NASA Jet Propulsion Laboratory California Institute of Technology JPL HOME EARTH SOLAR SYSTEM STARS & GALAXIES SCIENCE & TECHNOLOGY JPL Email News RSS Mobile Video
Follow this link to skip to the main content
JPL banner - links to JPL and CalTech
left nav graphic Overview Science Technology The Mission People Spotlights Events Multimedia All Mars
Mars for Kids
Mars for Students
Mars for Educators
Mars for Press
+ Mars Home
+ Rovers Home
Features
List of All Features
About Face: Rover Engineers Change the Rules for Driving

July 16, 2004

During testing, the right front wheel of a vehicle similar to the Spirit rover on Mars digs a trench in the sand when it is not turning.
Spirit front wheel digs a trench
During testing, the right front wheel of a vehicle similar to the Spirit rover on Mars digs a trench in the sand when it is not turning.
Image credit: NASA/JPL
Link to Full Res
Designing New Ways to Drive on Mars

When Joe Melko started design work four years ago on the six-wheeled rovers that are now on Mars, he didn't plan on steering one of them like a six-person river raft.

Five 1/2 months after landing the robots on Mars, that's what he and a team of engineers are doing. Now that the right front wheel on the Spirit rover is showing signs of wear, Melko and a team of assistants have been testing a surrogate rover at NASA's Jet Propulsion Laboratory to see how it performs on five wheels. No matter that the rover has gone six times the distance it was designed to drive on Mars. Scientists still have a lot of rock outcrops they'd like to investigate. And this team will keep that rover going as long as humanly possible.

To test the rover's performance on five wheels, they disable the sixth wheel. That sixth wheel drags rather than rolls, causing the whole vehicle to veer to the right.

"This is the equivalent of putting six people in a boat," says Melko, "three on one side, three on the other. One person on the right holds the paddle in the water and has the biggest paddle. Everybody else is rowing.

"The two other people on the right side need to row faster than the people on the left side, or the boat just turns to the right. On the left side, the person in front needs to reach ahead and row to the right and the person in back needs to reach behind and row to the left to keep the boat from going around in circles."


JPL engineers Joe Melko, left, and Eric Aguilar get a close-up look at the stuck right front wheel on one of the Mars rovers in JPL's In-Situ Instrument Lab.
Melko and Aguilar on ground
JPL engineers Joe Melko, left, and Eric Aguilar get a close-up look at the stuck right front wheel on one of the Mars rovers in JPL's In-Situ Instrument Lab.
Image credit: NASA/JPL
Link to Full Res
A similar situation exists on Mars. Engineers tell the rover to turn its wheels to the left to prevent it from veering to the right. They tell it that its wheels are straight even when they're not to prevent the rover's autonavigation system from straightening the wheels to go straight. If the rover straightens the wheels, it will veer to the right. Engineers then get together with software experts and create new commands to get the rover to do these things by itself. That's because Earthlings on this mission only get to send commands to Mars once a day.

If that sounds complicated, it is. But it gets more complicated if the rover is on a slope. Or if the disabled wheel encounters a rock. The wheel was designed to cling to rocks. That feature is great for climbing but when the wheel is not rolling, it hangs on to rocks. During tests at JPL's other-worldly In-Situ Instrument Lab, which looks a lot like Mars, Melko and his team test this by lying flat on the ground and holding loose rocks in place to see if the wheel can be dragged over them. They also test command sequences to see if the rover responds appropriately by freeing the wheel. Even a perfectly healthy rover with all six wheels working gets stuck on a sand-covered grade of 20 degrees. To minimize resistance, engineers drive the five-wheeled rover in reverse, pulling the stuck wheel behind the rover rather than letting it burrow into the soil ahead. Even so, the inoperative wheel gradually digs a trench and slows to a crawl. At that point, they re-engage the right front wheel to get the rover out of the trench.


JPL engineer Rich Petras shows University of Michigan engineering student Jessica Brooks how to operate a sensitive tracking device known as a theodolite.
Petras and Brooks with theodolite
JPL engineer Rich Petras shows University of Michigan engineering student Jessica Brooks how to operate a sensitive tracking device known as a theodolite.
Image credit: NASA/JPL
Link to Full Res
Steady at the Helm

Melko is a low-key, friendly, highly focused individual with broad shoulders, sandy hair, and a receding hairline. He's also the patient sort. Patience is an important quality in someone who has spent weeks conducting hundreds of tests on a limping vehicle to figure out how a similar vehicle will behave 240 million miles away. While running the tests, he gets several phone calls from mission planners asking him what to do next.

Melko also has lots of help. Fellow engineers and even students get in on the act. Jessica Brooks, from the University of Michigan, and Sarah Hornbeck, from Georgia Tech, both aerospace engineering students, take turns operating laser ranging devices, video monitors, and computer programs that record test data. When Melko is needed somewhere else, Chris Voorhees fills in to keep the tests moving. JPL avionics design engineer Eric Aguilar keeps a written record of observations about the rover's behavior and physically helps Melko hold the rover in place or pick it up and move it when needed. Mechanical engineers Rich Petras and Lee Magnone provide expertise with a theodolite, a highly sensitive, precision tracking system. Navigation experts Mark Maimone and Jeff Biesadecki translate the five-wheel driving performance into mathematical logic the rover's on-board computers can understand.


From left, JPL engineers Joe Melko, Eric Aguilar, and Georgia Tech engineering student Sarah Hornbeck go over data showing how the rover behaves on five wheels.
Melko, Aguilar, Hornbeck
From left, JPL engineers Joe Melko, Eric Aguilar, and Georgia Tech engineering student Sarah Hornbeck go over data showing how the rover behaves on five wheels.
Image credit: NASA/JPL
Link to Full Res
Melko, meanwhile, makes sure that the team runs enough tests of all the relevant numerical variations of distance, steepness, bearing, and surface friction to provide a useful indicator of how the rover will behave on Mars. A change in any one of these parameters means rover drivers need to change the amount by which they compensate for the rover's tendency to veer sideways. When they are only driving on five wheels, driving the rover backwards down a hill is the easiest task. Driving uphill is the toughest. Driving on flat, smooth rock surfaces is easier than driving on sand. Driving laterally across a slope presents a special challenge because the vehicle tends to slip slightly downward while moving in a relatively straight line.

Every one of these conditions requires a different set of commands. The commands are essentially if-then statements in computer logic – if this is your position, then do this; if this is not your position, then do something else.


JPL engineers Eric Aguilar, left, and Joe Melko monitor the rover's performance on a sandy slope outside JPL's In-Situ Instrument Lab.
Aguilar and Melko monitor rover
JPL engineers Eric Aguilar, left, and Joe Melko monitor the rover's performance on a sandy slope outside JPL's In-Situ Instrument Lab.
Image credit: NASA/JPL
Link to Full Res
Mars Mechanics from Millions of Miles Away

Engineers knew they had a problem in June when Spirit's right front wheel began using more energy than the other five wheels.

"It's taking up to 2 1/2 times as much energy to turn the wheel," says Melko. "And that tells us there's something definitely wrong. It's not an issue that it takes too much energy out of the rover. The rover has enough energy to handle it. It's more like hearing a noise in your car and thinking, ‘How much longer can I wait before I have to take it to the repair shop?' In this case, there's no repair shop."

Engineers first tried to fix the problem in July by heating the lubricant in the gear assembly to get it to spread to all the gears. This produced only a small improvement. It turns out there's only one kind of lubricant on the market that could be used for this mission. The lubricant can both be heated to more than 100 degrees C., which was necessary to bake out any impurities for planetary protection before landing the rover on Mars, and still survive the martian cold. It's not your typical 10W-40 motor oil. At temperatures of –50 degrees C., the lubricant, known as Bray Oil, becomes very sticky. So designers applied only a very thin layer of the lubricant to the wheel parts, just enough to provide a smooth surface but not thick enough to congeal and hamper the drive train.

That meant there was less lubricant available. During Spirit's trek across the martian surface, it's possible that a piece of debris got past the seals into the drive train. It's possible the gears created their own debris as they began to wear down. It's possible the drive train is running out of lubricant. Engineers don't know the specific cause of the problem without actually seeing the hardware, but all indicators are that the drive is just gracefully wearing out.


Engineering student Jessica Brooks lines up the targeting system in a theodolite with four reflectors on the rover outside JPL's In-Situ Instrument Lab.
Brooks takes theodolite measurement
Engineering student Jessica Brooks lines up the targeting system in a theodolite with four reflectors on the rover outside JPL's In-Situ Instrument Lab.
Image credit: NASA/JPL
Link to Full Res
"We knew when the wheel drives were designed that they were not being designed for a tremendously long life," explains Melko. "Other criteria, like moving when it's really cold, packaging it into a small area, reducing its weight, working within the amount of time we had for the projects, using the kind of gear drives we could get, all of those had to be coupled and balanced together in how we did things."

The other five wheel drives show no change in performance. "It's a lot like car performance," he adds. "Some cars last 200,000 miles. Some fall apart at 50,000 miles. Some last 400,000 miles. And that wheel on that particular rover has decided to show itself as the outlier. The good news is this wheel has driven the rover more than six times the amount pledged in the warranty and still has more life left."


The Spirit rover continues to explore Mars on five wheels as needed, leaving a shallow trench in the wake of the disabled sixth wheel.
Spirit's right front wheel on Mars
The Spirit rover continues to explore Mars on five wheels as needed, leaving a shallow trench in the wake of the disabled sixth wheel.
Image credit: NASA/JPL
Link to Full Res
Helping Spirit Last for the Long Haul

Melko now estimates that the right front wheel will continue operating for another 200 meters (650 feet) to 1,700 meters (1 mile) if used on all rover drives. To extend the life of the wheel, the team will drive the rover on five wheels whenever possible and use the sixth wheel for maneuvers that can't be done efficiently without it. Examples include driving uphill and placing the rover's scientific tool kit accurately over a target.

Engineers will do their best to keep the rover going as long as possible. That endeavor has defined their careers for the past few years and is their signature contribution to a mission that has changed the history of exploration on Mars. When the Spirit rover landed on Mars in early January and took a look around, everyone was awestruck.

"It was great," says Melko. "We saw the hills in the background when we landed and we managed to get all the way there. While we were sitting at the mechanical operations station, (principal investigator) Steve Squyres came over and said, ‘See those hills over there? That's where I want to go.'

"We had already figured out that the hills were about 3 kilometers (1.8 miles) away. So we said, ‘Well, let's get some data and then we'll give it a try.' We're glad we got him there, and we think there's still a good amount of life left in that wheel. And if begins to lose more capability, we'll still do science, just not on as many targets on high slopes."


USA.gov
PRIVACY    |     FAQ    |     SITEMAP    |     CREDITS