Self-driving-car for Udacity simulator

Project that simulates a self-driving car for the Udacity simulator

Overview

We're going to use Udacity's self driving car simulator as a testbed for training an autonomous car.

Dependencies

If you have anaconda or miniconda installed on your machine, you can create and install all dependencies on a dedicated virtual environment, by running one of the following commands

# Use TensorFlow without GPU
conda env create -f environments.yml 

# Use TensorFlow with GPU
conda env create -f environment-gpu.yml

Alternatively you can manually install the required libraries using pip.

virtualenv udacity-self-driving-car

source udacity-self-driving-car/bin/activate 

pip install -r requirements.txt

Suggested Setup

We have implemented and tested the simulator on Windows and MacOS Mojave. When executed on the Mac, the simulator seems to suffer from a performance bug (described in this issue), which seems to be due to Unity v.5.5.6f1. However, after exausting testing, we have found this configuration to be optimal for execution on Mac and Windows alike:

• keras==2.2.4
• tensorflow=1.14.0
• numpy==1.16.0

Usage

Run a pretrained model in the original Udacity simulator

Start up the original Udacity self-driving simulator, choose a scene, and press the Autonomous Mode button. Then, run the prea pretrained as follows:

python drive-original-simulator.py  # runs the model within models/model.h5

Run a pretrained model in the improved Udacity simulator

Alternatively, you may want to use the improved Udacity self-driving simulator developed in the work "Misbehaviour Prediction for Autonomous Driving Systems" by Stocco et al. and published at ICSE 2020.

First, download the improved Udacity self-driving simulator. Second, download this autoencoder, and place it in the sao folder. Then choose a scene, and press the Autonomous Mode button. Then, run the model as follows:

python drive.py  # improved Udacity simulator with CTE and effects

If you want to record the data of the simulation in a CSV file run as follows:

python drive.py -t="folder-name"  # records the simulation in the specified directory under the simulations folder

To train the model

You'll need the data folder which contains the training images.

python model.py

This will generate a file model-<epoch>.h5 whenever the performance in the epoch is better than the previous best. For example, the first epoch will generate a file called model-000.h5.

Credits

The credits for this code go to his original creator naokishibuya, llSourcell who created a nice wrapper to get people involved.