Autonomous-Driving-with-Udacity-Simulation

Low-cost deep learning-based project capable of autonomous driving in Udacity Simulation

How do you run this repo

• Firstly, run the notebook (you need to collect data first!) and train the model.
• Secondly, If you want, you can tune the functions and model on the notebook for a better driving. Also, It is okay for you to execute the notebook directly because it can drive the car in its current state.
• After training, now you have good and powerful a model. You can execute drive.py and after that just your need is open Udacity Similator in Autonomous Mode.
• You can also tune the drive.py code. But note that the preprocesses must be the same in both notebook and drive.py!

Collecting Data

• You can collect data in Udacity Simulator. 3 laps in either direction is sufficient, but more makes for a much better driving.

Importan note!

• You need to use python-socketio==4.6.1 and python-engineio==3.13.2 for communication between drive.py and simulation.