Moving goalposts
So the aim was to build a prototype boat simulator. The motion platform was to have 6 degrees of freedom, and the software was to be a rigid inflatable boat that you could pilot around a choppy sea. It was to be ready to show by the end of November. It was a big ask to be honest as neither the motion platform or software existed at the time, and the end of November was around 3 weeks time. We at Blueflame Digital and our partners at Phase Four Engineering aren’t afraid of a challenge, so all was fine….until it wasn’t.
I got a call from James who had just been chatting with the guys at VRSS. “Hi James, what’s up?” I say. “Nothing’s up as such, apart from that prototype needs to be ready by a week Wednesday” he replies.
Ah.
So the big ask just became as James put it, a “very big ask”. Apart from he used an alternative word for “very”.
My main role in this project was to update our middleware software that takes game telemetry and converts it into data appropriate to be able to feed to the motion platform. Rich’s was to create some kind of boat-on-a-choppy-sea demo, and James’s was to build the motion platform.
James let us know that the basic design will be a hexapod, also called a stewart platform. This was enough to know that we needed to calculate the leg lengths to produce the desired motion. We set to work. My first job was to build some kind of visualisation tool to try to verify that the data I’m calculating is at least halfway right, then after that I figured out some calculations and came up with what I thought were reasonable leg lengths.
Meanwhile Rich set about make the boat demo…
Two days to go
On the Monday prior to the demo day, I hear from James. He’s made good progress and has figured out how the actuators will be driven. As this is a prototype, we are using another piece of software called LinuxCNC to create the low level signals needed to drive the motors on the actuators, and our software will feed it the higher level position data. I figured out how to send information to this software, but I needed some help with the set up and how to format the information in the correct way, so I whizzed over to James’s workshop.
Meanwhile Rich continued with the boat demo…
The first job is to get a computer up and running that has a parallel port. James’s go to guy Moe was on hand to rummage through the old computers that they have and came up with a viable machine, it just needed to be constructed from a couple of parts.
I’m not too bad with simple plugging stuff together tasks, so at least I could construct the computer and feel helpful in some small way. Then Moe and I worked through LinuxCNC to try to figure out how to set it up as it was new to both of us. Moe did most of this to be perfectly honest, but I followed along so that I knew at least enough to be able to tweak any settings on demo day.
By the end of the day, I had a bunch of tasks I needed to do ready for the demo, so I cracked on with that. I managed to get most things done that night, it was a little late, but progress was good.
The day before
On the Tuesday prior to the demo day, I went over to the James’s workshop to continue with my tasks. The guys were running around like crazy building things. Unfortunately I hit on a series of bugs. Up to this point I had been developing on Linux, but I needed the demo to run on a Windows machine. I got most of the way there, but there was one outstanding bug that I just couldn’t see.
Come 8pm that night, the motion platform was ready for first construction. Which was a really big reveal for me. I had seen parts of the platform, and a 3d model, but it isn’t quite the same as seeing the prototype for real.
Unfortunately there was no time to bask in it’s glory as we still had a demo to get done the next day, and the demo was in Telford which is a 3 1/2 hour drive for us. A quick test was performed, and it worked! FIRST TIME. The platform moved no problems at all. We even had a quick test of feeding the data to the platform from our software, albeit from Linux
So after the aforementioned 3 1/2 hour drive, we met up with the VRSS guys. The only thing we can really do at this point is unload and go to bed. Well. Apart from that damn bug I left outstanding. The demo is at 9am the next morning, so there will only really be time to set up and test, certainly no time for bug fixing. The only thing to do is to get settled in the hotel room and get out the metaphorical bug spray.
I got to bed at 2:30am that night, but I was content. I’d found the bug, and I’d fixed it! Hurrah! Also, I’d heard from Rich who had pulled out all the stops. He didn’t have a boat demo to show. Oh, no, he had three! Each one was a little different, but it covered all angles. Oh, and he’d set it up for use with the 3 projector screen that we’d be using. Nice.
Show time
The next morning after a chilly 7:00am start, we got set up. After a very very manic week, we’d managed to get everything constructed and going ready for the demo. Result!