Interview | |||
Q 2: | We've all heard about the 15-minute delay of images from Mars. What
was that like? | ||
A 2: ![]() |
That's a very important point. One thing we cannot do is expect instantaneous action from our
control inputs. Because of the speed-of-light delay was about 10 to 11
minutes at the very beginning of the mission, and it got longer as we went
on.
It's critical that we compensate for it. If we didn't have the delay, we would drive the Rover like we drive a radio control car on Earth, where you get a live video image, and we'd see it was coming close to a cliff, and tell it to do an emergency stop. That would work fine on Earth with no time delay. But with the time delay, realize that information is 10 to 15 minutes old, and unless the Rover is going very, very slowly, by the time you say stop, it's now too late, the rover has gone over the cliff. So, we found a scheme to get around that. We developed a system, called computer-aided remote driving, or CARD for short. We're telling the Rover where to go, but we're not micro-managing it. We give the Rover discrete waypoints (or goalpoints) to achieve, described in terms of meters in X and Y from the Lander center. But, how it gets from one point to the other, we're leaving up to the Rover. It has better local sensing. It knows what dangers are being inflicted upon it. Is it tilted too much, is it headed toward the edge of a cliff, is it trying to go over a rock that's too large? These are things that, although we can sense with the initial camera staring at viewpoint, we have to rely on the Rover to make the low-level decision, and keep it from getting into trouble. Other people have done experiments on Earth where they would remotely drive the a robot in Hawaii from a control center in California, going over the lava flows, but unless you take that time delay into account and fundamentally deal with it, you really can't drive the rover on another planet that way. | ||
|