Research

CURRENT RESEARCH

  • Intelligent Tutoring
  • Serious Games
  • Interactive Agents
  • UI / HCI for Rapid Training Content Generation

PUBLISHED RESEARCH

  • Real-time fractal path analysis has the potential to determining degradation in mental capacity due to fatigue, brain injury or disease. In games research it can be used to examine the effects of extended play times. In robotics it can be used to examine how user mental fatigue affects the operation of tele-operated vehicles. One of my current projects using RTFPA is in the field of elder care. Real-time fractal path analysis may be utilized as an early warning monitor for dementia. The Florida Mental Health Institute (FMHI) is currently using my real-time fractal path analysis algorithm and application for tracking dementia patients.
  • SmartHome concept. The VA is interested in pursuing research into “smart home” technologies and prototypes. We were awarded $3,000,000 to implement a smart home within the James A. Haley Veterans’ Hospital Polytrauma Transitional Rehabilitation Program (PTRP) for TBI patients. The goal of the project is to create a system that provides patients with interactive, location based memory aids and prompts designed by the PTRP therapists to improve the rehab outcomes. As part of the prompting system we’ll be using Skinner’s stimulus fading technique to help the patients to relearn specific tasks. Our industry partner is Ubisense, which makes an ultra-wide-band (UWB) real-time location system (RTLS). Over 150 UWB RTLS sensors have been installed in the PTRP. Custom RTLS tags are being manufactured into stylish wristwatches for the patients and staff. 65 LCD touch screen systems will be installed in the PTRP shortly. Finally the software that drives the thing is in development by Ubisense. I’ll post some video as soon as it’s deployed for beta testing.
  • The research prior to my dissertation resulted in several publications in the areas of interface design, artificial intelligence, and sensor evaluation. In [Barnes2006] I developed an interface framework and the script manager used to control and coordinate semi-autonomous actions of robots. In [Craighead2006a] I was the project lead for evaluating Canesta’s time-of-flight range sensors as robot-mounted units for confined space searches inside a building collapse exercise in Kansas City, KS. As part of this project I also developed the noise reduction algorithm documented in [Craighead2006a]. In [Valavanis2007 and Ernst2006] I developed the networking interface between MATLAB and the X-Plane simulator as well as the X-Plane UAV models. This was used to test Fuzzy and PID UAV controllers before running the conversion algorithm, which converted a controller created in MATLAB and SIMULINK to microcontroller assembly code. In [Craighead2007a and Craighead2007b] I developed a subjective rating scale for simulators and game engines used for robotics research. This scale, a modified version of the FAA’s simulator rating scale, assigned a letter grade to a simulator or engine based on its capabilities. The simple letter grade ranking will allow robotics researchers to easily identify a simulator or engine that will meet their needs for cost, ease of use and simulation fidelity. In [Craighead2006b] I developed a fuzzy-logic based, software-only system for assessing a game players mood in real time. The system used a combination of rate of input and scoring over time in UnrealTournament 2004 to influence two fuzzy emotions, fear and joy. The combination of these emotions produced a range of moods from frustrated to gloating.

PRIOR RESEARCH

  • I led an effort between Silent Partner Technologies and the electrical engineering department at the University of South Florida to develop some novel compact quadrifilar and 3D antennas for RFID readers and tags. This was a great crash course in antenna theory and design for me. It turns out making an antenna that is both compact, efficient, and reasonably immune to environmental changes is incredibly difficult.
  • I’ve been developing a new, wireless force sensing glove for use by the Prevention and Management of Disruptive Behaviors (PMDB) trainers at the VA. They’ve been looking for something that gives students feedback on grip and exertion force as the goal of PMDB is to mange disruptive patients without injuring them. Unfortunately some students are a bit aggressive and end up injuring the trainers, thus the glove as a first step towards a real-time “That’s Hard Enough” feedback system. For this project I’ve really gained some good experience in embedded electronics and PCB design. As a side project on this I created a SRAM expansion module for the mbed microcontroller.
  • Current commercial tools for measuring gait and balance are large, expensive, and difficult to use resulting in minimal use outside of academic settings. My main research focus for the last few months has been on compiling a feature list for a small, portable, easy-to-use tool set for clinicians. The requirements list for this tool set has been gathered by conducting interviews with our in-house physical therapists and through testing a series of prototypes, again with our in-house PTs. This work has gained interest from one of our commercial partners and I am pursuing that as a means to secure future funding. This project has the potential to significantly increase the use of measurement technology by PTs in clinical settings.
  • I’ve been awarded $20,000 to order some equipment to develop a prototype VR rehab system. My goal is to emulate the capabilities of much more expensive systems such as the CAREN and demonstrate that these commercial systems are overkill (in terms of size and money) for use in day to day rehabilitation clinics.

DISSERTATION RESEARCH

  • My dissertation research is focused on the use of video games for training robot operators. This topic was inspired by my fieldwork with first responders who often have little to no exposure to robots prior to using them. There is a lack of research investigating how to train mobile robot operators. Typically training involves extensive stick time operating an actual robot. First responders often do not have access to robots for this type of training. Developing an open source simulation game provides a potential solution to this problem. I developed the Search and Rescue Game Environment (SARGE) [Craighead2008a, Craighead2008b and Craighead2008c] to provide the research community with a robot simulator and to provide first responders with a high fidelity training tool based on an up-to-date game engine. Using SARGE, my dissertation attempts to quantify the utility of using games for training. Specifically, I am investigating how training in distributed games, where players are not physically collocated, affects team performance. This is similar to the operation of robots used in practice for urban search and rescue. There is an increasing demand by many managerial personnel (who are often not in the field) to have real-time video and sensor data from field assets (robots). This either requires constant communication between the interested parties and the robot operator or a method of sharing control of the robot. Distributed games are a good proxy for distributed robot operation and training in these games will provide robot operators with much needed practice. In addition, I am developing a real-time fractal path tortuosity tool to determine player skill level. Previous research has been limited to utilizing this measure as an off-line tool. Fractal path tortuosity quantifies the directedness of a path. Skilled robot operators are likely to have a more directed path and therefore lower tortuosity.