highway-env自定义高速路环境
highway-env自定义高速路环境
问题描述
highway-env自车（ego vehicle）初始状态（位置，速度）可以根据给出的API进行设置，但周围车辆（other vehicles）初始状态为随机生成，不可设置（环境开发作者说的，见下图）。
问题测试
import gym
import highway_env
# Create environment
env = gym.make("highway-v0")
env.configure({
'observation': {
'type': 'Kinematics',
"absolute": False,
"normalize": False
},
'action': {'type': 'DiscreteMetaAction'},
'simulation_frequency': 15,
'policy_frequency': 10,
'other_vehicles_type': 'highway_env.vehicle.behavior.IDMVehicle',
'screen_width': 600,
'screen_height': 150,
'centering_position': [0.3, 0.5],
'scaling': 5.5,
'show_trajectories': False,
'render_agent': True,
'offscreen_rendering': False,
'manual_control': False,
'real_time_rendering': False,
'lanes_count': 2, # 车道数
'vehicles_count': 1, # 周围车辆数
'controlled_vehicles': 1,
'initial_lane_id': None,
'duration': 100,
'ego_spacing': 2,
'vehicles_density': 1,
'collision_reward': -1,
'right_lane_reward': 0.1,
'high_speed_reward': 0.4,
'lane_change_reward': 0,
'reward_speed_range': [20, 30],
'offroad_terminal': False,
'lane_from': 1
})
eposides = 10
rewards = 0
for eq in range(eposides):
obs = env.reset()
env.render()
done = False
while not done:
action = env.action_space.sample()
obs, reward, done, info = env.step(action)
env.render()
rewards += reward
print(rewards)
测试视频如下：highway_setting。上述代码中，定义了两辆车：一辆为自车，一辆为周围车辆，想要实现的场景为，两辆车初始在同一条车道，前车速度慢，后车根据驾驶场景，采用强化学习来实现超车然后换回本车道。然后视频可看出，自车初始化在id为1的车道，周围车辆初始化随机出现在两车道中。
依赖包版本
gym == 0.21.0
highway-env == 1.5
单车控制环境
根据上述无法自定义设置周围车辆，解决方案为：按照自车定义方式，然后加入到道路中，但不加入到控制车辆内，这里重写了奖励函数，代码如下：
import highway_env
from highway_env.envs.common.abstract import AbstractEnv
from highway_env.road.road import Road, RoadNetwork
from highway_env.envs.common.action import Action
from highway_env.vehicle.controller import ControlledVehicle
from highway_env.vehicle.kinematics import Vehicle
LANES = 2
ANGLE = 0
START = 0
LENGHT = 200
SPEED_LIMIT = 30
SPEED_REWARD_RANGE = [10, 30]
COL_REWARD = -1
HIGH_SPEED_REWARD = 0
RIGHT_LANE_REWARD = 0
DURATION = 100.0
class myEnv(AbstractEnv):
@classmethod
def default_config(cls) -> dict:
config = super().default_config()
config.update(
{
'observation': {
'type': 'Kinematics',
"absolute": False,
"normalize": False
},
'action': {'type': 'DiscreteMetaAction'},
"reward_speed_range": SPEED_REWARD_RANGE,
"simulation_frequency": 20,
"policy_frequency": 20,
"centering_position": [0.3, 0.5],
}
)
return config
def _reset(self) -> None:
self._create_road()
self._create_vehicles()
def _create_road(self) -> None:
self.road = Road(
network=RoadNetwork.straight_road_network(LANES, speed_limit=SPEED_LIMIT),
np_random=self.np_random,
record_history=False,
)
# 创建车辆
def _create_vehicles(self) -> None:
vehicle = Vehicle.create_random(self.road, speed=23, lane_id=1, spacing=0.3)
vehicle = self.action_type.vehicle_class(
self.road,
vehicle.position,
vehicle.heading,
vehicle.speed,
)
self.vehicle = vehicle
self.road.vehicles.append(vehicle)
vehicle = Vehicle.create_random(self.road, speed=30, lane_id=1, spacing=0.35)
vehicle = self.action_type.vehicle_class(
self.road,
vehicle.position,
vehicle.heading,
vehicle.speed,
)
self.road.vehicles.append(vehicle)
# 重写的奖励函数，仅考虑车辆碰撞影响
def _reward(self, action: Action) -> float:
reward = 0
lane = (
self.vehicle.target_lane_index[2]
if isinstance(self.vehicle, ControlledVehicle)
else self.vehicle.lane_index[2]
)
if self.vehicle.crashed:
reward = -1
elif lane == 0:
reward += 1
reward = 0 if not self.vehicle.on_road else reward
return reward
def _is_terminal(self) -> bool:
return (
self.vehicle.crashed
or self.time >= DURATION
or (False and not self.vehicle.on_road)
)
if __name__ == '__main__':
env = myEnv()
obs = env.reset()
eposides = 10
rewards = 0
for eq in range(eposides):
obs = env.reset()
env.render()
done = False
while not done:
action = env.action_space.sample()
obs, reward, done, info = env.step(action)
env.render()
rewards += reward
print(rewards)
测试视频如下：highway_env_single，从视频可看出，两辆车均初始化在同一个车道。但存在颜色没法区分问题，有知道的小伙伴可以留言讨论一下哦。
多车控制环境
import random
import highway_env
from highway_env.envs.common.abstract import AbstractEnv
from highway_env.road.road import Road, RoadNetwork
from highway_env.envs.common.action import Action
from highway_env.vehicle.controller import ControlledVehicle
from highway_env.vehicle.kinematics import Vehicle
LANES = 2
ANGLE = 0
START = 0
LENGHT = 200
SPEED_LIMIT = 30
SPEED_REWARD_RANGE = [10, 30]
COL_REWARD = -1
HIGH_SPEED_REWARD = 0
RIGHT_LANE_REWARD = 0
DURATION = 100.0
class myEnv(AbstractEnv):
@classmethod
def default_config(cls) -> dict:
config = super().default_config()
config.update(
{
"observation": {
"type": "MultiAgentObservation",
"observation_config": {
"type": "Kinematics",
"vehicles_count": 2,
"features": ["x", "y", "vx", "vy"],
"absolute": True,
},
},
"action": {
"type": "MultiAgentAction",
"action_config": {
"type": "DiscreteMetaAction",
},
},
"reward_speed_range": SPEED_REWARD_RANGE,
"simulation_frequency": 20,
"policy_frequency": 20,
"centering_position": [0.3, 0.5],
}
)
return config
def _reset(self) -> None:
self._create_road()
self._create_vehicles()
def _create_road(self) -> None:
self.road = Road(
network=RoadNetwork.straight_road_network(LANES, speed_limit=SPEED_LIMIT),
np_random=self.np_random,
record_history=False,
)
def _create_vehicles(self) -> None:
self.controlled_vehicles = []
vehicle = Vehicle.create_random(self.road, speed=23, lane_id=1, spacing=0.3)
vehicle = self.action_type.vehicle_class(
self.road,
vehicle.position,
vehicle.heading,
vehicle.speed,
)
self.controlled_vehicles.append(vehicle)
self.road.vehicles.append(vehicle)
vehicle = Vehicle.create_random(self.road, speed=30, lane_id=1, spacing=0.35)
vehicle = self.action_type.vehicle_class(
self.road,
vehicle.position,
vehicle.heading,
vehicle.speed,
)
self.controlled_vehicles.append(vehicle)
self.road.vehicles.append(vehicle)
def _reward(self, action: Action) -> float:
reward = 0
lane = (
self.vehicle.target_lane_index[2]
if isinstance(self.vehicle, ControlledVehicle)
else self.vehicle.lane_index[2]
)
if self.vehicle.crashed:
reward = -1
elif lane == 0:
reward += 1
reward = 0 if not self.vehicle.on_road else reward
return reward
def _is_terminal(self) -> bool:
return (
self.vehicle.crashed
or self.time >= DURATION
or (False and not self.vehicle.on_road)
)
if __name__ == '__main__':
env = myEnv()
obs = env.reset()
eposides = 10
rewards = 0
for eq in range(eposides):
obs = env.reset()
env.render()
done = False
while not done:
# action = env.action_space.sample()
action1 = random.sample([0,1,2,3,4], 1)[0]
action2 = 1
action = (action1,action2)
obs, reward, done, info = env.step(action)
env.render()
rewards += reward
print(rewards)
测试视频如下：highway_env_muti
后记
highway-env: 手册