SuperDriverAI: Towards Design and Implementation for End-to-End Learning-based Autonomous Driving
This work addresses the problem of autonomous driving for public road use, but it appears incremental as it builds on existing end-to-end learning approaches with added modules for robustness and interpretability.
The paper tackles the challenge of achieving fully autonomous driving on public roads by presenting SuperDriver AI, an end-to-end learning-based system that uses DNNs to learn from human drivers and determine driving maneuvers while ensuring safety, and it was tested with 150 runs in real-world scenarios in Tokyo.
Fully autonomous driving has been widely studied and is becoming increasingly feasible. However, such autonomous driving has yet to be achieved on public roads, because of various uncertainties due to surrounding human drivers and pedestrians. In this paper, we present an end-to-end learningbased autonomous driving system named SuperDriver AI, where Deep Neural Networks (DNNs) learn the driving actions and policies from the experienced human drivers and determine the driving maneuvers to take while guaranteeing road safety. In addition, to improve robustness and interpretability, we present a slit model and a visual attention module. We build a datacollection system and emulator with real-world hardware, and we also test the SuperDriver AI system with real-world driving scenarios. Finally, we have collected 150 runs for one driving scenario in Tokyo, Japan, and have shown the demonstration of SuperDriver AI with the real-world vehicle.