Published Paper Details:

Human activity detection is one of the leading applications of the machine learning algorithm today. It is used in the field of biomedical engineering, game development, developing better statistics for sports training, etc. The data is collected from the embedded accelerometer, gyroscope and other sensors of the XiaomiRedmi Note 7 Pro smartphone. In this project we will use data available in the WSIDM Lab. It includes data generated by smartphone accelerometers, gyroscopes, and other sensors to train supervised predictive models using Long Short-Term Memory (LSTM) machine learning techniques to generate a model. The experimental results show that the proposed approach can help improve the detection rate up to 95.87% and the transient detection rate over 80%, which is better than most existing similar models used to predict the type of motion that can be performed by the person, which is divided into the eight categories of standing, sitting, walking, running, sitting down, getting up and climbing stairs.