Vinayak Ravi

Vinayak Ravi

Perth, Western Australia, Australia
925 followers 500+ connections

Join to follow

About

I’m Vinayak — a passionate, forward-thinking engineer who sees every challenge as an…

Articles by Vinayak

See all articles

Activity

Join now to see all activity

Experience

  • CapabilityX Graphic

    CapabilityX

    Perth, Western Australia, Australia

  • -

    Perth, Western Australia, Australia

  • -

    Perth, Western Australia, Australia

  • -

    Chennai, Tamil Nadu, India

  • -

    Melbourne, Victoria, Australia

  • -

    Canberra, Australian Capital Territory, Australia

  • -

    Chennai, Tamil Nadu, India

  • -

    Chennai, Tamil Nadu, India

  • -

    Abu Dhabi

Education

  • The Australian National University Graphic

    The Australian National University

    -

  • -

    Activities and Societies: Quality management team: Conducted tutorial facilations and was part of green belt of quality programme, Symposium - Ignition (2017) - organised robot cars event, UAV tutoring. Robotics Group: part of robotics group and completed a project titled "scavenger robot", with fully fabricated robot chassis and mountable robotic actuator arm.

Publications

Projects

  • Embedded Motion Tracking system:

    Programmed FPGA and used UART serial communica-
    tion module (PMOD-NAV), to transfer real-time orien-
    tation of the sensor’s data on the terminal Which was
    visually verified using the Python GUI program.

  • Improved Contingency modelling and performance recognition of football players using Markov Chain

    Have utilised Markov chain & Bayesian network analysis to predict and determine a player’s impact to a team (Chelsea Football Club), by analysing his performance over a span of years and representing the parameters in terms of Discrete Event Systems (DES). Linear discriminant analysis (LDA) was also used to relate player’s attributes with the players.

  • Modelling of Photo voltaic (PV) module in solar panels:

    Have modelled photo-voltaic module by considering 4 parameters (open-circuit current, Irradiance, Temperature and output voltage) in Simulink

  • Preventing rolling action of HMVs, Cars by controlling longitudinal, lateral velocities and yaw rate of the system

    HMVs
    lateral, longitudinal velocities and yaw rate were controlled to regulated value (matched to reference), by inputting all wheel’s slip
    ratios and steering angle. Further, considering various frictional torques and inertias involved, for various radius of curvatures.

  • Proposal of adaptive solution for planning study programs for undergrads/ postgrads in College of Arts & Science (CASS), Australian National University (ANU):

    Proposed an Artificial Intelligence (AI) planner for the students to choose their majors/minors in their corresponding
    programs, abiding their course requirements and matching their credit points. Furthermore, facilitating integration with the
    local ANU server.

  • Proposal of PID controller for cruise control in Mercedes Benz CLA 180 Blue efficiency 2013

    Used MATLAB & Simulink
    to model a PID controller for the mentioned car and verified using Bode and Nichols plot. Furthermore, proposing Genetic
    Algorithm (GA) to augment the controller design, by choosing effective gain values.

  • Real Time Appearance based (R-TAB) Mapping:

    Worked on turtle-bot 2 to simultaneously localize and map (SLAM) an enclosed area and facilitated navigation autonomously. Real time generation of the map, from which loop-closure detection was enabled using memory management techniques.

    Built and programmed using ROS and python, a bag-
    of-words based loop closure detector, devising graph-
    based incremental update of map, helping in localization
    and path-planning for the robot with visual feedback to
    the operator.

  • Autonomous scene labelling of a car:

    Developed Convolutional Neural Networks (CNNs) to
    semantically segment 12 classes for a car to recognize
    different labels autonomously while driving, using Py-
    Torch/Keras. The validation and testing accuracy was
    found to be 87 % and 86% respectively.

  • Intelligent Street Lighting System:

    Compiled a thesis report, proposing augmented street lighting system for Canberra’s layout
    using adaptive LED street lights with power back up using renewable energy. The report contained detailed stakeholder,
    functional, performance and technical analysis.

  • Design of seating posture for passengers travelling in HMV

    -

    Analysing various subject’s seating posture  risk, while travelling in College bus by using RULA and REBA and quantifying it using FSR’s (Force resistive sensors). Furthermore, devising an optimal seating posture, which deemed to reduce the amount of pain indu

  • Building CNN to learn Fashion MNIST

    -

    Built Convolutional Neural Network using PyTorch/Keras with
    TensorFlow backend, altering various hyperparameters (batch
    sizes, optimizers, learning rates) to import and classify Fashion
    MNIST to their corresponding labels and achieved a validation/
    test accuracy of 91 % and 90.8% respectively.

  • Digitization of Tamil Text

    -

    Built a classifier to recognize Tamil text (MNIST) using Convolu-
    tion Neural Network (CNN) and other text recognition methods
    Support Vector Machines (SVD),composing various data augmen-tations. Validation accuracy was found to be 67% due to the lack of computational power and complex writing structure of the script.

  • Game Anaytics on Narrative-based videogames

    -

    Devised a comprehensive study on minimising ’Ludo-narrative
    dissonance’, thereby understanding the balance between narrative and gameplay in story-oriented games.

  • Gas leakage detection robot

    -

    Modelled the design of the robot (wall-attaching
    wheeled) using Solidworks and analysed in-flow
    stress/strain cosntraints of the robot in Ansys. De-
    bugged code in Arduino IDE. aiding in detecting hole
    inside the pipe by checking pressure/temperature gra-
    dients (Pressure sensor, PMOD NAV, Infra-red thermal
    camera) and localise the hole causing the leakage (LI-
    DAR).

  • Semantic Scene Classification and labelling

    -

    Trained and architected a Convolutional Neural Network using
    Keras to help it in recognising/classifying various scenes from
    Places365 dataset. Formulated three different scenarios (Trans-
    fer learning, altering loss functions/optimisers and data pruning) to study on optimal convergence on the learning process.

Languages

  • English

    Elementary proficiency

  • Hindi

    Elementary proficiency

  • Tamil

    Native or bilingual proficiency

  • Japanese

    Elementary proficiency

More activity by Vinayak

View Vinayak’s full profile

  • See who you know in common
  • Get introduced
  • Contact Vinayak directly
Join to view full profile

Other similar profiles

Explore more posts

Explore collaborative articles

We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.

Explore More

Add new skills with these courses

See all courses