It was a long time (almost three weeks!) since we made a space project. Our last project was on “TRAPPIST-1: Lights and Music” which was displayed at the Ontario Science Centre, Canada for 3 days over the Family Day weekend (17-19 February 2018).
This time we opted for something in our solar system, namely the Asteroid belt located roughly between the orbits of the planets Mars and Jupiter. There are millions of asteroids in this belt and some of them make a close approach to the Earth.
As always, we turn to NASA for the open-source data. The Center for Near Earth Object Studies (CNEOS) maintains the Sentry database. It is a highly automated collision monitoring system that continually scans the most current asteroid catalog for possibilities of future impact with Earth over the next 100 years. CNEOS also calculates the motion of all Near Earth Objects (NEOs) forward to 2200 A.D. and backward to 1900 A.D., and makes them available as NEO Earth Close Approaches database.
We decided to create a project which will use Artificial Intelligence / Neural Networks to predict the “Palermo Technical Impact Hazard Scale” for the collision of an asteroid with Earth. Our system would then convert the values of this scale into musical notes which would then be used to blink lights at the appropriate frequency on a custom designed hardware for displaying Asteroid data.
An ambitious goal to be completed in 40 hours (over the weekend). Very challenging but we wanted to do it not because it was easy but it was hard!
How our team members Artash (Grade 6) and Arushi (Grade 3) achieved it in 10 steps?
All our projects start with lots of planning, sketching and brainstorming. Once we finalise our ideas and plans, we divide the tasks so that each one of us has main responsibility on one task while giving support on other ones.
Step 1: Artash downloaded dataset from the Sentry and the NEO Earth Close Approaches database as csv file.
Step 2: He modified a Neural Network code written in Python by Milo-Spencer-Harper to read CSV file and increased the number of parameters it could model.
Step 3: We identified three parameters on which we would use Artificial Intelligence algorithm to predict Risk Index. These were “Proximity of Asteroid to Earth in Lunar Distance”, “Diameter of Asteroid”, and the “Velocity of the Asteroid”.
We added one more parameter outside the algorithm, namely Asteroid Type: whether it is a C, S or M-Type Asteroids which are rich in Carbon, Silica and Metals respectively.
Step 4: Artash used the Sentry database for training and testing his algorithm. The algorithm was then deployed on the NEO Earth Close Approaches database to come with the Asteroid Collision Hazard Scale values.
Step 5: Arushi loves piano. She wrote a Python programme which would convert the frequency data into specific musical notes.
Step 6: She took the Hazard Scale values from Artash, multiplied them with minus hundred to bring them into audible frequency range. She then used her programme to come up with musical notes corresponding to the values.
Step 7: Vikas and Arushi built a physical display over an old gramophone record (appropriately titled the Golden Record) using LED light strips, relays and Arduino. Arushi wired the relays and LED lights.
Step 8: Arushi modified the Arduino code so that it would be able to receive music data from Reaper (a music synthesis software) as midi files and control the relays to blink the LED strips.
Step 9: Artash embedded the musical notes obtained from Arushi’s Python programme as musical files to be played over different channels in Reaper.
Step 10: Artash and Arushi configured the Loop Midi and Hairless MidiSerial to read data from Reaper and port it to Arduino. They then tested it for a range of Asteroids data.
The project was ready within 39 hours and it worked great!
We used the remaining one hour to document our learnings and create a short presentation about it.
We were able to present our present at the Mission Hack in Toronto and gave several demonstrations of our project (as well as a presentation) to many people who were curious about our invention.