1. MapPrentice: Driving Assistance for Elderly

As we age, the way we drive changes; we drive more slowly, drive to fewer destinations, and feel less comfortable driving, particularly in unexpected situations such as traffic jams, construction, and bad weather. In addition, older drivers tend to have different preferences about they way they drive: avoid left turns, city streets vs. highway, etc. Current navigation systems, both on the web and in-car, do not adapt to the way that elders actually drive, but provide generic directions independent of who is doing the driving. In the MapPrentice project, we are investigating techniques for learning the driving preferences of elders in different situations, and are designing algorithms that can model these preferences and produce driving directions that are more appropriate for an individual elder. We intend to design and implement interfaces on mobile platforms and the Web to provide access to these improved driving directions.

Faculty advisor:  Anind Dey
Graduate student mentor:  Brian Ziebart
Location:  Carnegie Mellon
Project activities:

• Weekly meetings with research team
• Data collection and conducting user studies, including interviewing and interacting with subjects
• Design and development of information visualization systems for the Web
• Mobile phone/platform software development

Skills needed: Not a requirement, but programming ability is a plus
Skills acquired:

• Information visualization skills
• Ability to conduct studies with human subjects

^top

2. Ubiquitous Sensing Dystem for Evidence-based Care of Age-related Decline

This project will explore the application of sensing technology for evidence-based care for elders aging in place. To diagnose and assess medical conditions (e.g., Alzheimer’s disease, motor impairments) associated with physical or cognitive decline, clinicians must rely on three information sources: objective performance testing that often lacks real-world validity, self-report of patients that may be a selective account, and proxy-report of family members or caregivers who may lack contact with the patient. Clinicians would ideally have access to an accurate, validated account of patient’s behaviors from their everyday lives. 
The project will first identify opportunities where monitoring and recording real-life behaviors (such as Instrumental Activities of Daily Living) from people’s everyday lives will aid the diagnosis and assessment process. We will build sensing and recording systems that can collect the relevant data and automatically (or interactively) process it such that it is consumable to clinicians. We will evaluate the system first in a controlled setting and then follow with real-world deployments.
Faculty advisor: Anind Dey
Graduate student mentor: Matthew Lee
Location: Carnegie Mellon
Project activities:

• Assist in programming software to support the evidence-based care
• Sensing system
• Participate on field visits to study participant’s homes to assist in deployment and evaluation of the sensing system
• Attend weekly meetings with the research team
• Read background material on the project to gain an understanding of the domain and available techniques

Skills needed: Proficiency in programming in Java, C, C++, C#, or equivalent.
Special requirements: The student will accompany graduate student researchers to local people’s homes to help set up and maintain sensors. Though gaining this exposure to real world environments is educational, it is not strictly required for this project should the student be unable or not choose to take part in the field visits.

Skills acquired: The student will work closely with faculty and graduate mentors in setting up a formal
^top

3. Disparity in Quality of Wheelchairs provided to veterans with SCI

This is a survey study being sent to 488 veterans who are members of the Paralyzed Veterans of America. We are collecting information on the type of wheelchair they are currently using to determine whether wheelchairs purchased through the VA are better than wheelchairs purchased through other system. In addition, we would like to determine if different groups (such as ethnic minorities and women) receive the same quality wheelchairs as the Caucasian, male veterans.
Faculty advisor: Diane Collins, PhD  
Graduate student mentor: Ana Souza  
Location: HERL  
Project activities:

• Assist with data entry, bibliographic searches, preparation of data and data analysis,
• Prepare a RESNA paper
• Meet weekly with the research team
• Participate in other studies around HERL to get more research experience

Skills needed: Microsoft Office  
Skills acquired:

• Data management and entry
• Basic statistical analysis and research methods
• Technical

^top

4. Development of a Virtual Job Coach for Veterans with TBI

This pilot grant will develop and evaluate a Virtual Job Coach (VJC), a novel, computer-based method of delivering multimedia guidance to workers while engaged in industrial assembly tasks. The VJC is intended to meet the cognitive needs of traumatic brain injury (TBI) survivors who frequently are unable to sustain competitive employment due to distractibility, inappropriate demeanor and difficulties remembering and following instructions. The VJC is essentially an occupational teleprompter; analogous to the text teleprompter used by public speakers.
Faculty advisor: Don Spaeth, PhD  
Graduate student mentor: Harshal Mahajan  
Location: HERL  
Project activities:

• Software Development using Solid Works
• Computer aided design and Solidworks plugins to create animated 3D assemblies.
• Development may including researching commercial software code such as voice recognition or switch input so that the end-user can operate the system without a keyboard

Skills needed:

• Some experience with CAD software (doesn't have to be Solid Works).
• Knowledge of PC computer hardware and software at application programming level.
• Ability to write and debug code in C++ language and use compilers such as Microsoft Visual Studio highly preferred.
• Knowledge of video file formats such as AVI helpful.

Special requirements: Student must be able to work independently (although support will be available). Needs creative skills in software development and be willing to carefully document results. Ability to track down software support on the Internet is a key skill.
Skills acquired: PC architecture, Project Management, progress update communications via email.

^top

5. Assisted Living - Huma-Centered Design

This project focuses on using a intelligent interpretation of data from wireless sensor networks to monitor activities for elderly at home to improve their safety and quality of living.
Faculty advisor: Dr. Soundararajan Srinivasan 
Location: Robert Bosch LLC, 2 NorthShore Center 
Project activities: Interact with seniors living in geriatric care facilities to identify the special needs and constraints involved in the design of systems for use by older adults. The intern will also work closely with researchers at our Pittsburgh office and get familiar with the technology being developed. The person will therefore explore the usability and the anticipated effects of our monitoring systems on the aging population and the caregiver performance. He or she must elicit behaviors of interest within the bounds of a realistic geriatric care management, and develop methods that capture those behaviors. 
Skills needed:

• Thorough understanding of the user-centered design process, preferably gained through previous project experience.
• Previous experience with geriatric care, nursing, physical education or similar is a plus.

Special requirements: Limited visits to local locations (e.g.,Zelienople, PA).
Skills acquired: Experience in human centered design; designing for older adults.

^top

6. Assisted Living - Database Design

The project focuses on using a intelligent interpretation of data from wireless sensor networks to monitor activities for elderly at home to improve their safety and quality of living.
Faculty advisor: Dr. Soundararajan Srinivasan 
Location: Robert Bosch LLC, 2 NorthShore Center 
Project activities: Data collection, storage, management and visualization
Skills needed: MySQL, JSP 2.0; Java Servlets, Applets, Struts; XML, HTML, J2EE (preferred)
Special requirements: Limited visits to local locations (e.g.,Zelienople, PA).
Skills acquired: Database design and management of multisensor and human activity data.

^top

7. Client-Side Backup for a Distributed Task Guidance System

ICue is a task guidance system for people with cognitive impairments. For people who have difficulty remembering, or independently initiating, the steps in a task, ICue provides step-by-step guidance. Tasks are defined by a caregiver an uploaded to a central server. The server-side software delivers prompts based on the user’s schedule and progress on the current task. Currently, the user receives prompts on a standard web browser on a PDA, smart phone, desktop, or laptop. However, local (client-side) software is needed in case the user loses Internet connection. The local software will be responsible for delivering high-priority cues (e.g. to take medicine) until Internet connection is restored. The summer project will entail exploring possible software tools for use in implementing the local backup; and designing and developing an initial prototype.
Faculty advisor: Ed Lopresti 
Location: AT Sciences 
Project activities: Benchmarking software tools; software development; software testing
Skills needed:

• Programming experience or relevant coursework - required
• Experience with XML, PHP, JavaScript, and/or Ajax - preferred
• Experience with Java programming - preferred

Skills acquired: Exposure to design criteria based on the needs of people with cognitive impairments

^top

8. Robot 250 and Quality of Life 

We have a joint art-robotics activity throughout this summer that culminates in a series of productions and exhibits at museums across Pittsburgh. The student would conduct research in the use of the Canary environmental sensor for detecting and commenting on the quality of the environment in neighborhoods in Pittsburgh.
Faculty advisor: Illah Nourbakhsh 
Location: Carnegie Mellon 
Project activities: Student will do basic research, curriculum execution, weekly meetings, etc.
Skills needed: No required skills; but preferred is basic HTML and/or programming proficiency
Skills acquired: Use of embedded electronics in community health programs

^top

9. Virtual Physical Therapy Coach

The project involves a computer system that will watch a user doing physical therapy exercises and provide feedback - both specific (e.g., how to improve the exercise) and general (e.g., offering encouragement). The ultimate aim is to have a system that the user can take home to help them adhere to the physical therapy regimen between visits to the therapist. The idea is to provide assistance when the therapist is not available, not to replace the human therapist. We anticipate an increase in adherence and better performance of the exercises when using the virtual coach system. The coach is a graphical avatar that can speak to the user and portray emotional expressions on its face. The avatar and its control software have already been developed for other projects (notably, the roboceptionist). The virtual therapist project consists of three major parts: 1) developing methods for classifying how well the exercise is being done, based on vision and accelerometers; 2) developing methods for providing appropriate feedback to users, based on their performance and personality/preferences; 3) testing on users.

We are looking for a student to engage in #2 - the part relating to generating feedback. The general idea is for the system to have a variety of feedback strategies (e.g., nurturing vs. "drill sergeant") and to learn automatically when to use each strategy. The virtual coach would try different strategies at different times, and correlate how well the feedback worked in terms of getting the user to improve the exercise and/or continue exercising. Helpful feedback would be used again; unhelpful feedback would be less likely to be chosen in the future. In this way, the virtual coach would learn to adapt its feedback to the personalities of the users. We believe that this form of personalized interaction will be much more effective than generic feedback.
Faculty advisor: Reid Simmons 
Location: Carnegie Mellon 
Project activities:

• Develop software for having an avatar-based coach generate feedback to users based on perceptual indications as to how well the person is performing a particular exercise
• Interact with a physical therapist to understand the range of feedback that is typically used, and to determine when that feedback is appropriate
• Participate in regular team meetings

Skills needed: Student needs solid programming skills, preferably in C++. Student should have some background in human-robot or human-computer interaction. Some knowledge of psychology and design is helpful
Skills acquired: The student will acquire skills relating to assistive robotics.

^top

10. Personalized Smart Walker 

The project involves a walker that has been augmented with sensors that enable it to perceive the state of both the environment and the user. The goal is to encourage elderly to walk more by providing technology that can reduce their fears of getting lost or confused while walking around. The idea is that the smart walker monitors the environment and the user, providing directions, warnings about upcoming obstacles, offering encouragement, and providing advice to improve the person's walking style.
Currently, the walker is equipped with laser sensors that enable it to map and perceive the environment. We want to add sensors to provide feedback as to the state of the user (e.g., galvanomic skin response, walking speed, weight distribution). We are also adding a screen that displays a graphical avatar that can speak to the user and portray emotional expressions on its face. The avatar and its control software have already been developed for other projects (notably, the roboceptionist).
We are looking for a student to integrate the avatar with the existing walker software, to generate both verbal and gestural directions for the user, as well as providing generic feedback (e.g., encouragement to continue walking). It is important that the directions be natural sounding, concise, and easy to follow. The student will test the avatar-based system with users (at this point, most likely other students) to ascertain the ability of the users to follow the directions. Time permitting, the student will begin to integrate additional sensors into the walker and use those modalities to provide additional, personalized, interactions.
Faculty advisor: Reid Simmons 
Graduate student mentor: Rachel Gockley 
Location: Carnegie Mellon 
Project activities: Student will develop software for having an avatar generate natural-sounding directions to users for navigating around buildings. Student will be testing the software on users (most likely other students and team members). Student will participate in periodic team meetings.
Skills needed: Student needs solid programming skills, preferably in C++. Student should have some background in human-robot or human-computer interaction. Some knowledge of psychology and design is helpful.
Skills acquired: The student will acquire skills relating to assistive robotics

^top

11. Assessment of Wheelchair Maintenance and Repairs Issues and Impact on Mobility of Older Adults in Nursing Care Facilities

Specific Aims: To develop and standardized a clinical assessment tool (wheelchair maintenance checklist) that could determine maintenance and repairs issues of manual wheelchairs in nursing care facilities. Secondary Aim to determine impact of wheelchair maintenance and repairs on wheelchair-related mobility in older adults in nursing care facilities.
Methods: Residents (n=125) age 65 and older from four nursing care facilities (Southwestern Veterans Center, Hollidaysburg Veterans Home, Presbyterian Senior Care at Southmont, Presbyterian Senior Care at Willows) will be recruited for the study. Wheelchairs for the first 50 participants will be evaluated independently by two investigators at two time points. The rest of the wheelchairs for residents will be evaluated once by study investigators. Wheelchair dataloggers will be attached for two weeks to determine impact of maintenance/repairs problems on mobility.
Faculty advisor: Rory A. Cooper, PhD 
Graduate student mentor: Amol Karmarkar 
Location: HERL
Project activities: Literature. review, data collection, data management, data analysis, results dissemination
Special requirements: Able to travel to different nursing homes, along with Grad student mentor and other investigators
Skills acquired: Microsoft Excel, Access, Optional: MATLAB, SPSS or SAS

^top

12. Human Brain-Computer Interface with ECoG Recording

The central theme in my group is to restore volitional movement with neuroprosthetic devices, such as prosthetic arms that are controlled by myoelectric signals (EMG) and brain signals, such as ECoG and MEG.
Faculty advisor: Wei Wang 
Graduate student mentor: Gustavo Sudre 
Location: Pitt, PM&R BST 
Project activities: Student will participate various aspects of the projects, including meetings with my team, building some basic hardware for our experiments, developing software programs for our experiments, data collection, and data analysis.
Skills needed:

• Some basic electrical engineering/computer knowledge are required.
• Preferred skills: computer programming, such as C, software tools such as Matlab, LabVIEW.
• Some basic knowledge of human nervous system, mostly the brain.

Special requirements: Should be willing to work with patients with movement disorders. Occasionally, we may also need to work with animal models.
Skills acquired: Hardware/software setup for physiological experiments

^top