Research Interests

Engineering design thinking

PD_compositeImage

I am interested in engineering design thinking - specifically how engineers think divergently and generate new ideas. My research deals with identifying strategies, techniques and processes that expert designers employ when they solve an engineering design problem. As an outcome of these investigations, I have designed a design pedagogy - Flare-Fork collaborative strategy, for a collaborating team of novice designers.
We have developed a technology enhanced learning environment - Prolific Designer, based on Flare-Fork collaborative strategy. The intervention provides strategies, and scaffolds to a collaborating team of novice designers as it takes them through productive design steps.
At the end of solving the engineering design problem using Prolific Designer tool, novice designers would have extensively explored the problem and solution space of the design problem and identified multiple, diverse conceptual designs.

Prolific Designer Demo Video Related Publication

Learning 3D Geometry

Geometry-via-Gestures

How do children learn 3 dimensional geometry? Do they just memorize the properties of 3D objects? Do they build on their understanding of 2D geometrical objects? These were some of the questions that we explored in the project Geometry via gestures.
As an outcome of our investigations, we developed a gesture based learning environment to learn about the properties such as volume and surface area of 3D geometrical objects.

GvG Demo Video Related Publication-1 Related Publication-2

Basic electronics lab - with a twist

electronic-lab

Laboratory courses are expensive to conduct - what with innumerable precision instruments, components for students to use, and more demanding hands on guidance to give for each student. However, do the laboratory courses in their current format, meet the course expectations? How do we make laboratory courses more productive, engaging and at the same time meet the prescribed laboratory goals such as developing experimental problem solving skills?
We investigated the efficacy of a laboratory course with an innovative pedagogy geared to address some important laboratory goals. These included design of experiments and analysis of experimental designs to solve electronic design problems and expert like usage of sophisticated instruments.

Related Publication


Past Projects

Hardware Implementation of Extended Kalman Filter

Place Electrical Engineering, IIT Bombay
Designation M.Tech - Research Assistant
Duration 2014 - 2015
Software/Platform Matlab, VHDL
Project Report
Description

Extended Kalman filter is one of the standard non-linear estimation algorithm used in realtime target tracking. The objective of the project was to design and realize a scalable extended Kalman filter (EKF) for non-linear tracking applicA performance comparison of the VHDL implementation with that of Matlab implementation indicated a slightly faster response of the hardware implementation.

Design and Development of a chat BOT for e-learning

Place Excelsoft Technologies Pvt Ltd., Mysore
Designation Research Associate
Duration 2008 - 2009
Software/Platform Visual Basic
Description

Design and development of a BOT with the ability to interact with the user in an intelligent manner involved use of Natural language processing, Machine learning and intelligent interpretation of data. Spin-offs include integration of BOT module into a Semantic Learning environment and a 3D Virtual world school environment. Automatic BOT knowledge base enhancement from multiple sources such as ontology (OWL and RDF based), and e-learning data (SCORM, QTI), was also accomplished.

Design and Development of Serial to Ethernet converter

Place Crystaline Infotek Pvt Ltd. Pune
Designation Software Engineer
Duration September 2006 to February 2007
Software/Platform
  • Firmware development: Keil compiler, CodeVisionAVR compiler
  • Hardware development: Orcad Capture for schematic entry, Orcad PSpice for circuit simulation, View Plot and GC Preview for gerber verification
  • Testing: Ethereal for packet transfer verification
Description

This project dealt with the design and development of serial to Ethernet converter. The project revolved around efficient transfer of data from serial side to Ethernet side and vice versa at user configurable data rates and parameters. The configurable parameters included baud rate, inter-character delay, number of data bits, stop bits and parity on serial side and source and destination IP, port number, MAC address setting, gateway setting, timeout setting etc on the network side.
The serial side includes RS232 and RS485 (2 wire and 4 wire) protocol support with baud rates from 2400 baud to 115.2 Kbaud. The Ethernet side comprises of a hardware stack and Phy and supports 10 / 100 Mbps data rate options. The controller takes care of providing a variety of configuration options namely configuration through serial, telnet, UDP utility and also HTTP.
Firmware for configuration of and data handling by the serial to Ethernet converter card was developed. In-System and In-Application programming of the device was also implemented thereby enabling the use of the flash for storing configuration data, and eliminating the need to use external EPROM to store the configuration data.

Application of Artificial Neural Networks in detection of defects from radiographic images.

Place Department of Electronic Science, University of Pune
Designation Junior Research Fellow
Duration and Funding April 2005 to May 2006; ISRO sponsored project
Software / Platform Matlab 6
Related Publication
Description

This project involved the automatic detection, identification and classification of defects as seen in an industrial radiographic image. For this, neural networks were employed as pattern recognition system. Image processing algorithms were used to enhance the image and extract features for better defect detection. Features extracted included co-occurrence matrix features and wavelet features. Image enhancement using neural network paradigms such as PCNN was also implemented.

Application of Artificial Neural Networks to mesoscale meteorological studies using microwave data from satellites like NOAA.

Place Department of Electronic Science, University of Pune
Designation Junior Research Fellow
Duration and Funding December 2003 to December 2004; ISRO sponsored project
Software / Platform Matlab 6, C
Related Publication
Description

Different neural network paradigms such as self-organizing map and Back-Propagation was used for detection of temperature inversion layer and classification of clouds using microwave data (AMSU-B) obtained from NOAA satellite. Radio Sonde data was used for ground truth validation. In addition to Artificial Neural Networks, statistical clustering techniques were used in identification of temperature inversion layers. Successful clustering of cyclone and monsoon clouds was also achieved.

Social Network

Linkedin