Path Planning and Control
Path Planning and Control

This project contains implementations and demos of Path Planning and Control algorithms for Robotics. It features the Potential Field, Rapidly Exploring Random Trees (RRTs) and Probabilistic Roadmap (PRM) approaches to path planning and, the Pure Pursuit and Stanley controller for motion control.

Python
Advanced Lane Finding
Advanced Lane Finding

This project looks at tracking lane lines in a video stream under challenging lighting conditions. A Python notebook is used to demonstrate the steps required to achieve good tracking performance

PythonJupyter Notebook
Traffic Sign Classification
Traffic Sign Classification

A look into Traffic Sign Classification using Deep Learning. A Python notebook is used to outline the stepwise instructions to train a classifier on the German Traffic Sign Recognition Benchmark (GTSRB) using Tensorflow

PythonTensorflowJupyter Notebook
Turtlebot Mapping
Turtlebot Mapping

Implementation of a mapping algorithm using Log-Odds Updates and Bresenham's Line algorithm. A kinect-equipped Turtlebot is used with an Indoor Positioning System for localization

C++ROSSLAM
Reinforcement Learning for Sequential Robotics Tasks
Reinforcement Learning for Sequential Robotics Tasks

The core of my Master's Thesis, a novel approach to Hierarchical Reinforcement Learning and Reward Decomposition for tasks comprising of sequentially composed subtasks

PythonPyTorchReinforcement Learning
Behavior Cloning for Car Simulation
Behavior Cloning for Car Simulation

In this project, we train a car to drive around a racetrack using Behavior Cloning. Convolutional Neural Networks (CNNs) are trained with augmented simulator image frames and expert data obtained through human drivers

PythonTensorflow