• Share
  • Email
  • Embed
  • Like
  • Save
  • Private Content
1203
 

1203

on

  • 665 views

 

Statistics

Views

Total Views
665
Views on SlideShare
665
Embed Views
0

Actions

Likes
0
Downloads
1
Comments
0

0 Embeds 0

No embeds

Accessibility

Upload Details

Uploaded via as Microsoft PowerPoint

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    1203 1203 Presentation Transcript

    • Learning and Inferring Transportation Routines
    • Outline
      • Introduction
      • Hierarchical Activity Model
      • Inference, Learning, and Error Detection
        • Inference
        • Learning
        • Detection of User Errors
      • Experimental Results
    • Introduction
      • A hierarchical Markov model that can learn and infer a user’s daily movements through the community
      • The model allows :
        • Infer the locations of usual goals, such as home or workplace
        • Infer a user’s mode of transportation, such as foot, car, or bus, and predict when and where she will change modes
        • Predict her future movements, both in the short term and in terms of distant goals
        • Infer when a user has broken his ordinary routine in a way that may indicate that he has made an error, such as failing to get off his bus at his usual stop on the way home
        • Robustly track and predict locations even in the presence of total loss of GPS signals and other sources of noise
    • Hierarchical Activity Model
      • Locations and transportation modes
      • Trip segments
      • Goals
    •  
    • Inference
      • GPS-based tracking on street maps
      • Goal and trip segment estimation
      • The factorization (1) separates the state space of our estimation problem into its continuous and discrete parts
      • The left term on the right hand side of (1) , (2) follows by applying Bayes rule and the independences in our estimation problem
    •  
    • Learning
      • Finding goals
      • Finding mode transfer locations
      • Estimating transition matrices
    • Finding mode transfer locations
    • Detection of User Errors
      • The hierarchical tracker has “expensive” particles, each containing much state information, but requires few particles for accurate tracking
      • The flat, untrained tracker needs more particles to maintain tracking, but each particle is cheaper to compute
    • Experimental Results
      • Activity model learning
      • Detection of user errors
      • (a) Probability of the true goal (work place) during an episode from home to work, estimated using the flat and the hierarchical model
      • (b,c) Probability of an activity being normal or abnormal, estimated by the concurrent trackers
      • (b) Normal activity: driving from home to work
      • (c) Abnormal activity: missing to get off the bus at time t2
    •  
    •  
    •  
    • Thank you