Table of Contents
Yu Hang He Tesla 2019 Journal
Author: Yu Hang He, Email: hey6@unlv.nevada.edu
Date Last Modified: 10/17/2019
Week 10
Project Status
This week, I focused on reviewing the results from the scan last week in Buffalo and pinpointing the next set of payloads that we can purchase and mount on Spot. Early in the week, I received the scan data from Buffalo. I imported the point cloud data into AutoDesk Recap. After some manual registration process, I was able to build the 3D point cloud mesh within Recap. I do not have much experience with point cloud mesh data, but there some points that immediately stand out to me. First, the impact of roll-cage on the scan quality. It is apparent that during each scan, the roll-cage blocked a section of point cloud data. Imagine an object was placed very close in the light of a candle. The magnified shadow behind is the section of point cloud data that became absent. The obvious solution is to reduce the size of roll-cage. However, that increases the design and manufacturing difficulty.
Another issue is the lowered elevation from mounting the scanner on Spot vs. a tripod. The laser scan is typically conducted at eye level to see past low objects. Using the scanners while on Spot's back, a major portion of scan was blocked out by objects on the ground level that a scanner on a tripod would be able to see past. However, there is no easy solution to mount any linear actuator on Spot's back to raise the scanner without adding significant weight and stability concerns. Unfortunately, we do not have access to Spot's autonomous navigation capability currently.
This week, we decided on the next set of payloads to pursue. We settled on purchasing a Ricoh Theta V 360 Camera, a Intel RealSense D435 Camera, and a pico projector. The main focus would be placed on the 360 Camera since Matt is interested in a company called HoloBuilder that uses 360 camera for construction monitoring. I believe it is a simple payload that can be integrated quickly with Spot.
Project Agenda
In addition to the settle payloads, I need to look into building a pan-tilt unit and integrate the system with the Spot's electrical system. The pan-tilt unit would need to be able to be controlled wirelessly. Once the payloads arrived at the factory, I need to prototype mechanical mounts to attach them to Spot. I am hoping to design a modular system that can be easily detached and reattached.