Dr. Nirmala Vasudevan

In many Indian schools, a high student-teacher ratio makes it an uphill struggle for teachers to assess the knowledge of individual students and deficiencies in the students' understanding. Teachers should have a clear picture on what concepts each student has mastered, and which concepts the teacher needs to review in greater detail. This paper investigates the students' concept knowledge, based on the interaction of the students with an intelligent tutoring system. The Feature-Aware Student knowledge Tracing (FAST) algorithm was used, since the algorithm facilitates the separation of lesson-specific skills from concept knowledge. Data from 2400 first-grade students from 28 schools were used for the analysis. Findings include a moderate fit model and an easy interpretation of the model parameters.

Rainfall induced landslides are a common phenomena in the Western Ghat region of Southern India and result in numerous fatalities and damage to property. In order to collect the most relevant and useful data, at the time it is most needed, a wireless sensor network is being used for landslide monitoring. The advantage it gives to landslide monitoring is that it is an inexpensive and reliable way to communicate rapidly over a long distance and inhospitable terrains, collect data in real-time, and alter the data collection rate remotely to suit current environmental conditions. We have implemented a real time landslide monitoring system over a seven acre active complex landslide site. An array of geological sensors (piezometers, tiltmeters, strain gauges, rain gauges, dielectric moisture sensors, geophones) has already been deployed and the data is being automatically collected and forwarded via the wireless sensor network. The geotechnical data is then transferred over 300 km via a satellite link to a remote monitoring station for further analysis. This will give us a better understanding of landslides in this region and prevent the loss of human life.