Commit 8fedf4ed authored by Ibrahim's avatar Ibrahim
Browse files

optimize.py:Enabled altitude controller optimization.

parent 29706e00

README.md

+1 −1
Original line number Diff line number Diff line
@@ -4,7 +4,7 @@ Simulation of multi-rotor unmanned aerial vehicles in python. 


This package provides an object-oriented interface for modeling and simulating motors, propellers, and airframe of a UAV. Additionally, an OpenAI gym-compatible environment is provided for Reinforcement Learning experiments.



See the [Detailed Demo](/Detailed%20Demo.ipynb) jupyter notebook in the repository for a walkthrough.

See the [Detailed Demo](./Detailed%20Demo.ipynb) jupyter notebook in the repository for a walkthrough.



Code repository: [Github](https://github.com/hazrmard/multirotor), [Gitlab](https://git.isis.vanderbilt.edu/ahmedi/multirotor)

multirotor/optimize.py

+10 −12
Original line number Diff line number Diff line
@@ -22,8 +22,6 @@ from .controller import ( 




DEFAULTS = Namespace(

    # ntrials = 1000,

    # nprocs = 5,

    bounding_box = 20,

    max_velocity = 5,

    max_acceleration = 2.5,
@@ -228,16 +226,16 @@ def make_controller_from_trial(trial: optuna.Trial, args: Namespace=DEFAULTS, pr 
            max_acceleration = np.asarray((r_pitch_roll_max_acc, r_pitch_roll_max_acc, r_yaw_max_acc)),

            max_err_i = np.asarray((r_pitch_roll_max_acc, r_pitch_roll_max_acc, r_yaw_max_acc)),

        ),

        # ctrl_z = dict(

        #     k_p = trial.suggest_float('z.k_p', 0.1, 50),

        #     k_i = trial.suggest_float('z.k_i', 0.1, 10),

        #     k_d = trial.suggest_float('z.k_d', 1, 250),

        # ),

        # ctrl_vz = dict(

        #     k_p = trial.suggest_float('vz.k_p', 0.1, 50),

        #     k_i = trial.suggest_float('vz.k_i', 0.1, 10),

        #     k_d = trial.suggest_float('vz.k_d', 1, 250),

        # ),

        ctrl_z = dict(

            k_p = trial.suggest_float('z.k_p', 0.1, 50),

            k_i = trial.suggest_float('z.k_i', 0.1, 10),

            k_d = trial.suggest_float('z.k_d', 1, 250),

        ),

        ctrl_vz = dict(

            k_p = trial.suggest_float('vz.k_p', 0.1, 50),

            k_i = trial.suggest_float('vz.k_i', 0.1, 10),

            k_d = trial.suggest_float('vz.k_d', 1, 250),

        ),

    )

    if args.scurve:

        params.update({prefix + 'feedforward_weight': trial.suggest_float(prefix + 'feedforward_weight', 0.1, 1.0, step=0.1)})