Overview / Usage

This project covers important concepts in autonomous driving for eg., in terms of perception: lane detection, traffic sign recognition, vehicle detection etc, with an aim to further improve them in the future.

Digging deep: Sensor fusion, Semantic segmentation, Path Planning, Localization

Methodology / Approach

Perform an OpenCV technique like Canny to detect the edges of lane lines
Train a deep neural net based on LeNet architecture using TensorFlow or Keras to classify different traffic signs efficiently (a stop sign, a go sign, or maybe no sign at all!)
Different image processing techniques like HOG transforms and a sliding window for vehicle detection, YOLO deep neural network etc
Model predictive control & non-linear optimization for path planning, Unscented & extended Kalman filter for Localization

Technologies Used

Signal processing, Deep learning, OpenCV, TensorFlow, Keras, C++, Python

Repository

https://drive.google.com/open?id=1qwfXfAWyf0XtixPZCHocKsNyTbZL26xt