Deep Learning for Vision and Decision Making in Self Driving Cars-Challenges with Ethical Decision Making

Engineering notes

Key topics: autonomous driving, self-driving vehicle, self-driving car, self-driving, lidar, radar. See the paper for implementation details and experimental results.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为端到端自动驾驶、BEV感知、3D目标检测、轨迹预测、路径规划、LiDAR感知等高价值论文补充中文说明。

Original abstract

In recent times self-driving vehicle is a revolutionary Idea. Self-Driving cars are the driver less cars where the car runs by itself called as Autonomous cars. There are lot of underlying Technologies like Machine Learning and Deep Learning. This paper focuses on vision and decision making capabilities of self-driving cars using Deep Learning and also ethical challenges while making sudden decisions. Deep Learning is a sub part of Machine Learning which is inspired by the working process and functionality of biological neurons. Vision is the important and strong aspect of self-driving car that contribute a lot i.e., sensing the environment to identify obstacles, reading traffic signs, understanding traffic signal light status, traffic light count down time recognition and finally making appropriate decision on what it sees. Decision making is a complex task when it comes to real time scenarios, where the autonomous driving agent need to make decisions based on the data generated in real world. There is huge amount of data generated by various sensors, radars and LIDAR’s. Every time it is like a new experience for the autonomous driving agent, it is difficult to make different policies for different driving scenarios because of existing rules and amount of previous data available. Decision making depends on different technologies like computer vision and deep learning. Ethical decision making, sacrificial decision making is really a tough job for autonomous driving agent in unavoidable accident situations where people lives are at stake.

Links and sources

• Official / arXiv page

Comments