A Data Driven PID Control System

Proportional, Integral, Derivative control systems are used in aviation, automotive industry, robotics, mechatronics and many other application areas. The proportional, integral, and derivative coefficients are defined deterministically in a generic way regardless of the process they control. In this research paper we compute these coefficients using mean square error analysis. The data used in the analysis, are produced by setting up proper experiments, using stratification, pre-sampling, experimental, design, sampling and finally computation of the coefficients. The method outlined in this research paper can be used in any control application where PID controllers are to be considered. The application emphasized in this paper is UAVs (Unmanned Air Vehicles), and cruise control in automobiles. Our Data driven PID system is modeled like a finite state machine. The PID coefficients are optimized separately for each state. A PID control system uses sensors to obtain the needed input from its environment, processors to process the input and decide which action is needed, and actuators to control the components of the system that is designed to control. The processors we use are FPGAs. The reason for that is because they make it easy to upgrade, debug, and improve the software driving the hardware.