Cluster 11: Introduction to Autonomous Vehicles

Week 3

Friday - Miramar Lake

On this day, we graciously took a trip to Miramar Lake, so both our cluster and Cluster 7 (Hacking for Oceans) could collect data. During our time at the lake, everyone in our cluster had been diligently working on their team's car so that we could record GPS data. However, many, if not all of us had struggled with errors, lacking dependencies, Wi-Fi issues, or simply hardware issues.

In the end, our goal was to drive our cars around the parking lot to collect data, and then have our cars autonomously drive following the recorded path. But due to our struggles, many of us decided to cool off from the blazing hot sun, and visit the lake. While there may have been teensy incidents (refer to video), everyone had a great time!

Our CAs working hard

A strong example of the struggles we faced

IMG_9108.mov

A teeny accident... (he is perfectly safe and well)

A happy cluster photo after a tiring, but fun day at the zoo!

Saturday - San Diego Zoo

Our second field trip was to the San Diego Zoo. We all speed-walked around the zoo to see all of the animals during our limited time, and some of us even had the chance to ride the Skyfair Aerial Tram. Overall, our time at the zoo was a nice break from the brain-wrenching struggles we faced at the lake.

Monday - Beginning the Final Project

In the morning, Professor Silberman went over how exactly our autonomous vehicles are working, giving us insight into how the programs we use help to make our vehicle autonomous. Afterwards, we worked on our DonkeySimulator models and even got to see some of our models racing on the screen. We also continued working on our physical cars, making sure that all of the programs are properly set up, which involved lots and lots of troubleshooting with the help of our CAs.

Professor Silberman going over the remaining schedule for our cluster

CA Manu soldering a VESC

donkeycarsim.mov

Two of our teams racing their models on DonkeyCarSimulator!

Tuesday - Seeing the World

On Tuesday, we set up the configuring of the OAK-D-Lite camera and continued troubleshooting everything. We explored different files for the camera, such as the spatial detector, edge detector, and more!

Wednesday - conda activate donkey

On Wednesday, we all took a trip to the Mechanical and Aerospace Engineering (MAE) building, where there is a track for us to train our vehicles. Similar to the DonkeySimulator, where we had driven our car while it had been taking visual data, we finally moved on to applying the same concept to our real cars.

Using our mounted OAK-D Lites, we drove the car around the track while the camera would periodically take visual records, in which we would then train it to create a model. Unfortunately, we ran out of time to test our models as we await for Monday to come for us to finally see if our car safely completes an autonomous lap around the track.