# Self Training Autonomous Driving Agent

## TL;DR

This paper introduces a novel reinforcement learning-based autonomous driving agent that uses difference images in its auto-encoder, significantly improving training efficiency while maintaining performance in a simulated environment.

## Contribution

The paper presents a new auto-encoder architecture with difference images that enhances latent space representation for autonomous driving learning.

## Key findings

- Achieves same accuracy with 96% fewer agents
- Requires 87.5% fewer agents per generation
- Uses 70% fewer generations and 90% fewer rollouts

## Abstract

Intrinsically, driving is a Markov Decision Process which suits well the reinforcement learning paradigm. In this paper, we propose a novel agent which learns to drive a vehicle without any human assistance. We use the concept of reinforcement learning and evolutionary strategies to train our agent in a 2D simulation environment. Our model's architecture goes beyond the World Model's by introducing difference images in the auto encoder. This novel involvement of difference images in the auto-encoder gives better representation of the latent space with respect to the motion of vehicle and helps an autonomous agent to learn more efficiently how to drive a vehicle. Results show that our method requires fewer (96% less) total agents, (87.5% less) agents per generations, (70% less) generations and (90% less) rollouts than the original architecture while achieving the same accuracy of the original.