Chavers: The actual test of the Mighty Eagle is very exciting. It’s a big adrenaline rush to know that something you’ve worked on, something that we’ve built on very low budget, is actually operating. It’s exciting when the system works. But the best part of the day is the team that I have to work with. Everybody is very optimistic and wants to contribute. It’s nice being a part of something that’s bigger than you -- a team that’s bigger than the sum of the parts.
To download this interview as a podcast, click here