cv

Masters in Robotics (Electrical and Computer Engineering)

Northeastern University, Boston, MA - 02120

• Coursework - Mobile Robotics, Robot Mechanics and Control, Robotics Sensing and Navigation, Computer Vision.
• GPA - 3.9/40.

B.Tech. in Instrumentation and Control Engineering

Vishwakarma Government Engineering College, Ahmedabad, India - 382424

• Coursework - Process Control, Industrial Instrumentation, Control Theory.
• GPA - 8.2/10.0.

Research Assistant

Northeastern University, Boston, MA

• Here we are working on High vehicle speed and off-road environments each presenting exciting challenges for modern autonomy pipelines. To achieve high speed, the algorithms must quickly reason about the massive amount of data coming from various sensor streams. To navigate off-road, it is critical to understand both the geometry and semantics of the scene (e.g., a robot may be able to drive through soft bushes but not a fallen log). Our work is focused on learning a cost-to-go estimator to enable aggressively, online model predictive control (MPC)-based motion planning.
• We aimed to develope a 3D semantic map of the environment by fusing the RGB frame of the camera with a 3D lidar. This fused data was capable of segmenting the point cloud of object classes.
• Research work on High-Speed, off-road Autonomy in robots using online MPC (Model Predictive Control) based motion planning.
• Autonomy & Intelligence Laboratory

Embedded Software Intern

Studio Carbon, Gandhinagar, India

• As an Embedded software intern I was reponsible for programming and developing the mathematics of a smart fitness device. This device utilised a small laser pointer on top of a small laser based distance sensor. This whole assembly rotated on top of a 3d printed cylinder with the help of a DC encoded motor that gave the Yaw motion to the device.
• The pitch motion was acheived with the help of a small stepper motor. Here I designed a feed-forward and PID control system, reducing motor positioning error from 20° to 5° increasing the accuracy of the motor positioning, and programmed the firmware for a smart stepper motor.
• Apart from this I was also reponsible for programming a BLE stack for a smart fitness device, enabling seamless communication with a mobile app. I also developed a closed-loop control system for a laser-based distance sensor, capable of accurately counting the number of taps.

Embedded R&D Intern

WeHear Hearing Solutions, Ahmedabad, India

• Developed a BLE-based hearing aid device employing bandpass filtering on MEMS I2S microphone data.
• Utilized RTOS to optimize data communication and filtering, reducing latency from 400ms to 90ms while enhancing audio quality.
• Created multi-layer PCBs to industry standards for product level, increasing battery runtime from 2.5 hours to 5.5 hours.
• Programmed a Spectral Analysis devices for drug authentication and verification and tested around 50 different drugs with 89% accuracy

Management and Technical Core Team Member (EECE)

GTU Robotics Club, Gandhinagar, India

• This is the place where it all started. For a tenure of 1 year (2019 - 2020) I was a part of a small software group with a team of 45 members. I was responsible for programming the robot, actuating its actuators on time, making the robot perform the tasks as per the theme given by the host country, and completing the task within the time frame of 3 mins. Here we implemented and also made use of smart closed-loop motor drivers.
• For the second year, we mentored a team of 45 members by providing them with training in basic robotics (electronics and mechanical). This year we implemented PID control on IMU and wheel odometry for robot motion and used Free-RTOS to control various actuators of the robot. Programmed a Swerve Drive robot and designed an in-house VNH-5019-based motor driver. Also, statistical learning was implemented in arrow throwing as per the theme.
• Over the time period of 3 years in this club we developed and designed two semi-autonomous and manual robots as per the theme given by the host country.
• Secured 3rd rank out of 32 domestic teams and was awarded the best design award by Toyota Motors (2020).
• Acquired 10th rank at the National Level ABU-Robocon 2022.
• Secured 2nd rank out of 28 teams in domestic and 7th out of 21 teams in ABU Robocon 2021, China.

Structure From Motion

• Implemented factorization method for SFM based on Cholesky decomposition and KLT-based feature tracker for given 100 images.

Image Mosaicing and Stereo Vision

• Devised an algorithm for detecting and aligning features in multiple images through Homography estimation and subsequent panorama generation. A depth estimation algorithm was also performed to determine the disparity between stereo images.

Navigation stack using IMU and GPS

• Formulated a ROS driver using IMU and GPS for autonomous navigation. Employed Allan variance analysis to assess IMU stability and integrated Complimentary Filter and Dead reckoning algorithms to accurately estimate the vehicle’s true position.

Image Filtering and Noise Estimation

• Developed a noise estimation algorithm on additive Gaussian noise images.

Motion detection using simple Image Filtering

• Created a motion detection algorithm leveraging temporal evolution and a derivative of Gaussian Filter over image sequences. Also developed a noise estimation algorithm with a 98% accuracy rate.

Autonomous Navigation and SLAM

• Developed an autonomous system for reconnaissance in simulated disaster zones using Cartographer-based SLAM and frontier exploration. Achieved accurate victim detection (13 out of 15) through extrinsic camera calibration.

Lagori Piling Robot

• Gameplay and Robots

Throwing arrow in the pot

• Gameplay and Robots

Robo Rugby

• Gameplay and Robots