Bachelor and Master Theses

Title: Master Thesis: Implementation of a Decision Support System for an Ambient Assisted Living System
Subject: Computer Science
Level: Advanced
Description: The society is now witnessing a demographic shift, the birth rates are reducing and life expectancy is increasing. This leads to a larger proportion of elderly population in the society. According to the statistics of the World Population Ageing Report 2015, the world's elderly population (that is, people older than 60 years of age) is increasing at a fast pace and is predicted to reach 2.1 billion by 2050, which is more than double of the population of elderly adults in 2015. This change in demography has many consequences, the major being the insufficient number of caregivers. Therefore, we are in need of techniques that will assist the elderly in their daily life, while preventing their social isolation. The recent developments in Ambient Intelligence (AmI) and Information and Communication Technologies (ICT) have facilitated a technological revolution in the field of Ambient Assisted Living (AAL).

Successful decision making is one of the most prominent characteristic of any AAL system as it needs to adapt to highly dynamic environments and respond to critical scenarios in time.

As part of this thesis, you have to design and implement an event-based Decision Support System (DSS) with an associated knowledge base (KB) for an integrated AAL system (integrated implies multiple functionalities are tackled by the system, such as health monitoring, fall detection and raising alarm, managing supervised physical exercises, advices/reminders, enabling communication with caregivers, friends and family, monitoring home parameters, e.g., automatic detection of fire and communication to firefighters etc.). As the environment for decision making is highly dynamic, we need to ensure that the data is up-to-date and real-time and hence the need to identify whether the current data is enough for decision making or one needs to explore the environment further. The DSS should also incorporate intelligence for taking a decision when multiple events occur simultaneously, like a fall and fire etc. State-of-the-art artificial intelligent algorithms need to be incorporated in the DSS, in order to effectively tackle the scenarios/events described below. A combination of Case Based Reasoning and Planning (CBR, CBP) techniques together with fuzzy logic based reasoning can be used for non-critical scenarios. For critical scenarios like fire, fall, cardiac arrest etc., we need to use expert system-based decision making.

The scenarios for decision support are described below. The care taker is Jim, a man older than 60 years, who lives alone in his home. Jim has a heart condition, hence vital monitoring of health parameters (e.g. ECG), and fall detection are critical.

The scenarios are described below:

1) Fall detection (Critical)
Jim wakes at night and needs to go to the bathroom. Jim feels dizzy after getting out of bed. The light turns on automatically, but Jim loses his balance and falls. The wearable fall detector sensor senses the fall. It sends an event to the system gateway, informing of a potential fall. The sensor is only partly aware that a fall has taken place, and reports with low confidence the detected fall. Jim is not able to return to bed, and no further movement is detected in the home. The ambient fall detection sensors also report that a suspected fall might have occurred. The DSS takes a decision to alert Jim's caregivers and family members, who receive an alert on their smartphone. It can also be extended to the case where the family and caregivers communicate with the system (via telepresence) and DSS, after having received the fall.

2) Health parameter variation due to cardiac arrest, and fall detection (Critical)
Jim has unusual ECG variations detected by the heart monitoring sensor due to a cardiac arrest, and falls. The ECG variations and fall event are communicated to the DSS through the gateway, and the DSS analyzes the situation and infers that a fall has occurred due to cardiac arrest. Next, it informs caregivers (possibly hospital) and family of the events, for immediate rescue.

3) Planning of Jimís daily activities and issuing reminders (Non critical)
The case-based planning of DSS should be able to propose daily activity plans for Jim based on his health condition and input from caregivers. It should also send reminders to Jim if he has forgotten to do any of scheduled activities, like missing a medication or reminding of an appointment with doctor etc.

The choice of programming language for implementation is flexible, yet the preferred ones are C or Java.

Expected Outcomes:
1) Literature Review of existing DSS.
2) Selection and adaptation of Intelligent DSS algorithms.
3) C/Java implementation of the selected/proposed algorithms.
4) Evaluation of correctness (logical and real-time if the case) of outcomes for the selected scenarios via simulation.
5) Thesis report (in English).
Company: IDT, kontaktperson: Ashalatha Kunnappilly
Proposed: 2016-11-03
Prerequisites: 1) Programing skills in C or Java. 2) Communication and writing skills in English. 3) Knowledge in AI is a plus (but not required).
IDT supervisor: Ashalatha Kunnappilly
ashalatha.kunnappilly@mdh.se,
Examinator: Cristina Seceleanu
Cristina Seceleanu
cristina.seceleanu@mdh.se, +46-21-151764

Rapport och bilagor

Size

Senaste uppdatering


  • Mälardalen University |
  • Box 883 |
  • 721 23 Västerås/Eskilstuna |
  • 021-101300, 016-153600 |
  • webmaster |
  • Latest update: 2017.08.23