/self-driving-car-behavioral-cloning

The goal of this project is to train an AV to drive around a simulated race track using and end-to-end deep learning driving model.

Primary LanguageASP.NETMIT LicenseMIT

Behavioural Cloning

Udacity - Self-Driving Car NanoDegree

The goals of this project were the following:

  • Generate and augment a behavioural cloning dataset by driving in a simulator.
  • Build a deep-learning end-to-end driving model that predicts driving actions from camera data.
  • Test, train and validate the model using the simulator driving data.
  • Apply the model in the simulator, recording a video of the completion of one lap of the track.

Demo

Requirements

  • numpy
  • scipy
  • pandas
  • sklearn
  • matplotlib
  • seaborn
  • opencv
  • jupyterlab
  • tensorflow

In addition, this project requires the Term 1 Udacity simulator: https://github.com/udacity/self-driving-car-sim

Usage

To train the model:

  1. Download the relevant release of the Udacity simulator for your platform.
  2. Use the simulator to drive vehicle in manual mode and record data.
  3. python preproc.py
  4. python model.py

To run the trained model:

  1. Run the driving model: python drive.py models/model.h5 run1.
  2. Run the simulator in autonomous mode.