Description
Hello.
I use ifopt for path planning for self-driving vehicles. Make a nonlinear predictive problem with ifopt.
I'm not familiar with ifopt yet....
How can I make custom equality constraints for a discretized kinematic model?
I'm trying to modify test_vars_constr_cost.h file now.
For example, I want to define a problem with a prediction horizon for 5 steps, 4 states, and 2 control inputs, put 30 variables total for now. (In ExVariable Class, define x0_ to x24_). So if I have a discretized kinematic vehicle model as,
x_{k+1} = x_{k} + dt * v_{k} * cos(phi_{k})
y_{k+1} = y_{k} + dt * v_{k} * sin(phi_{k})
phi_{k+1} = phi_{k} + dt * v_{k} * tan(delta_{k}) / L
v_{k+1} = v{k} + dt * a{k}
then, I can define like this way.
x_{0} -> x(0), y_{0} -> x(1), ..., x_{1} -> x(6), y_{1} -> x(7),...
However, is there a way to define the time step of variables and constraints (prediction horizon in this case), not by modifying the header file in advance? I don't want to remove and add all these variables, constraints, and costs when I change the prediction horizons.
Thank you so much!