Information about Learning And Inferring Transportation Routines

Published on July 28, 2008

Author: mstjsw

Source: slideshare.net

Outline 1. Introduction 2. Hierarchical activity model 3. Inference 4. Learning 5. Experimental results 6. Application: Opportunity Knocks 7. Conclusions

1. Introduction

2. Hierarchical activity model

3. Inference

4. Learning

5. Experimental results

6. Application: Opportunity Knocks

7. Conclusions

1. Introduction Track and predict a user’s location Infer a user’s mode and predict when and where Infer the locations of transportation destinations. Indicate he or she has made an error.

Track and predict a user’s location

Infer a user’s mode and predict when and where

Infer the locations of transportation destinations.

Indicate he or she has made an error.

2. Hierarchical activity model Locations and transportation modes Trip segments Goals Novelty

Locations and transportation modes

Trip segments

Goals

Novelty

two time slices k − 1 and k novelty mode next goal and current trip segment GPS sensor measurement

Locations and transportation modes x k = 〈 l k , v k , c k 〉 z k ： GPS sensor measurements generated by the person carrying a GPS sensor m k ： transportation mode BUS , FOOT , CAR , and BUILDING . The domain of τ k is the set of outgoing neighbors of the current edge; θ k is picked from the edges close to z k

x k = 〈 l k , v k , c k 〉

z k ： GPS sensor measurements

generated by the person carrying a GPS sensor

m k ： transportation mode

BUS , FOOT , CAR , and BUILDING .

The domain of τ k is the set of outgoing neighbors of the current edge;

θ k is picked from the edges close to z k

Trip segments A trip segment is defined by its start location, t k .start , end location, t k .end , and the mode of transportation, t k .mode . ： trip switching ： a counter that measures the time steps until the next transportation mode is entered.

A trip segment is defined by its start location, t k .start , end location, t k .end , and the mode of transportation, t k .mode .

： trip switching

： a counter that measures the time steps until the next transportation mode is entered.

Goals & Novelty A goal represents the current target location of the person. g k ： goal ： goal switching node n k ： indicating whether a user’s behavior is consistent with historical patterns.

A goal represents the current target location of the person.

g k ： goal

： goal switching node

n k ： indicating whether a user’s behavior is consistent with historical patterns.

3. Inference Flat model Rao–Blackwellized particle filter for estimation in the flat model RBPF algorithm for the flat model Hierarchical model RBPF algorithm for the hierarchical model Detecting novel behavior

Flat model

Rao–Blackwellized particle filter for estimation in the flat model

RBPF algorithm for the flat model

Hierarchical model

RBPF algorithm for the hierarchical model

Detecting novel behavior

Rao–Blackwellized particle filter for estimation in the flat model Sampling step Kalman filter step Importance weights histories location of the person car location

Sampling step

Kalman filter step

Importance weights

RBPF algorithm for the flat model Sampling step Kalman filter step Importance weights

RBPF algorithm for the hierarchical model

Detecting novel behavior

4. Learning Finding mode transfer locations estimate the mode transition probabilities using the expectation maximization (EM) algorithm Finding goals a person typically spends extended periods of time whether indoors or outdoors. Estimating transition matrices use EM to estimate the transition matrices

Finding mode transfer locations

estimate the mode transition probabilities using the expectation maximization (EM) algorithm

Finding goals

a person typically spends extended periods of time whether indoors or outdoors.

Estimating transition matrices

use EM to estimate the transition matrices

5. Experimental results Activity model learning Empirical comparison to other models Error detection

Activity model learning

Empirical comparison to other models

Error detection

Activity model learning six most common transportation goals frequently used bus stops and parking lots, the common routes using different modes of transportation

six most common transportation goals

frequently used bus stops and parking lots,

the common routes using different modes of transportation

Empirical comparison to other models

Error detection

6. Application: Opportunity Knocks The name of our system is derived from the desire to provide our users with a source of computer generated opportunities it plays a sound like a door knocking to get the user’s attention.

The name of our system is derived from the desire to provide our users with a source of computer generated opportunities

it plays a sound like a door knocking to get the user’s attention.

Opportunity Knocks(1)

Opportunity Knocks(2)

Opportunity Knocks(3)

Opportunity Knocks(4)

Opportunity Knocks(5)

Opportunity Knocks(6)

7. Conclusions provide predictions of movements to distant goals, and support a simple and effective strategy for detecting novel events that may indicate user errors system limitation ： it uses fixed thresholds to extract goals and mode transfer locations.

provide predictions of movements to distant goals, and support a simple and effective strategy for detecting novel events that may indicate user errors

system limitation ： it uses fixed thresholds to extract goals and mode transfer locations.

Learning and Inferring Transportation Routines Lin Liao a, Donald J. Pattersonb, Dieter Fox , Henry Kautza, aDepartment of Computer Science and Engineering

Read more

Learning and inferring transportation routines ... daily movement patterns using unsupervised learning ... mode of transportation (i.e ...

Read more

Learning and Inferring Transportation Routines Lin Liao Dieter Fox Henry Kautz Department of Computer Science & Engineering University of Washington

Read more

Learning and inferring transportation ... 311â€“331 www.elsevier.com/locate/artint Learning and inferring transportation routines ... Learning ...

Read more

I will give the presentation I gave at the AAAI conference this year. We introduce a hierarchical Markov model that can learn and infer a user’s daily ...

Read more

Title: Learning and Inferring Transportation Routines: Publication Type: Journal Article: Year of Publication: 2007: Authors: Liao L, Patterson D J, Fox D ...

Read more

Learning and inferring transportation routines. Lin Liao, Donald J. Patterson, Dieter Fox, Henry A. Kautz; AI; 2004; View PDF; Cite; Save ...

Read more

This paper introduces a hierarchical Markov model that can learn and infer a user's daily movements through an urban community. The model uses multiple ...

Read more

@MISC{A_learningand, author = {Lin Liao A and Donald J. Patterson B and Dieter Fox A and Henry Kautz A}, title = {Learning and Inferring ...

Read more

BibTeX @MISC{Liao04learningand, author = {Lin Liao and Dieter Fox and Henry Kautz}, title = {Learning and inferring transportation routines}, year ...

Read more

## Add a comment