2017 Hackathon

Projects

Mind control video games

Electroencephalography (EEG) involves the measurement of electrical signals in your brain. Two commercial EEG headsets were hacked to allow custom input to an original video game, with the result that the team were flying virtual paper planes by mind power alone. For good measure, they also so set it up so they could browse their files with their thoughts.

Arduino Pong

Using an Arduino, the classic video game Pong was remade. The control was provided by variable resistors and the display by an LED array. This creation was complete with retro sound effects and could be simultaneously linked up to and displayed on a computer, allowing the audience to watch the judges to battle it out during the team's final presentation.

Muscle control robotics

Electromyography (EMG) is similar to EEG except it measures activity in your muscles. This team hooked up an EMG armband to a robotic arm allowing them to control it with their movements from afar. They also added speech function and got it to introduce them during their presentation.

Tweeting greenhouse

Temperature and humidity sensors were used with an Arduino to create a device that tweets atmospheric measurements -- perfect for monitoring your greenhouse! They also used machine learning with this data to predict the weather, and built an outside extension with an ultrasonic distance sensor to detect and scare away rabbits.

Political neural networks

By collecting hundreds of tweets by Labour and Tory MPs, this team used Google TensorFlow to train an artificial neural network to identify the political sympathies behind twitter accounts. They tested it out live on suggestions from the audience: apparently Jeremy Corbyn is 75% Labour and the tweeting greenhouse 51% Tory!

Instant stock prices

This team built a device for your wrist using an Arduino that accessed live stock prices from the internet, displaying them on a screen. It would tell you if the stock was up or down for the day and an alarm would go off if it reached some desired (or feared!) price.

Evolving camouflage

Genetic algorithms use random mutations and a fitness function to iterate towards an optimal solution. Like animals evolving to hide in their environment, this team used a genetic algorithm to disguise shapes against arbitrary backgrounds.