Random Thoughts:

Here's a link to my blog. I don't write blog posts frequently, but nevertheless, this is a recount of some of my travels and a medley of random thoughts.

My Travel Map. I love travelling. And, I would like to travel a lot more.

Coding is Fun!

  • Here's a visualization of how my research topics evolved over time :) I used python, pdfminer to convert pdfs to texts, nltk and pyenchant for some post processing, and word_cloud to generate the word clouds.

Before Ph.D. During Ph.D.

  • I like to read PHD Comics. Just for fun, I thought it would be cool to see if some of the trends in the comic strips hold true for me as well. Here's one comic strip which shows the Grad Student Work Output (left). And here's my Work Output (right). I used RescueTime API and urllib to get the data. The graph should automatically update everyday.



  • Here's another fun example from PHD Comics. Its regarding average time spent per email from advisor to grad student and from grad student to advisor. Apparently, the no. of words per email is a clear indicator. Assuming time spent in writing a word is the same in both the cases (which, I agree, is a big assumption ;) ), here's a comparable graph of the average no. of words per email from my Ph. D. advisor to me and the other way round. I used Gmail Python API to get the data. Looks like, my advisor has been more kind to me :)

Other Fun Projects:

  • During my undergrad, we developed a mobile robot which is capable of carrying small loads and reaching a particular destination avoiding all the obstacles in its way. We completed the design, fabrication, and control of this wheeled robot and verified the effectiveness of our algorithm using real-world experiments. We were interested in fabricating the robot using old hardware components and used car windshield wiper motors to drive the wheels of the robot and bicycle chains for transmission. Here is the Project Report.
  • Cricket is a very popular game in India! During my undergrad, as a part of another project, we developed a simulated batting robot that is capable of playing conventional cricketing shots to a ball being bowled. Our main objectives in this project were to develop intelligent control methodologies for a batting robot using neural networks and fuzzy logic such that for each bowl, the robotic batsman can hit the ball to maximize the scores based on the placement of the fielders and the ball trajectory. We studied cricket ball aerodynamics, taking into consideration the Drag and Magnus effects, so that its trajectory can be estimated. We designed a 7-DOF batting robot using a CAD software and performed kinematic analysis of the batting robot. We created a Cricket field visualization and simulated the game using C++. Here is the Project Report.