Saturday, November 07, 2015

C@SScot 15: Opening Keynote

C@SScotland Conference - Keynote: Dr Iain Martin, University of Dundee
How Can You Test an Autonomous Planetary Lander?
Live captured

{Biggest attendance at C@SS conference so far. Reflecting a growing interest in Computing or confidence in our subject?}

University of Dundee has a space technology centre which collects and archives loads of data. Among other things, they have a space systems research group which looks at designing autonomous planetary landers. Computational Thinking underpins their work.

Problem Analysis:
Major task and very difficult to prove you can do it. Very high stakes, many possible reasons it can fail and very expensive to try. The lander has to be able to land autonomously (time delays and limited knowledge of what you are landing on!). You have to balance amount of fuel needed for manoeuvring and landing with the sensors and science stuff you want the lander to carry. The lander will need ”a whole bunch of sensors" to detect position and hazards. Cameras are a low cost, lightweight sensor (with no moving parts - a good thing!). Need to process the images. Difficult to prove the tech works. Best Mars landing so far is MSL which still had a 6km landing eclipse - would really like to get better! One way of testing is to create simulated data. This is difficult. Comparison was made with Apollo 11 mission. Landing site was Boulder strewn but Armstrong was able to steer to a plain just beyond the crater with seconds of fuel left!

{Loads more stuff showing how Computational Thinking underpins major engineering projects like this but I occidentally deleted it. Oops!}

No comments: