Getting robots to carry out repetitive tasks precisely and untiringly is relatively easy. Getting them to do it in the most efficient way possible is another thing entirely. Unai Ugalde-Olea, lecturer in the Department of Electronics Technology of the UPV/EHU-University of the Basque Country is working on a new way of controlling robot joints that promises energy savings of up to 40 percent by changing the way the robot handles instructions.
In one sense, robots seem like simple things. A computer gives an instruction and the robot carries it out. What could be simpler? In fact, industrial robots are designed to be very flexible, yet have to operate operate with a high degree of precision, so they can’t use simple stop/go instructions.
Instead, the joints are operated by digital controllers that send orders to the joint motors in a series of pulses, which ensures that the robot maintains the correct position and angle. Effectively, this means that the robot keeps executing the same order over and over again in what’s called zero-order reconstruction.
Ugalde-Olea’s approach is to design a control system that can handle not just the last order, but also the one before that, so that the orders given to the robot aren't a jerky string of move, pause, move commands, but a sequence that forms a constant line called a fractional-order reconstruction. This allows the robot to anticipate its path, so the joints can maintain precision while providing energy savings of up to 40 percent.
Ugalde-Olea says that this approach has been used in the laboratory on small prototypes, with the approach having potential applications for robots with flexible joints such as wrists and elbows.
"The next step would be to check whether it is possible to achieve similar savings on an industrial scale," says Ugalde-Olea. "In any case, even if the savings are lower, in view of the cost of electrical power, it would be a significant step forward."
Ugalde’s findings were published in Control Engineering Practice.
See the stories that matter in your inbox every morning