| Autor: |
Chakraborty, Goutam, Dendou, Tetsuhiro, Kikuchi, Daigo, Chiba, Kyohei |
| Zdroj: |
2012 IEEE International Instrumentation & Measurement Technology Conference Proceedings; 1/ 1/2012, p1963-1967, 5p |
| Abstrakt: |
Data collected from pressure sensors attached to shoe insole is a rich source of information. (1) We can detect faults in walking and balancing problems for old people. (2) The pressure sensor data can be used to design personalized foot orthoses. (3) We can calculate the calorie burnt, even when walking and jogging are mixed, and the road slope changes. (4) We can use the data to train sprinters or tennis players. (5) We can even use the data for person identification. In addition, it can be used for alarms for situations arising from mis-handling of machines, like accelerator pedal in a car. We attached very thin pressure sensors on top of a shoe insole and collected data. A few important and readily detectable features from the time series data collected by those sensors are extracted and used for person identification, or to classify whether a person is walking or jogging. Nearly 100% classification accuracy was achieved. Thus, the target to classify whether the person is walking or jogging or climbing up or down the stairs is possible. This success also encouraged us to investigate whether it is possible to find the body weight and the step-length from this data. Once that is possible, the system can accurately deliver the calorie burnt at the end of the day. We will further explore the possibility, using sophisticated feature extraction and classification techniques, to detect faults in walking and predict the probability of fall for elderly people, or people with problem in balancing due to various diseases or caused by accidents. [ABSTRACT FROM PUBLISHER] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
