From 6e2c1c9d136859faf3e24c29d97142a1538ee570 Mon Sep 17 00:00:00 2001
From: Erik Nygren <erik.nygren@sbb.ch>
Date: Wed, 24 Jul 2019 18:48:15 -0400
Subject: [PATCH] updating local obs for rail env to be better suited for the
task
---
examples/training_example.py | 10 ++-
flatland/core/env_observation_builder.py | 1 +
flatland/envs/observations.py | 110 ++++++++++++++++-------
3 files changed, 86 insertions(+), 35 deletions(-)
diff --git a/examples/training_example.py b/examples/training_example.py
index 3139209..67e0c74 100644
--- a/examples/training_example.py
+++ b/examples/training_example.py
@@ -1,9 +1,10 @@
import numpy as np
from flatland.envs.generators import complex_rail_generator
-from flatland.envs.observations import TreeObsForRailEnv
+from flatland.envs.observations import TreeObsForRailEnv, LocalObsForRailEnv
from flatland.envs.predictions import ShortestPathPredictorForRailEnv
from flatland.envs.rail_env import RailEnv
+from flatland.utils.rendertools import RenderTool
np.random.seed(1)
@@ -12,12 +13,14 @@ np.random.seed(1)
#
TreeObservation = TreeObsForRailEnv(max_depth=2, predictor=ShortestPathPredictorForRailEnv())
+LocalGridObs = LocalObsForRailEnv(view_height=10, view_width=2, center=2)
env = RailEnv(width=20,
height=20,
rail_generator=complex_rail_generator(nr_start_goal=10, nr_extra=1, min_dist=8, max_dist=99999, seed=0),
- obs_builder_object=TreeObservation,
+ obs_builder_object=LocalGridObs,
number_of_agents=2)
+env_renderer = RenderTool(env, gl="PILSVG", )
# Import your own Agent or use RLlib to train agents on Flatland
# As an example we use a random agent here
@@ -66,6 +69,7 @@ for trials in range(1, n_trials + 1):
# Reset environment and get initial observations for all agents
obs = env.reset()
+ env_renderer.reset()
# Here you can also further enhance the provided observation by means of normalization
# See training navigation example in the baseline repository
@@ -80,6 +84,8 @@ for trials in range(1, n_trials + 1):
# Environment step which returns the observations for all agents, their corresponding
# reward and whether their are done
next_obs, all_rewards, done, _ = env.step(action_dict)
+ env_renderer.render_env(show=True, show_observations=True, show_predictions=False)
+
# Update replay buffer and train agent
for a in range(env.get_num_agents()):
agent.step((obs[a], action_dict[a], all_rewards[a], next_obs[a], done[a]))
diff --git a/flatland/core/env_observation_builder.py b/flatland/core/env_observation_builder.py
index 060785f..4acdf16 100644
--- a/flatland/core/env_observation_builder.py
+++ b/flatland/core/env_observation_builder.py
@@ -74,6 +74,7 @@ class ObservationBuilder:
direction[agent.direction] = 1
return direction
+
class DummyObservationBuilder(ObservationBuilder):
"""
DummyObservationBuilder class which returns dummy observations
diff --git a/flatland/envs/observations.py b/flatland/envs/observations.py
index 498d98e..fb8e107 100644
--- a/flatland/envs/observations.py
+++ b/flatland/envs/observations.py
@@ -698,71 +698,115 @@ class LocalObsForRailEnv(ObservationBuilder):
The observation is composed of the following elements:
- transition map array of the local environment around the given agent,
- with dimensions (2*view_radius + 1, 2*view_radius + 1, 16),
+ with dimensions (view_height,2*view_width+1, 16),
assuming 16 bits encoding of transitions.
- - Two 2D arrays (2*view_radius + 1, 2*view_radius + 1, 2) containing respectively,
+ - Two 3D arrays (view_height,2*view_width+1, 2) containing respectively,
if they are in the agent's vision range, its target position, the positions of the other targets.
- - A 3D array (2*view_radius + 1, 2*view_radius + 1, 4) containing the one hot encoding of directions
+ - A 3D array (view_height,2*view_width+1, 4) containing the one hot encoding of directions
of the other agents at their position coordinates, if they are in the agent's vision range.
- A 4 elements array with one hot encoding of the direction.
"""
- def __init__(self, view_radius):
+ def __init__(self, view_width, view_height, center):
"""
:param view_radius:
"""
super(LocalObsForRailEnv, self).__init__()
- self.view_radius = view_radius
+ self.view_width = view_width
+ self.view_height = view_height
+ self.center = center
+ self.max_padding = max(self.view_width, self.view_height - self.center)
def reset(self):
# We build the transition map with a view_radius empty cells expansion on each side.
# This helps to collect the local transition map view when the agent is close to a border.
-
- self.rail_obs = np.zeros((self.env.height + 2 * self.view_radius,
- self.env.width + 2 * self.view_radius, 16))
+ self.max_padding = max(self.view_width, self.view_height)
+ self.rail_obs = np.zeros((self.env.height + 2 * self.max_padding,
+ self.env.width + 2 * self.max_padding, 16))
for i in range(self.env.height):
for j in range(self.env.width):
bitlist = [int(digit) for digit in bin(self.env.rail.get_full_transitions(i, j))[2:]]
bitlist = [0] * (16 - len(bitlist)) + bitlist
- self.rail_obs[i + self.view_radius, j + self.view_radius] = np.array(bitlist)
+ self.rail_obs[i + self.view_height, j + self.view_width] = np.array(bitlist)
def get(self, handle):
agents = self.env.agents
agent = agents[handle]
+ agent_rel_pos = [0, 0]
- local_rail_obs = self.rail_obs[agent.position[0]: agent.position[0] + 2 * self.view_radius + 1,
- agent.position[1]:agent.position[1] + 2 * self.view_radius + 1]
-
- obs_map_state = np.zeros((2 * self.view_radius + 1, 2 * self.view_radius + 1, 2))
+ # Correct agents position for padding
+ agent_rel_pos[0] = agent.position[0] + self.max_padding
+ agent_rel_pos[1] = agent.position[1] + self.max_padding
- obs_other_agents_state = np.zeros((2 * self.view_radius + 1, 2 * self.view_radius + 1, 4))
+ # Collect the rail information in the local field of view
+ local_rail_obs = self.field_of_view(agent_rel_pos, agent.direction, state=self.rail_obs)
- def relative_pos(pos):
- return [agent.position[0] - pos[0], agent.position[1] - pos[1]]
+ # Locate observed agents and their coresponding targets
+ obs_map_state = np.zeros((self.view_height + 1, 2 * self.view_width + 1, 2))
+ obs_other_agents_state = np.zeros((self.view_height + 1, 2 * self.view_width + 1, 4))
- def is_in(rel_pos):
- return (abs(rel_pos[0]) <= self.view_radius) and (abs(rel_pos[1]) <= self.view_radius)
+ # Collect visible cells as set to be plotted
+ visited = self.field_of_view(agent.position, agent.direction)
- target_rel_pos = relative_pos(agent.target)
- if is_in(target_rel_pos):
- obs_map_state[self.view_radius + np.array(target_rel_pos)][0] += 1
+ # Add the visible cells to the observed cells
+ self.env.dev_obs_dict[handle] = visited
- for i in range(len(agents)):
- if i != handle: # TODO: handle used as index...?
- agent2 = agents[i]
+ direction = self._get_one_hot_for_agent_direction(agent)
- agent_2_rel_pos = relative_pos(agent2.position)
- if is_in(agent_2_rel_pos):
- obs_other_agents_state[self.view_radius + agent_2_rel_pos[0],
- self.view_radius + agent_2_rel_pos[1]][agent2.direction] += 1
+ return local_rail_obs, obs_map_state, obs_other_agents_state, direction
- target_rel_pos_2 = relative_pos(agent2.position)
- if is_in(target_rel_pos_2):
- obs_map_state[self.view_radius + np.array(target_rel_pos_2)][1] += 1
+ def get_many(self, handles=None):
+ """
+ Called whenever an observation has to be computed for the `env' environment, for each agent with handle
+ in the `handles' list.
+ """
- direction = self._get_one_hot_for_agent_direction(agent)
+ observations = {}
+ for h in handles:
+ observations[h] = self.get(h)
+ return observations
- return local_rail_obs, obs_map_state, obs_other_agents_state, direction
+ def field_of_view(self, position, direction, state=None):
+ # Compute the local field of view for an agent in the environment
+ data_collection = False
+ if state is not None:
+ temp_visible_data = np.zeros(shape=(self.view_height, 2 * self.view_width + 1, 16))
+ data_collection = True
+ if direction == 0:
+ origin = (position[0] + self.center, position[1] - self.view_width)
+ elif direction == 1:
+ origin = (position[0] - self.view_width, position[1] - self.center)
+ elif direction == 2:
+ origin = (position[0] - self.center, position[1] + self.view_width)
+ else:
+ origin = (position[0] + self.view_width, position[1] + self.center)
+ visible = set()
+ for h in range(self.view_height):
+ for w in range(2 * self.view_width + 1):
+ if direction == 0:
+ if 0 <= origin[0] - h < self.env.height and 0 <= origin[1] + w < self.env.width:
+ visible.add((origin[0] - h, origin[1] + w))
+ if data_collection:
+ temp_visible_data[h, w, :] = state[origin[0] - h, origin[1] + w, :]
+ elif direction == 1:
+ if 0 <= origin[0] + w < self.env.height and 0 <= origin[1] + h < self.env.width:
+ visible.add((origin[0] + w, origin[1] + h))
+ if data_collection:
+ temp_visible_data[h, w, :] = state[origin[0] + w, origin[1] + h, :]
+ elif direction == 2:
+ if 0 <= origin[0] - h < self.env.height and 0 <= origin[1] + w < self.env.width:
+ visible.add((origin[0] + h, origin[1] - w))
+ if data_collection:
+ temp_visible_data[h, w, :] = state[origin[0] + h, origin[1] - w, :]
+ else:
+ if 0 <= origin[0] - h < self.env.height and 0 <= origin[1] + w < self.env.width:
+ visible.add((origin[0] - w, origin[1] - h))
+ if data_collection:
+ temp_visible_data[h, w, :] = state[origin[0] - w, origin[1] - h, :]
+ if data_collection:
+ return temp_visible_data
+ else:
+ return visible
--
GitLab