143
Aspects of Ambient Assisted Living and Its Applications
• Cameras-static or wearables (Rougier et al. 2011; Tabar et al. 2006)
• Combination of several strategies (Grassi et al. 2010)
In spite of all the examinations devoted to fall recognition, there isn’t a 100% solid calculation that captures all falls without issuing false cautions.
8.3.2 Activity Classification
The goal of this classification is to develop classifiers to recognize human activities and
exercises. The distinction between activity and action is that activity is a basic development or change in the stance of the client (e.g., gets up, lies, and strolls), while movement
could be a mix of a few activities speaking to a complex, more elevated amount of reflection, for example, cooking, cleaning, and eating. Distinctive machine learning strategies
are used for activity classification (Jalal et al. 2011).
TABLE 8.1
Sensor Selection Criteria for Home
Sensor Type
Location
Targeted Activity
Robustness
Efficiency
FSR
Under bed
Lying, sleeping
High
High
Under couch
Sitting, lying
High
High
Under chair
Sitting
Low
Medium
Photocell
In drawer
Kitchen activities
High
High
Cupboard/
wardrobe doors
Bathroom activities, changing
clothes
High
High
Digital distance
Back of chair
Sitting
Medium
High
Toilet seat cover
Bathroom activities
Medium
High
Above water tap
Bathroom/kitchen activities
Medium
Medium
Sonar distance
Walls
Activity related to presence in
a room
High
Medium
Contact
Regular door
Activity related to leaving/entering
room/house, showering
High
High
Sliding door
Showering, changing clothes
Medium
Medium
Drawer
Bathroom/kitchen activities
Low
Medium
Temperature
Above oven
Cooking
High
Medium
Near stove
Cooking
Medium
Low
Infrared
Around TV
Watching TV
High
Medium
Humidity
Near shower sink
Showering
Medium
Low
Pressure mat
On bed
Lying, sleeping
Medium
High
On couch
Sitting, lying
High
High
On chair
Sitting
Medium
Medium
Vibration
In drawer
Kitchen activities
Medium
Low
Source: Tunca, C., et al., Sensors (Basel), 14(6), 9692–9719, 2014. Doi: 10.3390/s140609692.
Note: Infrared (IR) sensors—to track multiple people in a closed environment.
Sensors in smart phones—accelerometer, gyroscope, magnetometer, and barometer—can capture the multiple activities like walking, climbing, running, and moving. It can locate the person in the environment.
RGB-D sensor—senses a person’s skeleton in a RGB video.
Motion sensor, door sensors, light switch sensor, power usage sensors are used in a smart home.
Aspects of Ambient Assisted Living and Its Applications
• Cameras-static or wearables (Rougier et al. 2011; Tabar et al. 2006)
• Combination of several strategies (Grassi et al. 2010)
In spite of all the examinations devoted to fall recognition, there isn’t a 100% solid calculation that captures all falls without issuing false cautions.
8.3.2 Activity Classification
The goal of this classification is to develop classifiers to recognize human activities and
exercises. The distinction between activity and action is that activity is a basic development or change in the stance of the client (e.g., gets up, lies, and strolls), while movement
could be a mix of a few activities speaking to a complex, more elevated amount of reflection, for example, cooking, cleaning, and eating. Distinctive machine learning strategies
are used for activity classification (Jalal et al. 2011).
TABLE 8.1
Sensor Selection Criteria for Home
Sensor Type
Location
Targeted Activity
Robustness
Efficiency
FSR
Under bed
Lying, sleeping
High
High
Under couch
Sitting, lying
High
High
Under chair
Sitting
Low
Medium
Photocell
In drawer
Kitchen activities
High
High
Cupboard/
wardrobe doors
Bathroom activities, changing
clothes
High
High
Digital distance
Back of chair
Sitting
Medium
High
Toilet seat cover
Bathroom activities
Medium
High
Above water tap
Bathroom/kitchen activities
Medium
Medium
Sonar distance
Walls
Activity related to presence in
a room
High
Medium
Contact
Regular door
Activity related to leaving/entering
room/house, showering
High
High
Sliding door
Showering, changing clothes
Medium
Medium
Drawer
Bathroom/kitchen activities
Low
Medium
Temperature
Above oven
Cooking
High
Medium
Near stove
Cooking
Medium
Low
Infrared
Around TV
Watching TV
High
Medium
Humidity
Near shower sink
Showering
Medium
Low
Pressure mat
On bed
Lying, sleeping
Medium
High
On couch
Sitting, lying
High
High
On chair
Sitting
Medium
Medium
Vibration
In drawer
Kitchen activities
Medium
Low
Source: Tunca, C., et al., Sensors (Basel), 14(6), 9692–9719, 2014. Doi: 10.3390/s140609692.
Note: Infrared (IR) sensors—to track multiple people in a closed environment.
Sensors in smart phones—accelerometer, gyroscope, magnetometer, and barometer—can capture the multiple activities like walking, climbing, running, and moving. It can locate the person in the environment.
RGB-D sensor—senses a person’s skeleton in a RGB video.
Motion sensor, door sensors, light switch sensor, power usage sensors are used in a smart home.