*2.2. HAR System*

HAR is a way toward recognizing common human activities in daily living. It is turning into an attractive research field because of numerous areas of application. Physiological signals, environmental signals, location data and acceleration signals are the categories of input data that are acquired from wearable sensors in an HAR setting. Physiological signals data such as skin conductivity, heart rate, skin temperature, respiration rate and electrocardiography signals have also been considered in a few research studies to improve the recognition accuracy [34]. Environmental signals, for example, audio level, temperature, humidity are proposed to provide information exposing the individual's environment [35,36]. The Global Positioning System (GPS) empowers all kinds of area-based information. Current smartphones are equipped with integrated GPS, making it exceptionally advantageous for context-aware applications such as the recognition of the individual's transportation mode [37]. Accelerometers can be categorized into two types: either as body wearable sensors or incorporated with new models of mobile devices such as smartphones and smart watches which record the body movement [38,39]. The majority of research work in HAR applications is performed using wearable sensors. Figure 1 shows the HAR with different wearable sensors.

**Figure 1.** Wearable sensors for human daily activities.

The various types of wearable sensors are used to identify different human activities in three categories: gesture, low-level activities and high-level activities as shown in Figure 2.

**Figure 2.** Categories of human daily activities.

Gestures activities are classified as extremely short activities such as bending the arm, or opening and closing the hands. Human daily activities such as sitting, standing and running which typically last between seconds and a few minutes are known as low-level activities. Activities like working at the office, cleaning the house and sightseeing normally last for more than a few minutes up to a few hours are known as high-level activities [40]. Detailed information about the general process for training and testing the data of HAR systems based on wearable sensors can be found in [41].
