This repository is for the project of Udacity Nanodegree - Self-driving Car Engineer : Finding Lane Lines Proejct. It is forked from https://github.com/udacity/CarND-Behavioral-Cloning-P3).
The goals / steps of this project are the following:
- Use the simulator to collect data of good driving behavior
- Build, a convolution neural network in Keras that predicts steering angles from images
- Train and validate the model with a training and validation set
- Test that the model successfully drives around track one without leaving the road
- Summarize the results with a written report
It needs us to apply everything we learned from the lecture including collecting training proper data, set model architecture, tuning hyper-parameters, and check overfitting etc. I used CNN architecture to train model, especially architecture from NVIDIA which has 5 CNN layers and 4 fully connected layers. To reduce training time, I used AWS EC2 GPU instance with Udacity AMI.
I cannot say my final model is perfect, it still falls out from the center lane sometimes, but at least it learned how to recover to the center lane. I will refine this model to perfectly perform on autonomous mode even after submission.
This project has 3 types of outputs:
- model.py : script used to train model. It needs argv to indicate where the training images are
python model.py images- writeup.md : writeup file that specify details on how I completed this project.
- model-final.h5 : final model I got from the project and used to record the autonomous driving
- final.mp4 : autonomous driving video
- drive.py : script for test driving (I didn't modify this file)
python drive.py model-final.h5