Prajwal's website

Optimal Design of eVTOLs for Urban Mobility using Analytical Target Cascading (ATC)

Tools: MATLAB | Python

Implemented a Multidisciplinary Optimization technique called Analytical target cascading on design of electric vertical take-off and landing aircrafts (eVTOLs). The project mainly involved modelling different subsystems such as motors, wings, propellers and batteries and then optimizing all the subsystems simultaneously while using individual optimization solvers.

Outcomes

Demonstrated convergence to optimum for a set of different ranges and payloads
Presented this work at Applied Energy Syposium which was organized by MIT and Harvard
Developed a generic ATC MATLAB package

Autonomous landing in the presence of heavy cross winds using nested MPC

Tools: MATLAB | Simulink | Flight gear

Implemented a nested MPC architecture in MATLAB Simulink to control and execute a cross wind landing of aircraft with variable wind gusts. This work was primarily done to test the effectiveness of nested MPC to respond to sudden disturbances. The results show that nested MPC performs well with satisfactory performance metrics for 200 iterations. Ofcourse, a realistic control algorithm should need more than a million iterations to comply with FAA standards but this work demonstrates a potential solution to an open problem till date.

Outcomes

Aircraft lands perfectly with even wind gusts upto 30knots
Recieved an A+ in this course
Validating using experimental data of motor performance

Design of Hybrid HydroKinetic Turbine

Tools : ANSYS Fluent | AutoCAD

This was my submission for Bachelor's project at IIT Madras. This project focused on designing and testing a hybrid hydrokinetic turbine. This design innovatively combines the best of Savonius and Darreius turbines to get the best operating efficiency over a long range of RPMs. This design selectively operates Savonius turbine at low RPMs and Darrieus turbine at high RPMs using a spring-inertia based mechanism.

Outcomes

Developed a physical prototype to demonstrate the concept of hybrid turbine switching at a desired RPM
Achieved a higher efficiency by selectively enabling only the more efficient turbine at different operating RPM
Published a thesis and presented this work in BTP open house

Machine Learning in Finance (multiple projects)

Tools | Sci-kit Learn | Keras

Completed multiple projects relating to prediction of financial instrument prices, economic indicies etc. These are set of projects that I was involved with in my ML for finance and High Frequency Trading courses. Prediction of finacial datasets is quite tricky in the sense that the data is usaually hight random with factors affecting these datasets not included in the features of the data. This makes these datasets very challenging to predict.

Roles and Responsibilities

Economic index percentage change prediction: Implemented Random Forest regression on dataset representing spread of treasury bond yield and commercial paper to predict economic activity change in next 3-, 6- and 9-months’ timeframe
Predicted the Moody’s rating of different corporate accounts based on total sales, EBITDA, cash flow etc. using ADABOOST on Decision trees and KNN
Implemented Kalman filter for stock price level movement using a constant velocity model and compared with performance of industry standard techniques like LSTM and Random Forest

Projects and Course Work

Optimal Design of eVTOLs for Urban Mobility using Analytical Target Cascading (ATC)

Outcomes

Autonomous landing in the presence of heavy cross winds using nested MPC

Outcomes

Design of Hybrid HydroKinetic Turbine

Outcomes

Machine Learning in Finance (multiple projects)

Roles and Responsibilities

Menu

Work experience

Projects & Course Work

Education & Training

Skills & Hobbies

Get in touch