Scientists Create New, More Versatile Control Algorithms for Autonomous Systems
Work at the Universidad Carlos III de Madrid has resulted in a new class of algorithms that are able to obtain a greater number of design specifications, which should make them suitable for numerous industrial applications.
By Robotics Trends' News Sources - Filed Dec 06, 2011

In addition to evaluating applications in industrial robotics, scientists at Universidad Carlos III de Madrid are applying the new class of algorithms they have developed to TEO (Task Environment Operator), the humanoid robot created at the university’s RoboticsLab. (Credit: UC3M RoboticsLab)


Standard controllers used in industrial processes generally have derivative and integrative control actions with whole, preset orders. In contrast, the new type of controller that scientists at the Universidad Carlos III de Madrid (UC3M) propose is based on a generalization of those whole orders and others which are not whole, known as fractionary, which allow for an increase in the number of control parameters that can be auto-tuned. "This type of controller is characterized as having fractionary derivatives and integrals (0.7 or 1.8, for example), meaning they are more versatile controllers, which allow us to obtain a greater number of design specifications", says the main author of the research report, Concepción A. Monje Micharet, a professor UC3M Department of Systems Engineering and Automation. 

The methodology Monje has proposed allows these controllers to be auto-tuned for a wide variety of systems. "The most innovative aspect of this method is that it is experimental and it can be tuned using just a few simple steps", explains the researcher, who, for this work, recently received IFAC’s (International Federation of Automatic Control) prize for the Best Paper in the last three years (2008-2010) from the journal Control Engineering Practice.

The study, carried out by the researcher at the Universidad de Extremadura in collaboration with the Universidad de Castilla-La Mancha and Utah State University (USA), has received over fifty awards since its publication. 

There are countless applications for this type of control, according to the researchers, who have implemented it in areas such as the regulation of liquid level systems, the control of servomotors, the operation of flexible robotic arms, the operation of unmanned aircraft, and the regulation of pneumatic systems, among other examples. "These controls can compete in any field of application," concludes Monje, who is currently trying to apply this development to controlling the stability of the humanoid robot TEO, whose prototype is taking shape in the Robotics Lab at UC3M. 

When starting this research, the scientists did a complete review of the existing publications and research on fractional control and calculus. Following this initial analysis, they defined their objectives, which were focused on proposing a generalist method for auto-tuning fractional controls that would work for very diverse types of systems. "The most complicated part was developing a tuning method which, in addition to working adequately, could do so even with serious computational restrictions," they explain; because one of the advantages of this method is that its resolution uses very simple lineal equations that can be computed quickly, making it easy to implement in any programmable logic device. 

SOURCE: Universidad Carlos III de Madrid



<< Return to story