A Novel Complete-Surface-Finding Algorithm for Online Surface Scanning with Limited View Sensors

Area of Science:

• Robotics and Automation
• Non-Destructive Testing (NDT)
• Materials Science

Background:

• Robot-assisted Non-Destructive Testing (NDT) offers speed and access to hazardous areas.
• Current methods struggle with unknown or deforming surfaces, lacking accurate digital models for robotic path planning.
• Existing solutions require pre-scans with camera systems to generate CAD models, limiting flexibility.

Purpose of the Study:

• To develop a novel algorithm for robot-assisted ultrasonic testing of unknown surfaces in a single pass.
• To enable autonomous surface profiling using small-aperture sensors on surfaces of any scale.
• To overcome limitations of digital model availability and accuracy in robotic path planning for NDT.

Main Methods:

• Development of a novel algorithm for autonomous robotic path planning and surface scanning.
• Utilizing local sensor information to map surfaces of unknown geometry.
• Implementing single-pass scanning for robot-assisted ultrasonic testing.

Main Results:

• The algorithm successfully mapped complex and curved surfaces in simulation using only local data.
• Practical demonstrations confirmed full mapping of curved and non-simple surfaces with no operator intervention.
• The method generated traditional raster patterns for surfaces with single-direction curvature.

Conclusions:

• The developed algorithm enables autonomous, single-pass robotic ultrasonic testing of unknown surfaces, regardless of scale.
• This approach is applicable to diverse scenarios, including drone/crawler inspections and confined spaces.
• Future work will integrate low-accuracy environmental sensing to address collision prevention and expand practical applications.