Author : Filip Dymczyk

Project aim was to:

• practice C++17 skills as well as OOP, design patterns and best practices,
• increase knowledge about the behavior of dynamical control systems through their logic implementation,
• learn to validate components with the use of GoogleTest framework,
• incorporate responses visualization through Matplotlibcpp,
• design a GUI using the Qt library and connect it to the dynamical systems simulation.

Logic features:

• support of two representations of dynamical objects:
  • differential equation (SISO), e.g. $\dot{x}$ = $-a_1 x + a_2 u$,
  • state space (MIMO): $$\dot{x} = Ax + Bu \ y = Cx + Du $$
• support of two controller types:
  • PID - with possible derivative input low pass filtering,
  • Bang-Bang.
• possibility of simulating measurement noise on the object output,
• ability to create whole control loops, configured as open or closed ones with or without controllers with a selected object representation,
• different supported input signal types, such as:
  • Heaviside,
  • Ramp,
  • Rectangle,
  • Sine Wave,
  • Pulse Wave.
• additionally, an experimental feature of PID Tuner using Recursive Linear Regression was introduced (requires further validation).

There are plans to expand the control logic features with:

• different integration (solvers) methods - for now only the simplest forward Euler method is utilized,
• support of discretization of continuous objects and simulating their behavior taking the sampling time into account,
• other types of controllers, e.g. LQR,
• introducing of observers and/or estimators, such as Kalman filter,
• output or full state feedback,
• pole-placing functionalities, e.g. root-locus, regular pole-placement or using optimization (LQ problem).

GUI features:

• option to select object representation and entering parameters/matrices in Matlab-like syntax,
• possibility to simulate it's response on the previously mentioned input signal types with an ability to change their parameters,
• option to create more advanced control loops (open/closed) with a selected controller type (with changeable parameters),
• possibility to plot control signal value (optional), simulate measurement noise with a tunable standard deviation and low pass filter the PID derivative input with desired coefficient,
• ability to specify operation time and time step.
• added experimental PID Tuner using RLS with tunable parameters (requires further validation).