Robotic Vision

Robotic fabrication workflows incorporating vision-based feedback loops present a unique opportunity to begin to facilitate change within the architecture and construction industry.  However, the flexibility and autonomy of current robotic construction systems is limited. This restricts the introduction and use of robotic systems as mainstream tools for the construction industry. This project investigates the beginnings of a robotic workflow that combines common architectural tools, robot control systems and vision-based feedback. The intent of this workflow is to develop a system for object detection based on known object geometry.

The outcome of this project resulted in a system for controlling robot motion through the integration of a vision sensor in a robotic vision application. As a way to determine the robustness of the system, the robot was presented with a block/brick with known dimensions. The robotic feedback system was then asked to identify the location of the object in the environment, manipulate the resulting sensor data and execute a set of motion instructions in order to confirm the object's location.