Presentation 2008-07-18
Pedestrian Context Inference Using a 3-Axis Accelerometer
Naoki IKETANI, Masaaki KIKUCHI, Kenta CHO, Masanori HATTORI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We introduce a new method to recognize pedestrian contexts using a handheld mobile computer and a three-axis accelerometer. Without being affected by the position and direction of the accelerometer, we can recognize four kinds of pedestrian contexts: "still (not moving)", "walking", "running", and "on a vehicle (on a bus or train)." Our context recognition method consists of gravity estimation, feature value extraction from sensor information, context classification through a three-layer neural network, and the post-process context correction based on our original heuristic state-transition model. We evaluated the accuracy of the context recognition through cross validation. The accuracy reached 81% before the post-process correction and 84-87% after the post-process correction.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Context-aware / accelerometer
Paper # USN2008-24
Date of Issue

Conference Information
Committee USN
Conference Date 2008/7/10(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Ubiquitous and Sensor Networks(USN)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Pedestrian Context Inference Using a 3-Axis Accelerometer
Sub Title (in English)
Keyword(1) Context-aware
Keyword(2) accelerometer
1st Author's Name Naoki IKETANI
1st Author's Affiliation Research and Development Center, Toshiba Corp.()
2nd Author's Name Masaaki KIKUCHI
2nd Author's Affiliation Research and Development Center, Toshiba Corp.
3rd Author's Name Kenta CHO
3rd Author's Affiliation Research and Development Center, Toshiba Corp.
4th Author's Name Masanori HATTORI
4th Author's Affiliation Research and Development Center, Toshiba Corp.
Date 2008-07-18
Paper # USN2008-24
Volume (vol) vol.108
Number (no) 138
Page pp.pp.-
#Pages 6
Date of Issue