Author: Yu Hang He, Email: hey6@unlv.nevada.edu
Date Last Modified: 09/23/2019

This week, I met Shiv Mysore, a manufacturing equipment engineer at Tesla. His team is responsible for building a Automated Guided Vehicle for Tesla's Gigafactory in house. Tesla employs AGV during many processes of manufacturing, delivering materials from one point to another. However, these AGVs are slow, prone to failure and very expensive. Tesla wants to build their own AGVs that is better and cheaper than the ones on the market. During the training session last week, we demonstrated Spot for a hour or so inside their AGV area and communicated with their engineers. They were fascinated by Spot. In their own development of AGV, they are hoping to implement automated navigation. They hope to collaborate with our team in the future.

As I mentioned before, I have never worked extensively with GUI or computer vision before this point. It was a challenging and educational experience to implement them on Spot. I find the sense of accomplishment after I was able to successfully implement them a very rewarding experience. The feeling of figuring something out after puzzling over it for a long time is enough to motivate me.

This week, I successfully finished the GUI for the Spot. The GUI is built using Qt, the same platform that was used to built Hubo's interface. The GUI have some basic commands to connect to Spot, start its power, motion commands, and E-stop. In addition to passing basic commands, I was able to stream and monitor various data from Spot. I was able to stream and record the position, velocity, and load of each joints. Finally, I was able to stream camera feed from the 5 cameras that Spot used for its computer vision algorithm. I was to integrate Opencv's face detection algorithm with Spot's camera data and detect people's faces.

Next week, the Faro laser scanner will finally return to Gigafactory 1 for 1 week (It is used in the other factories for scanning as well). I plan to use Faro's API to remotely control the scanner using WLAN, and finally build the laser scanner mount and roll-cage.