commonroad-agent 1.2.0

Agent simulation with behavior models for CommonRoad

CommonRoad Agent Simulation

This repository comprises module to perform multi agent simulations with CommonRoad scenarios. Dynamic obstacles in a scenario can be replaced by agents with certain behavior models implemented in this repository. This allows to perfom simulations with interacting traffic participants that do not follow predefined trajectories but react on other traffic participants.

Supported Behavior Models

If nothing specified the dynamic obstacles in a scenario move according to the predefined trajectories. If assigned like shown in the examples, the scenario can be simulated forward with the step() while the considered dynamic obstacles behave according to the specified behavior model. Currently, the following behavior models are implemented:

IDM

The IDM is a well known car following model. It does not include lane changes and is suited best for highway scenarios. An example can be found here

MOBIL

MOBIL is an extension of the IDM with lane change decisions. An example can be found here

LEVEL_0_POLICY

The level 0 policy defines a basic behavior of an interaction unaware traffic participant. This policy will be used to train higher level agents with reasoning hierarchy.

Installation

The following need to be installed on your machine:

Requirements

• Python >= 3.6

Installation via pip:

pip install commonroad-agent