Machine learning model for building Self Driving Cars using Behavioral Cloning. For generating the dataset and testing the model I have used the udacity self driving car simulator.
python run.py --path="path-to-model-file"
For eg: python run.py --path=modelv6_0.1730.h5
The dataset is generated using udacity self driving car simulator. It gives information about the steering angle,throttle and left,right,center images as seen from the front of the car.
-
One out of the three images (left,right,center) is chosen at random from each sample of the dataset.
-
The image is flipped at random.
-
The angle is increased by 0.2 degree if the image chosen is left and decreased by 0.2 degree if the image chosen is right.