ITRA

Information Technology Research Academy

Human Sense

Towards context aware sensing, inference and actuation for applications in Energy and Healthcare.

a. Sensing systems for non-mobile data collection:

  • Integration or extension of commercial off the shelf sensing solutions for monitoring energy related parameters: Proposed sensors for energy and healthcare will be used for this task.
  • Deployment of initial prototype systems in campus/homes for small scale data collection: Proposed sensors for energy and healthcare will be used for this task.

b. Development and deployment of mobile based systems for preliminary data collection to guide future directions of the project:

  • Automated detection of radio interfaces present at any moment that can be used to create ad hoc networks: Proposed access points in the budget are to be used for this task.
  • Integrating data collection from external ECG and SpO2 sensors and in-built accelerometer sensors to create a mHealth kit: Bluetooth based health sensors to be used for this task.
  • Extensive data collection of available radio interfaces and the quality of the communication channel for each one of them for empirical determination of the feasibility of ad hoc networks for mobile based data collection and communication: Access points and network interfaces to be used for this task.
  • Developing suitable mechanisms to deal with unreliable communication channel and constrained energy for reliable data fusion within the mHealth kit: Some mobile phones will be used to collect data for this task.
  • Develop suitable mechanisms for on device computation for minimal data transmission: Some mobile phones will be used to collect data for this task.
  • Porting some of these mobile based systems for diverse mobile environments (e.g. Symbian, Blackberry, Windows) while the primary environment will be Android: Some mobile phones will be used to collect data for this task.
  • Develop a cloud-based participatory sensing system so that mobile phone users can contribute data easily: Server available with IIIT Delhi will be used for this purpose.
  • Extending data collection methodologies for supporting interfaces available with low cost phones (e.g. SMS/missed call): : Some mobile phones will be used to collect data for this task.

 c. Understanding user requirements and incentive structures:

  • Study different ways to collect user data for the applications in healthcare and energy as proposed in the project; take a detailed look at the previous work in this domain for energy and healthcare applications
  • Develop an interview protocol and collect data from representative set of potential users
  • Develop Focus group discussion script and collect data from representative set of potential users

d. Development of robust and scalable middleware for data aggregation and user feedback:

  • Architecture design of the middleware with details of modules and APIs to enable data collection from sensing devices with diverse interfaces, mobile based sensors, human agent intermediaries and external data sources
  • Development of the beta prototype of the middleware with modules for efficient management and retrieval of data

e. Understanding the data provenance requirements of the target application:

  • Create the data flow model of the target application
  • Identify the provenance data that should be collected across different layers of data collection
  • Analyze different approaches of modeling data provenance. Designing a provenance model for the target application and evaluating it with respect to different performance metrics like storage and querying flexibility in the simulated environment

f. Modeling using collected data:

  • Use collected data to create channel models for the target environments, assuming typical usage of the phone (for example, in a user’s pocket) or the hardware used for measurement: Some mobile phones will be used to collect data for this task
  • Use the empirically determined channel models and existing network topologies to estimate (via simulation) available network capacity in terms of throughput, delay and metrics that may be suitable for applications of interest to areas of healthcare and energy: Proposed Access Points to be used for this task
  • Extend the on-device computation mechanisms to evaluate a mix of distributed and centralized strategies with an aim of coming up with heuristics that can be used in practice: Some mobile phones will be used to collect data for this task
  • Analyze collected data to evaluate the feasibility of using of user context, as collected by sensors like GPS and accelerometers, to optimize communication: Some mobile phones will be used to collect data for this task
  • Build models of the data for non-pathological and pathological subjects on a resource rich server and identify differences for cardiovascular diseases (CVD)
  • Extend the CVD model with same or similar functionality to work on the mobile device with minimal help from the server
  • Develop algorithms to both detect the onset and monitor cardiovascular diseases on the mobile unit with minimal help from the server: Some mobile phones will be used to collect data for this task

g. User modeling:

  • Use the understanding developed through interviews and focus group discussions held in the first year to develop (low- and hi-fidelity) prototypes of the systems that will be developed for field deployment
  • Evaluation of the prototypes amongst small set of representative users

h. GIS based spatial analysis: GIS Software and Server at IIIT Delhi to be used for this task:

  • Spatial data modeling using collected data and a Desktop GIS application
  • Mining of spatial data collected from sensors about energy to find out trends and spatial dependency
  • Identification of effective Web-GIS technology for spatial data visualization and analyses for energy data
  • Development of Web-GIS application for near real time visualization of energy data and analysis tools for decision makers and cluster institutes
  • Mining of spatial data collected from sensors about healthcare to find out trends, spatial dependency and spatial autocorrelation
  • Desktop GIS tool for decision support system
  • Identification of effective Mobile-GIS technology for spatial data visualization and basic analysis of the data
  • Development of Mobile GIS app for visualization of data and analyses tools

i. Evaluation and validation of project outcomes using real world deployments on campus of participating institutions as well as with end user collaborators: Multiple mobile phones proposed in the project to be used for this task.

j. Development of multiple applications built over the RESTful interface of middleware that use the collected and developed middleware functionalities and using these applications for detailed data analysis:

  • Understand the constrains that must be honored (e.g. consistency) when sharing data across multiple end consumers
  • Build mechanisms for easy integration of domain specific (particularly for energy and healthcare) data sharing requirements through modular design
  • Enhance the middleware providing RESTful API interface, supporting third party applications
  • Use the middleware for data collection in medium scale on-campus deployments for energy and healthcare
  • Development of multiple applications built over the RESTful interface of middleware that use the collected and developed middleware functionalities and using these applications for detailed data analysis

k. Work with industrial collaborators for commercialization of research outcomes: Money for entrepreneurial activities proposed for this task.

Summary and Vision

Consider a world where every human is empowered with information that helps her (and therefore eventually the whole ecosystem) in making optimal use of available resources. Such a vision can be realized using recent technology advancements, in low cost and pervasive sensing, on-device and cloud computation, and ubiquitous communication, allowing for collection of individual-specific data about her activities and the corresponding context in which these activities are performed, which can be environmental, social, economic, cultural and educational, among others. The HumanSense project proposes to monitor and adapt to individual and context specific requirements to enable systemic studies, thus paving the way for data-driven research to deliver public goods. It will build upon proven technologies and platforms, while exploiting the research and field deployment capabilities of the team, collaborators and advisory board members. Wider participation and long term sustainability will be ensured through open dissemination of collected data and project outcomes.

Principal Investigators
Dr. Amarjeet Singh, IIIT Delhi amarjeet@iiitd.ac.in Link to homepage
Dr. Debopam Acharya, SNU Debopam.Acharya@snu.edu.in Link to homepage
Dr. S. R. N. Reddy, IGDTUW rammallik@yahoo.com Link to homepage

Human Sense

Towards context aware sensing, inference and actuation for applications in Energy and Healthcare.


Posted in Human Sense, ITRA-Mobile Projects, Project

|

Comments Off on Human Sense

PhD Students

Human Sense

Towards context aware sensing, inference and actuation for applications in Energy and Healthcare.


Posted in Human Sense, ITRA-Mobile Projects, Project

|

Comments Off on Human Sense

 

Collaborators

Human Sense

Towards context aware sensing, inference and actuation for applications in Energy and Healthcare.


Posted in Human Sense, ITRA-Mobile Projects, Project

|

Comments Off on Human Sense

 

Institutions

Indraprastha Institute of Information TechnologyShiv Nadar University,Indira Gandhi Delhi Technical University for Women

humansense

For more details, check the project website.

                                                   Project Status
PhD Student(s) 5
Publications   1
Number of Workshop/Training(s)  0
Labs Created 0

 

Publications

EnergyLens: Combining Smartphones with Electricity Meter for Accurate Activity Detection and User Annotation

Presented at Conference: ACM eEnergy, June 11 – 13, Cambridge, UK Authors: Manaswi Saha (Indraprastha Institute of Information Technology – Delhi (IIITD)); Shailja Thakur (Indraprastha Institute of Information Technology – Delhi (IIITD)); Amarjeet Singh (Indraprastha Institute of Information Technology – Delhi (IIITD)); Yuvraj Agarwal (Carnegie Mellon University (CMU))


Posted in 2014, Human Sense

|

Comments Off on EnergyLens: Combining Smartphones with Electricity Meter for Accurate Activity Detection and User Annotation

NILMTK: An Open Source Toolkit for Non-intrusive Load Monitoring

Presented at Conference: ACM eEnergy, June 11 – 13, Cambridge, UK Authors: Nipun Batra (IIIT Delhi); Jack Kelly (Imperial College, London); Oliver Parson (University of Southampton); Haimonti Dutta (Columbia University); William Knottenbelt (Imperial College, London); Alex Rogers (University of Southampton); Amarjeet Singh (IIIT Delhi); Mani Srivastava (UCLA)


Posted in 2014, Human Sense

|

Comments Off on NILMTK: An Open Source Toolkit for Non-intrusive Load Monitoring

Human Sense

Towards context aware sensing, inference and actuation for applications in Energy and Healthcare.


Posted in Human Sense, ITRA-Mobile Projects, Project

|

Comments Off on Human Sense