upkie.envs.wrappers.differentiate_action.DifferentiateAction Class Reference

Act on the derivative of the action. More...

Public Member Functions
def	__init__ (self, env, np.ndarray min_derivative, np.ndarray max_derivative, float action_penalty=0.0)
	Initialize wrapper. More...
def	reset (self, **kwargs)
	Reset the environment. More...
Tuple[np.ndarray, float, bool, bool, dict]	step (self, np.ndarray action)
	Step the environment. More...

Public Attributes
	action_space
	Action space.
	action_penalty
	Weight for an additional penalty on the differential action added to the reward.

Detailed Description

Act on the derivative of the action.

Constructor & Destructor Documentation

__init__()

def upkie.envs.wrappers.differentiate_action.DifferentiateAction.__init__	(		self,
			env,
		np.ndarray	min_derivative,
		np.ndarray	max_derivative,
		float	action_penalty = 0.0
	)

Initialize wrapper.

Parameters

env	Environment to wrap.
min_derivative	Lower bound on the derivative of the original action.
max_derivative	Upper bound on the derivative of the original action.
action_penalty	Weight for an additional penalty on the differential action added to the reward.

Note

We assume the original action lives in a vector space.

Member Function Documentation

reset()

def upkie.envs.wrappers.differentiate_action.DifferentiateAction.reset	(		self,
		**	kwargs
	)

Reset the environment.

Parameters

kwargs

Keyword arguments forwarded to the wrapped environment.

step()

Tuple[np.ndarray, float, bool, bool, dict] upkie.envs.wrappers.differentiate_action.DifferentiateAction.step	(		self,
		np.ndarray	action
	)

Step the environment.

Parameters

action

Action from the agent.

Returns

Tuple with (observation, reward, terminated, truncated,info). See upkie.envs.upkie_base_env.UpkieBaseEnv.step for details.

---

The documentation for this class was generated from the following file:

• upkie/envs/wrappers/differentiate_action.py