---
title: "Discussion on \"Building an RL Traffic Agent for Indian Roads, Part 4: The Architecture\n\" | Hashnode"
description: "Discussion on \"Building an RL Traffic Agent for Indian Roads, Part 4: The Architecture\n\". Before writing a line of code, I had to pick an architecture — LIDAR vs structured state, discrete vs continuous, and why SAC won over DDPG and TD3."
url: "https://hashnode.com/posts/building-an-rl-traffic-agent-for-indian-roads-part-4-the-architecture/6a76b1ed56b1d360a07c88ba"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [Building an RL Traffic Agent for Indian Roads, Part 4: The Architecture](https://local-minima.hashnode.dev/building-an-rl-traffic-agent-for-indian-roads-part-4-the-architecture?utm_source=hashnode&utm_medium=feed)

Before writing a line of code, I had to pick an architecture — LIDAR vs structured state, discrete vs continuous, and why SAC won over DDPG and TD3.

Aug 8 · 19 min read · local-minima.hashnode.dev  
Read the article

-   [#reinforcement-learning](https://hashnode.com/tag/reinforcement-learning)
-   [#self-driving-cars](https://hashnode.com/tag/self-driving-cars)
-   [#deep-learning-ai](https://hashnode.com/tag/deep-learning-ai)

## Written by

[Jeevika S](https://hashnode.com/@jeevikasrinath)

@jeevikasrinath · Bangalore, India

  
[View profile](https://hashnode.com/@jeevikasrinath)

3rd-year AIML undergrad | writing technical deep-dives on ML & RL research

Documenting research and projects across deep learning and reinforcement learning

## Responses

Markdown supported · ⌘↵ to post

No responses yet.
