Abstract: The most important and challenging research area in proactive and ubiquitous computing is Human activity recognition. Automatic classification of human activity milieu (from simple activities to more complex ones) are crucial for applications like monitoring elderly people in assisted living, activity aware media content delivery, designing smart homes and appliances, quantified self, smart health care etc. Physical activity recognition using wearable sensors is capable of providing priceless information regarding an individual’s degree of operative ability and lifestyle. During the old age periods, falls are a major problem.so they are forced to depend on others. To monitor the way of walking of elderly people development of a technology that analyzes the relationship between the possibility of fall with the fitness and the total number of daily living activities of the elderly person that looks for precursors to falls. As we are surrounded by a lot of IoT devices, and efficient communication between man and machine are important for the proper working of these devices. So, activity recognition can be used to make our life easier through these IoT devices. There are several models for recognizing activities that use different techniques. We aim to analyze various models of human activity recognition and to propose a model for secure activity monitoring.
Keywords: IoT, Automatic Classification, Ubiquitous Computing, Quantified Self
| DOI: 10.17148/IJARCCE.2019.81218