Object classification, detection and tracking for autonomous vehicle navigation in Indian road scenarios

Worked on algorithms such as YOLO, R-CNN to detect, classify and track objects like car, pedestrian etc. Estimated the velocity of vehicles present in a video. Worked on KITTI LiDAR to get performance analysis of algorithms such as ‘‘Gaussian Mixture Models“,‘‘Fast Segmentation”. Created a dataset suitable for Indian roads which can be used for future research towards the development of autonomous vehicles in India.

A Blind Assistive Device

Designed a hat which can detect obstacles and warns according to the distance from it. An improved version of this can be used by blind people replacing traditional stick thus improving reliability.

Number plate detection from traffic surveillance footage

As part of winter internship, worked under Dr.Deepak Mishra at Indian Institute Of Space Science Technology(IIST), Trivandrum worked on traffic video footage to detect, recognise and match number plates of vehicles.

Research Assistant, Electrical Line Inspection Robot, Ph.D(Mech)

Worked as a research assistant for a Mechanical PhD project and as a 3-member team, developed a robot (which has 14 servo motors and 10 DC motors) which can traverse through electrical transmission lines and cross junctions, poles.

Online Shopping skillset for Alexa - NITCkathon.ai (24hr event)

As part of NITC hackathon conducted by JMR InfoTech, worked as a 3-member team and developed a skillset for Amazon Alexa to understand different skills related to online shopping.

Smart Energy Meter - MITS Hackathon, Kochi (24hr event)

Worked as a 4-member team and developed a module which can be installed to existing energy meters(instead of replacing the entire energy meter) thus providing details about electricity consumption, thefts etc to a remote website through Wi-Fi module.

Cloud Based Attendance using Biometric Scanner

As a 3-member team, we installed a biometric scanner in our lab (Robotics Interest Group - RIG) replacing the register for entry/exit timings. Data is sent to local website so that it is easy to know whether a person is present or not. The relevant statistics are displayed in the local website