Software

Anaerobic Bioreactor

This Matlab Toolbox provides routines for simulation, identification, estimation and control of anaerobic bioreactors. It implements classical Runge-Kutta methods for simulation of dynamic models. State estimation is done using Kalman Filter (KF), Extended Kalman Filter (EKF) and Particle Filter (PF). Both dynamic output feedback and state feedback (static and dynamic) are implemented.

educopter

educopter is a complete firmware for multirotors based on 8-bit microprocessors. It implements drivers for gyroscopes, accelerometers, barometers and compass. A linear sensor fusion algorithm estimates angles to be controlled by PID controllers. Pilot reference signals are sent to the vehicle using 4 channel remote controllers.

px4tuning

px4tuning is a python package that computes optimal PID parameters for quadcopters running PX4 firmware using the data-driven control technique Virtual Reference Feedback Tuning. The package extracts the dynamical information of the vehicle from a flight log .ulg file without the need of a mathematical model. Six PID controllers are computed: roll, pitch and yaw angles and roll, pitch and yaw angle rates.

pyvrft

pyvrft is a python package that implements the Virtual Reference Feedback Tunging control method. The optimal gains of fixed-order controllers are computed using only input/output data collect from experiments, without the need of a mathematical model of the process. The algorithm computes MIMO controllers using both classical least-squares and instrumental variables technique.

SIB - System Identification Toolbox

SIB is a Matlab toolbox for system identification of discrete dynamic systems. It implements the Prediction Error Method (PEM) with classical model structures as FIR, ARX, ARMAX, OE and BJ. The toolbox has specific optimization solvers that have proved efficient to solve the typical non-convex problems.

TED - The Experiment Design

TED is a Matlab toolbox for minimum energy input signal design. It solves LMI problems to find the optimal spectrum that should be applied in order to minimize the covariance of system identification estimates. Both continuous and discrete spectra can be used resulting in realizations of filtered white noise and sums of sines.