Loading presentation...

Present Remotely

Send the link below via email or IM


Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.


Ambient Assisted Living - Internet of Things

No description

Shweta Seth Mehndiratta

on 14 May 2014

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Ambient Assisted Living - Internet of Things

Ambient Assisted Living - Internet of Things
Internet of Things
"Humanization of Technology"
Reduced costs, improved quality and safety, Speed
Market Research
Smart Diapers
Infant Monitor
Insulin Injection Tracker
Prescription Pills
IoT Architecture
Assistive Tools and Technologies
AAL Tools and Technologies
Smart Home
Mobile and Wearable Sensors
Sensors in a Smart Home
Mobile and wearable sensors
Sensor Communication
TinyOS operating System
IEEE 802.15.4 (ZigBee) as the radio interface
Network architecture - BAN (Body Area Network)
BAN - enables three tiers of wireless communication around a human being
Assistive technologies are based on "ambient intelligence"
Aimed at empowering people's capabilities with the help of digital environments which are
responsive to human needs
Assisted living technologies using ambient intelligence - Ambient-Assisted Living tools (AAL)
Organization of mutual assistance community
Assist elderly to overcome physical limitations
Three main categories
ADL - Activities of Daily Living
(Feeding, Dressing, Grooming etc.)
EADL - Enhanced Activities of Daily Living
(Engaging in hobbies, social activities etc.)
IADL - Instrumental Activities of daily living
(Using telephone, preparing food etc.)

AAL Algorithms
Activity Recognition
Context Modeling
Mobile Activity Recognition
Ambient Activity Recognition
Vision-Based Activity Recognition
Anomaly Detection
Location and Identity Identification

Any Questions !!
Thank You
Challenges and Conclusion
AAL promise many future opportunities for the elderly with the advancing technologies

There are many challenges which must be overcome to unleash the full power of computational technologies

In sensor technology, care should be taken to make devices wireless and electromagnetic energy radiated by the devices should not harm human tissue

Security and privacy of devices is critical, user authentication should be based on biometric and physiological features to safeguard user privacy.

Simplifying assumptions in algorithms should be relaxed

AAL systems should not reply on users effort

Activity Recognition Algorithm
Most important component - Human Activity Recognition (HAR)

HAR recognizes human activity pattern using low level sensor data.

Data is different from different sensors
wearable sensors like accelerometer or gyroscope - time series data (shown in figure below)
ambient sensors such as motion sensors - numerical/ categorical data
cameras/ thermographic devices - image/video data

Fig : Processing activity time series data
Mobile Activity Recognition
Data is in the form of time series - sequence of data points measured at regular intervals

Processing time series data is a multi stage process
sensor data are recorded at a particular frequency
data is precessed to remove high frequency noise and segmented into shorter segments
Statistical & Morphological data is extracted from signal segments
Feature selection and dimensionality reduction techniques are applied
Internet of things and AAL - ZigBee Protocol
Based on IEEE802.15.4 networking standard

Zig bee creates a WAN for M2M communication so sensors and devices can communicate and control

Zig Bee also collects the data securely over the network

Devices can wireless connect within a range to 70m indoors and 400m outdoors

Also has facility to display device control and monitor the information gathered

Supports decentralized mesh networking, can support upto 65,000 nodes

Organization of mutual assisted community
Ambient Activity Recognition
Rely on labeled data for training, decision trees, neural networks, core based reasoning, graphical models

Graphical models are able to deal with sequential nature of data
Markov chains
Dynamic Bayesian network
Hidden Markov Model

But consistent predefined activity is hard to achieve, to overcome this challenge, data mining methods are employed
Frequent sensor mining
Activity stream mining
Activity episode discovery
Sequential activity mining

Vision Based Activity Recognition
Figure :Vision Based Activity Approaches
1. Rashidi, Parisa, and Alex Mihailidis. "A Survey on Ambient-Assisted Living Tools for Older Adults." IEEE Journal of Biomedical and Health Informatics 17 (2013): 579-589. Print.

2. Bassi, Alessandro, Martin Bauer, Martin Fiedler, Thorsten Kramp, Rob Van Kranenburg, Sebastian Lange, and Stefan Meissner, eds. Enabling Things to Talk: Designing IoT Solutions with the IoT Architectural Reference Model. Berlin, Heidelberg: Springer, 2013.

3. Logvinov, O. (2014, Jan 28). Healthcare and the ‘Internet Of Things’. [Web Log Post] Retrieved April 01, 2014 from http://www.mdtmag.com/blogs/2014/01/healthcare-and-‘internet-things’

4. Couturier, j., Sola,, d., & Borioli, g. (2012). How Can the Internet of Things Help to Overcome Current Healthcare Challenges, Digiworld Economic Journal. Retrieved from http://innovation-regulation2.telecom-paristech.fr/wp-content/uploads/2012/10/CS87_COUTURIER_et_al.pd

5. Hu, F. Dan, X. & Shen, S. (2013). On the Application of the Internet of Things in the Field of Medical and Health Care. In the Proceedings of 2013 IEEE International Conference on Green Computing and Communications and IEEE Internet of Things and IEEE Cyber, Physical and Social Computing. 2053-2058, doi:10.1109/GreenCom-iThings-CPSCom.2013.384.

Provide detailed context info

Challenge is variation in natural setting, algorithmic complexity and privacy concerns

Data is first pre processed by applying foreground segmentation techniques and then activity recognition from video frames
Full transcript