---

Nonlinear factors such as air compressibility, leakage and friction make the control of pneumatic systems complex. Model-based robust control strategies are appropriate candidates for pneumatic systems, however in such controllers the measurement of state variables of the system are needed. In a pneumatic system the state variables are position and velocity of the actuator, and pressure in both sides of the cylinder. Pressure measurement is usually obtained by means of costly and low response sensors. A better way to deal with the measurement problem is to use observers to reconstruct the missing velocity and pressure signals. However the problem in a pneumatic system is that the system is not observable and pressure signals could not be observed by means of position signals only. To deal with this problem, in this paper, the pneumatic actuator is modeled as two separate chambers and the resulting subsystems are observable independently. High gain observers are designed for mentioned subsystems and for each chamber the pressure of the other chamber is considered as a disturbance. The input signal for each observer is the actuator position signal only. Finally a sliding-mode control strategy is designed for position tracking and experimental results verify that both controller and observer objectives are satisfied.

---

In this paper, position control is addressed for a pneumatically actuated 6-DoF Gough-Stewart parallel robot. At first, dynamic model of the pneumatic system of each link of the robot which comprises a pneumatic actuator and a proportional electrical control valve is extracted. Unknown parameters of the obtained dynamic model consisting friction force, viscous coefficient and the parameters of the valve are identified by employing an evolutionary algorithm. Then, position control of the robot’s pneumatic actuator is performed based on designing Backstepping-Sliding Mode controller according to the nonlinear dynamic model of the pneumatic system. Moreover, kinematic equations of the 6-DoF parallel robot are achieved and a novel method is proposed, the so-called Geometry-based Quasi-Forward Kinematic, to the end of calculating the position of the end-effector of the robot without using expensive position sensors. Accordingly, kinematic closed-loop control of the parallel robot, which is based on simultaneous joint space and task space control, is investigated for trajectory tracking using potentiometers, a rotation sensor, and based on the computed position of the end-effector by the proposed method. Desired sinusoidal trajectories with pure motions and also complicated trajectories are tracked in which error of positions and rotations are lower than 2 (cm) and 3 (deg), respectively. The results reveal that the trajectory tracking control of the pneumatic 6-DoF Gough-Stewart parallel robot is performed properly based on the proposed control strategies and the novel method for calculating the position of the end-effector.