Files
Gymnasium/gym/envs/box2d/lunar_lander.py

585 lines
20 KiB
Python
Raw Normal View History

__credits__ = ["Andrea PIERRÉ"]
2021-07-29 02:26:34 +02:00
import math
import sys
Seeding update (#2422) * Ditch most of the seeding.py and replace np_random with the numpy default_rng. Let's see if tests pass * Updated a bunch of RNG calls from the RandomState API to Generator API * black; didn't expect that, did ya? * Undo a typo * blaaack * More typo fixes * Fixed setting/getting state in multidiscrete spaces * Fix typo, fix a test to work with the new sampling * Correctly (?) pass the randomly generated seed if np_random is called with None as seed * Convert the Discrete sample to a python int (as opposed to np.int64) * Remove some redundant imports * First version of the compatibility layer for old-style RNG. Mainly to trigger tests. * Removed redundant f-strings * Style fixes, removing unused imports * Try to make tests pass by removing atari from the dockerfile * Try to make tests pass by removing atari from the setup * Try to make tests pass by removing atari from the setup * Try to make tests pass by removing atari from the setup * First attempt at deprecating `env.seed` and supporting `env.reset(seed=seed)` instead. Tests should hopefully pass but throw up a million warnings. * black; didn't expect that, didya? * Rename the reset parameter in VecEnvs back to `seed` * Updated tests to use the new seeding method * Removed a bunch of old `seed` calls. Fixed a bug in AsyncVectorEnv * Stop Discrete envs from doing part of the setup (and using the randomness) in init (as opposed to reset) * Add explicit seed to wrappers reset * Remove an accidental return * Re-add some legacy functions with a warning. * Use deprecation instead of regular warnings for the newly deprecated methods/functions
2021-12-08 22:14:15 +01:00
from typing import Optional
2016-05-03 22:27:42 +03:00
import numpy as np
import Box2D
2021-07-29 02:26:34 +02:00
from Box2D.b2 import (
edgeShape,
circleShape,
fixtureDef,
polygonShape,
revoluteJointDef,
contactListener,
)
2016-05-03 22:27:42 +03:00
import gym
from gym import error, spaces
from gym.utils import seeding, EzPickle
2016-05-03 22:27:42 +03:00
FPS = 50
2021-07-29 02:26:34 +02:00
SCALE = 30.0 # affects how fast-paced the game is, forces should be adjusted as well
MAIN_ENGINE_POWER = 13.0
SIDE_ENGINE_POWER = 0.6
2016-05-03 22:27:42 +03:00
2021-07-29 02:26:34 +02:00
INITIAL_RANDOM = 1000.0 # Set 1500 to make game harder
2016-05-03 22:27:42 +03:00
2021-07-29 02:26:34 +02:00
LANDER_POLY = [(-14, +17), (-17, 0), (-17, -10), (+17, -10), (+17, 0), (+14, +17)]
2016-05-03 22:27:42 +03:00
LEG_AWAY = 20
LEG_DOWN = 18
LEG_W, LEG_H = 2, 8
LEG_SPRING_TORQUE = 40
SIDE_ENGINE_HEIGHT = 14.0
SIDE_ENGINE_AWAY = 12.0
2016-05-03 22:27:42 +03:00
VIEWPORT_W = 600
VIEWPORT_H = 400
2016-05-03 22:27:42 +03:00
class ContactDetector(contactListener):
def __init__(self, env):
2016-05-25 11:16:15 +03:00
contactListener.__init__(self)
self.env = env
2016-05-03 22:27:42 +03:00
def BeginContact(self, contact):
2021-07-29 15:39:42 -04:00
if (
self.env.lander == contact.fixtureA.body
or self.env.lander == contact.fixtureB.body
):
2016-05-03 22:27:42 +03:00
self.env.game_over = True
for i in range(2):
if self.env.legs[i] in [contact.fixtureA.body, contact.fixtureB.body]:
self.env.legs[i].ground_contact = True
def EndContact(self, contact):
for i in range(2):
if self.env.legs[i] in [contact.fixtureA.body, contact.fixtureB.body]:
self.env.legs[i].ground_contact = False
2016-05-03 22:27:42 +03:00
class LunarLander(gym.Env, EzPickle):
"""
### Description
This environment is a classic rocket trajectory optimization problem.
According to Pontryagin's maximum principle, it is optimal to fire the
engine at full throttle or turn it off. This is the reason why this
environment has discreet actions: engine on or off.
There are two environment versions: discrete or continuous.
The landing pad is always at coordinates (0,0). The coordinates are the
first two numbers in the state vector.
Landing outside the landing pad is possible. Fuel is infinite, so an agent
can learn to fly and then land on its first attempt.
To see a heuristic landing, run:
```
python gym/envs/box2d/lunar_lander.py
```
<!-- To play yourself, run: -->
<!-- python examples/agents/keyboard_agent.py LunarLander-v2 -->
## Action Space
There are four discrete actions available: do nothing, fire left
orientation engine, fire main engine, fire right orientation engine.
## Observation Space
There are 8 states: the coordinates of the lander in `x` & `y`, its linear
velocities in `x` & `y`, its angle, its angular velocity, and two boleans
showing if each leg is in contact with the ground or not.
## Rewards
Reward for moving from the top of the screen to the landing pad and zero
speed is about 100..140 points.
If the lander moves away from the landing pad it loses reward.
If the lander crashes, it receives an additional -100 points. If it comes
to rest, it receives an additional +100 points. Each leg with ground
contact is +10 points.
Firing the main engine is -0.3 points each frame. Firing the side engine
is -0.03 points each frame. Solved is 200 points.
## Starting State
The lander starts at the top center of the viewport with a random initial
force applied to its center of mass.
## Episode Termination
The episode finishes if:
1) the lander crashes (the lander body gets in contact with the moon);
2) the lander gets outside of the viewport (`x` coordinate is greater than 1);
3) the lander is not awake. From the [Box2D docs](https://box2d.org/documentation/md__d_1__git_hub_box2d_docs_dynamics.html#autotoc_md61),
a body which is not awake is a body which doesn't move and doesn't
collide with any other body:
> When Box2D determines that a body (or group of bodies) has come to rest,
> the body enters a sleep state which has very little CPU overhead. If a
> body is awake and collides with a sleeping body, then the sleeping body
> wakes up. Bodies will also wake up if a joint or contact attached to
> them is destroyed.
## Arguments
To use to the _continuous_ environment, you need to specify the
`continuous"=True` argument like below:
```python
import gym
env = gym.make("LunarLander-v2", continuous=True)
```
<!-- ### Version History -->
<!-- - v2: -->
<!-- - v1: -->
<!-- ### References -->
## Credits
Created by Oleg Klimov
"""
2021-07-29 02:26:34 +02:00
metadata = {"render.modes": ["human", "rgb_array"], "video.frames_per_second": FPS}
2016-05-03 22:27:42 +03:00
def __init__(self, continuous: bool = False):
EzPickle.__init__(self)
2016-05-03 22:27:42 +03:00
self.viewer = None
2016-05-25 11:16:15 +03:00
self.world = Box2D.b2World()
2016-05-03 22:27:42 +03:00
self.moon = None
self.lander = None
self.particles = []
self.prev_reward = None
self.continuous = continuous
# useful range is -1 .. +1, but spikes can be higher
2021-07-29 15:39:42 -04:00
self.observation_space = spaces.Box(
-np.inf, np.inf, shape=(8,), dtype=np.float32
)
if self.continuous:
# Action is two floats [main engine, left-right engines].
# Main engine: -1..0 off, 0..+1 throttle from 50% to 100% power. Engine can't work with less than 50% power.
# Left-right: -1.0..-0.5 fire left engine, +0.5..+1.0 fire right engine, -0.5..0.5 off
self.action_space = spaces.Box(-1, +1, (2,), dtype=np.float32)
else:
# Nop, fire left engine, main engine, right engine
self.action_space = spaces.Discrete(4)
2016-05-03 22:27:42 +03:00
def _destroy(self):
2021-07-29 02:26:34 +02:00
if not self.moon:
return
2016-05-25 11:16:15 +03:00
self.world.contactListener = None
2016-05-03 22:27:42 +03:00
self._clean_particles(True)
self.world.DestroyBody(self.moon)
self.moon = None
self.world.DestroyBody(self.lander)
self.lander = None
self.world.DestroyBody(self.legs[0])
self.world.DestroyBody(self.legs[1])
def reset(self, *, seed: Optional[int] = None, options: Optional[dict] = None):
Seeding update (#2422) * Ditch most of the seeding.py and replace np_random with the numpy default_rng. Let's see if tests pass * Updated a bunch of RNG calls from the RandomState API to Generator API * black; didn't expect that, did ya? * Undo a typo * blaaack * More typo fixes * Fixed setting/getting state in multidiscrete spaces * Fix typo, fix a test to work with the new sampling * Correctly (?) pass the randomly generated seed if np_random is called with None as seed * Convert the Discrete sample to a python int (as opposed to np.int64) * Remove some redundant imports * First version of the compatibility layer for old-style RNG. Mainly to trigger tests. * Removed redundant f-strings * Style fixes, removing unused imports * Try to make tests pass by removing atari from the dockerfile * Try to make tests pass by removing atari from the setup * Try to make tests pass by removing atari from the setup * Try to make tests pass by removing atari from the setup * First attempt at deprecating `env.seed` and supporting `env.reset(seed=seed)` instead. Tests should hopefully pass but throw up a million warnings. * black; didn't expect that, didya? * Rename the reset parameter in VecEnvs back to `seed` * Updated tests to use the new seeding method * Removed a bunch of old `seed` calls. Fixed a bug in AsyncVectorEnv * Stop Discrete envs from doing part of the setup (and using the randomness) in init (as opposed to reset) * Add explicit seed to wrappers reset * Remove an accidental return * Re-add some legacy functions with a warning. * Use deprecation instead of regular warnings for the newly deprecated methods/functions
2021-12-08 22:14:15 +01:00
super().reset(seed=seed)
2016-05-03 22:27:42 +03:00
self._destroy()
self.world.contactListener_keepref = ContactDetector(self)
self.world.contactListener = self.world.contactListener_keepref
2016-05-03 22:27:42 +03:00
self.game_over = False
self.prev_shaping = None
2021-07-29 02:26:34 +02:00
W = VIEWPORT_W / SCALE
H = VIEWPORT_H / SCALE
2016-05-03 22:27:42 +03:00
# terrain
CHUNKS = 11
2021-07-29 02:26:34 +02:00
height = self.np_random.uniform(0, H / 2, size=(CHUNKS + 1,))
chunk_x = [W / (CHUNKS - 1) * i for i in range(CHUNKS)]
self.helipad_x1 = chunk_x[CHUNKS // 2 - 1]
self.helipad_x2 = chunk_x[CHUNKS // 2 + 1]
self.helipad_y = H / 4
height[CHUNKS // 2 - 2] = self.helipad_y
height[CHUNKS // 2 - 1] = self.helipad_y
height[CHUNKS // 2 + 0] = self.helipad_y
height[CHUNKS // 2 + 1] = self.helipad_y
height[CHUNKS // 2 + 2] = self.helipad_y
2021-07-29 15:39:42 -04:00
smooth_y = [
0.33 * (height[i - 1] + height[i + 0] + height[i + 1])
for i in range(CHUNKS)
]
2021-07-29 02:26:34 +02:00
2021-07-29 15:39:42 -04:00
self.moon = self.world.CreateStaticBody(
shapes=edgeShape(vertices=[(0, 0), (W, 0)])
)
2016-05-03 22:27:42 +03:00
self.sky_polys = []
2021-07-29 02:26:34 +02:00
for i in range(CHUNKS - 1):
p1 = (chunk_x[i], smooth_y[i])
2021-07-29 02:26:34 +02:00
p2 = (chunk_x[i + 1], smooth_y[i + 1])
self.moon.CreateEdgeFixture(vertices=[p1, p2], density=0, friction=0.1)
self.sky_polys.append([p1, p2, (p2[0], H), (p1[0], H)])
2016-05-03 22:27:42 +03:00
self.moon.color1 = (0.0, 0.0, 0.0)
self.moon.color2 = (0.0, 0.0, 0.0)
2016-05-03 22:27:42 +03:00
2021-07-29 02:26:34 +02:00
initial_y = VIEWPORT_H / SCALE
2016-05-03 22:27:42 +03:00
self.lander = self.world.CreateDynamicBody(
2021-07-29 02:26:34 +02:00
position=(VIEWPORT_W / SCALE / 2, initial_y),
2016-05-03 22:27:42 +03:00
angle=0.0,
2021-07-29 02:26:34 +02:00
fixtures=fixtureDef(
2021-07-29 15:39:42 -04:00
shape=polygonShape(
vertices=[(x / SCALE, y / SCALE) for x, y in LANDER_POLY]
),
2016-05-03 22:27:42 +03:00
density=5.0,
friction=0.1,
categoryBits=0x0010,
2021-07-29 02:26:34 +02:00
maskBits=0x001, # collide only with ground
restitution=0.0,
), # 0.99 bouncy
)
self.lander.color1 = (0.5, 0.4, 0.9)
self.lander.color2 = (0.3, 0.3, 0.5)
2021-07-29 02:26:34 +02:00
self.lander.ApplyForceToCenter(
(
self.np_random.uniform(-INITIAL_RANDOM, INITIAL_RANDOM),
self.np_random.uniform(-INITIAL_RANDOM, INITIAL_RANDOM),
),
True,
)
2016-05-03 22:27:42 +03:00
self.legs = []
for i in [-1, +1]:
2016-05-03 22:27:42 +03:00
leg = self.world.CreateDynamicBody(
2021-07-29 02:26:34 +02:00
position=(VIEWPORT_W / SCALE / 2 - i * LEG_AWAY / SCALE, initial_y),
angle=(i * 0.05),
fixtures=fixtureDef(
2021-07-29 02:26:34 +02:00
shape=polygonShape(box=(LEG_W / SCALE, LEG_H / SCALE)),
2016-05-03 22:27:42 +03:00
density=1.0,
restitution=0.0,
categoryBits=0x0020,
2021-07-29 02:26:34 +02:00
maskBits=0x001,
),
)
leg.ground_contact = False
leg.color1 = (0.5, 0.4, 0.9)
leg.color2 = (0.3, 0.3, 0.5)
2016-05-03 22:27:42 +03:00
rjd = revoluteJointDef(
bodyA=self.lander,
bodyB=leg,
localAnchorA=(0, 0),
2021-07-29 02:26:34 +02:00
localAnchorB=(i * LEG_AWAY / SCALE, LEG_DOWN / SCALE),
2016-05-03 22:27:42 +03:00
enableMotor=True,
enableLimit=True,
maxMotorTorque=LEG_SPRING_TORQUE,
2021-07-29 02:26:34 +02:00
motorSpeed=+0.3 * i, # low enough not to jump back into the sky
)
if i == -1:
2021-07-29 15:39:42 -04:00
rjd.lowerAngle = (
+0.9 - 0.5
) # The most esoteric numbers here, angled legs have freedom to travel within
2016-05-03 22:27:42 +03:00
rjd.upperAngle = +0.9
else:
rjd.lowerAngle = -0.9
rjd.upperAngle = -0.9 + 0.5
leg.joint = self.world.CreateJoint(rjd)
self.legs.append(leg)
self.drawlist = [self.lander] + self.legs
return self.step(np.array([0, 0]) if self.continuous else 0)[0]
2016-05-03 22:27:42 +03:00
def _create_particle(self, mass, x, y, ttl):
2016-05-03 22:27:42 +03:00
p = self.world.CreateDynamicBody(
2021-07-29 02:26:34 +02:00
position=(x, y),
2016-05-03 22:27:42 +03:00
angle=0.0,
2021-07-29 02:26:34 +02:00
fixtures=fixtureDef(
shape=circleShape(radius=2 / SCALE, pos=(0, 0)),
2016-05-03 22:27:42 +03:00
density=mass,
friction=0.1,
categoryBits=0x0100,
maskBits=0x001, # collide only with ground
2021-07-29 02:26:34 +02:00
restitution=0.3,
),
)
p.ttl = ttl
2016-05-03 22:27:42 +03:00
self.particles.append(p)
self._clean_particles(False)
return p
def _clean_particles(self, all):
while self.particles and (all or self.particles[0].ttl < 0):
2016-05-03 22:27:42 +03:00
self.world.DestroyBody(self.particles.pop(0))
def step(self, action):
if self.continuous:
action = np.clip(action, -1, +1).astype(np.float32)
else:
assert self.action_space.contains(
action
), f"{action!r} ({type(action)}) invalid "
2016-05-03 22:27:42 +03:00
# Engines
2021-07-29 02:26:34 +02:00
tip = (math.sin(self.lander.angle), math.cos(self.lander.angle))
side = (-tip[1], tip[0])
[WIP] add support for seeding environments (#135) * Make environments seedable * Fix monitor bugs - Set monitor_id before setting the infix. This was a bug that would yield incorrect results with multiple monitors. - Remove extra pid from stats recorder filename. This should be purely cosmetic. * Start uploading seeds in episode_batch * Fix _bigint_from_bytes for python3 * Set seed explicitly in random_agent * Pass through seed argument * Also pass through random state to spaces * Pass random state into the observation/action spaces * Make all _seed methods return the list of used seeds * Switch over to np.random where possible * Start hashing seeds, and also seed doom engine * Fixup seeding determinism in many cases * Seed before loading the ROM * Make seeding more Python3 friendly * Make the MuJoCo skipping a bit more forgiving * Remove debugging PDB calls * Make setInt argument into raw bytes * Validate and upload seeds * Skip box2d * Make seeds smaller, and change representation of seeds in upload * Handle long seeds * Fix RandomAgent example to be deterministic * Handle integer types correctly in Python2 and Python3 * Try caching pip * Try adding swap * Add df and free calls * Bump swap * Bump swap size * Try setting overcommit * Try other sysctls * Try fixing overcommit * Try just setting overcommit_memory=1 * Add explanatory comment * Add what's new section to readme * BUG: Mark ElevatorAction-ram-v0 as non-deterministic for now * Document seed * Move nondetermistic check into spec
2016-05-29 09:07:09 -07:00
dispersion = [self.np_random.uniform(-1.0, +1.0) / SCALE for _ in range(2)]
m_power = 0.0
2021-07-29 15:39:42 -04:00
if (self.continuous and action[0] > 0.0) or (
not self.continuous and action == 2
):
# Main engine
if self.continuous:
2021-07-29 02:26:34 +02:00
m_power = (np.clip(action[0], 0.0, 1.0) + 1.0) * 0.5 # 0.5..1.0
assert m_power >= 0.5 and m_power <= 1.0
else:
m_power = 1.0
2021-07-29 02:26:34 +02:00
ox = (
tip[0] * (4 / SCALE + 2 * dispersion[0]) + side[0] * dispersion[1]
) # 4 is move a bit downwards, +-2 for randomness
oy = -tip[1] * (4 / SCALE + 2 * dispersion[0]) - side[1] * dispersion[1]
2016-05-03 22:27:42 +03:00
impulse_pos = (self.lander.position[0] + ox, self.lander.position[1] + oy)
2021-07-29 02:26:34 +02:00
p = self._create_particle(
3.5, # 3.5 is here to make particle speed adequate
impulse_pos[0],
impulse_pos[1],
m_power,
) # particles are just a decoration
p.ApplyLinearImpulse(
(ox * MAIN_ENGINE_POWER * m_power, oy * MAIN_ENGINE_POWER * m_power),
impulse_pos,
True,
)
self.lander.ApplyLinearImpulse(
(-ox * MAIN_ENGINE_POWER * m_power, -oy * MAIN_ENGINE_POWER * m_power),
impulse_pos,
True,
)
s_power = 0.0
2021-07-29 15:39:42 -04:00
if (self.continuous and np.abs(action[1]) > 0.5) or (
not self.continuous and action in [1, 3]
):
# Orientation engines
if self.continuous:
direction = np.sign(action[1])
s_power = np.clip(np.abs(action[1]), 0.5, 1.0)
assert s_power >= 0.5 and s_power <= 1.0
else:
2021-07-29 02:26:34 +02:00
direction = action - 2
s_power = 1.0
2021-07-29 15:39:42 -04:00
ox = tip[0] * dispersion[0] + side[0] * (
3 * dispersion[1] + direction * SIDE_ENGINE_AWAY / SCALE
)
oy = -tip[1] * dispersion[0] - side[1] * (
3 * dispersion[1] + direction * SIDE_ENGINE_AWAY / SCALE
)
2021-07-29 02:26:34 +02:00
impulse_pos = (
self.lander.position[0] + ox - tip[0] * 17 / SCALE,
self.lander.position[1] + oy + tip[1] * SIDE_ENGINE_HEIGHT / SCALE,
)
p = self._create_particle(0.7, impulse_pos[0], impulse_pos[1], s_power)
2021-07-29 02:26:34 +02:00
p.ApplyLinearImpulse(
(ox * SIDE_ENGINE_POWER * s_power, oy * SIDE_ENGINE_POWER * s_power),
impulse_pos,
True,
)
self.lander.ApplyLinearImpulse(
(-ox * SIDE_ENGINE_POWER * s_power, -oy * SIDE_ENGINE_POWER * s_power),
impulse_pos,
True,
)
self.world.Step(1.0 / FPS, 6 * 30, 2 * 30)
2016-05-03 22:27:42 +03:00
pos = self.lander.position
vel = self.lander.linearVelocity
state = [
2021-07-29 02:26:34 +02:00
(pos.x - VIEWPORT_W / SCALE / 2) / (VIEWPORT_W / SCALE / 2),
(pos.y - (self.helipad_y + LEG_DOWN / SCALE)) / (VIEWPORT_H / SCALE / 2),
vel.x * (VIEWPORT_W / SCALE / 2) / FPS,
vel.y * (VIEWPORT_H / SCALE / 2) / FPS,
self.lander.angle,
2021-07-29 02:26:34 +02:00
20.0 * self.lander.angularVelocity / FPS,
1.0 if self.legs[0].ground_contact else 0.0,
2021-07-29 02:26:34 +02:00
1.0 if self.legs[1].ground_contact else 0.0,
]
assert len(state) == 8
2016-05-03 22:27:42 +03:00
reward = 0
2021-07-29 02:26:34 +02:00
shaping = (
-100 * np.sqrt(state[0] * state[0] + state[1] * state[1])
- 100 * np.sqrt(state[2] * state[2] + state[3] * state[3])
- 100 * abs(state[4])
+ 10 * state[6]
+ 10 * state[7]
) # And ten points for legs contact, the idea is if you
# lose contact again after landing, you get negative reward
2016-05-03 22:27:42 +03:00
if self.prev_shaping is not None:
reward = shaping - self.prev_shaping
self.prev_shaping = shaping
2021-07-29 15:39:42 -04:00
reward -= (
m_power * 0.30
) # less fuel spent is better, about -30 for heuristic landing
2021-07-29 02:26:34 +02:00
reward -= s_power * 0.03
2016-05-03 22:27:42 +03:00
done = False
if self.game_over or abs(state[0]) >= 1.0:
done = True
reward = -100
2016-05-03 22:27:42 +03:00
if not self.lander.awake:
done = True
reward = +100
return np.array(state, dtype=np.float32), reward, done, {}
2016-05-03 22:27:42 +03:00
2021-07-29 02:26:34 +02:00
def render(self, mode="human"):
from gym.utils import pyglet_rendering
2021-07-29 02:26:34 +02:00
2016-05-03 22:27:42 +03:00
if self.viewer is None:
self.viewer = pyglet_rendering.Viewer(VIEWPORT_W, VIEWPORT_H)
2021-07-29 02:26:34 +02:00
self.viewer.set_bounds(0, VIEWPORT_W / SCALE, 0, VIEWPORT_H / SCALE)
2016-05-03 22:27:42 +03:00
for obj in self.particles:
obj.ttl -= 0.15
2021-07-29 02:26:34 +02:00
obj.color1 = (
max(0.2, 0.2 + obj.ttl),
max(0.2, 0.5 * obj.ttl),
max(0.2, 0.5 * obj.ttl),
)
obj.color2 = (
max(0.2, 0.2 + obj.ttl),
max(0.2, 0.5 * obj.ttl),
max(0.2, 0.5 * obj.ttl),
)
2016-05-03 22:27:42 +03:00
self._clean_particles(False)
for p in self.sky_polys:
self.viewer.draw_polygon(p, color=(0, 0, 0))
2016-05-03 22:27:42 +03:00
for obj in self.particles + self.drawlist:
for f in obj.fixtures:
trans = f.body.transform
if type(f.shape) is circleShape:
t = pyglet_rendering.Transform(translation=trans * f.shape.pos)
2021-07-29 15:39:42 -04:00
self.viewer.draw_circle(
f.shape.radius, 20, color=obj.color1
).add_attr(t)
self.viewer.draw_circle(
f.shape.radius, 20, color=obj.color2, filled=False, linewidth=2
).add_attr(t)
2016-05-03 22:27:42 +03:00
else:
2021-07-29 02:26:34 +02:00
path = [trans * v for v in f.shape.vertices]
2016-05-03 22:27:42 +03:00
self.viewer.draw_polygon(path, color=obj.color1)
path.append(path[0])
self.viewer.draw_polyline(path, color=obj.color2, linewidth=2)
for x in [self.helipad_x1, self.helipad_x2]:
flagy1 = self.helipad_y
2021-07-29 02:26:34 +02:00
flagy2 = flagy1 + 50 / SCALE
self.viewer.draw_polyline([(x, flagy1), (x, flagy2)], color=(1, 1, 1))
2021-07-29 02:26:34 +02:00
self.viewer.draw_polygon(
[
(x, flagy2),
(x, flagy2 - 10 / SCALE),
(x + 25 / SCALE, flagy2 - 5 / SCALE),
],
color=(0.8, 0.8, 0),
)
2016-05-03 22:27:42 +03:00
2021-07-29 02:26:34 +02:00
return self.viewer.render(return_rgb_array=mode == "rgb_array")
def close(self):
if self.viewer is not None:
self.viewer.close()
self.viewer = None
def heuristic(env, s):
"""
The heuristic for
1. Testing
2. Demonstration rollout.
Args:
env: The environment
s (list): The state. Attributes:
s[0] is the horizontal coordinate
s[1] is the vertical coordinate
s[2] is the horizontal speed
s[3] is the vertical speed
s[4] is the angle
s[5] is the angular speed
s[6] 1 if first leg has contact, else 0
s[7] 1 if second leg has contact, else 0
returns:
a: The heuristic to be fed into the step function defined above to determine the next step and reward.
"""
2021-07-29 02:26:34 +02:00
angle_targ = s[0] * 0.5 + s[2] * 1.0 # angle should point towards center
if angle_targ > 0.4:
angle_targ = 0.4 # more than 0.4 radians (22 degrees) is bad
if angle_targ < -0.4:
angle_targ = -0.4
2021-07-29 15:39:42 -04:00
hover_targ = 0.55 * np.abs(
s[0]
) # target y should be proportional to horizontal offset
2021-07-29 02:26:34 +02:00
angle_todo = (angle_targ - s[4]) * 0.5 - (s[5]) * 1.0
hover_todo = (hover_targ - s[1]) * 0.5 - (s[3]) * 0.5
if s[6] or s[7]: # legs have contact
angle_todo = 0
2021-07-29 15:39:42 -04:00
hover_todo = (
-(s[3]) * 0.5
) # override to reduce fall speed, that's all we need after contact
if env.continuous:
2021-07-29 02:26:34 +02:00
a = np.array([hover_todo * 20 - 1, -angle_todo * 20])
a = np.clip(a, -1, +1)
else:
a = 0
2021-07-29 02:26:34 +02:00
if hover_todo > np.abs(angle_todo) and hover_todo > 0.05:
a = 2
elif angle_todo < -0.05:
a = 3
elif angle_todo > +0.05:
a = 1
return a
2021-07-29 02:26:34 +02:00
def demo_heuristic_lander(env, seed=None, render=False):
total_reward = 0
steps = 0
Seeding update (#2422) * Ditch most of the seeding.py and replace np_random with the numpy default_rng. Let's see if tests pass * Updated a bunch of RNG calls from the RandomState API to Generator API * black; didn't expect that, did ya? * Undo a typo * blaaack * More typo fixes * Fixed setting/getting state in multidiscrete spaces * Fix typo, fix a test to work with the new sampling * Correctly (?) pass the randomly generated seed if np_random is called with None as seed * Convert the Discrete sample to a python int (as opposed to np.int64) * Remove some redundant imports * First version of the compatibility layer for old-style RNG. Mainly to trigger tests. * Removed redundant f-strings * Style fixes, removing unused imports * Try to make tests pass by removing atari from the dockerfile * Try to make tests pass by removing atari from the setup * Try to make tests pass by removing atari from the setup * Try to make tests pass by removing atari from the setup * First attempt at deprecating `env.seed` and supporting `env.reset(seed=seed)` instead. Tests should hopefully pass but throw up a million warnings. * black; didn't expect that, didya? * Rename the reset parameter in VecEnvs back to `seed` * Updated tests to use the new seeding method * Removed a bunch of old `seed` calls. Fixed a bug in AsyncVectorEnv * Stop Discrete envs from doing part of the setup (and using the randomness) in init (as opposed to reset) * Add explicit seed to wrappers reset * Remove an accidental return * Re-add some legacy functions with a warning. * Use deprecation instead of regular warnings for the newly deprecated methods/functions
2021-12-08 22:14:15 +01:00
s = env.reset(seed=seed)
while True:
a = heuristic(env, s)
s, r, done, info = env.step(a)
total_reward += r
if render:
still_open = env.render()
2021-07-29 02:26:34 +02:00
if still_open == False:
break
if steps % 20 == 0 or done:
print("observations:", " ".join([f"{x:+0.2f}" for x in s]))
print(f"step {steps} total_reward {total_reward:+0.2f}")
steps += 1
2021-07-29 02:26:34 +02:00
if done:
break
if render:
env.close()
return total_reward
class LunarLanderContinuous:
def __init__(self):
raise error.Error(
"Error initializing LunarLanderContinuous Environment.\n"
"Currently, we do not support initializing this mode of environment by calling the class directly.\n"
"To use this environment, instead create it by specifying the continuous keyword in gym.make, i.e.\n"
'gym.make("LunarLander-v2", continuous=True)'
)
2021-07-29 02:26:34 +02:00
if __name__ == "__main__":
demo_heuristic_lander(LunarLander(), render=True)