Researchers with Huazhong University of Science and Technology and the University of California at Berkeley have developed a new feedforward method that improves on conventional feedforward techniques.
The “feedback” mechanism, a more widely known process is where a machine corrects its actions after observing an error in the performance. For example, when a thermostat exceeds a set point, the systems is turned off.
Such systems provide crude mechanisms of control and if the preferred range is set narrowly the machine will have to bounce between turning on and off, increasing probability of it being worn out faster.
“Feedforward” control systems are less well known, but they overcome the shortcomings of feedback mechanisms. Feedforward mechanisms measure changes in the overall system in which a given machine is operating in, and directs the machine to take these changes into account before they have a chance to have an undesired impact.
Engineer Zhang Dailin explains that the feedforward control is akin to a baseball player tracking a ball’s trajectory, adjusting his movements in anticipations of where it will land.
Feedforward mechanisms used to control machines are much more accurate than feedback mechanisms, but they can be computationally hard. This is due to the need for a good understanding of whole-system modelling when developing feedforward control systems.
The team of researchers from Huazhong University of Science and Technology and the University of California at Berkeley developed a new method for feedforward mechanisms with improvements and is set to be tried out on industrial robots and machine vision.
Normally, the parameters that are fed into such a technique are fixed, but the new technique obtains such parameters from an uncertain environment to develop an updated model of the system in real time via a series of steps.
Tested on a simulation of a motor, the researchers found that the technique did indeed achieve better performance than the traditional feedforward mechanism.
With specialists in machine vision and improving the calibration of industrial robots in the research team, they are looking to trial their technique on real-world challenges. [APBN]