All content is written, photographed, and edited by our amazing COSMOS students!
Monday 7/21
Today was an exciting day! Prof. Silberman asked us for updates on our cars with most teams being ready to take data down at the track for their deep learning model of the car. After hearing that we needed a bit more time to work on making our cars ready, Prof. Silberman gave us the whole day to work on the cars. This allowed most teams, at the end of the day, to be able to train their cars on the track in the MAE building in Warren College. And so, it was very crowded there.
Tuesday 7/22
We started off our day with the Final Presentation overview by Mr. Mike, in which he explained how our presentation is going to be like. Then the rest of the day, we had more time to collect data and train our model to do 3 autonomous laps. Not only was the track crowded but the sun played a major role in our deep learning models. Since the models were trained on certain times of the day and if those same conditions were not met, the model would deviate from the track, leading to many frustrations in all teams. After lunch, we all had a group photo and continued working on making a better model
All teams + the staff who guided us through the process of making an autonomous vehicle.
(From left to right) CA Daniel, Professor Silberman, Mr. Mike, CA Madison, CA Jingli
Wednesday 7/23
Today, we had more time working on our final project and RC car for the physical race on Friday! We also worked on getting the deep learning model working on the track, which we needed to complete 3 autonomous laps with. Team 4 figured out that the current version of our donkey car library cannot be used for OpenCV because dependencies that cannot be accessed. So, most teams worked on improving their models and final projects.
Thursday 7/24
Today was another workday for everyone. Professor Silberman and the other CAs came out to help us with our cars. Of course, there were still issues with the cars that kept needing to get fixed. After fixes and training our models, most teams went out to the track to test and gather more data for their deep learning models.