Xombie Test Flight

Successful Test Flight for Mars Landing Technology

So we've got a robot space probe approaching Mars, and it wants to land at a particular spot. How do you find that spot? Use GPS you say, but that only works on Earth. You could use inertial navigation, but that is only accurate to a certain point. The farther you travel from your last known point, the farther off you could be. If your last known point is on Earth, by the time you got to Mars you could be a zillion miles off course. Besides, I am not sure inertial navigation would even work in deep space. Most of the way to Mars you are going to be in free-fall and subject to the gravitic whims of the solar system. Your inertial navigation system will think you are in free fall, not subject to any acceleration in any direction.
    Anyway, getting to Mars isn't the problem, we know how to do that. The problem is locating the spot on the planet where we want to land. These guys are using a camera to look at the ground and compare it to a picture they have on board. It will probably have to be a picture taken by a probe, nothing from Earth is going to have enough resolution. It's weird how much processing power it takes to recognize an image, considering that you and I do it instantly and continuously.
     The best part of the video is watching the rocket motor twitch this way and that to keep the spacecraft balanced. All big rockets do this but you hardly ever get to see it being done.

     Just for grins, codingame dot com had a programming puzzle that mimics this situation, though vastly simplified. It only operates in 2-D, the terrain mapping has already been done, and we have a fairly large area to land in. You can watch my solution play out here.