Reinforcement Learning to Drive a Car by Pattern Matching
Proceedings of the 24th DAGM Symposium on Pattern Recognition
| Hi-index | 0.00 |
This abstract describes the current research in the area of autonomously driving of a vehicle along different road courses [1]. The focus of this paper are two main aspects: firstly, parameters of the environment are being extracted from a video image coming from one single camera which is installed in or in front of the vehicle which is to drive along the road course; secondly, the incoming images from the camera need to be processed by a computer system that way, that not only Steering Commands for the vehicle are being generated (for accelerator / brake as well as the steering wheel) but the appropriateness of those Steering Commands is being constantly weighed and continuously improved over time. Consequently, the current work focuses on a system which is able to learn and to develop completely on its own the ability to steer different vehicles in different environments and combines research in the areas of Intelligent Image Processing, Machine Learning and Robotics.