Authors: Weize Zhang, Mohammed Elmahgiubi, Kasra Rezaee, Behzad Khamidehi, Hamidreza Mirkhani, Fazel Arasteh, Chunlin Li, Muhammad Ahsan Kaleem, Eduardo R. Corral-Soto, Dhruv Sharma, Tongtong Cao

Published on: May 03, 2024

Impact Score: 7.6

Arxiv code: Arxiv:2405.01394

Summary

• What is new: The application of language-assisted perception models in conjunction with Inverse Reinforcement Learning (IRL) for the motion planner in an autonomous driving context.
• Why this is important: Improving the reliability and performance of autonomous driving systems in challenging traffic scenarios.
• What the research proposes: A modular architecture consisting of five main components (sensing, localization, perception, tracking/prediction, and planning/control) enhanced by language-assisted perception models and IRL.
• Results: Achieved first place in the CARLA Leaderboard 2.0 Autonomous Driving challenge 2023, demonstrating the effectiveness of the proposed solution.

Technical Details

Technological frameworks used: Modular architecture with components for sensing, localization, perception, tracking/prediction, and planning/control.

Models used: Language-assisted perception models, Inverse Reinforcement Learning (IRL).

Data used: Open-source driving datasets.

Potential Impact

Automotive industries, particularly companies focused on the development of autonomous driving technologies; ADAS (Advanced Driver-Assistance Systems) vendors; navigation and mapping services.

Want to implement this idea in a business?

We have generated a startup concept here: MapGear.