Swapnil Das

Developing software for the Jacobi robotics platform to control and maneuver robot arm motion planning and execution, in order to handle scenarios like palletizing boxes in a factory environment.

Researched under Professor Raj Rajkumar in Autonomous Vehicles. Developed an audio sensing suite to classify and localize sounds like sirens and horns around a car for path planning using CNNs. Developed novel filtering estimation techniques.

Working on a system to detect emergency vehicles and flashing lights in a vehicle’s vicinity, using fused camera and LiDAR data. Implemented using a combination of novel CNN-based techniques and traditional computer vision methods.

Serving as a teaching assistant for graduate-level embedded real-time systems course, heading lab sections where students modify existing platform kernels to support tasksets with different priorities and scheduling, as well as demonstrate their power efficiency. Assisting students with implementational issues and theory-related questions on cyber-physical systems.

Led a project under Professor Raj Rajkumar and his class “Wireless Sensor Networks” to create a network of wirelessly connected sensors around a vehicle including microphones, driving wheel pressure sensors and cameras to assess a driver’s emotional state and quantify their annoyance using self-developed heuristics and psychoacoustic studies. Designed the sensors to communicate wirelessly to a central computer system and relay an ‘annoyance score’ that could then be used by other parts of the vehicle system to enhance driver experience.

Worked on 4 semester-long video games in different genres for PC and VR as game designer, writer, musician and programmer. Served GCS in an advisory capacity to improve member recruitment and retention, and conduct talks related to gaming. For more details, refer to https://www.gamecreation.org/members/SwapnilDas.

Programmed a Baxter robot with an inference-based and ML-based setup to interact with a human, tracking the motion of their hand and moving its robotic arm in response to predict the human’s future movement and meet in the middle.

Researching under Professor James Demmel and the Berkeley Benchmarking and Optimization Group (BeBOP) in Reproducible Basic Linear Algebra Subprograms (ReproBLAS) to ensure floating point associativity. Targeting existing BLAS performance standards, as well as the introduction of parallelism, increased functionality, and extension to other platforms.

Researching under Professor Erickson Rangel do Nascimento and Professor Ruzena Bajcsy to model human driver behavior given external video feeds, internal automobile data and pressure sensors, with the aim to determine driver annoyance in context of the environment.

Programmed a Baxter robot to successfully model the pendulum motion of a ball and grab it mid-flight. Implementational details and demonstrational videos are provided at https://cpkurotori.github.io/pendulum-bot-website/.

Currently working as a Tutor and Lab Assistant for EE16B - assisting students with circuit design, project debugging as well as assisting students mathematical concepts related to circuit analysis, transfer functions, filters, system controls, SVD and PCA.

Assisted in data gathering and analysis on 3D printer efficiency, energy and material consumption through extraction of printer power data from specialized sensor logs via webscraping in Python. Also aided in categorizing failed prints by their respective failure mode.

Among a staff of high-scoring students from CS 61A - in charge of instructing, and assisting students in lab sessions with general coding and project debugging. Also part of after-class office hours to help students with understanding of programming paradigms and the Python language.

Assisted Airtech Consultancy Services, an engineering consultancy business based in Kolkata, with brainstorming, identifying market challenges, developing market strategies as well as the development of a professional website for the business.

Researched mechanics of autonomous vehicles, including sensor development and motion. Delved into the commercial aspects of autonomous vehicles, visited relevant manufacturing businesses. Discussed factors to make them practical and user-friendly.

Investigated Surface Plasmon Resonance (S.P.R) and its role in the development of biological sensors, through simulations on FDTD Solutions software, followed by later experimentation with high-index prisms and monochromatic light set-up.