FLOWX Engineer 86-21-54150349
Abstract— pneumatic actuators configuration is simple and they are robust, and have good energetic properties because of the air compressibility, and are relatively cheap. Despite these advantages these cheap pneumatic actuators are dif?cult to control – pressure dynamics have typical timescales on the order of 100ms, and this delay can severely harm simplistic control methods. To solve this problem, a model based controller with a good model of the pressure dynamics is used. Here we provide a general parametric model of these mechanics based on both a theoretical analysis and an empirical study with man-like robot.
I. INTRODUCTION AND RELATED WORK
Pneumatic actuators are popular for several reasons.
Naturally back-drivable, low- friction, adjustable compliance and very robust, they have a high strength to-weight ratio – for example a typical cylinder of 5 cm diameter weighing 100 grams, running at a standard 85 psi (=590 KPa) above room pressure, produces 1160 Newtons or260 pounds of force. In addition, the simple configuration of pneumatics makes them cheap pneumatic actuators.
The biggest disadvantage and difficulty of this cheap pneumatic actuator is that they operate much more slowly than electric motors or hydraulics, with dynamic timescales on the order of 100ms. To properly control a cheap pneumatic system ,a good model of these dynamics is required. With regard to Models of such systems they can generally be categorized into physical or parametric models. The former models are constructed based on ?rst principles and endeavors to be consistent as closely as possible to the underlying physical system. Parametric models function with unknown constants that equals to using a curve-?tting procedure. Although both types of models are predicted to have good qualities, the design objectives are different. The physical model aims at accurately obtain all the physical properties, regardless of the miscellaneous importance they hold for prediction. Apart from focusing on predictive power, the design of a parametric model must also take into consideration secondary target. For example, ensuring good convergence and eliminating local minima in the parameter space.
Previous study has paid focal attention either on precise physical models of pneumatic systems or on linear parametric models. By this research, first, we construct a physical model based on ?rst principles, and then employ such model in the guidance of the design of a non-linear parametric model. Now the research most close to ours is the study where the basis for the non-linear parametrization is quadratic polynomial. Instead of general polynomials, we utilize specially elaborated functions, chosen to conform to the prediction the original physical model.
II. PHYSICAL PNEUMATICS MODEL
A. Ports, Valves and Chambers
A pneumatic cylinder is a device with two chambers with a sliding hole in the middle of them. Whereas, in the chamber the air pressure is controlled by valve connecting the chamber to one of the two ports: the port of supply connects the chamber to a compressor and the port of emission connects the chamber to pressure room. By some installation a single valve with two output ports is in joint with both chambers of a cylinder, letting high pressure in either chamber, but not both.
Another design was to use proportional valves rather than binary valves with a Pulse Width Modulation scheme. Proportional valves offer ?ne-grained control of the port size, and are also less noisy than a PWM setup.