Control and Optimization of a Sensor Manufacturing Process

This paper presents a feedback control optimization framework for a sensor development and fabrication process. The existing sensor adaptive neuro-fuzzy inference system intelligent model was configured in a closed-loop feedback control framework to optimally automate the sensor manufacturing process. Three main sensor manufacturing components were assumed to be the process inputs while the sensor relative optical density was considered as the output in the sensor intelligent model. The optimization and automation of the phenol sensor development were investigated by designing two controllers in a closed loop control tracking problem. Although the closed loop framework was inherently nonlinear, sensor design parameters with practical significance were successfully achieved for both types of controllers. The fuzzy logic controller receives the closed loop system error and the error rate and generates the incremental changes on the three sensor process inputs. The second controller is a proportional-integral-derivative type that processes each manufacturing process input and that manipulates the closed loop system error to produce the actual sensor process inputs. The numerical results for the proposed frameworks demonstrate that the sensor automation and design optimization can be enormously improved in closed loop control frameworks, eliminating the need of tedious trial-and-error sensor development process.