jkleiber/control-blocks

control-blocks

The goal of this project is to make a node based control system block diagram software. This will enable open source development and simulation of control systems.

Current Capability

• Basic GUI and simulation for simple math / constants
• Block types:
  • Constant
  • Sum
  • Gain (only faciliator of feedback at the moment)
  • Mux (can add ports but not remove)
  • Display
• Can add/remove blocks and wires
• Saving and loading the diagram (only one filename supported right now)

Dependencies

See third_party for a list of dependencies and how to install them.

Active Work

• ODE solver
  • Using Boost for this.
• Improve feedback systems
  • This should be ready theoretically if each Dynamic system block actually implements ApplyInitial()
  • Unclear if basic blocks like sum and mux should have this feature, but it is likely they should
• Variables in a workspace

Roadmap

The current work is focusing on Steps 1-3.

1. Basic GUI and scripting
   • Scripting will tentatively be in python, but might end up in C++
   • C++ will be used in the backend for speed optimization in the future.
2. Simulation for general dynamical systems (x_dot = f(x,u,t))
   • Series simulation only, no summing junctions
3. Simulation for linear time invariant (LTI) systems
   • State-space LTI systems
   • S-domain transfer functions
4. Diagnostics for LTI systems
   • Bode plot
   • Nichols chart
   • Nyquist plot
   • Pole-zero map
   • Step response
   • Impulse response
5. Basic block diagram algebra
   • Summing junctions
   • Feedback systems
   • Multiplexing
6. Advanced block diagram manipulation
   • I/O linearization
7. Controllers
   • PID
   • LQR f
8. Advanced control
   • H-infinity
   • Model predictive control (MPC)