End2EndAD

Models

Branch	Model
resnet	use resnet train AS NN
cnn	use cnn train AS NN
lstm	use cnn+lstm train AS NN
to be added...	to be added...

Datasets

All data is collected using the "CitySample" simulator.

You can download the dataset at:
Saved_ordered.zip

Contents

The dataset contains three scenarios: go_straight, turning_right, turning_left.
Each scenario includes the following folders:

Folder Name	Description	Example Data
drivings	Contains car status information (Throttle, Brake, Steering, Wheel Angle, Heading, Position, Speed, Acceleration, Direction). Each timestamp is stored in a CSV file named yyyymmdd_hhmmss_mmm.	Filename: 20241103_170827_000.csv Content: Throttle: 0 Brake: 0 Steering: 1 Wheel Angle: 6.058597 Heading: -171.15765 Position: X=-49698.427 Y=13564.725 Z=3.951 Speed: 18.020468 Acc: -0.562561 Direction: Forward
images	Contains left, mid, and right camera images. Each timestamp is stored as a .png image named yyyymmdd_hhmmss_mmm_(L/M/R) (L/M/R are camera tags).	Filename: 20241103_170801_000_M.png
labels	Contains relevant labels for objects in the images. Each timestamp is stored in a CSV file named yyyymmdd_hhmmss_mmm_(L/M/R) (L/M/R are camera tags).	Filename: 20241103_172137_000_L.csv Content: car: 177, 175, 189, 184 car: 128, 169, 161, 194 Human: 0, 171, 8, 186

Data Naming Convention

All data is named as yyyymmdd_hhmmss_mmm or yyyymmdd_hhmmss_mmm_(L/M/R) where:

• yyyymmdd: Date in the format Year-Month-Day.
• hhmmss: Time in the format Hour-Minute-Second.
• mmm: Milliseconds.
• (L/M/R): Camera tags representing Left, Mid, and Right cameras.