HR: 12:05h
AN: ED52A-08    [Abstracts]
TI: EarthTutor: An Interactive Intelligent Tutoring System for Remote Sensing
AU: * Bell, A M
EM: bell@stottlerhenke.com
AF: Stottler Henke Associates, Inc., 951 Mariner's Island Blvd Suite 360, San Mateo, CA 94404 United States
AU: Parton, K
EM: parton@stottlerhenke.com
AF: Stottler Henke Associates, Inc., 951 Mariner's Island Blvd Suite 360, San Mateo, CA 94404 United States
AU: Smith, E
EM: exsmith@odu.edu
AF: Department of Ocean, Earth and Atmospheric Sciences Old Dominion University, 768 W52nd Street, Crittenton Hall, Room 124 Old Dominion University, Norfolk, VA 23529 United States
AB: Earth science classes in colleges and high schools use a variety of satellite image processing software to teach earth science and remote sensing principles. However, current tutorials for image processing software are often paper-based or lecture-based and do not take advantage of the full potential of the computer context to teach, immerse, and stimulate students. We present EarthTutor, an adaptive, interactive Intelligent Tutoring System (ITS) being built for NASA (National Aeronautics and Space Administration) that is integrated directly with an image processing application. The system aims to foster the use of satellite imagery in classrooms and encourage inquiry-based, hands-on earth science scientific study by providing students with an engaging imagery analysis learning environment. EarthTutor's software is available as a plug-in to ImageJ, a free image processing system developed by the NIH (National Institute of Health). Since it is written in Java, it can be run on almost any platform and also as an applet from the Web. Labs developed for EarthTutor combine lesson content (such as HTML web pages) with interactive activities and questions. In each lab the student learns to measure, calibrate, color, slice, plot and otherwise process and analyze earth science imagery. During the activities, EarthTutor monitors students closely as they work, which allows it to provide immediate feedback that is customized to a particular student's needs. As the student moves through the labs, EarthTutor assesses the student, and tailors the presentation of the content to a student's demonstrated skill level. EarthTutor's adaptive approach is based on emerging Artificial Intelligence (AI) research. Bayesian networks are employed to model a student's proficiency with different earth science and image processing concepts. Agent behaviors are used to track the student's progress through activities and provide guidance when a student encounters difficulty. Through individual feedback and adaptive instruction, EarthTutor aims to offer the benefits of a one-on-one human instructor in a cost-effective, easy-to-use application. We are currently working with remote sensing experts to develop EarthTutor labs for diverse earth science subjects such as global vegetation, stratospheric ozone, oceanography, polar sea ice and natural hazards. These labs will be packaged with the first public release of EarthTutor in December 2005. Custom labs can be designed with the EarthTutor authoring tool. The tool is basic enough to allow teachers to construct tutorials to fit their classroom's curriculum and locale, but also powerful enough to allow advanced users to create highly-interactive labs. Preliminary results from an ongoing pilot study demonstrate that the EarthTutor system is effective and enjoyable teaching tool, relative to traditional satellite imagery teaching methods.
UR: http://www.earthtutor.com
DE: 0805 Elementary and secondary education
DE: 0820 Curriculum and laboratory design
DE: 0825 Teaching methods
DE: 0845 Instructional tools
DE: 0855 Diversity
SC: Education and Human Resources [ED]
MN: Fall Meeting 2005