IMPLEMENTATION OF ALGORITHM-3 FIXED-POINT ITERATIVE QUADRATIC PROGRAMMING MODEL PREDICTIVE CONTROL USING AVR ATMEGA8535

• Main Author: (NIM 13203095), ANDY
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/9935
• Institution: Institut Teknologi Bandung

<p align="justify">MPC is an advance control technique which uses process model of the system explicitly to predict the behavior of the process in the future to obtain the optimum control action by minimizing the objective function. The weakness of MPC is the complexity of its computation which needs considerable resource to be done. The availability of a new MPC algorithm with simpler computation procedure, allows MPC to be implemented with microcontroller ATmega8535, of which resources are relatively limited compared with more advance controller such as FPGA, PIC, etc. All computation in microcontroller is done using fixed-point computation to overcome the limitation of microcontroller, both memory and time of computation. The purpose of this final project is to design MPC controller which gives the optimum control signal to the plant so the plant can reach the steady state in a short period of time without suffering of integrator windup caused by the constraints of the actuator.<p align="justify"><p>Firstly, the implementation of MPC is made with several lengths of horizon to obtain the horizon which result the best MPC performance. Then, MPC is compared with PI and PI anti windup type of controllers to compare the performance of the controllers on how they handle the constraints of the actuator. The plant which is used on this final project is DC motor while the actuator is pre-amplifier which has constraint on input voltage. The implementation results show that MPC gives the most satisfying performance compared with PI and PI anti windup controller. <br />