A video demo and explanation of the EOG-based speller
Electrodes to track horizontal component of eye movements
Electrodes to track vertical component of eye movements
English characters sorted for Huffman coding
Advisors: Prof. A. P. Vinod & Dr. L. N. Theagarajan
Location: IIT Palakkad, India
Spring 2019
Devised an efficient electrooculography (EOG) based speller to assist paralysed and speech-impaired individuals in typing on a virtual keyboard using eye movements.
Tracked eye movements by acquiring signals from muscles surrounding the eye, conditioned them using analog circuits and classified them on MATLAB.
Novelty lies in the implementation of Huffman coding to increase the average typing speed by nearly 30% as compared to the benchmark.
Monitoring the vertical and horizontal components of the acquired electrooculogram
A screenshot of the virtual keyboard GUI developed on MATLAB
Advisor: Dr. Arvind Ajoy
Location: Smart India Hackathon, Jaipur
Devised a radio repeater system using a Raspberry Pi and an SDR.
The SDR demodulates a live (on the air) analog radio broadcast.
The demodulated stream is digitised and live-streamed to devices on a network over Wi-Fi.
Built a web-based GUI using HTML, CSS, Bootstrap, PHP and Python to allow change in demodulation mode and remote tuning of the SDR’s local oscillator frequency.
Led a team of five to the Grand Finale of 2018’s Smart India Hackathon. Received appreciation and travel awards from the Ministry of Information and Broadcasting, Govt. of India.
Advisor: Prof. Hossein Hashemi
Location: University of Southern California, USA
Summer 2019
Developed a phase retrieval based algorithm to calibrate large-scale phased arrays efficiently.
Algorithm estimates the phases of complex feed coefficients of a phased array from its measurable parameters: far-field radiation pattern amplitudes and near-field intensities.
Potential applications include phased array antenna synthesis (estimation of phased array’s feed coefficients, given the desired radiation pattern), characterisation of non-linear phase shifters on optical phased array (OPA) chips, and beam-forming in phased arrays and lidars.
Presented the results of this work in a research event at the Viterbi School of Engineering, University of Southern California, Los Angeles, USA.
Advisor: Prof. A. P. Vinod
Location: IIT Palakkad, India
Spring 2019
Built a non-invasive system with a GUI to continuously monitor blood pulse wave.
Designed signal conditioning circuitry to interface an infrared plethysmograph with an ADC.
An adaptive calibration-free algorithm was developed and used to extract heart rate from the pulse waveform generated by the plethysmograph when worn by a subject.
System detecting an abnormally high heart rate and alerting
System monitoring a normal heart rate
Location: IIT Palakkad, India & NTU, Singapore
Fall 2019 - Present
Building a robust ML-based deception detection system that can identify lies by analysing a person's physiological signals & behavioural features.
Extracting novel features from the time-series waveforms and performing statistical analysis to test various hypotheses.
Using machine learning models for predicting the guiltiness of a person from his/her physiological signals and behavioural features.