Human Activity recognition using Deep Learning

Abstract:

Machine learning research is heavily focused on human activity detection since it has various applications in a variety of fields, including security, entertainment, ambient supported living, and health management and monitoring. Researchers' interest in human daily activities is seen from studies on human activity recognition (HAR). As a result, the general architecture of the HAR system and a description of its key elements are described in this work. Review of the state-of-the-art in accelerometer-based human activity recognition According to this survey, the majority of recent research that employed CNN for HAR identification relied on it even though other deep learning models also showed acceptable accuracy. The paper suggests a 2 different classification depending on the kind of machine learning (conventional or deep learning) and the manner of execution (online or offline). Comparing 48 studies prediction performance, algorithms, activity categories, and used equipment. The study concludes by contrasting the difficulties and problems associated with identifying human movement based on accelerometer sensors utilizing deep learning versus conventional machine learning, as well as online versus offline.

Introduction:-

Human activity recognition using smartphone sensors like accelerometer is one of the hectic topics of research. HAR is one of the time series classification problem. In this project various machine learning and deep learning models have been worked out to get the best final result. In the same sequence, we can use LSTM (long short term memory) model of the Recurrent Neural Network (RNN) to recognize various activities of humans like standing, climbing upstairs and downstairs etc. 

LSTM model is a type of recurrent neural network capable of learning order dependence in sequence prediction problems. This model is used as this helps in remembering values over arbitrary intervals.
Human Activity Recognition dataset can be downloaded from the link given below: HAR dataset 
Activities: 
 

• Walking
• Upstairs
• Downstairs
• Sitting
• Standing

Accelerometers detect magnitude and direction of the proper acceleration, as a vector quantity, and can be used to sense orientation (because direction of weight changes). GyroScope maintains orientation along a axis so that the orientation is unaffected by tilting or rotation of the mounting, according to the conservation of angular momentum. 
Understanding the dataset: 
 

• Both the sensors generate data in 3D space over time. 
  (‘XYZ’ represents 3-axial signals in X, Y, and Z directions.)
• The available data is pre-processed by applying noise filters and then sampled in fixed-width windows ie., each window has 128 readings.

Train and Test data were separated as 
The readings from 80% of the volunteers were taken as training data and remaining 20% volunteers records were taken for test data. All the data is present in the folder downloaded using the link provided above.
Phases 
 

• Choosing a dataset 
   
• Uploading the dataset in the drive to work on google colaboratory
• Dataset cleaning and data Preprocessing
• Choosing a model and building deep learned network model
• Exporting in Android Studio.