This YouTube video shows the amazing progress that the golf cart team was able to make for the rest of the semester. Not only does the video recognition software work, but the team was able to move the tech onboard to the actual cart!
We are back and better than ever this semester. Major progress has been made and wheels are on the ground. After countless hours of grueling work, mashing random computer keys, and hooking up wires to other wires, we are finally able to revel in the carts ability to move within the boundaries of a makeshift track. While it is only able to turn right (for now), this was a major win in our books. This video was only scratched the surface for what is to come.
Other accomplishments we had today were: We implemented the new Jetson on board computer in order to ensure the cart has the most up-to-date hardware. We designed a new track in order to train the cart to turn both right AND left. We are able to control the cart through a virtual steering wheel remotely from online.
With the help from our fearless leader, Addisu, we were able to create a map of the Wond’ry Second Floor using LIDAR (as seen in the picture above). Additionally, we worked on debugging code for our IMU.
One issue that we have been struggling with is the fact that out ZED camera was not being recognized by our Jetson. Through intense detective work, we were able to discover the root of the problem. We suspect that the camera’s firmware had been mysteriously upgraded while being used on a separate computer (shady). In order to fix this, we had to downgrade the camera firmware, allowing us to install and finally use the ZED SDK.
Today we figured out how to power the LIDAR sensor without connecting to the Jetson pins. We are now successfully powering the LIDAR from the battery pack. This made us very happy. If you want to know how happy, just look at the above picture of Mr. Gustave beaming from ear to ear.
Today was mostly spend testing the LIDAR manually to see how sensitive it is to its surroundings (VIDEO OF LIDAR TESTING BELOW). We wanted to confirm that the position of the sensor is in the ideal location on the model cart. We also discussed the trajectory of the group’s progress and decided that we would reach out to more faculty for advice. The rest of the session was spent celebrating Taco Tuesdays a day late while eating delicious food from Bar
Today was a productive day for the team. We set out to: troubleshoot and test the LIDAR, play with ZED Camera, and reassemble the car. Though we did not get to the last step, we were able to successfully clear up the headaches caused by Bridge. We were able top recompile kernel with “CONFIG_BRIDGE” enabled. The LIDAR data, is now officially being received, allowing us to hopefully ramp up the detection function, and move forward in writing scripts.
In the later portion of the meeting, we were able to play around with the ZED camera. We projected the image from the camera onto a big screen, and were able to see the visualized data in real time. The images shown in the above slideshow are the images and depth data captured by the ZED camera.
With an intimate group of just five people today, we looked to work out some kinks in the model cart. We tested the switches and we finally got them to WORK. We spent the rest of the meeting troubleshooting the LIDAR sensor but we figured out that, at least for now, we cannot connect to the Jetson computer while the LIDAR is running.