Once you've got some idea of where you want to be you've got to come up with some clever ideas about how to get there. There's plenty of different ways to come up with a decent set of concepts like brainstorming, the five Ws and thinking-hats, just for starters. It's worth thinking about which method of concept generation is going to be best for your design and for the group of people that are doing it.
Design isn't a one-size-fits-all process.
As this was quite a small project I opted for sitting down with a pencil and a few sheets of paper. The concepts that I generated were very fluid and didn't go through a stage of ideas, then combinations of those ideas and iterations. Everything just happened at once, usually in a mess of scribbled notes and drawings.
There's all sorts of ways to add something that can run software to a robot. Probably the two most popular options right now are Raspberry Pi and Ardiuno. Now the RPi is a very competent device, but its capabilities and complexity were well beyond what I'd be looking for in this robot. Ah, Arduino!
That leaves us with the small matter of how The Luggage will physically fit together. This is mechanical engineering, and I'm not great at it.
From my experiences with Hexy, it was immediately obvious that I didn't want to have The Luggage using its own legs for locomotion. Instead I opted for some method of wheeled movement and its little legs slaved off these wheels somehow...
"Somehow" being the operative term. At the outset, I really didn't have any good idea about how to make the legs move when The Luggage moved around. Of course, from our requirements, this was an essential feature.
My afore mentioned poor ability to do mechanical engineering is slightly harsh: I can mechanical engineer (to some level), but I'm terrible at doing KISS1 mechanical engineering. No, not that KISS.
KISS stands for: Keep It Simple, Stupid.
This design principle means that for a given problem, very often the simplest solution will be the best solution. If you think that your current candidate solution is looking too complicated, perhaps there's a simpler solution out there2. Finding it is usually a pretty tricky creative exercise. Fortunately, my friendly mechanical engineering colleagues were around again to steer me in the right direction.
Over a nice brew, we came up with a concept that would see the legs mechanically linked to the movement of the drive wheels. This should only3 need a few gears and a crank, rather than any of the outlandish concepts that I'd manage to produce involving individual motors per leg or hydraulic actuation.
At this stage, the design has now progressed sufficiently to allow us to move into
building the robot detailed design of the various components. The next few posts will go into how each of the components actually came into being.
1Definitely mentioned this before.
2Of course, sometimes the correct solution seems like something totally mad, but it's still the best answer. In these situations you've got to be very sure about your calculations and design decisions. Fortunately NASA is populated by rocket scientists.
3This a warning word that you should look out for when talking to engineers. It's often a prefix for a doubling of the budget or timescales. Or both.