Sure we could try that. I'm really not worried at all about the release mechanisms though. I've had plenty of time to think it over and I have a few mechanisms that can work alongside each other without interfering. As long as one works each time that's all that will matter. If failure occurs in a mechanism more than once in a hundred test runs we won't use it. That way the chances of real failure will be less than (1\100)^N where "N" = the number of active fail-safes. I think a minimum of 5 fail-safes should be used, making the chance of total failure less than one in ten billion. Safe enough for ya? 🙂
I've been thinking of the robotic hand and I just remembered one of the reasons I hadn't seriously considered it before: programming. The great advantage of a hand is that even a simple one is capable of more complex movements than an electric toothbrush on a stick, but that means you can't just rely on random movement for it to work; I'll have to program finger strokes. Given that the rest of the programming is easy enough and I can insert any kind of module coding into the program without having to rethink the overall design, I'm prepared to give it a go. If it works it should be awesome, but if not it may cost a lot of time, money and effort. But ain't that the philosophy we work by.
What I think I'll do is program stroke patterns both for individual fingers and the whole hand, so that you can get both one finger doing its thing or all of them working together in sync. It should just be a matter of timed servo commands. I can cook up a bunch of different actions or templates to stick in the primary coding level for that module, then they'll be activated and coordinated from the secondary level. If I'm right, the only difficulty will be in working out how to make the hands tickle and how to turn that into a sequence of code.
I've been thinking of the robotic hand and I just remembered one of the reasons I hadn't seriously considered it before: programming. The great advantage of a hand is that even a simple one is capable of more complex movements than an electric toothbrush on a stick, but that means you can't just rely on random movement for it to work; I'll have to program finger strokes. Given that the rest of the programming is easy enough and I can insert any kind of module coding into the program without having to rethink the overall design, I'm prepared to give it a go. If it works it should be awesome, but if not it may cost a lot of time, money and effort. But ain't that the philosophy we work by.
What I think I'll do is program stroke patterns both for individual fingers and the whole hand, so that you can get both one finger doing its thing or all of them working together in sync. It should just be a matter of timed servo commands. I can cook up a bunch of different actions or templates to stick in the primary coding level for that module, then they'll be activated and coordinated from the secondary level. If I'm right, the only difficulty will be in working out how to make the hands tickle and how to turn that into a sequence of code.



