Design And Analysis Of Context Aware System

Published Date: 02 Nov 2017

Support for people with onset of dementia

Dementia is a chronic, progressive disease that is generally visible in the older population. Dementia leads to the loss of global cognitive ability in a person, which gradually progresses with age. People who are suffering from this disease find it difficult to perform their daily activities .This can be due to one of many factors such as the inability of the patient to maintain recent recall or inability to remember to remember which is known as prospective memory. They may also lack the ability to maintain a working memory which entails that they cannot retain and handle information that is newly acquired .They have difficulties in maintaining orientation with respect to time and place. In this paper we will examine one of the scenarios faced by patients inflicted with dementia who are trying to carry out their daily activities. Further, this paper will try to design a pervasive computing system to aid these patients.

Scenario

The best living environment for patients diagnosed with dementia would be one which helps them to be independent as much as possible. People suffering from dementia, after the initial stages, lack the cognitive abilities to finish most of their daily tasks. Tasks such as using the microwave oven, making tea, watching their favorite TV programs become hard for people who are affected by dementia, as they find it hard to handle various devices and the procedures involved in using them. The patient needs to be reminded to take his food/water and medicines on a regular basis. In this scenario, the selected aspect of this design will concentrate on a patient trying to cook food in the microwave oven for lunch without the help of any care giver. Context information available from the environment would be patients schedule from accumulated patterns, operational state of the oven, location of the individual, and steps done to complete the tasks.

Design

The patient’s ability to complete tasks in this scenario can be aided by reminders and alerts in their current environment which in our case is the kitchen. The system is scheduled to go active in case the person has forgotten to have food or hydrate himself/herself at specified intervals. Once active, the system will locate the individual through RFID (3.1–10 GHz) and motion detection. Different RFID tags are placed in commonly used objects of the patient such as shoes, watches, key chains and belts. These sensing tags hold the patients identification codes. Barcode is an alternative to RFID but it requires direct line of sight.

Prior to setting up the proposed system, a study similar to Almudevar et al will be carried out, where a patient was asked to wear RF (radio frequency ) transmitting watch for a period of 3 weeks .This will help us get the regions of high occupancy in the house. Based on these findings appropriate monitors and sensors will be put in place. The monitors are used to deliver contextually appropriate prompts to the patient. It is used to guide the patient through various activities.

Fig1: Example of activity workflow

Prior to launch of the system, it is loaded with activity workflows, videos and guides which are triggered based on the activity the patient is performing or looses track in performing (fig1). The patient is triangulated using RFID and passive infrared sensors (PIR). PIR is sensitive to the movements occurring in the vicinity (6 meters 110 degrees * 70 degrees). RFID sensors are preferred in case of multiple patients in one home so as to identify them differently. The house is fitted with RFID readers, PIR and other sensors, a microprocessor and a GSM module. The reader identifies the patient and his location is ascertained using the PIR and passes this data to the microprocessor where it is further analyzed. The GSM module is utilized to send emergency messages to the care giver in times of emergency such as, if the gas is kept on for a prolonged period or the patient is leaving the home proximity.

"Sensors may capture data about the environment and then send it to a processing unit for analysis using various methods that may directly examine the events or infer them through predictive models" (Evans et al. 2011).The house is also fitted with door/window entry point sensors (a 2.4GHz IEEE 802.15.4 standard wireless sensor), which detects the change in status (open or closed) of the door. To monitor the activity of electrical devices the system uses electric/power sensor (IMS-4000 Wireless Power Sensor), and the activity of each device has a specific power stamp which can be recognized by the system. Sitting areas are fitted with pressure sensors to detect the presence of the individual. Additionally doors can be fitted with soft lights to guide the patient.

We use a system similar to the French kitchen system developed by Hopper et al .The kitchen is fitted with S&R system (Sensing and recognition) where instead of camera’s (as this effects privacy and robustness) or Body worn sensors (too complex for patient to use) we put wireless acceleration sensors (40-50Hz) (e.g.: Open movement WAX3) on instruments used for cooking.

In the scenario at hand, a patient is lying in his bed and has forgotten to eat his lunch at the scheduled time pre-entered by the care giver. The system is able to monitor this by using pressure sensors installed under the bed. The individual is identified using the RFID placed on his/her watch. The system provides a peaceful reminder on the display in the bedroom TV/monitor .This alerts the patient who then proceeds to kitchen; this is monitored by the system using the PIR sensors placed in the walkway and the bedroom door sensors changing state. If the person has a loss of memory on the way he/she would be guided using lights placed on the doors. In case of further distraction a message is displayed on the watch with an alarm via wireless communication by the system. Once in the kitchen, the patient is provided with the guide on operating the oven and a walk through for preparation of the food. In case certain steps are forgotten, which is detected by the Sense and recognition system, the display highlights and prompts the step. Conditions required by the oven for specific food could be preloaded in the system so as to avoid complexity in the situation. When the food is almost cooked and ready to eat, it will be indicated by the timer going off and the state of the oven will switch to standby recognized by the power sensors. The system then locates the individual and sends a prompt to the closest display, along with the one to the watch. If the patient goes out of the home proximity during this process, the system will send an emergency message to the care taker using the GSM module.

PART 2

Based on Marc Langheinrich, 2001."Privacy by Design ‐Principles of Privacy--‐Aware Ubiquitous Systems." In Proceedings of the 3rdinternational conference on Ubiquitous Computing (UbiComp'01), Springer--‐Verlag, 273--‐291.

To make our system a privacy-respecting ubiquitous system, we will need to ensure that our system follows the principle of privacy. This system does not handle high level user sensitive data. Despite this, there has to be a certain amount of privacy to ensure security of the system, because it might be vulnerable to attacks from malicious sources.

Our system as per the scenario does not collect a lot of sensitive data other than the location of the user which once utilized to complete a particular task is deleted at the end of the day. To ensure that the data is not leaked due to unwanted accidents, we can use an authorized and encrypted system so as to ensure that the data is not being tampered with or leaked. This does not imply total security or total privacy, as Marc Langrich says:

"Undoubtedly, professional surveillance by spies and private investigators

will continue to happen, just as it has happened in the past. New technologies may be

found that will be able to (partially) sniff out such surveillance devices "

To deal with visitors privacy in our system, we should include a notice or an announcement which informs the visitor that they are under surveillance and that this is been done to help a patient in the home that they are entering. As our system records only the patient details and triangulates the location based on RFID and PIR there shouldn’t be privacy issues related to visitors. So there will be no need of a publicly accessible website to show the data gathered. However, if need be, under the supervision of the caretaker the visitor could see the data gathered on the system console.

As Langrich quotes, "it also requires collectors to receive explicit consent from the data subject."

Our system needs to have explicit consent from the patient or their care taker, in the form of a written contract as the subject of our data collection is the patient.

In our scenario based design we will not provide anonymity; rather we provide pseudonymity to a certain level. Pseudonymity is maintained by changing the ID or tag because the patient can be identified with the RFID tag. This is trivial because the user set is small.

The system follows the principle of collection and use limitations as it collects data only for the purpose of helping the user if it detects there has been a change in schedule or the person is having an episode. The only data recovered is the location and motion detection which are recorded so as to give the care taker a general outlay of the patients activities during the day. These records are then deleted to maintain privacy of the patient.