And there was much rejoicing
I finally got my research platform to work predictably today. Since the major problem (not sure if I mentioned what it was before) was significant lag between the robots' perceived positions and their actual positions, I thought I should probably try running the system on a faster machine in order to verify where the problem might be. If the other machine was able to run the system smoothly, then I would be able to tell that the problem was with the old machine-- if not, well, then it's still the same old can of worms.
Fortunately, one of the guys we work with just got a job somewhere else, so his nice quad-core, nearly brand-new computer was open for use. I snagged it, got it hooked up, and started fiddling around with getting the system set up on it. No small task, unfortunately. The platform has about 4 cross dependencies, none of which are documented, and only one of which is obvious (mostly). Fortunately I had already figured out about two of the cross dependencies, the 3rd was just a matter of going to download the drivers from the manufacturer's website and hope I got the right version, and the fourth involved a whole lot of poking and testing.
However, I finally managed to beat it into submission, and the thing worked flawlessly. It was really good to be able to fix it and have it work-- now I just have to see if my advisor wants to just give me the test computer permanently, or if I'm going to have to try to convince him to buy a new one, which could be a bit of a pain.
Fortunately, though, now I have an installation document detailing how to install the system onto a new machine. Now nobody should ever have to go through that again.
Now maybe I can get this academic paper started/written before April 1st...
Fortunately, one of the guys we work with just got a job somewhere else, so his nice quad-core, nearly brand-new computer was open for use. I snagged it, got it hooked up, and started fiddling around with getting the system set up on it. No small task, unfortunately. The platform has about 4 cross dependencies, none of which are documented, and only one of which is obvious (mostly). Fortunately I had already figured out about two of the cross dependencies, the 3rd was just a matter of going to download the drivers from the manufacturer's website and hope I got the right version, and the fourth involved a whole lot of poking and testing.
However, I finally managed to beat it into submission, and the thing worked flawlessly. It was really good to be able to fix it and have it work-- now I just have to see if my advisor wants to just give me the test computer permanently, or if I'm going to have to try to convince him to buy a new one, which could be a bit of a pain.
Fortunately, though, now I have an installation document detailing how to install the system onto a new machine. Now nobody should ever have to go through that again.
Now maybe I can get this academic paper started/written before April 1st...