Abstract

This talk will cover some recent work on sensing and information processing with smartphones on the phone, and real in time. The first part of the talk will be on sensing the “context” of a mobile phone: whether the user is indoor or outdoor. Our method uses a semi-supervised framework and learns from environmental data in real time. It uses only low power sensors and is thus lightweight and suitable for use on a phone.

The second part will be on simplifying sensor data, with focus on locations and trajectories. GPS and other sensors can produce large volumes of data that can be costly to store, transmit and process. We show a method that uses topological persistence to filter redundant data points on-the-fly, and computes multi-resolution summaries.

Bio: Dr. Rik Sarkar is a Lecturer at University of Edinburgh. His main research interests are Sensor and mobile networks, Social networks, Computational geometry and Algorithms. He completed PhD from Stony Brook University, USA and M.Tech from IIT Bombay, India. http://homepages.inf.ed.ac.uk/rsarkar/