REU Summer Projects 2015
PROJECT: Towards a Portable, Clinic Friendly Gait Analysis System: Utility of the Wii Balance Board
The goal of this project is to develop a software program capable of collecting gait data in a clinical environment. Previous research has utilized the Wii balance board in postural stability and performance (jumping) applications. This project will focus on gait because it is the end goal of many rehabilitation and training protocols. The intern will review literature for existing Wii balance board applications, write a computer program in Visual Basic (or other program) to simultaneously visualize Wii balance board data and video from a webcam, collect pilot data with control participants, and compare to a conventional gait analysis system.
LEVEL: Advanced
REQUIRED BACKGROUND: Computer Science; Engineering (e.g. electrical, biomedical, mechanical)
REQUIRED SKILLS: Programming experience (e.g. Visual Basic, Matlab); Excellent written and oral communication skills
MENTOR: Jon Akins, PhD
PROGRAM: ASPIRE REU
PROJECT: Dual-task Gait Analysis
This project will develop dual-task gait analysis applications for the computer assisted rehabilitation environment (CAREN). Dual-task refers to performing cognitive and motor tasks simultaneously. The motor task will be walking and the cognitive task will include several from the literature. Previous research has demonstrated significant dual-task decrements in gait parameters in several populations. Dual-task gait assessments are important for mobility research since walking in the community is often performed concurrently with cognitive tasks. The intern will develop applications for the CAREN system to collect single and dual-tasks and collect pilot data on control and potentially lower limb amputees.
LEVEL: Advanced
REQUIRED BACKGROUND: Engineering (e.g. biomedical, mechanical, electrical)
REQUIRED SKILLS: Programming experience (e.g. Matlab, LabVIEW); excellent written and oral communication skills
MENTOR: Jon Akins, PhD
PROGRAM: ASPIRE REU
PROJECT: Tendon Mechanobiology And Wheelchair Activities Performed By People With Spinal Cord Injury
We will be measuring interstitial biomarker concentrations in rotator cuffs of wheelchair users and able-bodied individuals to evaluate chemical characteristics of overuse pathology. We will also measure biomarker responses to wheelchair activities to better understand how these activities contribute to overuse pathology. Quantitative ultrasound markers of tendinopathy will be compared with biomarkers to relate physical tendon characteristics (fiber alignment, fluid accumulation, etc.) to chemical characteristics. The intern will collect and analyze wheelchair propulsion biomechanics and transfer skill information; conduct ultrasound image analysis; compile and enter data; and assist with subject testing.
LEVEL: Intermediate
REQUIRED BACKGROUND: Physiology, Biology, Kinesiology, Bioengineering
REQUIRED SKILLS: Preferably exposure to biochemistry, mammalian biology, and/or physiology.
MENTOR: Michael Boninger, MD, PhD and Nathan Hogaboom, MS
PROGRAM: ASPIRE REU
PROJECT: Smart Technology for Online Education
Technology provides a tremendous opportunity to revitalize education. Massive Online Open Courseware (MOOC) has the potential to make world-class educational opportunities widely available. However, with a capacity for hundreds of thousands of students in a single course, teachers and teaching assistants cannot be relied upon to assist every student in need of help. There is an opportunity to build artificial intelligence systems that help to connect students with peers to provide the help they need, as well as assist them in parsing the hundreds of thousands of forums posts that can occur in a MOOC setting. Previous research has demonstrated that students receiving personalized assistance in their education can see improvements as high as two standard deviations compared to a lecture-oriented learning experience. This project attempts to close that gap by facilitating conversations between students, and providing automatic interventions when a student is struggling. The ultimate goal is to combine the ubiquity of online courses with the social benefits of a classroom experience. The technical approach is to use machine learning to learn natural language discourse structures, which encode hierarchical semantic relationships between clauses or sentences. This allows us to better understand the nature of conversions between students, which may allow us to identify students that are struggling as well as students that are able to act as mentors to their peers. The student’s responsibilities will include implementing and evaluating one or more existing discourse parsing algorithms. These algorithms can be implemented in an environment of the student’s choosing. The methods the student implements will be based on work presented at recent machine learning and natural language processing conferences. Although the student will not be expected to design or modify new algorithms, any original ideas put forth by the student are welcomed and encouraged. In addition, the student will be asked to help collect new datasets to facilitate future research. This will entail annotating the structure of student writing and tutoring logs (such as indicating when two sentences share a common subject). The student can then use the algorithms implemented over the summer to obtain results with the newly collected data. These experiments will likely produce publishable material.
LEVEL: Advanced
REQUIRED BACKGROUND: Computer Science, Mathematics, or Statistics.
REQUIRED SKILLS: Students should be familiar with object oriented programming languages, specifically Java or C++. Students should also have a basic understanding of concepts in discrete mathematics and probability theory, such as combinatorial equations and probability density functions. Some familiarity with a technical computing environment such as Matlab is recommended. Knowledge of machine learning and artificial intelligence is not required but welcomed.
MENTOR: Reid Simmons, PhD and Robert Fisher, MS
PROGRAM: QoLT REU
PROJECT: Robot Escort for the Elderly
We are developing a mobile robot that can escort elderly in assisted living facilities, particularly aimed at those with way-finding difficulties. The innovation is that the robot will be operating side-by-side with the people, rather than leading them, as is currently commonly done. We have already done a teleoperated user study that indicated the elderly would accept walking next to a robot; this summer, we intend to continue development of the software that will enable the robot to do this autonomously -- keeping pace and appropriate distance with a person and providing directions at suitable times through both verbal and non-verbal cues.
Intern will help develop software to reliably detect people and estimate their pose and velocity using an on-board Kinect sensor, decide how to position the robot to match pace and maintain an appropriate distance, especially if the presence of obstacles, and plan how to move in socially acceptable ways when turning or going through narrow openings, such as doors. Alternately, the student might investigate how the robot can non-verbally signal directions based on the goal and perceived engagement of the walker.
LEVEL: Basic
REQUIRED BACKGROUND: Computer science or electrical engineering
REQUIRED SKILLS: Knowledge of C++ and Linux is mandatory. Good programming and math skills are essential. Previous experience with mobile robots and the Kinect is a big plus.
MENTOR: Reid Simmons, PhD
PROGRAM: QoLT REU
PROJECT: StrongArm
The Strong Arm is a specialized wheelchair attachment designed to help wheelchair users transfer in and out of their electric powered wheelchair to other surfaces (e.g. chairs and beds) with the assistance of a caregiver. Unlike current technologies used to transfer people such as a floor lift, which is bulky and less easy to transport, the Strong Arm attaches directly to the EPW for transportation and use in the community.
The Strong Arm uses direct interaction to allow a caregiver to lift the person using the wheelchair with literally one finger by directing a handle located at the end of the arm. It moves around the wheelchair on a track to transfer people on either side, and for easy stowing at the back of the wheelchair. Designed for transfers in confined spaces, it runs on the same battery source provided by the wheelchair without drawing much power. The intern will assist with stability testing and analyze study data.
LEVEL: Basic
REQUIRED BACKGROUND: Engineering
REQUIRED SKILLS: Knowledge of Statics, Dynamics, and Statistics preferred
MENTOR: Rory Cooper, PhD, Jessica Burkman, and Garrett Grindle, MS
PROGRAM: ASPIRE REU
PROJECT: Wheelchair Testing
The intern will develop and execute test plans and perform testing in the testing facility.
LEVEL: Intermediate
REQUIRED BACKGROUND: Bioengineering, Engineering, clinical background preferred
REQUIRED SKILLS: Device testing, clinical work, Engineering
MENTOR: Jon Pearlman, PhD and Anand Mhatre
PROGRAM: ASPIRE REU
PROJECT: Development of a Reliable Biomeasure for Accurate Detection of Behavioral Dysregulation in Adults with TBI
This study aims to evaluate the feasibility of a suite of biomeasures to reliably detect self- regulation behavior in the moment via physiological changes when compared to direct behavioral observation during task performance within a laboratory setting. The proposed development project is the initial and critical stage for developing a TBI- specific EMA/ EMI using biomeasures for increasing behavioral self-regulation.
Pilot testing will be done with healthy volunteers (n ≤ 10) within the laboratory to measure the sensitivity and specificity of selected biomeasures for accurate detection of the presence or absence of behaviors commonly associated with dysregulation when compared to baseline physiological data. Testing will be done through having participants complete a series of increasingly challenging problem- solving tasks selected due to their potential to illicit negative responses indicative of dysregulation (i.e. frustration, agitation, and irritability) and result in behaviors associated with task breakdown (i.e. failure to correctly complete task, task interruptions, or refusal to continue task). The selected biomeasures will be continuously recorded via a portable biofeedback device worn throughout all research activities—measuring physiological states during any given moment. To determine the ability for the biomeasures to accurately detect behavioral dysregulation, direct behavioral observation by two trained clinicians will be used for comparison. Ultimately, biomeasures that correlate to a high degree with clinician judgment within the laboratory will provide insight to the ability for these same biomeasures to also detect behavioral dysregulation in the natural environment where self- regulation deficits truly manifest. If anticipated results are achieved with healthy volunteers, future research aims will be carried out with adults with traumatic brain injury who commonly experience behavioral self- regulation difficulties during daily living. The intern will be involved with research protocol development, assistance with data collection, data analysis, manuscript writing, and other dissemination activities.
LEVEL: Intermediate
REQUIRED BACKGROUND: Fields related to computer science, information science, engineering, social sciences, and rehabilitation as they may fit in line with the scope of the project are preferred.
REQUIRED SKILLS: A student with skills in the following areas would be a good fit for this project: Computer programming (preferably r- programming), machine learning and predictive modeling experience, background in biosignal processing, knowledge of traumatic brain injury rehabilitation, prior research experience
MENTOR: Dan Ding, PhD, Michael McCue, PhD, and Ashlee McKeon, MS, CRC
PROGRAM: QoLT REU
PROJECT: Neural representation of objects during Brain-Computer Interface
The student will process neural data recorded while a subject controlled a robotic arm to interact with objects using the signals recorded on electrodes implanted in primary motor cortex. The goal will be to identify reliable patterns in the data that predict the proximity of objects to the hand. Additional analysis will include generation of signal quality metrics. Project activities will include analysis of recorded neural data in Matlab, using a combination of pre-written and self-generated code. May additionally help with creation of objects for the BCI user to interact with to generate new data.
REQUIRED BACKGROUND: Biomedical Engineering, Neuroscience, or Computer Science
REQUIRED SKILLS: Matlab, Intro to Neuroscience
MENTOR: Jennifer Collinger, PhD and John Downey, PhD
PROGRAM: ASPIRE REU
PROJECT: Effects of Wheelchair Seat Position and Footprint Length on Ramp Propulsion Biomechanics
This project is looking at how changing seat position and footprint length on a manual wheelchair will alter a persons biomechanics on ramps. The Clinical Practice Guidelines on ultralight manual wheelchair setup recommend setting up the wheelchair to minimize forces by moving the seat more posterior with respect to the larger, rear wheels and shortening the footprint length (e.g. distance separating the front casters and rear wheels). For most propulsion activities this setup works really well because it allows the user to propel over level surfaces, pop a wheelie, maneuver in areas where space is limited, negotiate curbs and navigate rough terrains more easily. However, the CPG recommendations for UMW setup were based on studies that were conducted on level surfaces. This UMW setup may result in problems when propelling up ramps because it inherently makes the UMW more unstable (e.g. 'tippy'). This study is looking to collect pilot data on the effects of set up to clarify the clinical practice guidelines and support new ultralight manual wheelchair design.
The intern will operate the VICON Nexus software to label data; assist in data analysis; running Matlab programs and performing statistical analysis; assist with literature searching. At the end of the summer a conference paper will be written. There may also be some assistance/shadowing of subject testing occurring during the summer.
LEVEL: Intermediate
REQUIRED BACKGROUND: Engineering preferred, specifically biomedical, mechanical or computer science/engineering. Other possible backgrounds possible but are not limited to: biology, rehabilitation science, kinesiology etc.
REQUIRED SKILLS: Familiarity with Matlab is strongly recommended, as well as with Microsoft word and excel. A basic understanding of anatomy would also be useful.
MENTOR: Alicia Koontz, PhD and Sarah Bass
PROGRAM: ASPIRE REU
PROJECT: Scaffolded Design
SCaD (Scaffolded Design) is a web-based peer-evaluation tool, that allows students in large design classes peer-evaluate each other improving their learning and lowering instruction burden in providing feedback. SCaD has been using a modified version of another web-based peer evaluation tool from Peerceptiv, but we hope to develop a new prototype this summer that implements all the needs/changes required for a design evaluation tool. SCaD will have to facilitate teachers to create and clone classes, create different types of assignment, easily create and manage individuals and groups in class, multiple randomization routines that can match individual/group raters with projects, and most importantly allow commonly used design artifacts (video, cad models etc) be implemented into the webpage. We Anticipate SCaD would require a front-end coded in either HTML/CSS, HTML5, ASP or other related language, back-end running MySQL and related DBMS along with other scripting languages. Project work would involve attending weekly meetings with mentors to both receive instructions and provide updates on progress. Independent work on implementing user-interface suggestions and scripting the webtool. Writing up the results for a peer-refereed journal along with a poster for internship program requirements.
LEVEL: Advanced
REQUIRED BACKGROUND: Computer Science, graphic design
REQUIRED SKILLS: Demonstrated experience using web-development tools (HTML5, HTML-CSS, PhP, Javascript, etc)
Demonstrated experience in database management using MySQL and other server side scripting languages.
Experience with interface design and graphics design would be beneficial
Experience or ability to work independently under deadlines
MENTOR: Jon Pearlman, PhD and Mahi Mandala, MS
PROGRAM: ASPIRE REU
PROJECT: Electric Power Wheelchair Driving Assessment
The number of veterans with disabilities has increased by 25 percent since 2001. Today, over 2.9 million have significant disabilities and represent almost 50% of those treated yearly in the VA Healthcare system. Twenty percent of the over 80,000 wheelchairs prescribed by the VA each year to veterans with disabilities are power wheelchairs. In addition, it is estimated that every year 27,000 veterans desire mobility but fail the initial assessment. These veterans may never achieve independence in mobility if not at least given the opportunity to undergo an individualized training program. Of those who do receive power wheelchairs, it is estimated that over 6,000 have marginal driving skills. Impaired driving skills may lead to inadequate mobility for self-care tasks or reduced community participation or quality of life. Yet, more concerning is that over half of all power wheelchair users have at least one major accident every year, some even resulting in serious injuries or death, and many of these are due to problems controlling the wheelchair in various environments. However, due to lack of standardized assessment tools, developing evidence based training protocols for EPW users to improve driving skills has been a challenge.
To address the limitations of the currently existing tools and quantify the cognitive and sensory functions necessary to operate a wheelchair, the Power Mobility Screening Tool (PMST) in conjunction with its “on the road” counterpart the Power Mobility Clinical Driving Assessment (PMCDA) was developed. The goal of this study is to test the psychometric properties of these assessment tools and quantitatively validate these scores using motion capture technology. Further, we will also study the cognitive load of the EPW user while they are performing different driving tasks to develop training strategies that could be targeted to improve this cognitive load.
LEVEL: Intermediate
REQUIRED BACKGROUND: Biomedical Engineering
REQUIRED SKILLS: Good proficiency in MATLAB and SOLIDWORKS, excellent inter-personal and communication skills (this would be absolutely essential during data collection, since the student would have the opportunity to interact with participants)
MENTORS: Brad Dicianno, PhD and Deepan Kamaraj, MD
PROGRAM: ASPIRE REU
PROJECT: Design of an Assistive Robotic Manipulator Evaluation Tool (ARMET)
The goal of this project is to design and develop an assistive robotic manipulator evaluation tool (ARMET). The intern will design and develop a task board for the ARMET, which will include multiple daily activity tasks for users to complete using ARM. The task board needs to be portable and easy to operate by clinicians. Project activities include designing the layout of the task board using Solidworks; building a prototype in the machine shop; and testing the task board.
LEVEL: Intermediate
REQUIRED BACKGROUND: Mechanical Engineering
REQUIRED SKILLS: Solidworks, basic knowledge of electronics and microcontrollers
MENTOR: Hongwu Wang, PhD
PROGRAM: ASPIRE REU
PROJECT: 3D Printing Assistive Technology
Computers’ flexibility and exponentially increasing power have revolutionized advanced manufacturing. The vast promise of these new fabrication technologies may, in President Obama’s words, `revolutionize the way we make almost everything.’ We propose a set of innovations needed to make this promise a reality: computationally enhanced tools to make physical device construction accessible for ordinary users and new concepts for expertise amplification through virtual service teams, which combine a range of human and automated capabilities. We will explore these concepts in the domain of assistive technology (AT). AT's potential is currently under-realized because the expertise needed to create the right AT is in short supply and the custom nature of AT makes it difficult to deliver inexpensively. Our primary goal is to develop new algorithms, tools, and computer assisted work strategies to catalyze on-line communities of skilled experts to dramatically scale up the ability of online communities to perform work, and apply these advances to making dramatic changes to the real-world task of producing customizable AT devices. To revolutionize the production of AT we must advance the creation of highly customizable devices while increasing coordination and expertise within distributed, diverse communities. The REU intern on this project will develop fundamental abstractions for computationally customizable physical device production with the goal of enabling high-level control and testing of the semantics and properties of physical devices by end users; advance the model of service delivery by introducing mechanisms for ongoing iteration; and follow up which produce a more detailed understanding of AT use abandonment in comparison to more traditional AT provision models.
LEVEL: Intermediate
REQUIRED BACKGROUND: Computer Science, Engineering, Rehabilitation Technology
REQUIRED SKILLS: Programming experience in at least one language required; 3-D modeling experience preferred
MENTORS: Jenn Mankoff, PhD and Scott Hudson
PROGRAM: QoLT REU
PROJECT: Development of the International Society of Wheelchair Professionals (ISWP)
The University of Pittsburgh (Pitt) is proposing to develop the International Society of Wheelchair Professionals (ISWP), a self-sustaining network of international and regional partners dedicated to the professionalization of wheelchair (WC) services around the world. ISWP will be built around a federation of regional and international Affiliate Members which will help ensure ISWP activities are culturally relevant, timely, and focused on the most important WC-related issues. ISWP will initially be led by a group of WC experts at Pitt, with strategic partnerships that have already been established with USAID & WHO and will be the basis for two outcomes. In the near-term, ISWP will be started and an Affiliate network will grow; in the long term, Pitt plans to use ISWP as a starting point for a WHO Collaborating Centre on WCs. The intern will develop target competencies by interviewing clinicians trained in the WHO Wheelchair Skills Basic and Intermediate Training courses; draft the Intermediate Knowledge and Skills assessment; review the assessment for face validity and perform test/ retest reliability measurements
LEVEL: Basic
REQUIRED BACKGROUND: Experience working in a clinical setting; interest in rehabilitation science
REQUIRED SKILLS: MS Excel, time management, critical thinking, strong communication skills
MENTOR: Mary Goldberg, PhD and Yohali Burrola
PROGRAM: ASPIRE REU
PROJECT: ToT Wheelchair Service Provision Evaluation
The Training of Trainers (ToT) Wheelchair Service Provision Evaluation is a project that uses a mixed methods formative and summative assessment to evaluate the impact of a wheelchair service training course on rehabilitation practices of trainers and their trainees. Activities for the project will include the following: revising a semi-structured interview/focus group guide based on data from previous ToT focus groups investigating the effectiveness of the training course; following up with ToT participants and conducting interviews to measure the effectiveness of the ToT course on their current practice; obtaining information on previous ToT participants' trainees and interviewing them to determine the impact of the training course on their current practice.
LEVEL: Intermediate
REQUIRED BACKGROUND: Rehabilitation, Rehabilitation Counseling, Occupational Therapy, Physical Therapy, Biomedical Engineering, Rehabilitation Engineering, Physical Education, Sports Medicine, Athletic Training, Disability Studies
MENTOR: Mary Goldberg, PhD and Alexandra Miles
PROGRAM: ASPIRE REU
PROJECT: Designing and testing an adapter for incremental alignment changes in lower limb prosthetics.
In clinical practice, leg prostheses are aligned by means of a dedicated pyramid adapter system. This allows continuous scale perturbations of joint angles and component translation with respect to each other. While the step-less adjustability is a clear advantage for most applications, it has its limitations in situations where a clearly defined and repeatable alignment change is desired, such as in research interventions. This often limits the efficiency and rigidity of research protocols and is part of the reasons why the evidence base for prosthetic interventions is notably thin. In this REU project, we will work on an adapter design that allows step-wise adjustment of joint angles in prosthetics, starting by analyzing the state of the art, including the development of different design concepts, simulation and optimization of designs on the computer, production of a prototype (or prototypes) and ending with the bench testing (and possibly clinical testing) of such a prototype.
LEVEL: Intermediate
REQUIRED BACKGROUND: (Medical) Engineering
REQUIRED SKILLS: Computer-aided design; finite element analysis; biomechanics
MENTOR: Goeran Fiedler, PhD
PROGRAM: ASPIRE REU
PROJECT: Using Prosthesis-Integrated Sensors to Investigate Kinetics Problems of General Interest
Artificial legs have many disadvantages when compared to the physiological sound limbs they are replacing. One of the very few advantages, however, is the possibility to include data collection equipment directly into the weight bearing structure of the prosthesis. This promises a quality of kinetics data that is very difficult to attain in able-bodied subjects. The aim of this REU project is to utilize a prosthesis-integrated load cell to investigate a problem that is of interest in biomechanics research in general (e.g. not limited to amputee biomechanics). Applicants are invited to contribute to the development of the actual research question, which may be based on personal interest/hobbies. For example, the intra-limb forces experienced during certain sports activities could be investigated. The intern will complete the following activities:
Upon application - define hypothesis; prior to start of program and led by advisor - devise research plan, apply for IRB approval, conduct literature review, prepare data collection, statistical plan, recruit subjects, collect and analyze data, disseminate findings as poster and journal manuscript.
(Realizing an entire such project from IRB application through analysis and dissemination in the available 10-week-span was successful before. However, it may be advantageous to schedule some preparatory work prior to the actual internship as detailed above.)
LEVEL: Advanced
REQUIRED BACKGROUND: Rehabilitation Science, Athletic Training, Kinesiology, Engineering
REQUIRED SKILLS: Physics, Statistics, Research Writing
MENTOR: Goeran Fiedler, PhD
PROGRAM: ASPIRE REU
Support for this program is provided by the National Science Foundation, Grant EEC 1852322.
Education > Undergraduates > ASPIRE > Projects > REU Projects 2015
