Chaitanya Ganesh

• Worked on the ADAS unit to build the next generation CAN communication interface through a Raspberry pi. Also to develop a robust server framework to enable wireless communication to the car from remote client. • Deployed TinyYOLO, Squeezenet, MobileNet SSD models for on board object recognition on the ADAS node on NVIDIA Jetson Nano. • Evaluating safety of Deep Learning Algorithms for Autonomous Driving on NVIDIA TX2 in collaboration with RISE, Sweden.

Battery behavior data was collected from 32 Volvo V60 cars in a highly compressed format and analysed in order to create a battery aging model in real world scenarios. This was achieved through a data science pipeline which was responsible for decompressing, cleaning, extracting and selecting significant features. The processed data was used to perform predictive survival analysis. This was achieved through the statistical methods and through memory-based recurrent neural networks.

Worked on high accuracy robust indoor localization system for a mobile robot using only visual input. Used a LEGO EV3 brick as the control module and a Raspberry Pi model 3B+ as a compute module. It was equipped with ROS Kinetic, OpenCV3 for image processing and scikit-learn for machine learning. A web server was setup on the robot to stream the processed frames to a remote desktop client.

• Worked on collective Environment Sensing project in collaboration with Singapore-MIT Alliance for Research and Technology, Singapore and Massachusetts Institute of Technology • Implemented Hadoop and Distributed Tensorflow on a raspberry pi cluster as a fault tolerant message passing interface for distributed computing • Developed an algorithm using python for distributed scientific calculations on large clusters and scaled up the operation to test a cluster of 8 to run complex fluid dynamics library

Machine Learning Artificial Intelligence Data Science Machine Vision