based on the particular device, regardless of the order in which theywere detected. The instruments’ data collected by the python programs were stored in a local MySQLdatabase on the Raspberry Pi. This local database on the Raspberry Pi had seven data tables; twofor the acoustic Doppler current profiler, four for the weather transmitter and one for the waterquality Sonde. Some instruments used multiple tables to store data in a way that reflected howdata was retrieved from them. All of the tables in the local database had an index column thatserved as the primary key, a column with time stamps from the system time and columns of datavalues. The data values were stored as floating point numbers to preserve precision.6.0 System
: Scatterplot of Average Test Score vs. Class Participation PointsFigure 9 shows the average quiz score, compared to the total of all other points in the course.Both totals were adjusted to 100 points to simplify comparison. There is a surprisingly highcorrelation between the two scores (R = 0.59, P = 0.002). However, this probably reflects the factthat those students who skipped quizzes also tended to skip classes and skipped turning in someassignments. The fact that three students did very well in the quizzes, but got a “C” in the courseindicates that there is no clear cause and effect. Figure 9: Quiz Score vs. Total Non-Quiz Course ScoreStudents who participated in the questions with instructor feedback between attempts
paper is organized into the following sections: Background: The Need for a MobileRobotics Course, Mobile Robotics Course Goals, Course Innovations, Analysis of StudentFeedback, Reflections, and Conclusion. Page 26.460.2Background: The Need for a Mobile Robotics CourseThe Mobile Robotics course was developed as part of a progression of educational roboticsinitiatives birthed on our campus from 2005 to 2013. A brief overview of these initiatives is firstgiven to provide the motivation and context for the creation of this course and its designelements. Figure 1In 2005, the idea of using robotics to
. Therefore, after segmentation, these features were extracted by the featureextractor. Then, these features were input into the classifier. Basically, the classifier can recognizethese 3 objects with very high accuracy (89.1% for the digital scale, 91.3% for the pump and98.4% for the Xplorer GLX. The relatively low accuracy of the recognition is attributable to theKinect’s inability to cope with reflective surfaces which reduces the scanning accuracy. (a) (b) Figure 9: Step motor (a) photograph of physical step motor; (b) model in GBVL Page