open research topics for the model car - AutoModelCar/AutoModelCarWiki GitHub Wiki

GPS-Simulation

The cars has a camera pointing up, to the ceiling. Use markers on the ceiling, or features on the ceiling, to locate the car on the floor. Define an xy system of coordinates, with an origin, and provide the xy-position of the car in meters according` to the camera image.

Robust line detection considering light reflection on the road

Changes of contrast, from sun light, for example, can make line detection difficult.

Develop an illumination-adaptive algorithm for line detection.

Create map editor to make an automatic map.

Use SLAM methods to build a map of the road after one or more manual drives.

Car Localization

Using Odometry and camera data/Lidar data find the car position on the map.

Use simulated GPS to locate the car on the road`

Apply everything and get the best possible localization in the map.

Trajectory planning

implement Astar and RRT for following the road and overtaking, also for parking

Obstacle recognition

Recognize important obstacles using the map: cars obstructing the road, cars coming from left or right at intersections, cars coming from behind before overtaking.

Communication between cars.

Develop car-to-car communication.

Recognition of traffic lights

Develop a traffic light recognition module.

Autocharging

Develop a hardware module for autocharging. When the battery is running low the car goes to an autocharging station.

Autonomous Sensor Calibrations

Each sensor should be calibrated automatically. For example: servo motor angle could be calibrated using Lidar, the angle of the camera could be detected using straight lines on the ground, ...

Drive the car using an Oculus Rift and a wheel joystick

lane keeping usging deep learning