Gait analysis with using accelerometer sensor of smart phones

The ability to measure human movements in sports fields is among the important issues for performance measurement and development. This situation is also an important part of clinical evaluations. Electromagnetic systems are among the most widely used methods to evaluate human movements. In this study, walking data of 50 different people were used in a 100-meter-long corridor. The walking data was obtained from the accelerometer sensor with a software developed for the smartphone. Local Binary Pattern (LBP) method was applied to the data and 256 features were removed. Tests were carried out with different classification algorithms and 90% successful classification was achieved with Cubic SVM. With this method, it is thought that more economical methods will be developed instead of high-cost devices in detecting gait disorders.