Probabilistic Models for Mobile Phone Trajectory Estimation
Speaker: Arvind Thiagarajan , CSAILContact:
Date: August 11 2011
Time: 10:00AM to 11:00AM
Location: 32-D463 (Star)
Host: Hari Balakrishnan, CSAIL
Sheila Marian, x3-1996, email@example.com
Abstract: This thesis is concerned with the problem of determining the trajectory of a mobile device --- a sequence of road segments on an outdoor map, or a sequence of rooms visited inside a building, in an energy-efficient and accurate manner. GPS, the dominant location technology on mobile phones today, has two major limitations: it is energy-intensive, making it impractical for continuous
monitoring, and it does not work indoors. This thesis explore two solutions: sub-sampling GPS to save energy, and using alternatives such as WiFi and cellular localization. In both cases, the challenge is to match a sequence of infrequent and/or inaccurate position samples to an accurate output trajectory. I will present and evaluate three related systems that use probabilistic
models to solve this problem.
VTrack uses Hidden Markov Models to match noisy or sparsely sampled geographic (lat,lon) coordinates to a sequence of road segments on a map.
CTrack improves on VTrack by using ``soft'' information in the form of raw WiFi or cellular signal strengths, rather than geographic coordinates, and by using movement and turn ``hints'' from the accelerometer and compass to improve accuracy.
iTrack uses a particle filter that combines data from the accelerometer and gyroscope with WiFi signals to track a mobile phone indoors accurate to less than a metre, with little manual training effort.
For more information, please contact: Sheila Marian, firstname.lastname@example.org
Thesis Supervisors: Prof. Hari Balakrishnan and Prof. Sam Madden
Thesis Committee: Prof. Hari Balakrishnan, Prof. Sam Madden, Prof. Seth Teller
See other events that are part of
See other events happening in August 2011