GNSS-Based Narrow-Angle UV Camera Targeting: Case Study of a Low-Cost MAD Robot.

Area of Science:

• Robotics and Automation
• Electrical Engineering
• Geospatial Navigation

Background:

• Precise camera targeting is crucial for Multi-Spectral Automatic Diagnostic (MAD) robots in power equipment inspection.
• Existing solutions often rely on expensive technologies like LiDAR SLAM and Ultra-Wideband (UWB).

Purpose of the Study:

• To propose and validate a novel, low-cost camera targeting procedure for MAD robots.
• To analyze Global Navigation Satellite System (GNSS) reception conditions during power equipment inspections.

Main Methods:

• Utilized a low-cost MAD robot with open-source ArduRover firmware and two Ublox F9P GNSS receivers.
• Implemented a new targeting procedure without modifying existing ArduRover firmware.
• Conducted experimental analysis on overhead power transmission lines and open switchgear.

Main Results:

• Achieved camera targeting with an error within 0.5 degrees for an 8x4 degree field of view.
• The proposed GNSS-only method outperformed more expensive LiDAR SLAM and UWB solutions.
• Provided a quantitative analysis of GNSS performance, including satellite reception, dilution of precision, and positioning accuracy.

Conclusions:

• A simple and effective GNSS-only camera targeting solution is feasible for MAD robots in power equipment inspection.
• The developed procedure offers a cost-effective alternative to complex sensor systems.
• This study presents the first quantitative assessment of GNSS performance in real-world power plant environments for robotic inspection.