adelbennaceur / self-driving-car-simulation

Behavioral cloning: end-to-end learning for self-driving cars.

self-driving-car-simulation

Self driving car simulation using Udacity.The goal of this project is to make a car drive autonomously using a deep learning approach by feeding an input image to a neural network and predicting the steering angle.

Code structure:

• main.py : contains the script to train the neural network.
• model.py: contains the model (neural network) architecture.
• utils.py: contains functions for preprocessing and loading the data.
• drive.py: contains the script to connect to the simulator and run your model.

Model architecture

the model is inspired from the nvidia End to End self driving car network.

Dataset:

The dataset is collected from the udacity self driving car simulator.

Dataset directory structure:

Directory
├── driving_log.csv
└─┬ IMG
  └── center_2019_03_12_17_11_43_382.jpg
  └── .....

Results

Install dependencies:

$pip install -r requirements.txt

Train on your own dataset:

run the command to see the available arguments:

$ python main.py -h

Example:

$ python main.py -dir path/to/dataset -optimizer rmsprop -lr 0.001 -batch_size 32 -epochs 20