Human Activity Recognition: Accuracy across Common Locations for Wearable Sensors

摘要: In recent years much work has been done on human activity recognition using wearable sensors. As we begin to deploy hundreds and even thousands of sensors regular workers, hospital patients, army soldiers, the question shifts more toward a balance between what information can be gained their broad immediate user acceptance. this paper compare classification accuracy four different configurations accelerometer placement body hidden Markov models (HMMs). We find single placed in three parts evaluate whether there is significant improvement by adding multiple accelerometers or not. also number states that best each achieving lowest test error K-fold cross-validation.