This project aims to design, implement, and evaluate a low-cost, robust and open assisted living environment that exploits inexpensive technologies to enable elderly people to regain their capability of independent living. Suppose such an assisted living environment were deployed in Walter's house. Every morning Walter wakes up, has breakfast, takes his medicine, takes a walk around the park, and returns home to work on his novel. The RFID readers installed in the house, along with the RFID tags attached on the various appliances/belongings, note when he exits his house, when he takes his medicine, when he returns, and which room he is in. The GPS chip knows the rate at which he walks around the park. On that particular day, Walter tripped and fell unconscious in his living room. He was alone. The RFID based localization techniques localized he was near the living room door, and that he was not moving for an unusually long time. A message flashed on the TV asking whether to call for help and was not answered. Within minutes, an ambulance arrived at his door and his son was notified of this critical event. Walter was treated and returned to his regular routine the next morning. From an economic perspective, such a networked assisted living environment represents a major potential saving in the financial drain of senior care (e.g., the manpower, money, and efforts taken to care for elderly people who are in fact capable of living independently with modest assistance). From a social perspective, a networked assisted living environment increases not only the ability to live independently, but also individual empowerment and the ability to communicate with family, friends, and health care professionals.

Thanks to rapid technology advances in sensing (e.g., Ubisense/RFID techniques for tracking, presence identification, and localization), computing, and wireless networking (e.g., Bluetooth/IEEE 802.11 enabled devices and infrared-equipped remote controls), many component technologies needed for realizing such an open environment are already available. Moreover, new and improved devices and applications are expected to emerge.

What is missing is the scientific and technology foundation for the design of an easy-to-use, low-cost, robust, secured and open environment for assisting elderly people with the minimum intrusion of privacy, e.g., without use of surveillance video camera in their houses. The main intent of our project is to develop and evaluate (with Ethnography-based studies and user group assessment) (i) a software infrastructure that allows disparate technologies, software components, and wireless devices of different protocol families to work together in a low cost, dependable, and secure fashion with predictable properties; and (ii) a user-friendly interface that abstracts this software infrastructure for technology-naïve users and provide a wide spectrum of services to facilitate elder people with their independent/assisted living.

To design, implement, and evaluate such a software infrastructure, layers of sensing, localization, communication, and event/data management have to be carefully designed to meet the following challenges (Figure 1):

1) Safety, robustness and availability: First, critical services must be failure safe, and be delivered in spite of the failures of useful but non-critical services. Moreover, the system as a whole must have high availability and robustness.

2) Low cost, evolvability and interoperability: The independent living infrastructure must be open and support the use of low-cost, third-party devices. They must have well defined interfaces with machine checkable QoS assumptions, as incompatible and incomplete real time, fault tolerance, and security assumptions have been reported to be a major source of unexpected interactions leading to high maintenance cost and service disruption.

3) Security and privacy: Medical and personal data have to be protected in the assisted living infrastructure where wireless networking will be the predominant communication medium. Security mechanisms have to be built in in both information storage and communication facilities.

4) Quality-of-service provisioning: In spite of the existence of various forms of workload dynamics, ranging from transmission of reminder messages, monitoring information, audio commands, to time-critical multimedia streams supporting tele-medicine, quality of service (QoS) has to be provided at different levels to applications subject to their timing, reliability and criticality requirements.

5) Wireless interference mitigation: Different wireless devices of disparate protocol families can interfere with each other when they are in range, e.g., Bluetooth versus IEEE 802.11b, and IEEE 802.11a versus microwave. However, quality of service (QoS) must be provided even in the presence of wireless interference and medium contention. Transmission control protocols must be designed and deployed to mitigate interference and to provide adequate temporal QoS.

6) Light-weight, easy-to-use HCIs for elderly population: The user interfaces must be designed to be easy-to-use, safe, accommodating with respect to user mistakes, and provide different control levels of information disclosure.

7) Thorough evaluation and user group studies: Before its eventual deployment, an assisted living environment should be evaluated in terms of the ease of use of various HCI components, the extent to which these technologies help elder people in the home or the assisted living facilities, and their attitudes toward deploying these technologies.

To develop a feel of the challenges, we visited in February 2004, with the help of JoAnne Carlin, Vice President of Residence Care for Classic Residence by Hyatt, several assisted living facilities in Illinois, and depicted quite a number of scenarios in which new technologies would be most helpful to residents and caregivers from both the customer and management perspectives.

We envision an open environment in which a security-enhanced, assisted-living ALD is co-located with the home PC, and is equipped with one or more wireless interface cards (IEEE 802.11, Bluetooth, Ultra Wide Band, and Infrared). The software infrastructure that supports assisted living is installed on the ALD (which then acts as the intelligence of the environment.) Such devices are nascent today but will be commonplace in future years.

Scenario 1: The ALD obtains updated medical and appointment records of residents through secure channels from their health care providers. When it is time for elderly people to carry out their time driven routines, such as taking medicine, taking vital signs, and/or making clinic appointments, the ALD detects the wireless-enabled devices (e.g., TVs, cell phones, wearable headsets, and/or active badges) in the environment, and sends reminder messages to one or more devices that are in the proximity of the resident(s). For example, if the resident is watching TV at the time when the reminder message is scheduled, the TV will be switched to an information channel (with the use of Infrared remote control) and display the reminder message. In this manner, elderly people can be reminded of their time driven routines. Whether or not these routines are followed as advised is detected by exploiting sensor localization technologies such as Ubisense or RFID --- the prescription bottles can be attached with light-weight RFID or Ubisense tags (with unique barcodes) and one or more RFID/Ubisense readers in the environment are scheduled to track location changes (if any) of these bottles. Note that each RFID tag costs approximately 40 cents today, and the cost is expected to further decrease in the future.

Scenario 2: Personal belongings such as eyeglasses, hearing aids, and key chains can be attached with tags, and located through the use of Ubisense/RFID readers. When a resident cannot find his/her belongings (because of forgetfulness), the ALD can schedule RFID/Ubisense readers to scan the environment and help locate the object upon receipt of a vocal command (through, for example, a light-weight, Bluetooth-enabled headset).

Scenario 3: With the same set of sensor localization techniques, the assisted living environment can profile elderly people's movement in a privacy preserving manner (e.g., without the use of surveillance video cameras) and detect early warning signs for depression (stop taking medicine regularly, giving up routine activities, or staying in bed for long periods of time) and/or other chronic diseases such as ParkinsonÕs disease and AlzheimerÕs disease. In Walter's example given in Section B, Walter will wear a RFID tag or an active badge (turned into, for example, a bracelet). The RFID readers installed in the environment will keep track of his location, and hence will be able to detect abnormal spatio-temporal movement, without intrusion of privacy.

Scenario 4: In case of the need for emergency attention (e.g., the blood pressure/sugar has been dangerously high/low, and/or the resident has been detected via localization techniques to be immobile on the floor for a unreasonably long time like in Walter's example), real-time wireless channels can be established to notify on-site caregivers (in the case of assisted living) and/or designated relatives, and facilitate transmission of EKG and other measures in real time.

Note that the listed example scenarios are by no means complete, and are simply provided to set the stage for subsequent discussions. In our preliminary work, we have laid an initial software architecture that provides the functionalities needed to realize these and other possible scenarios.

Figure 2 depicts our initial layout that realizes only the functionalities required for assisted living, but not all the important reliability, robustness, security, privacy, and device interoperability properties. (We will elaborate on these research issues in Section D.) For ease of exposition, we demonstrate the software architecture in Java, but the architecture can be implemented in any platform that supports mobile agents. The functionality of each software module is described below.

1. When a resident enters the environment, the person location monitoring application (labeled as 1a in Figure 2) identifies him/her (e.g., by reading the RFID tag the resident wears), and downloads (through secure channels from a mobile servant code library) the mobile servant (labeled as 1b) of the resident. A mobile servant is a mobile agent configured (according to the attended resident's preference) to move with the resident, and is responsible for executing the various tasks. At the same time, the person location monitoring application updates the person location DDB (labeled as 1c) of the current location of the resident. When a resident leaves the environment, the monitoring application updates the DDB to reflect the fact that the resident has left.

2. The device location monitoring application (labeled as 2a) identifies wireless devices (if any, by identifying its RFID tag or its NIC address) the resident carries, and downloads from a module driver code library the necessary module drivers (labeled as 2b). The module drivers present to the mobile servant (and other applications) unified programming interfaces, and shield the various device-specific details from the latter. For example, the module drivers associated with Bluetooth-enabled headsets and 802.11-enabled PDAs should all support the MediaPlayer application programming interface. The mobile servant developer thus needs only to know the MediaPlayer interface. At the same time, the device location monitoring application also updates the device/object location DDB (labeled as 2c) of the current location of the wireless device. When a mobile device leaves the environment, the monitoring application updates the DDB as well.

3. As every wireless LAN has its own protocol stack implemented in the native code, the device integration layer (labeled as 3) is provided to shield the various protocol-specific network stacks and present a unified programming interface to mobile module drivers. The device integration layer should also take care of QoS provisioning, device co-existence and interference issues. (We will discuss these research issues associated with the device integration layer in Sections D.4-D.5.)

4. All the downloaded mobile agents, such as mobile servants and mobile module drivers run in the Java security sandbox (labeled as 4). The Java security sandbox is in charge of carrying out all the security policies such as authentication, access control and encryption/decryption on incoming mobile agents.

Various services can be provided by writing various applications. For example, the time-driven reminder service (Scenario 1) can be provided by having a reminder application (labeled as 5 in Figure 2) find, at scheduled time instants, a proper wireless device in range and issue reminders (e.g., time to take medicine). The reminder application will also schedule the RFID/Ubisense readers to scan the environment and keep track of the resident's response (e.g., whether or not the medicine has been taken). Similarly, the object localization service can be provided as follows. An object localization application, along with its proper module driver, resides on top of the device integration layer and receives vocal commands (e.g., through the light-weight, Bluetooth-enabled headset of a resident). Upon receipt of a vocal command, the application either (i) queries the device/object location DDB or (ii) schedules RFID/Ubisense reader scanning events to find the whereabout of the requested object. Then the application can send a response message to a proper wireless device that is either in range or specified by the resident.

It should be clear that numerous applications can be supported under this software infrastructure, and new wireless devices can be plugged into the infrastructure to provide new capabilities. However, this is not sufficient --- the assisted-living software infrastructure will be truly useful only if it will (1) deal robustly with a wide range of failure scenarios, (2) be very reliable in diverse operating conditions, (3) communicate securely with well-authenticated parties who are granted proper access to the information, (4) respect the privacy of its users, (5) provide QoS even in the presence of wireless interference, and (6) be easily used by personnel with very little training and limited physical and mental capabilities. This list of requirements is beyond what most software is able to achieve today, and hence new technology that can contribute to its achievement is essential.To address these important issues, we are carrying out R&D tasks in the six following research thrusts:

• Embedded software robustness and system integration: Lui Sha, Professor, Department of Computer Science, University of Illinois at Urbana Champaign


• HCI and social computing: Karrie Karahalios, Assistant Professor, Department of Computer Science, University of Illinois at Urbana Champaign


• Security and privacy: Carl Gunter, Professor, Department of Computer Science, University of Illinois at Urbana Champaign
  
  
• QoS provisioning: Marco Caccamo, Assistant Professor, Department of Computer Science, University of Illinois at Urbana Champaign; Chang-Gun Lee, Assistant Professor, Department of Electrical and Computer Engineering, Ohio State University


• Wireless networking and communications: Jennifer Hou, Professor, Department of Computer Science, University of Illinois at Urbana Champaign


• Quality of care: Susan DesHarnais, Institute for Healthcare Studies, Northwestern University.