Autonomous Magnetic Maps Generation for Monitoring using Small and Autonomous Aerial Robots
Abstract
Traditional aerial surveys are performed by a manned operator using a spe- cial aircraft configured with expensive sensors. The cost and human risk of those studies increase with the size of the mapped area. A particular case of aerial surveys is magnetic surveying which records spatial variation in the Earth's magnetic field with a magnetometer. Those type of surveys aid profes- sionals in different areas: in geology, to detect minerals of commercial interest; in the military, to find unexploded ordnance (UXO); in archeology; in the min- ing industry, to find scrap metals in iron ore piles; among others. In this study, magnetic maps are produced using small and autonomous aerial robots. We present a novel approach for automatic mapping using a custom fluxgate mag- netometer that is small and light enough to be embedded into a rotary-wing aerial platform. The magnetic coverage method could be performed by multiple robots simultaneously. The coverage methodology for multiple robots is based on a hexagonal subdivision of the environment supporting various "non-flyable" regions or priorities. Those hexagonal segments are clustered, and by optimiza- tion step, groups of cells are allocated to the different robots. The robot is prepared to fly at a very low altitude ( 1.5 m) to maximize the quality of the magnetic data acquired. The data points are spaced in an equidistant grid to facilitate the process of interpolating the map by a multi quadratic Radial Basis Function (mRBF) interpolator. Our method was validated in simulation and real experiments. Results have shown an accurate estimated magnetic repre- sentation of metal object placed in a test setup and a robust operation in a harsh scenario such as an open-pit mine. Besides, the results achieved in the experiments demonstrate that this approach is feasible in real scenarios, and further pushes the small aerial platforms towards the position of ultimate tool for efficient, accurate, and secure magnetic surveying.
- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2019
- Bibcode:
- 2019AGUFMIN41C0872C
- Keywords:
-
- 3394 Instruments and techniques;
- ATMOSPHERIC PROCESSES;
- 1920 Emerging informatics technologies;
- INFORMATICS;
- 1942 Machine learning;
- INFORMATICS;
- 1972 Sensor web;
- INFORMATICS