Deep-learning plays a vital role in the development of Autonomous Vehicles. With Convolutional Neural Networks at the heart of vision tasks, objects are recognized and their placement mapped into the ‘mind’ of the car. Lidar is used to capture the environment around the vehicle and then instantaneously recognized by putting multiple views through a previously trained visual model.
Look at the articles cited below to better understand the science of this and how complex the process is.
In case you have not seen this video, imagine if all (but the Motor Cycle) vehicles were autonomous. I guess you could try navigating this intersection.
Good discussion of the fundamental principles behind Autonomous Vehicle vision.
Serious investigation into Convolutional Neural Networks using Pytorch to train the model.
The author mentions using Colab free GPUs to train the model and mentions the