OTM is a platform for simulating vehicular traffic scenarios. It offers multiple models that can be arbitrarily combined in a single simulation. You can use one of the native OTM models (CTM, 2Q, Newell), or you can create a plugin for your own model. You can similarly implement new controllers as plugins.

Learn more here.

Python API

Most users will interact with OTM through its Python API, which can be installed with pip:

pip install pyotm

This package provides a wrapper class called OTMWrapper, which creates a gateway to the OTM jar file using py4j. OTMWrapper also includes some convenience methods for working with OTM. These are illustrated in the following example. (instersection.xml)

from pyotm.OTMWrapper import OTMWrapper

# in case there is a lingering open gateway
if "otm" in locals():
	del otm
	
# load the configuration file
otm = OTMWrapper("intersection.xml")

# initialize (prepare/rewind the simulation)
otm.initialize(start_time=0.0)

# run step-by-step using the 'advance' method
time = 0.0  # in seconds
advance_time = 10.0
while(time<3600.0 ):
	otm.advance(advance_time)   # seconds, should be a multiple of sim_dt
	print(otm.get_current_time())
	time += advance_time;

# deleting the wrapper to shut down the gateway
del otm

Use run to execute a full simulation.

from pyotm.OTMWrapper import OTMWrapper
import numpy as np

if "otm" in locals():
	del otm

otm = OTMWrapper("intersection.xml")
otm.run(start_time=0,duration=2500,output_dt=10)
Y = otm.get_state_trajectory()
print(Y.keys())
del otm

Here get_state_trajectory has been used to retrieve the state trajectory. These are the flow and density profiles in each of the links in the network. You can extract more fine-grained data using ``output requests'' or through the Java API. An output requests prompts the simulator to dump specific outputs to text files. You can see which outputs are available here.

The OTMWrapper exposes OTM's Java API through its otm attribute, which is an instance of core.OTM. From here you can place requests to the Output object, or gain access to all of the scenario elements (links, nodes, lane groups, controllers, sensors, actuators, etc.) via the Scenario object. Both of these are otained with their getters.

The following example demonstrates some of the OTM API getter methods.

from pyotm.OTMWrapper import OTMWrapper

# in case there is a lingering open gateway
if "otm" in locals():
	del otm
	
# load the configuration file
otm = OTMWrapper("intersection.xml")

# initialize (prepare/rewind the simulation)
otm.initialize(start_time=0.0)

scenario = otm.otm.scenario()
actuators = {id:scenario.get_actuator(id) for id in scenario.actuator_ids()}
commodities = {id:scenario.get_commodity(id) for id in scenario.commodity_ids()}
controllers = {id:scenario.get_controller(id) for id in scenario.controller_ids()}

network = scenario.network()
link_ids = list(network.link_ids())
links = {id:network.get_link(id) for id in link_ids}
nodes = {id:network.get_node(id) for id in network.node_ids()}
source_link_ids = network.source_link_ids()
link1 = links[1]

# run step-by-step using the 'advance' method
time = 0.0  # in seconds
advance_time = 10.0
while(time<3600.0 ):
	otm.advance(advance_time)   # seconds, should be a multiple of sim_dt
	print(otm.get_current_time(), link1.get_veh())
	time += advance_time;

# deleting the wrapper to shut down the gateway
del otm