train_policy_gradient.PolicyGradientTrainingNode Class Reference

Inheritance diagram for train_policy_gradient.PolicyGradientTrainingNode:

Collaboration diagram for train_policy_gradient.PolicyGradientTrainingNode:

Public Member Functions
def	__init__ (self)
def	update_policy (self)
def	select_action (self, state)
def	get_reward (self)
def	on_complete_episode (self)
def	on_complete_step (self, state, action, reward, next_state)
Public Member Functions inherited from training_node.TrainingNode
def	__init__ (self, policy, actions, laser_sample_count, max_episode_length, learn_rate)
def	on_crash (self, _)
def	get_episode_summary (self)
def	on_complete_episode (self)
def	on_receive_laser_scan (self, message)
def	on_complete_step (self, state, action, reward, next_state)
def	check_car_orientation (self)
def	on_model_state_callback (self, message)
Public Member Functions inherited from reinforcement_learning_node.ReinforcementLearningNode
def	__init__ (self, actions, laser_sample_count)
def	perform_action (self, action_index)
def	convert_laser_message_to_tensor (self, message, use_device=True)
def	on_receive_laser_scan (self, message)

Public Attributes
	policy_history
	current_episode_rewards
Public Attributes inherited from training_node.TrainingNode
	policy
	max_episode_length
	episode_count
	episode_length
	total_step_count
	cumulative_reward
	is_terminal_step
	net_output_debug_string
	episode_length_history
	cumulative_reward_history
	state
	action
	car_position
	car_orientation
	drive_forward
	steps_with_wrong_orientation
	episode_start_time_real
	episode_start_time_sim
	optimizer
	episode_result_publisher
Public Attributes inherited from reinforcement_learning_node.ReinforcementLearningNode
	scan_indices
	laser_sample_count
	actions
	drive_parameters_publisher

Detailed Description

ROS node to train the Policy Gradient model

Definition at line 20 of file train_policy_gradient.py.

Constructor & Destructor Documentation

def train_policy_gradient.PolicyGradientTrainingNode.__init__

(

self

)

Definition at line 24 of file train_policy_gradient.py.

Member Function Documentation

def train_policy_gradient.PolicyGradientTrainingNode.get_reward

(

self

)

Definition at line 79 of file train_policy_gradient.py.

Here is the caller graph for this function:

def train_policy_gradient.PolicyGradientTrainingNode.on_complete_episode

(

self

)

Definition at line 101 of file train_policy_gradient.py.

Here is the call graph for this function:

Here is the caller graph for this function:

def train_policy_gradient.PolicyGradientTrainingNode.on_complete_step	(		self,
			state,
			action,
			reward,
			next_state
	)

Definition at line 105 of file train_policy_gradient.py.

Here is the caller graph for this function:

def train_policy_gradient.PolicyGradientTrainingNode.select_action	(		self,
			state
	)

Definition at line 67 of file train_policy_gradient.py.

Here is the caller graph for this function:

def train_policy_gradient.PolicyGradientTrainingNode.update_policy

(

self

)

Definition at line 40 of file train_policy_gradient.py.

Here is the caller graph for this function:

Member Data Documentation

train_policy_gradient.PolicyGradientTrainingNode.current_episode_rewards

Definition at line 38 of file train_policy_gradient.py.

train_policy_gradient.PolicyGradientTrainingNode.policy_history

Definition at line 37 of file train_policy_gradient.py.

---

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

• /home/travis/build/Autonomous-Racing-PG/ar-tu-do/docs/master/ros_ws/src/autonomous/reinforcement_learning/scripts/train_policy_gradient.py