This includes resources and other aids used in this task. Also add them in the google sheet if you will (It also has the Sensors - Cost and Accuracy sheet):
- List of some LiDARs
- Comparing LiDARs
- Not related to our project: Luminar LiDARs: A startup bought by Volvo. Inexpensive LIDARs (from $500). Too bad they're coming up in 2022.
Using Plugins For Sensors/Controller - Gazebosim
Steps to ROSify custom build robots - ROS Answers
Connecting robot with ros using gazebo plugins - Gazebosim
A Research Paper on Collision detection and obstacle avoidance - pdf
Motion control algorithms for sensor equipped robots - pdf
Lane keeping in autonomous driving with Model Predictive Control & PID - Article
Curvature of lane and offset from centre of lane - Github
Some thesis that we're following
- Fixed errors in all three world files, collision bugs, xml syntax bugs etc
- Assembeled virat from meshes
- Added cameras(x2) and a Differential drive
- Added teleop package
- Added lidar
- Added slam package for mapping surroundings
-
Choosing the relevant files from the resources given.
-
Spawning a "dry model" of the bot without any sensors, plugins or such.
-
Choosing the right sensors and plugins for the bot.
-
Setting up object and lane detection models and integrating them with ROS.
-
Setting up the navigation stack for the bot - the path-planners and localisation techniques.
-
Finding
or makinga controller. -
Going through the IGVC guide.
- added wheels and started the assembly
- created the xacro and gazebo files for better handling of files.
- added cameras
- created urdf file without sensors for dry sim
- added laser sensor
- created basic launch files for worlds, tweaked the igvc_basic.world a little to avoid the error
- add bot model to launch file
- teleop package complete
- make slam package
- MPC controller for differential drive
- Added fake node for controller diagnostics
- Path visualizer for controller in Rviz
- CMake workspace for controller diagnostics