baleshwar-mahto Goto Github PK

1. Shreem Autonomous Robot :
-- Co-implemented Epsilon Star algorithm in Python for online coverage path planning of unknown environment to navigate autonomously in an inclined rectangular island of solar panels .
-- Done research work and read research papers in related domains and suggested optimal method for development .
-- Derived mathematically the angle of inclined solar panels with ground in terms of roll and pitch angles provided by IMU ( Inertial Measurement Unit ) sensor to compute the required traction for auto-hold functionality on inclined solar panels .

2. Shreem Remote Controlled ( RC ) Robot :
-- Created the initial documentation for the Shreem Remote controlled robot through reverse engineering .

3. Unicorn Fully-automatic Robot :
-- Assisted Lead Engineer in troubleshooting 42 control panels ( Control Unit ) of the Unicorn fully-automatic robot .
-- Created root cause analysis ( RCA ) report on the failures of the control panels at that specific site and suggested optimal methods to further stablize it .
-- Contributed code to multiple Unicorn robot related projects for new features enhancement and enabling wireless communication between the robot and remote monitoring station . (Modbus RTU & LoRaWAN )

-- Working on the solar energy generation forecasting using Hybrid methods and its deployment in the Heroku cloud platform . The motive of this research is to enhance the accuracy of the existing state of the art forecasting models using new emerging techniques and create a user interface for others to interact with my model .

-- Worked on the project titled " Trajectory Planning of Robotic Arm Using Reinforcement Learning " using UR5 robotic Arm by the Universal Robots .
-- Worked and applied Q-learning algorithm to make the tip of UR5 to move in a circular trajectory using openai_ros package in Robotic Operating System ( ROS) Kinetic version .
-- Trained the reinforcement learning agent for 6-7 hours in the online the Contruct platform and found the agent to learn .

-- Designed and developed an Automatic License Plate detection and Recognition system for Rural India with specific design constraints of being deployable in a Raspberry Pi , working without internet connectivity and build using Open Source .

-- Applied Tesseract OCR for License plate recognition and was able to improve its accuracy by 30% - 40% using image processing techniques.

-- Deployed YOLO v4 model having weight size of 244 MB in Raspberry Pi 3 B of 1GB RAM using Darknet framework .

Discrete-time Sliding Mode Controller Design :
-- Researched about the latest papers on sliding mode control . Designed a discrete-time sliding mode controller for a two-wheeled mobile robot , which is a kind of autonomous guided vehicle , for path tracking of basic paths , namely circular path and straight line path .