Posture Identification Using Artificial Neural Network

This research shows that artificial neural networks (ANN) for classification divide the human position into six groups stand, sit, run, sleep and bending back and forth. Recognition of human posture has many health benefits analyses, including geriatric care, lifestyle analysis and patient monitoring. Most wireless (Wi-Fi) reception and analyzing sensor data on Raspberry Pi device with little lag, our method can identify the postures mentioned above time for neural network training and testing. A set of 44,800 samples from 3 patients gathered. The best nervous system architecture (6-9-6) with hyper parameters was found after extensive testing and experiment with different systems architecture that ensures overall accuracy 97.589%.