Behavioral Cloning

---

Behavioral Cloning Project

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

---

Files

This project includes the following files:

• model.py containing the script to train the model
• utils.py containing a Python generator
• drive.py for driving the car in autonomous mode
• model.h5 containing a trained convolution neural network
• writeup_report.md

Run

Using the Udacity provided simulator and the drive.py file, the car can be driven autonomously around the track by executing

python drive.py model.h5

The model.py file contains the code for training and saving the convolution neural network. The file shows the pipeline I used for training and validating the model. I used two model

Architectures

I used two model architecture for this assignment: the NVIDIA architecture (defined into NVIDIA_model.py) and a Squeeze model with simple bypass (defined into SqueezeNet.py).

The NVIDIA model is a classical architecture with convolutional layers and fully connected layers. The Squeeze model is composed only by convolutional layers and fire modules.

To reduce overfitting I applied dropout and data augmentation. I also applied batch-normalization. The model used an adam optimizer, so the learning rate was not tuned manually.

Training Data

Training data was chosen to keep the vehicle driving on the road. I used a combination of center lane driving, recovering from the left and right sides of the road (in particular for the second track).

Design

For the first track I decided to use the modified NVIDIA model. At the end of the process, the vehicle is able to drive autonomously around the track without leaving the road.

Final Model Architecture

My final model consisted of the following layers:

Layer	Description
Input	160x320x3 cropped and normalized
Convolution 5x5	2x2 stride, valid padding
Convolution 5x5	2x2 stride, valid padding
Convolution 5x5	2x2 stride, valid padding
Convolution 5x5	2x2 stride, valid padding
Convolution 5x5	2x2 stride, valid padding
Flatten
Dropout	0.5
Fully connected	hidden units 100
Fully connected	hidden units 50
Fully connected	hidden units 10
Fully connected	hidden units 1

...

Creation of the Training Set & Training Process

To capture good driving behavior, I first recorded one laps on track two using center lane driving. Here is an example image of center lane driving:

I then recorded the vehicle recovering from the left side and right sides of the road back to center so that the vehicle would learn to .... These images show what a recovery looks like starting from ... :

Then I repeated this process on track two in order to get more data points.

After the collection process, I had ~ 12k data points. I then preprocessed this data by cropping the images and normalizing.

I finally randomly shuffled the data set and put 20% of the data into a validation set.

Reference

NVIDIA model
SqueezeNet model