def setUp(self):
vehicles = VehicleParams()
vehicles.add("rl",
acceleration_controller=(RLController, {}),
num_vehicles=1)
vehicles.add("human",
acceleration_controller=(IDMController, {}),
num_vehicles=1)
self.sim_params = SumoParams()
self.network = MergeNetwork(
name="test_merge",
vehicles=vehicles,
net_params=NetParams(additional_params=MERGE_PARAMS.copy()),
)
self.env_params = EnvParams(additional_params={
'max_accel': 1,
'max_decel': 1,
"target_velocity": 25
})
HORIZON = 600
# number of rollouts per training iteration
N_ROLLOUTS = 20
# number of parallel workers
N_CPUS = 2
# inflow rate at the highway
FLOW_RATE = 2000
# percent of autonomous vehicles
RL_PENETRATION = [0.1, 0.25, 0.33][EXP_NUM]
# num_rl term (see ADDITIONAL_ENV_PARAMs)
NUM_RL = [5, 13, 17][EXP_NUM]
# We consider a highway network with an upstream merging lane producing
# shockwaves
additional_net_params = ADDITIONAL_NET_PARAMS.copy()
additional_net_params["merge_lanes"] = 1
additional_net_params["highway_lanes"] = 1
additional_net_params["pre_merge_length"] = 500
# RL vehicles constitute 5% of the total number of vehicles
vehicles = VehicleParams()
vehicles.add(veh_id="human",
acceleration_controller=(IDMController, {
"noise": 0.2
}),
car_following_params=SumoCarFollowingParams(
speed_mode="obey_safe_speed", ),
num_vehicles=5)
vehicles.add(veh_id="rl",
acceleration_controller=(RLController, {}),
def get_flow_params(exp_num=1,
horizon=6000,
simulator="traci",
multiagent=False):
"""Return the flow-specific parameters of the merge network.
This scenario consists of a single-lane highway network with an on-ramp
used to generate periodic perturbations to sustain congested behavior.
In order to model the effect of p% CAV penetration on the network, every
100/pth vehicle is replaced with an automated vehicle whose actions are
sampled from an RL policy.
This benchmark is adapted from the following article:
Kreidieh, Abdul Rahman, Cathy Wu, and Alexandre M. Bayen. "Dissipating
stop-and-go waves in closed and open networks via deep reinforcement
learning." 2018 21st International Conference on Intelligent Transportation
Systems (ITSC). IEEE, 2018.
Parameters
----------
exp_num : int
experiment number
* 0: 10% RL penetration, 5 max controllable vehicles
* 1: 25% RL penetration, 13 max controllable vehicles
* 2: 33% RL penetration, 17 max controllable vehicles
horizon : int
time horizon of a single rollout
simulator : str
the simulator used, one of {'traci', 'aimsun'}
multiagent : bool
whether the automated vehicles are via a single-agent policy or a
shared multi-agent policy with the actions of individual vehicles
assigned by a separate policy call
Returns
-------
dict
flow-related parameters, consisting of the following keys:
* exp_tag: name of the experiment
* env_name: environment class of the flow environment the experiment
is running on. (note: must be in an importable module.)
* network: network class the experiment uses.
* simulator: simulator that is used by the experiment (e.g. aimsun)
* sim: simulation-related parameters (see flow.core.params.SimParams)
* env: environment related parameters (see flow.core.params.EnvParams)
* net: network-related parameters (see flow.core.params.NetParams and
the network's documentation or ADDITIONAL_NET_PARAMS component)
* veh: vehicles to be placed in the network at the start of a rollout
(see flow.core.params.VehicleParams)
* initial (optional): parameters affecting the positioning of vehicles
upon initialization/reset (see flow.core.params.InitialConfig)
* tls (optional): traffic lights to be introduced to specific nodes
(see flow.core.params.TrafficLightParams)
Raises
------
AssertionError
if the `exp_num` parameter is a value other than 0, 1, or 2
"""
assert exp_num in [0, 1, 2], "exp_num must be 0, 1, or 2"
# inflow rate at the highway
flow_rate = 2000
# percent of autonomous vehicles
rl_penetration = [0.1, 0.25, 0.33][exp_num]
# num_rl term (see ADDITIONAL_ENV_PARAMs)
num_rl = [5, 13, 17][exp_num]
# We consider a highway network with an upstream merging lane producing
# shockwaves
additional_net_params = ADDITIONAL_NET_PARAMS.copy()
additional_net_params["merge_lanes"] = 1
additional_net_params["highway_lanes"] = 1
additional_net_params["pre_merge_length"] = 500
# RL vehicles constitute 5% of the total number of vehicles
vehicles = VehicleParams()
vehicles.add(veh_id="human",
acceleration_controller=(IDMController, {
"noise": 0.2
}),
car_following_params=SumoCarFollowingParams(
speed_mode="obey_safe_speed", ),
num_vehicles=5)
vehicles.add(veh_id="rl",
acceleration_controller=(RLController, {}),
car_following_params=SumoCarFollowingParams(
speed_mode="obey_safe_speed", ),
num_vehicles=0)
# Vehicles are introduced from both sides of merge, with RL vehicles
# entering from the highway portion as well
inflow = InFlows()
inflow.add(veh_type="human",
edge="inflow_highway",
vehs_per_hour=(1 - rl_penetration) * flow_rate,
depart_lane="free",
depart_speed=10)
inflow.add(veh_type="rl",
edge="inflow_highway",
vehs_per_hour=rl_penetration * flow_rate,
depart_lane="free",
depart_speed=10)
inflow.add(veh_type="human",
edge="inflow_merge",
vehs_per_hour=100,
depart_lane="free",
depart_speed=7.5)
return dict(
# name of the experiment
exp_tag="merge",
# name of the flow environment the experiment is running on
env_name=MultiAgentMergePOEnv if multiagent else MergePOEnv,
# name of the network class the experiment is running on
network=MergeNetwork,
# simulator that is used by the experiment
simulator=simulator,
# sumo-related parameters (see flow.core.params.SumoParams)
sim=SumoParams(
sim_step=0.2,
render=False,
restart_instance=True,
),
# environment related parameters (see flow.core.params.EnvParams)
env=EnvParams(
horizon=horizon,
sims_per_step=5,
warmup_steps=0,
additional_params={
"max_accel": 1.5,
"max_decel": 1.5,
"target_velocity": 20,
"num_rl": num_rl,
},
),
# network-related parameters (see flow.core.params.NetParams and the
# network's documentation or ADDITIONAL_NET_PARAMS component)
net=NetParams(
inflows=inflow,
additional_params=additional_net_params,
),
# vehicles to be placed in the network at the start of a rollout (see
# flow.core.params.VehicleParams)
veh=vehicles,
# parameters specifying the positioning of vehicles upon init/reset
# (see flow.core.params.InitialConfig)
initial=InitialConfig(),
)
def merge_example(render=None):
"""
Perform a simulation of vehicles on a merge.
Parameters
----------
render: bool, optional
specifies whether to use the gui during execution
Returns
-------
exp: flow.core.experiment.Experiment
A non-rl experiment demonstrating the performance of human-driven
vehicles on a merge.
"""
sim_params = SumoParams(render=True,
emission_path="./data/",
sim_step=0.2,
restart_instance=False)
if render is not None:
sim_params.render = render
vehicles = VehicleParams()
vehicles.add(veh_id="human",
acceleration_controller=(IDMController, {
"noise": 0.2
}),
car_following_params=SumoCarFollowingParams(
speed_mode="obey_safe_speed", ),
num_vehicles=5)
env_params = EnvParams(additional_params=ADDITIONAL_ENV_PARAMS,
sims_per_step=5,
warmup_steps=0)
inflow = InFlows()
inflow.add(veh_type="human",
edge="inflow_highway",
vehs_per_hour=FLOW_RATE,
departLane="free",
departSpeed=10)
inflow.add(veh_type="human",
edge="inflow_merge",
vehs_per_hour=100,
departLane="free",
departSpeed=7.5)
additional_net_params = ADDITIONAL_NET_PARAMS.copy()
additional_net_params["merge_lanes"] = 1
additional_net_params["highway_lanes"] = 1
additional_net_params["pre_merge_length"] = 500
net_params = NetParams(inflows=inflow,
additional_params=additional_net_params)
initial_config = InitialConfig(spacing="uniform", perturbation=5.0)
network = MergeNetwork(name="merge-baseline",
vehicles=vehicles,
net_params=net_params,
initial_config=initial_config)
env = MergePOEnv(env_params, sim_params, network)
return Experiment(env)