We made a small video explaining what the project is and some screenshots of what the final outcome of the project will look like. The video also gives details about our approach and the use of crowd sourcing in building our custom data set.
More details about the current work going on and the challenges we faced will be posted soon!
Please find the concept video at the below link.
Massive Open Online Courses (MOOCs) (such as EdX and Coursera) have initiated a revolution in higher education by providing opportunities for interested students to learn from the comfort of their individual locations at their desired pace. However, an important and highly successful aspect of traditional classroom education, which is modulating content delivery based on understanding real-time student feedback, is conspicuously missing in such e-learning environments.While existing e-learning environments provide a basic technology framework, the personalization of such environments with human-in-the-loop feedback is still missing. This project, e-DRIShTI, aims to bridge this gap by developing a system for automatic recognition of students’ engagement levels during e-learning sessions, using advanced computer vision and machine learning methods.
Considering the ubiquitous presence of cameras in consumer devices such as tablets and laptops, it is possible now to develop a system that can detect and recognize student engagement levels from the face images captured by the camera during the e-learning sessions. Such a system has several applications towards smarter, personalized e-learning environments: (i) it can allow for content to be modified based on a student’s engagement level; (ii) it can be used as feedback for curriculum development; (iii) it can potentially be used as a diagnostic for early detection of learning disabilities; and (iv) it can be stored and archived as part of students’ learning portfolios which can later be mined or analyzed for providing feedback to the students. In general, this project can lead to a more concerted, larger effort on automatically obtaining real-time student feedback – including other states such as confusion, boredom, excitement and interest – towards an effective personalized e-learning experience.
We were hoping this post would include details on our research trip to Belize during January. Unfortunately, a team member was in an accident, which caused us to postpone our trip until this spring. However, we do have some information to share on preparations to date and future plans.
Plenty of hours were spent preparing our equipment for the trip. Much of time was spent on completing the assembly of our “Game of Drones” build, affectionately known as Motley (see figure below).
I designed the dome/lid (white, seen in the middle of Motley) and had it printed on a 3D printer in our Center for Design and Fabrication. An idea from our friends at UAV-America suggests that we add a bit of plasti-dip to the dome edge and inner rim of the frame to secure the fitting, since the lid does not quite snap into place. All of the components were oriented and re-seated within this hardier frame. The newly-combined GPS receiver and magnetometer can now be mounted externally on the dome (note: the GPS/magnetometer are not shown in the above image). We are able to fit two LiPo (lithium polymer) batteries inside the cavity and we await warmer weather – or a trip to the indoor gymnasium – to test Motley’s flight characteristics.
We also invested additional time building our Aquacopter UAV (see picture below). The Aquacopter is a durable frame that, as the name implies, is specially built for water environments. The frame and components arrived in early December, not leaving much time to build or modify. With help from UAV-America and additional hours in our lab, we were to get the motors generating lift and the quad up in the air. However, we are still experiencing irregular flight characteristics, primarily with unpredictable and unresponsive yaw motion. The motors may be slightly pitched in their mounting, which would cause these issues.
Future plans include adding shims under the motors, performing test flights to find the cause of the problem (and therefore fix the yaw), and also determine if the Aquacopter can actually float. This would give it a significant edge over Motley in our data collection, because being able to retrieve the quad from a water landing is crucial.
We also spent time reading and understanding FAA regulations governing what types of batteries are allowed on commercial flights. Specifically, we reviewed the Delta Airlines baggage restrictions to understand what size and quantity LiPo batteries we can pack and transport in our carry on bags. As Christine writes:
I’ve looked at the restrictions on carrying on lithium batteries and the only restrictions I can find are that you can only have two batteries that are “large”, or between 8 and 25 grams of lithium and between 100 and 160 watt hours. None of our batteries are anywhere close to those restrictions, so according the guidelines you sent me we can take as many of our batteries as we want to. These can’t be damaged and must be put in their original packaging (or, if we don’t have that, then we just need to insulate the battery terminals from metal or other batteries). Also, we have to make sure that there won’t be a lot of pressure on the batteries at any time. All the batteries must be in carry on.
We are taking extra precautions to print and adhere to these regulations, and plan to adhere a print-out with an inventory of batteries stored in each battery safe bag to the outside of our carry-on luggage. Hopefully, this will save time and confusion as we proceed through the airport security area!
The growth study has begun at USF. Environmental Engineering students are researching the direct effects of fertilizer treatments on the growth of rye grass. Students are utilizing different combinations of struvite, MCP, clinoptilite and chabazite to see which fertilizer and/or combination produces the greatest positive difference from the conventional fertilizers used today. Different methods of production were used to make the some of the fertilizers along with different amounts – a half and full dose are used. It is unknown what the specific trends are going to be until more weeks pass, however, there is promise of a good ending in this beginning.
The students involved in this endeavor under the direction of Dr. Sarina Ergas are Adib Amini, Veronica Aponte, John Pilz, Lindsay Guntner and Andres Garcia. Andres Garcia is a high school student who is utilizing his experiences at a research university like USF to sharpen his scientific abilities with hopes of a future career in science. Just as the symbolic growth of these sprouts show promise for interesting results, the students in this study hope to be a part of this until the end. Please follow us as we continue to provide updates on this growth study here on “Students for A Smarter Planet.”