Pathfinder


Pathfinder is an interactive table, featuring an application with the goal of de-mystifying the AI blackbox.

Find Out More

Project background


How does Artificial Intelligence work? There are many different ways of creating AI of different complexity and most people interact with AI daily, but few understand how it actually works. Our goal is to show how a basic pathfinding AI used in map applications 'think'.

We have created our own touch-table with a projector, some mirrors, and a mobile phone. We have also created an application for this table that finds paths from random locations in Gothenburg to Universeum, and the AI is designed to show how it finds a path within the traffic network. You can also block roads yourself, and see how the AI finds new paths around the obstacle.

How does it work?

Technologies


OpenCV 3

An android phone is used as camera, by using OpenCV.

Projector

Projecting image ontop of frosted glass pane. The image reflects on mirrors twice to enlarge the picture.

Artificial intelligence

A pathfinding algorithm (A*) is used in the application.

Software

Combining the different softwares and the hardware.

Demo



In this demo, you can see the idle screen that is shown when no one is interacting with the table, to invoke interest from bypassing visitors. It is a randomized maze, and an AI trying to find the shortest path between two points with randomized locations. By placing your hand in the black, empty box in the middle of the maze, you wake up the program.

A map of central Gothenburg is shown, and the user is prompted to place their hands on the screen to block off roads. After a delay, the screen turns black and a photo is taken of the screen from beneath the table. This photo is manipulated and combined with an image of the roads, to create an image of allowed space for the AI to search on.

Then, the pathfinding algorithm is visualized, showing it 'walking' along the roads, and trying to find paths around obstacles.

The team


Our team consists of three members, studying on the master's programme Interaction Design and Technologies at Chalmers University of Technology.



Jenny Orell

orell student.chalmers.se

Hanna Römer

hannaro student.chalmers.se

Johan Swanberg

johansw student.chalmers.se