Force control in a pneumatic system using hybrid adaptive neuro-fuzzy model reference control

The dynamic model of a force control system depends not only on the actuator but also on the manipulated object. Generally the object dynamics are unknown and frequently change, and it therefore is difficult to identify the actual dynamics of the system. Designing a nonadaptive controller to perform well under unknown and variant conditions is not feasible. To enhance the controller performance for the pneumatic system, in this paper, a hybrid adaptive neuro-fuzzy model reference is proposed. The concept of multimode switching is applied to activate either a 'bang-bang' controller or an adaptive neuro-fuzzy model reference controller (ANFMRC). Bang-bang control is applied when the actual output is far away from set point, which results in high error value. In this mode, fast tracking of the output is required. ANFMRC is applied in medium and small error ranges to perform a good response. In this paper, two types of controller; ANFMRC and hybrid ANFMRC are investigated. Simulation and experiment of a force controlled pneumatic system is investigated to evaluate the efficiency of the algorithm.