CarND Term 1 P3: Behavioral Cloning with Deep Learning
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
examples
.gitignore
README.md
drive.py
helper_functions.py
model.h5.7z.001
model.h5.7z.002
model.h5.7z.003
model.py
video.mp4
video.py
viewer.py
writeup_report.md

README.md

Driving Behavioral Cloning with Deep Learning

Udacity - Self-Driving Car NanoDegree

About this repository

This repository is a project that simulates a behavioral cloning for autonomous driving using the deep learning approach.

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

Requirements

  1. Install the Udacity starting kit from https://github.com/udacity/CarND-Term1-Starter-Kit

  2. Get the Udacity simulator:

Installation

git clone https://github.com/ywiyogo/CarND1-P3-BehavioralCloning.git

Usage

  1. Merge the model.h5.7z.001, model.h5.7z.002, model.h5.7z.003 containing a trained convolution neural network to model.h5

  2. Go to the repository folder and activate the virtualenvironment:

     source activate carnd-term1
    
  3. Start the deep learning

     python drive.py model.h5
    
  4. Start the simulator program and choose the "Autonomous Mode"

  5. Optional: to record a video of the simulation, run the 3rd command above with an output directory name as the second argument:

     python drive.py model.h5 <output_dir_name>
    

    After the simulation convert the images in the output directory to a video by executing this command:

     python video.py <output_dir_name> --fps 48
    

Documentation

Please read this writeup