+ All Categories
Home > Documents > Hoelzl Gerold. Overview Motivation System design Summary Future work 17.09.2015 2 Hoelzl Gerold.

Hoelzl Gerold. Overview Motivation System design Summary Future work 17.09.2015 2 Hoelzl Gerold.

Date post: 28-Dec-2015
Category:
Upload: nathan-mcdowell
View: 227 times
Download: 4 times
Share this document with a friend
Popular Tags:
26
Hoelzl Gerold
Transcript

Hoelzl Gerold

Overview

Motivation System design Summary Future work

19.04.23 2Hoelzl Gerold

Motivation

19.04.23 3Hoelzl Gerold

Motivation

During training (jogging) it is fun to listening to music problem:

rhythm of the music ≠ rhythm of the runner

19.04.23 4

music rhythm (bpm) running rhythm

Hoelzl Gerold

Motivation

Idea adaptive selection of music according to the

running respectively the walking pace

additionally playing auditive status information like

heart rate speed, distance warnings if predefined limits are exceeded …

19.04.23 5Hoelzl Gerold

System design

19.04.23 6Hoelzl Gerold

System design

3 tasks movement analysis audio analysis mapping: movement<-> music

19.04.23 7Hoelzl Gerold

System design Movement analysis

19.04.23 8

[Scholich M., 1978, S.93]

acceleration maximum during heel- and toe ground contact g > 1,75 m/s²

recognition of a footstep: phase 7: „front support phase“phase 8,9: „rear support phase“

phase 7, 8 ,9 : „support phase“

[Larch R., 2001, S.3]

Hoelzl Gerold

System design Movement

acceleration values

extracted footsteps

distance calculation ->d(i)[ms]

19.04.23 9

d(0)d(1)

d(2)d(3) d(4)

1

0

])[(1 n

i

msidn

mftFeature for classifying running rhythm defined as mean footsteptime (mft)

Hoelzl Gerold

System design

Audio analysis audio feature

beats per minute (bpm) (beat-tracking) recognized as a binary regular pulse

underlaying the music

separate calculation too computational intensive on mobile

devices stored as metadata

19.04.23 10Hoelzl Gerold

System design

Mapping

O[bpm] = F(M,f[mft]) mapping function

M = {m1,…,mn} set of music piecesf[mft] feature

O[bpm] M set of fitting music pieces

19.04.23 11

O = F(M,f)mft [ms]

feature f output O

music pieces [bpm]

Hoelzl Gerold

System design - Hardware

19.04.23 12Hoelzl Gerold

System design

19.04.23 13Hoelzl Gerold

System design

Hardware sensors

acceleration-sensor (movement analysis) ECG-sensor (biofeedback) GPS-sensor (speed, route, tracking)

communication technology Bluetooth

controller mobile phone (JavaME)

19.04.23 14Hoelzl Gerold

System design Hardware (prototype)

acceleration-sensorSparkfun WiTilt v33-axis acceleration sensor with integrated Bluetooth capabilities

ECG-sensorPolar Heart Rate Monitor Interface +AMB2300-EV-Bluetooth-Board

GPS-sensor (integrated in mobile phone)

19.04.23 15Hoelzl Gerold

System design - Software

19.04.23 16Hoelzl Gerold

System design Software design

19.04.23 17

Sensor-Controller

ECGSensor

Acceleration

Sensor

GPSSensor

Controller-

View

Audioplayer

Feature-calculation

Mappingmft<->bpm

Actors(pulse,

speed,..)

Hoelzl Gerold

System design

Software JavaME

API´s JSR 82: Bluetooth JSR 179: Location (GPS) JSR 135: Media (Audio, Video)

SonyEricsson JP8 current JavaME platform implementation

19.04.23 18Hoelzl Gerold

Summary & Future work

19.04.23 19Hoelzl Gerold

Summary

adaptive selection of music according to the running- respectively the walking pace

additionally playing auditive status information like

heart rate speed, distance warning if predefined limits are exceeded …

19.04.23 20Hoelzl Gerold

Summary

Experimental results prototype showed high accuracy and

stability in both laboratory and naturalistic environments

short response time giving realtime “feeling”(~50ms)

Unexpected: crash of Bluetooth connections when built

in GPS sensor was used

19.04.23 21Hoelzl Gerold

Summary

mapping function difficult to define one function for many

users music rhythm / speed is perceived very

subjective learning „on the fly“

19.04.23 22Hoelzl Gerold

Summary

integrated sensors usage possible (less components) but restriction of device usage likely

(e.g. fixed position of acceleration sensor for movement analysis)

19.04.23 23Hoelzl Gerold

Future work

support training programs in adaptively selecting forcing / retaining music pushing music if beneath a defined limit assuasive music if above a defined limit

19.04.23 24Hoelzl Gerold

Future work

tracking with status information (using remote computer)

„control sportsman“ competitions (without borders)

19.04.23 25Hoelzl Gerold

Questions?

19.04.23 26Hoelzl Gerold


Recommended