/tactics2d

A 2D Multi-agent Reinforcement Learning Environment for Driving Decision-making

Primary LanguagePythonGNU General Public License v3.0GPL-3.0

Tactics2D

Codacy Codecov

Test Modules Read the Docs Github license

About

Tactics2d provides

Quick Start

Development

Folder Structure

.
├── data
│   └── maps
│       ├── configs
│       └── defaults
├── tactics2d
│   ├── common
│   ├── envs
│   ├── env_wrapper
│   ├── interface
│   ├── map_base
│   ├── map_converter
│   ├── map_parser
│   ├── object_base
│   ├── render
│   ├── utils
│   └── vehicle_kinematics
│       └── base
└── tests

Map Naming

<scenario type>_<data source>_<index>_<country>.osm/xodr

The available scenario types include

Notation Scenario
C Racing track, racing circuit
H Highway
I Intersection without traffic signs
P Parking lot
R Roundabout
SI Intersection with traffic signs
U Unknown scenario or custom scenario

Some modified maps from the following track datasets are available in Tactics2D. They are provided in the Lanelet2 format.