I’ve been thinking about a project that i would like to do for a few weeks now, and to when i brought the idea up in a lecture, i was told that the idea of mine is not currently being used in industry, which means that there might be quite a market for the product if its done well.
We have been learning how to control robots (namely robotic arms) using control theory and the techniques used.
This is not easy stuff, by the way, a robot arm with 6 degrees of freedom is very complex.
What got to me was the amount of times that because the robot is so complex (and it has to be to allow sufficent movement and a correct inverse kinematic model) that many calculations are done by assuming the worst case scenario, which reduces the performance of the robots performance (namely speed of response and overshoot) in every case which is better than the worst i.e the robot is assumed to be carrying maximum load for the calculations, so when it is not it performs worse because the transfer function isnt optimal anymore.
What i would like to do is try and produce a program which controls the robot, a software controller rather than a hardware controller, which takes in more inputs from the robot and calculates the optimal transfer function every time the robot moves, or picks up an item. Which would improve the performance of the robot considerably when the load is reduced, or the couples are higher.
Computers are certainly powerful enough to do this kind of task i believe, and i think this could be a very worthy project indeed, but i dont know how to go about it. I’m currently stuck into my final year project, so i’ll have to wait until after i finish my exams, but i would also need the resources like a robot and access to the infomation as well, i think uni would definatly be a great place to do this, but how?
A Phd? another degree? a masters? i dont really know so..
I’m going to speak to the head of robotics about it, see what he thinks, and hopefully he can advise on what to do, or how to go about doing it.