System Architecture

Class Section

Simulation

Class Definition

Class Name	Definition
SensorSystem	Collect sensor data
DataCenter	Process data
Router	Re-route the vehicle in certain road network
Controller	Control the vehicle, including steering, accelerating and braking
Env	Provide interface for interacting with various aspects of a traffic simulation
NetWork	Specify features of a network, including the positions of nodes, properties of the edges, and junctions connecting these nodes, properties of vehicles and traffic lights, and other features as well

Function Explanation

• SensorSystem

  Function Name	Explanation
  add_new_sensor(self, detect_type, sensor_name, noise_function=generate_no_noise)	Add new sensor to the SensorSystem
  detect_sensor_data_from_env(self, env, veh_id, detect_type)	Return sensor data from Env
  get_sensors(self)	Return all sensors
  get_fuse_function(self)	Return sensor fuse function
  get_detect_types(self)	Return types of all sensors

• DataCenter

  Function Name	Explanation
  update_data(self, data_center_name, t, veh_id=None,**kwargs)	Update the data named data_center_name of the vehicle with index veh_id at step t in DataCenter
  get_data(self, data_center_name, **kwargs)	Get data names data_center_name from DataCenter

• BaseRouter

  Function Name	Explanation
  choose_route(self, env)	Return the routing method implemented by the controller

• BaseController

  The detailed explanation is presented in Controller section.

• Env

  Function Name	Explanation
  restart_simulation(self, sim_params, render=None)	Restart an already initialized simulation instance
  setup_initial_state(self)	Store information on the initial state of vehicles in the network
  step(self, rl_actions)	Advance the environment by one step
  reset(self)	Reset the environment
  additional_command(self)	Additional commands that may be performed by the step method
  clip_actions(self, rl_actions=None)	Clip the actions passed from the RL agent
  apply_rl_actions(self, rl_actions=None)	Specify the actions to be performed by the rl agent(s)
  compute_reward(self, rl_actions, **kwargs)	Reward function for the RL agent(s)
  terminate(self)	Close the TraCI I/O connection
  render(self, reset=False, buffer_length=5)	Render a frame
  pyglet_render(self)	Render a frame using pyglet
  get_state(self)	Return the state of the simulation as perceived by the RL agent
  action_space(self)	Identify the dimensions and bounds of the action space
  observation_space(self)	Identify the dimensions and bounds of the observation space

• NetWork

  Function Name	Explanation
  specify_edge_starts(self)	Define edge starts for road sections in the network
  specify_internal_edge_starts(self)	Define the edge starts for internal edge nodes
  specify_nodes(self, net_params)	Specify the attributes of nodes in the network
  specify_edges(self, net_params)	Specify the attributes of edges connecting pairs on nodes
  specify_types(self, net_params)	Specify the attributes of various edge types (if any exist)
  specify_connections(self, net_params)	Specify the attributes of connections
  specify_routes(self, net_params)	Specify the routes vehicles can take starting from any edge

Controller

We adopt four different typical controller to evaluate the impact of noise.

We use IDMController and BandoFTLController to validate the effectiveness of our proposed margin in terms of traffic throughput and safety.

Experiment

Gaussian noise with $\mu=5$

Simulation with noise and without noise result:

Untrained model:

Trained model:

Comdrive Project Status

Overview

We're excited to share the Comdrive project with the community. Comdrive is an in-development simulation tool designed with specific goals in mind. It's important to note that the architecture of Comdrive is currently under development, and as such, it may not yet offer full functionality. Our team is diligently working to enhance its capabilities and stability.

Current State

• Development Phase: Please be aware that Comdrive is in a development phase. We're in the process of refining its architecture and expanding its feature set.

• Functionality: Some features within Comdrive are in the beta stage and may exhibit instability or limited functionality. These features are being improved with each update.

• Stability: Given its current development status, Comdrive may exhibit unstable behavior under certain conditions. We are committed to improving its stability and performance.

Usage Considerations

• Feedback: Users are encouraged to provide feedback on any issues encountered or suggestions for improvement. Your input is invaluable as we strive to develop Comdrive into a more robust tool.

• Updates: Regular updates are planned for Comdrive, aimed at addressing current limitations and introducing new features. We recommend staying updated with the latest version for the best experience.

Future Plans

• Open Source: While Comdrive is not yet open-sourced due to its preliminary state, we intend to make it publicly available once we achieve a more stable and comprehensive version. This approach ensures that the community receives a reliable and effective tool.

• Collaboration: We are open to collaboration and contributions from the community. Detailed guidelines will be provided as Comdrive progresses towards a stable release.

Contact and Support

• For any inquiries or support needs, please contact us. We appreciate your interest in Comdrive and look forward to growing with your support.