One of automotive industries most actual trend is developing autonomous vehicles. This can help decreasing the number of road accidents and the damage caused by them. It also makes driving more comfortable as cars are getting more and more advanced driver assistance functions. Driving needs less attention, and they can use their time of travelling for other, more comforting, or useful activities.
To build a car with such capabilities it needs to ’see’ its environment through sensors. In case of ’Lane Keeping assist’ automakers often use a camera as the sensor with a running image recognition software. This software calculates many parameters, of the visible lines, such as distance or angle to the centerline of the vehicle. From these data the controller software calculates a steering angle.
During my thesis work I have to implement a lane keeping assist software on a gokart that was previously prepared for the Go-Kart! Go-Bosch! competition. The Gokart has an MPC2 automotive camera, which can provide the needed data of the lines. Previously, for my projects, I have made a lane keeping software, which was tested in simulated environment. My task is now to improve the software, and implement it on the gokart.
The gokart had a telemetry module, which functioned as a gateway between the central computer and the radar, and the camera. In this setup the lane keeping assist did not work well so we decided to remove the telemetry and implement it on the central computer, so we can rewrite its software to get better performance. I also have to write a new core software for the gokart, that is functioning more like real cars. Finally I implemented the lane keeping function.