Prateek Sharma

I'm a highly motivated Automotive Engineer with a great passion for Autonomous vehicles.

I am always looking for new things to learn and solve new real-world problems to keep life a little

challenging and interesting. I think self-driving vehicles are going to be an integral part of our

lives in the near future and I am pushing forward to make it a reality.

Projects

  • Github
  • Linkedin
  • Email

Research in Motion Planning

Curvature Optimization and Path Smoothening for RRT

Designed 2 post-processing algorithms to smoothen the path obtained from RRT with Non-Holonomic Constraints

Default Map

Optimizes for the loop formed in the Path

Maze Map

Reduces path length

Scattered Square Map

Less computational time with Goal Check

Comparison Plots

- Reduced the average curvature by 90% and path length by 50% with only a 10% increase in the overhead time

- Path obtained can be directly used by Path tracking controller

IROS 2020

8th F1tenth Las Vegas Grad Prix (Scaled Autonomous vehicle Racing)

Code Performance

Disparity Extender implemented on the f1/10th

Map

Downsized MotoGP map used in the race

Docker Structure

Docker containerization stucture used for head to head race

Turtlebot 3 Burger Implementations

- Hands-on experience in developing and implementing Autonomous driving algorithms in Simulation and actual Embedded System
- Integrated Line following, Obstacle Avoidance, Stop Sign Detection, AprilTag detection, Human Detection and Tracking with Behaivor Planning to make the Turtlebot Fully Autonomously

AprilTag Tracking

Used YOLO for AprilTag Detection

Line Following

Used YOLO for line recognition

Wall Following

Used 2D Point cloud data from Lidar to navigate

Complete Autonomous Driving Stack

Used Finite State Machine for Behavior control

Obstacle Avoidance

Local Navigation in Obstacle Space

Line following in Gazebo

Detecting and Tracking Yellow line's center

Wall following in gazebo

Used 2D Lidar data to navigate

Obstacle Avoidance in gazebo

Local Navigation avoiding static obstacles

Square

Turtlebot basics- Controls

Wall detection

Turtlebot basics- Lidar use

Deep Reinforcement Learning on F1/10th Vehicle

- Enabling a F1/10th vehicle with autonomous driving and obstacle avoidance using Deep Reinforcemnt Learning

- Different action sets and reward strategies were implemented and compared

3 Actions with Obstacle Avoidance

15 Actions with Obstacle Avoidance

8 Actions

5 Actions

15 Actions

3 Actions

F1/10th Implementations

- Hands-on implementation of ADAS Algorithm on F1/10th vehicle
- Devloping and implementing Path Tracking Controllers on F1/10th and CARLA Simulator

ADAS Algorithms on F1/10th

Autonomous Lane Keeping and ACC

Model Predictive Control

Dynamic Path tracking algorithm

Stanley Implementation on CARLA

Geometric Path tracking algorithm

Stanley Controller

Stanley implementation in F1/10th simulator

Follow the Gap

Reactive Method of Obstacle Avoidance

Wall following in F1_tenth_gym_ros simulator

Basic algorithm on F1_10th vehicle

Motion Planning Algorithms

RRT*

Asymptotically optimal version of RRT

8 Agents

Force-based Local Navigation under uncertainity

RRT

Trajectory Planning

Crowd Crossing

Local Navigation under uncertainity for crowd

A* Search

Path Finding algorithms

Breadth First Search

Path Finding algorithms

Depth First Search

Path Finding algorithms

Probabilistic Roadmap- PRMs

Sampling Based path planning algorithm

QT5 Appliactions

Calculator

Calculator with memory option

NotePad

Notepad/ Text Editor

Skills

Programming/ Coding Languages: C++, Python, MATLAB, SIMULINK
Software Development Platforms : Linux, Windows, Arduino IDE, Raspberry Pi, OpenCR
Software : ROS, Git, QT5, VREP, PreScan, Siemens NX, SolidWorks, CarSim
Technical Areas of Interest : Robotics, Motion Planning and Control, ROS, Behavior Planning, ADAS

Get in touch with me

  • Github
  • Linkedin
  • Email
  • Resume