bolero.environment.OptimumTrajectory¶

class bolero.environment.OptimumTrajectory(x0=array([ 0., 0.]), g=array([ 1., 1.]), execution_time=1.0, dt=0.01, obstacles=None, obstacle_dist=0.1, penalty_start_dist=0.0, penalty_goal_dist=0.0, penalty_vel=0.0, penalty_acc=0.0, penalty_obstacle=0.0, log_to_file=False, log_to_stdout=False)[source]¶

Optimize a trajectory according to some criteria.

Parameters:

x0 : array-like, shape = (n_task_dims,), optional (default: [0, 0]) Initial position. g : array-like, shape = (n_task_dims,), optional (default: [1, 1]) Goal position. execution_time : float, optional (default: 1.0) Execution time in seconds dt : float, optional (default: 0.01) Time between successive steps in seconds. obstacles : array-like, shape (n_obstacles, n_task_dims) (default: None) List of obstacles. obstacle_dist : float, optional (default: 0.1) Distance that should be kept to the obstacles (penalty is zero outside of this area) penalty_start_dist : float, optional (default: 0) Penalty weight for distance to start at the beginning penalty_goal_dist : float, optional (default: 0) Penalty weight for distance to goal at the end penalty_vel : float, optional (default: 0) Penalty weight for velocities penalty_acc : float, optional (default: 0) Penalty weight for accelerations penalty_obstacle : float, optional (default: 0) Penalty weight for obstacle avoidance log_to_file: optional, boolean or string (default: False) Log results to given file, it will be located in the $BL_LOG_PATH log_to_stdout: optional, boolean (default: False) Log to standard output

__init__(x0=array([ 0., 0.]), g=array([ 1., 1.]), execution_time=1.0, dt=0.01, obstacles=None, obstacle_dist=0.1, penalty_start_dist=0.0, penalty_goal_dist=0.0, penalty_vel=0.0, penalty_acc=0.0, penalty_obstacle=0.0, log_to_file=False, log_to_stdout=False)[source]¶

get_acceleration()[source]¶

Get acceleration values during the performed movement.

Returns:	acceleration : array-like, shape (n_steps,) the total acceleration (scalar) at all previous timestamps

get_args()¶

Get parameters for this estimator.

Returns:	params : mapping of string to any Parameter names mapped to their values.

get_collision(obstacle_filter=None)[source]¶

Get list of collisions with obstacles during the performed movement.

Parameters:	obstacle_filter : array-like, shape (n_desired_obstacles, 1) specify which obstacles cause collisions, e.g. set (0, 2) to exclude the second of three obstacles Returns ——- collisions : array-like, shape (n_steps,) vector of values in range [0, 1] where distances above self.obstacle_dist result in 0 (no collision), and distance below are scaled linearly, so that 1 corresponds to an intersection.

get_feedback()[source]¶

Get reward per timestamp based on weighted criteria (penalties)

Returns:	rewards : array-like, shape (n_steps,) reward for every timestamp; non-positive values

get_goal_dist()[source]¶

Get distance of trajectory end and desired goal location.

Returns:	goal_dist : float goal distance

get_maximum_feedback()[source]¶

Returns the maximum sum of feedbacks obtainable.

get_num_inputs()[source]¶

Get number of environment inputs.

Returns:	n : int number of environment inputs

get_num_obstacles()[source]¶

Get number of obstacles in environment.

Returns:	n : int number of obstacles

get_num_outputs()[source]¶

Get number of environment outputs.

Returns:	n : int number of environment outputs

get_outputs(values)[source]¶

Get environment outputs.

Parameters:	values : array Outputs of the environment: positions, velocities and accelerations in that order, e.g. for n_task_dims=2 the order would be xxvvaa

get_speed()[source]¶

Get speed values during the performed movement.

Returns:	speed : array-like, shape (n_steps,) the speed (scalar) at all previous timestamps

get_start_dist()[source]¶

Get distance of trajectory start and desired start location.

Returns:	start_dist : float start distance

init()[source]¶

Initialize environment.

is_behavior_learning_done()[source]¶

Check if the behavior learning is finished.

Returns:	finished : bool Always false

is_evaluation_done()[source]¶

Check if the time is over.

Returns:	finished : bool Is the episode finished?

plot(ax)[source]¶

Plot a two-dimensional environment.

Parameters:	ax : Axis Matplotlib axis

reset()[source]¶

Reset state of the environment.

set_inputs(values)[source]¶

Set environment inputs, e.g. next action.

Parameters:	values : array, Inputs for the environment: positions, velocities and accelerations in that order, e.g. for n_task_dims=2 the order would be xxvvaa

step_action()[source]¶

Execute step perfectly.