This repository provides a list of relevant ROS2 packages and other open source packages in the autonomous driving space. There are other great lists with slightly different scope for ROS2 and general robotics.

Software stacks

Name	License	Link	Language	Short description
Autoware	Apache 2.0	here	C++	Full software stack
AWS DeepRacer	Apache 2.0	here	C++	Collection of packages for AWS DeepRacer including ros2 integration
Apollo	Apache 2.0	here	C++	Full software stack

Ground Filtering

Name	License	Link	Language	Short description
CSF	Apache 2.0	here	C++	Cloth Simulation Filtering from aerial lidar
Patchwork	MIT	here	C++	Filtering, 43Hz with Ouster 0S0
Cascaded Ground Segmentation	Apache 2.0	here	C++	Slope based filtering
GndNet	MIT	here	Python	ground removal with neural network
Fast Ground Segmentation of 3D Point Clouds	GPL 3.0	here	C++	makes use of scan line properties
Ground Removal	None	here	Python	also check out original paper, only python code

Sensor Calibration

Name	License	Link	Language	Short description
ILCC	BSD2	here	Python	3D-LiDAR and camera extrinsic calibration
mi-extrinsic-calib	LGPL	here	C++	3D-LiDAR and camera extrinsic calibration
automatic_lidar_camera_calibration	None	here	C++	3D-LiDAR and camera extrinsic calibration

Object detection

Name	License	Link	Language	Short description
Apollo	Apache 2.0	here	C++	Mask-Pillars Lidar Object Detection
Depth Clustering	MIT	here	C++	Clustering for Detection
mmdetection3d	Apache 2.0	here	Python	Variety of Lidar Object Detection Models
Open3D-ML	MIT	here	Python	Variety of Lidar Object Detection Models
OpenPCDet	Apache 2.0	here	Python	Variety of Lidar Object Detection Models
CFR-Net for Object detection	Apache 2.0	here	Python	Object detection with a neural net based on camera and radar data

Lane detection

Name	License	Link	Language	Short description
Papers with Code		here		Overview Page with Benchmarks
TuSimple Benchmark	Apache 2.0	here	Python	Lane Line Dataset Loader
CU Lane Dataset	Non-comercial	here		Lane Line Dataset
lanedet	Apache 2.0	here	Python	Collection of Multiple Lane Detectors
RESA	Apache 2.0	here	Python	RESA: Recurrent Feature-Shift Aggregator for Lane Detection
lanenet-lane-detection	Apache 2.0	here	Python	Multi-Stage Lane Detection
Ultra-Fast-Lane-Detection	MIT	here	Python	300+ FPS in Benchmark
Codes-for-Lane-Detection	MIT	here	Python	Several Networks
LaneAF	None	here	Python	Robust Multi-Lane Detection with Affinity Fields

3D segmentation

Name	License	Link	Language	Short description
lidar_bonnetal	MIT	here	Python	Lidar Range Image Segmentation
SalsaNext	MIT	here	Python	Uncertainty-aware Lidar Segmenation
Cylinder3D	Apache 2.0	here	Python	Cylindric Convolutions for Lidar Segmentation

General perception

Name	License	Link	Language	Short description
Awesome Multiple Object Tracking		here		Meta repository with papers and code
easy_perception_deployment	Apache 2.0	here	C++	Deployment of CV models
ros2_object_analytics	Apache 2.0	here	C++	General purpose object detection and tracking packages
vision_opencv	None	here	C++	Interface to OpenCV
slam_gmapping	None	here	C++	Wrapper for gmapping SLAM algorithm
cartographer	Apache 2.0	here	C++	Wrapper for cartographer SLAM algorithm
lidarslam_ros2	BSD-2-Clause	here	C++	NDT scan matching SLAM algorithm
slam_toolbox	LGPL-2.1	here	C++	Holistic library with a variety of SLAM options
perception_pcl	None	here	C++	Functions for handling pointcloud data with PCL
image_common	None	here	C++	Functions for handling images
ros_msft_onnx	MIT	here	C++	Wrapper for inference with ONNX models

Prediction

Name	License	Link	Language	Short description
Wale-Net Prediction Network	LGPL-3.0	here	Python	Encoder-decoder neural network for vehicle trajectory prediction with uncertainties

Planning algorithms

Name	License	Link	Language	Short description
Global racetrajectory optimization	LGPL-3.0	here	Python	Various approaches from minimum curvature to minimum time
Graph-based local trajectory planner	LGPL-3.0	here	Python	Local trajectory planner with independent graph-based path and velocity planning
Velocity optimization module	LGPL-3.0	here	Python	Optimization-based velocity planning module considering external parameters (e.g., the friction coefficient or a maximum power limit)
Trajectory supervisor	LGPL-2.1	here	Python	Online verification framework for autonomous vehicle trajectory planning

Control algorithms

Name	License	Link	Language	Short description
Vehicle motion control	LGPL-3.0	here	Simulink	Motion control (LQR) and state estimation for autonomous vehicles

Hardware drivers

Name	License	Link	Language	Short description
Ouster LIDAR	Apache 2.0	here	C++	Supports all models
Velodyne LIDAR	Apache 2.0	here	C++	Supports 64E(S2, S2.1, S3), the 32E, the 32C, and the VLP-16. Additional functionality available.
Smartmicro Radar	Apache 2.0	here	C++	Supports UMRR-96
CAN interface	Apache 2.0	here	C++	Interfaces the socketcan API of linux
OXTS INS	Apache 2.0	here	C++	Supports all OXTS units
Novatel INS	MIT	here	C++	Supports OEM7 units

Tools & Simulation

Name	License	Link	Language	Short description
CARLA	MIT	here		Full stack simulation environment based on the Unreal Game Engine
Foxglove	MPL-2.0	here		Web-based visualization and diagnosis tool
Webviz	Apache 2.0	here		Web-based visualization and diagnosis tool
PlotJuggler	LGPL-3.0	here		Time-series visualization tool

Papers & Articles

Name	Link	Short description
Racing overview	here	Meta repository with papers and code
Autonomous Vehicles on the Edge: A Survey on Autonomous Vehicle Racing	here	Survey paper on autonomous vehicle racing
Oops! It's Too Late. Your Autonomous Driving System Needs a Faster Middleware	here	Analysis of ROS1 / ROS2 / Cyber with respect to latency and performance
A Self-Driving Car Architecture in ROS2	here	Software architecture for a vehicle with ROS2
Real-time configuration with ROS2	here	Details on configuration of ROS2 and OS for RT performance
Latency Analysis of ROS2 Multi-Node Systems	here	Detailed latency analysis for different ROS2 middlewares