Collaborative UAV Bar Transport

Recreating centralized linear MPC results from the paper "Decentralized Model Predictive Control for Equilibrium-based Collaborative UAV Bar Transportation" ICRA'22

Running the Code

Requirements

pip install numpy matplotlib scipy casadi

Setup

Following lines on code execution defines initial state vector and target setpoint. Modify these with your desired values.

  thetal, thetaf = get_thetas() #define your thetas
  x0, u0 = get_eqb(pb = [1, 1, 2], thetal=thetal, thetaf=thetaf) #ref variables
  xinit, uinit = get_eqb(thetal=thetal, thetaf=thetaf) #initial variables

A small description of the used functions in the above is as follows:

• get_eqb(pb, yawb, yaw1, yaw2, thetaf, thetal) -> x, u: Takes desired equilbrium configuration and returns corresponding eqb. state and control inputs
• get_thetas([thetal, thetaf]) -> thetal, thetaf : Takes initial estimate of your desired theta values, and returns closest solution wherein equilbrium exists

You may also change value of thetal and thetaf at eqb by passing the a vector of initial estimate of your desired values in get_thetas() to get the closest solution to your given initial value. These can be then passed on to get_eqb function for getting corresponding state vector and control inputs.
Default value takes initial position of bar pb to be [0, 0, 2] and all yaws as zero, and both thetal and thetaf to be zeros.

Running Controller

Once you have defined your desired initial state vector and control input (xinit and uinit), and target setpoint (x0 and u0), you can simply run the python file to output controller results which includes 3D plot of trajectory and graphs showing tracking of state vector.

python centralized_controller.py

Implementation Details

The single python implements the entire dynamics, linearization, discretization and finally a linear MPC solver using ipopt solver using casadi symbolic library.