Repo for my masters thesis for a deep learning based self-driving car.
The goal is to create a self-driving car model that, rather than dirving the car, will assist the driver incase the driver is not driving up to standards, or appropriately. i.e. If the dirver is swaying within their lane, not slowing down for a intersection etc.
Stuff to do:
init CARLA environmentadd custom maps in CARLAneed to add more maps
add agents in CARLACollect training data from CARLAidentify the required training attributescollect the right amount of training and test data
- Create appropriate R-CNN model (or other)
possibly requires 2 models with a seperate controller.
- Figure out weight distributions for different data types
- simulate
- ???
- Graduate