Dakshina Fernando

Research Expertise

Thrust Vectoring

Hyper-sonic Air Travel

Intelligent Control

Process Optimization

Ongoing Research

Thrust Vectoring

This research project delves into the application of Thrust Vector Control (TVC) as an innovative attitude control mechanism in conventional aircraft design. Historically recognized for its effectiveness in the aerospace industry, TVC has predominantly been employed in vertical take-off and landing (VTOL) aircraft, rockets, and interceptors. However, its potential impact on lateral dynamics in conventional aircraft remains largely unexplored. The study focuses on integrating TVC into conventional aircraft for attitude control and proposes a robust Fault Tolerant Control (FTC) system for reconfiguring faults. The investigation explores two configurations - Multi-engine Single Thrust Vectored System (MESTV) and Multi-engine Dual Thrust Vector System (MEDTV). The initial phase concentrates on a tail-end single TVC engine with dual-engine forward thrusters. Results from the mathematical model indicate promising maneuvering capabilities in both yaw and pitch control deflections when compared to traditional static control systems, showcasing the potential advantages of this novel TVC system.

P.C. https://www.quora.com/Why-do-rockets-use-nozzles-to-control-jet-thrust-instead-of-vanes

Micro Gravity

Embarking on a unique exploration, this research project delves into the intersection of Pharma and Aerospace Engineering, seeking synergies that reside in these two diverse fields. This study aims to uncover the potential collaborations and innovations that can arise from the confluence of pharmaceutical sciences and aerospace engineering. By bridging these disciplines, we anticipate discovering novel approaches and applications, envisioning advancements that could revolutionize both sectors. Through this investigation, we aspire to shed light on the unexplored possibilities at the intersection of Pharma and Aerospace Engineering, fostering a deeper understanding of how their combination might yield groundbreaking solutions and contribute to advancements in both realms.

P.C. https://www.tohoku.ac.jp/en/press/microgravity_worms_help_solve_astronauts_muscle_troubles.html

Putting Wings on AI - What It Takes to Feel Safe

Coming soon!

P.C. https://readwrite.com/importance-of-human-interaction-as-ai-and-ml-rises/

Poster Presentations

Projects

Image Classifier

The objective of this study was to develop and design a CNN architecture using flatten and dense layers supported by batch normalization and dropout as appropriate for an algorithm with multiple convolutional layers.

Sentiment Analysis

The objective of this project was to preprocess the Sentiment Polarity Dataset Version 2.0 using self-created NLP functions and find the best Word embedding method with maximum classification accuracy. In this project TF-IDF embedding was identified as the most accurate classification with a Linear support vector machine and 10 - fold cross validation.

Deep Learning

The objective of this project was to evaluate the time series forecasting performance of ANN, RNN, and LSTM deep learning models using a fixed number of layers and equivalent hyperparameters. In this specific case it was recommended that a combination of ANN and RNN models to be considered in parallel for best fit and an LSTM model for minimum loss.

Automated Guidance Vehicle

The objective of this project was to develop an automated following load carrier capable of half a ton payload. The system was enhanced to navigate extreme terrain with gravel and complex ground topologies.

Invited Speeches