Working on large devices or surfaces in manufacturing presents a challenge. It is inefficient and risky for workers to climb high, and mounting robotic arms around these large entities is cumbersome, costly, and inflexible. This project introduces the use of Unmanned Aerial Manipulators (UAMs), prized for their expansive operational range and easy accessibility, as an efficient, flexible, and economic solution for large surface inspection such as defect localization and detection. As a replacement for the stationary robotic arm array, the mobile (potentially tethered) manipulator platform is more flexible and efficient for perception and manipulation tasks on large devices.

Another challenge in large device inspection is the tradeoff between detection accuracy and efficiency. The defects are commonly on a small scale, requiring detailed inspection. However, given the vastness of the surfaces, conducting meticulous inspections over the entire area can be slow. Ensuring thoroughness comes at the expense of efficiency. This project works to improve the detection efficiency without compromising detection accuracy by utilizing multi-modality: both vision and tactile sensing, where vision focuses on the large scale and tactile sensing focuses on the small scale.

Principal Investigator
Sebastian Scherer
Additional Investigator
Guanya Shi
Research Areas
Robotics
The graphid has three parts. The top part shows how the texture and tactile image differ. The second part shows the device against a wall with different textures. The third part shows the gelsight and the tactile roller.