My Projects
Below is a brief about my recent projects.
Pose Estimation, Categorization and Segregation using Robot Manipulator (POSCAR) [GitHub]
Nov 2020 - Dec 2020
- Used NVIDIA DOPE to identify the object and estimate its 6D Pose with the image input from a 2D RGB camera.
- Simulated a Robotic arm with a camera near the end-effector to identify the pose of the objects and then move the robotic arm to the pose estimated by DOPE to then manipulate the said objects in the scene.
Quadcopter
Jun 2020 – Jul 2020
- Built a Quadcopter powered by an ArduCopter microcontroller and operated by an 8 channel PWM Remote Control.
- Installed a GPS to exploit Return-To-Launch (RTL).
Industrial Goods Loading System [GitHub]
Apr 2020 – May 2020
- Implemented a cargo handling system where a 2DOF robotic Arm, operated by a Raspberry Pi, picks up cargo from the conveyor and places it onto a cart, operated by Arduino.
- Using OpenCV, identified the location of the cart at the loading area for the robotic arm to place the cargo.
- Established Bluetooth Communication between the Raspberry Pi and the Arduino using HC-06, to convey the arrival of the cart and successful placement of the cargo on the cart.
Optimal and Learning Control Assignment Projects [GitHub]
Jan 2020 – May 2020
Implemented a Series of Assignment Projects. Several concepts used in implementing these assignment projects include:
- Dynamic Programming
- Lable Correction Path Planning Algorithms (Breadth-First Search, Depth First Search, A-Star)
- Linear–Quadratic Regulator (with and without desired trajectory and constraints)
- Model Predictive Control
- Reinforcement Learning Check out my project’s Github page for more details about my Assignment Projects.
Walking a 1D-Linear Inverted Pendulum Model (LIPM) [GitHub]
May 2020 - May 2020
- Implemented ‘Model Predictive Controller’ to enable a 1D-LIPM to the desired steps by tracing the motion plan generated by a ‘Linear Quadratic Regulator.’
- Compiled a code to generate the desired steps of variable velocity for the Motion Planner.
- Implemented ‘Push recovery’ by identifying Instantaneous Capture Point at each time step to perform necessary stepping in order to maintain balance. The model reacts similarly to how a human would.
SONAR for Visually Challenged
Apr 2020 - Apr 2020
- Built a wearable device for the visually challenged to help perceive distance.
- The device gently vibrates to enable the user to sense the closeness of an object/obstacle (past a given threshold).
- Used the parallel processing capabilities of a Parallax Propeller to constantly monitor the distance using a PING ultrasonic sensor and enable the corresponding vibration using an oscillator.
Snake-Mic [GitHub]
Feb 2020 – Mar 2020
- Implemented a mobile robot, designed to look like a Mic for Artistic Performance as required by Dramatist and Director of Chilean theater, Professor Alejandro Moreno Jashes. It was operated by an Arduino Nano.
- Made an Android App to control the robot using Bluetooth Communication (HC-06 connected to Arduino Nano to enable Bluetooth Communication).
- The Snake-Mic can be controlled to move around the room by an artist using an Android Phone, while the Mic plays good quality music through its Bluetooth Enabled 3” Speaker Driver.
Model to avoid deaths in cars due to hypothermia and suffocation in cars
Nov 2019 – Dec 2019
- Developed a solution to avoid deaths in cars due to hyperthermia and suffocation of kids, disabled, or pets in left in cars for too long by accident.
- Simulated a hot car environment using a box.
- Used an Arduino to monitor the temperature, and the CO2 (and CO) levels using a Digital Temperature Sensor and an Analog CO2 Sensor.
- Used PIR sensor to detect any movement in the car indicating life.
- Established Bluetooth Connectivity between the Arduino and the user using an HC-05 module. Once the threshold is crossed the Arduino Communicates the predicted danger to the user and based on the feedback, lowers the windows of the car (to avoid suffocation) and/or turns on the Air Conditioning System of the car (to avoid hypothermia) based on the sensor outputs.
Implement a Kinematic Model of a SCARA Manipulator on MATLAB
Oct 2019 – Nov 2019
- Simulated the differential kinematics model of a SCARA manipulator using MATLAB and SIMULINK.
- Implemented the desired trajectory with minimum error and controlled using the Inverse Dynamics control.
Control an Inverted pendulum in ROS [GitHub]
Nov 2018 - Nov 2018
- Controlled the simulation of an Inverted Pendulum Model on Gazebo.
- Used and tuned a PID controller complied on Python as a ROS Node to balance the Inverted Pendulum.