An Android app was developed, which gets the room temperature and ambience using sensors connected to a Raspberry Pi. The Pi sends updates to the app for every 'x' degrees change in temperature, as specified by the user. The Pi automatically closes, opens, or keeps the blind half open, depending upon the Fuzzy rules provided. The same can be done by sending commands to the Pi to its Twitter account using tweets. In addition to this, the app can turn on/off the lights in the room, and also provide manual control for the blind.
This project involves Classifiers developed in Java, which can learn and detect whether a previously unseen network traffic is anomalous in nature or not. The output of the program (Anomalous or Normal) is based upon the output of a bunch of decision trees, and a Naive Bayes Classifier. The mode of the results is chosen as the final output. The Kyoto 2006+ dataset was used for training the classifiers obtained from http://www.takakura.com/Kyoto_data/data_with_IP/
This project was developed in C++, and involved usage of concepts like inheritance and template classes to optimally solve the water jug problem, Sam Lloyd's puzzle and correcting time on a clock.
This project was developed in MATLAB, and involved recognizing faces on the basis of Eigen vectors of the image matrix.
This project was developed in Java, and involved using concepts of distributed systems, multi-threading, synchronization and TCP & UDP communication. It involves a grid of size N x N and an object in one of the grids, which needs to be surrounded by an 'M' number of robots, in minimum number of moves, each co-ordinating with each other to avoid collision and optimality.
This project was developed in Java, and used a genetic algorithm which was able to find more than 8 character matches in the deciphering key after running the program for more than 500 generations. A better result was obtained for 0.5% mutation or by sometimes disabling mutation altogether.