Interpretable End-to-End Urban Autonomous Driving With Latent Deep Reinforcement Learning

By Jianyu Chen, Shengbo Eben Li, Masayoshi Tomizuka. IEEE transactions on intelligent transportation systems (Print), 2020

339 citations recorded by the source corpus.

Unlike popular modularized framework, end-to-end autonomous driving seeks to solve the perception, decision and control problems in an integrated way, which can be more adapting to new scenarios and easier to generalize at scale. However, existing end-to-end approaches are often lack of interpretability, and can only deal with simple driving tasks like lane keeping. In this article, we propose an interpretable deep reinforcement learning method for end-to-end autonomous driving, which is able to handle complex urban scenarios. A sequential latent environment model is introduced and learned jointly with the reinforcement learning process. With this latent model, a semantic birdeye mask can be generated, which is enforced to connect with certain intermediate properties in today’s modularized framework for the purpose of explaining the behaviors of learned policy. The latent space also significantly reduces the sample complexity of reinforcement learning. Comparison tests in a realistic driving simulator show that the performance of our method in urban scenarios with crowded surrounding vehicles dominates many baselines including DQN, DDPG, TD3 and SAC. Moreover, through masked outputs, the learned model is able to provide a better explanation of how the car reasons about the driving environment.

Subfields: Artificial Intelligence, Automotive Engineering, Control and Systems Engineering

View Paper: https://www.semanticscholar.org/paper/c66b647faa4f2970efb0aea2f961a8789cd97db3

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Hongsheng Li
2. Pravendra Singh
3. Hao Shao
4. Peng Wu
5. Yu Liu
6. Bernhard Jaeger
7. Li Chen
8. Pranav Singh Chib

In Automotive Engineering

1. Kashyap Chitta
2. Pravendra Singh
3. Pranav Singh Chib
4. Andreas Geiger
5. Hongyang Li
6. Li Chen
7. Peng Wu
8. Hongsheng Li

In Control and Systems Engineering

1. Pravendra Singh
2. Peng Wu
3. Li Chen
4. Pranav Singh Chib
5. Kashyap Chitta
6. Hongyang Li
7. Andreas Geiger
8. Hongsheng Li

Citations in the corpus

• End-to-End Autonomous Driving: Challenges and Frontiers (2023, methodology): “Early fusion combines sensory inputs before feeding them into shared feature extractors, where concatenation is a common way for fusion [32], [135], [136], [137], [138].”
• Recent Advancements in End-to-End Autonomous Driving Using Deep Learning: A Survey (2023, methodology): “…like Deep Deterministic Policy Gradient (DDPG) and Asynchronous Ad-vantage Actor Critic (A3C) [83], maximum entropy models [81] such as Soft Actor Critic (SAC) [84], and policy-based optimization methods such as Trust Region Policy Optimization (TRPO) and Proximal Policy Optimization (PPO) [85].”
• Safety-Enhanced Autonomous Driving Using Interpretable Sensor Fusion Transformer (2022, background): “Latent DRL [18] generates intermediate feature embedding from a top-2 down view image by training a variational auto-encoder.”