Autonomous
vehicle navigation
using reinforcement
learning

Autonomous Vehicle Navigation System brings to you a way to help achieve better route planning, path prediction, smoother manoeuvrability in dynamic environments and thereby achieving optimized fuel efficiency and
enhanced human comfort.


Reinforcement Learning Algorithms implemented

Deep Q networks
‍


ppo WITH CROPPED IMAGE VIEW
‍

double Dueling DQN
with per


IMITATION LEARNING
‍

proximal policy optimization WITH
VARIATIONAL AUTO ENCODER


ppo WITH IMITATION LEARNING
‍

project videos


Training agent episode 1


Training agent episode 30


training agent episode 100

Autonomous Vehicle Navigation using
double DQN + PER


PPO with Variational autoencoder


PPO with cropped image MIDTERM DEMO


imitation learning
‍


ppo WITH imitation learning - WIDE ROADS


Final Project Video


ppo WITH imitation learning - FINAL DEMO

our team

Aditya jain

Isha chaudhari

Harshita bhorshetti

aditi jain

saichand duggirala

devanshi desai

adwaita jadhav

Engineering Design Document

Technical Paper

“I’m working with these guys from now on!”

John Smith

“Top-notch job. Couldn’t have asked for better design and photography.

Phil Johnson