The prohibitive costs of excavating all geophysical anomalies are one of the greatest impediments to the efficient remediation of unexploded ordnance (UXO)- contaminated lands. Innovative discrimination techniques that can reliably distinguish between hazardous UXO and nonhazardous metallic items are required. To achieve reliable discrimination performance with electromagnetic induction (EMI) sensors, very high signal-to-noise ratios (SNR) and centimeter (cm)-level positional and degree-level orientation accuracy are required.
The objective of this SERDP Exploratory Development (SEED) project was to conduct a proof-of-principle demonstration that a single, fixed gradiometer or vector magnetometer could track an active EM sensor at sub-centimeter position and sub-degree accuracy over short baselines.
The basis of this project is that active EMI sensors generate large, predominantly, dipolar fields and that the direction, strength and gradients of these fields varies in a systematic way as a function of distance, bearing and orientation from the transmitter. The position and orientation of the transmitter can, in principal, be obtained by making either one 3- component vector or one tensor gradient magnetic measurements, assuming the transmitter moment is known.
In a series of field trials, a full tensor magnetic gradiometer, constructed from four 3-axis vector magnetometers (fluxgates) mounted on a wooden tripod, was used to accurately measure the field produced by a EM63 cart while both stationary and in motion. During the stationary trials, the cart was placed at various known positions ranging from between 2.9 m to 9 m away from the gradiometer. The transmitter waveform signal strengths were measured over a minimum interval of 10 seconds at each position. The measured fields were used to calculate the location of the cart, with respect to the sensor system, using three different dipole-inversion algorithms. Examining the accuracy and spread of the calculated inversions of each algorithm for all cart positions enabled a comparison of the effectiveness of the algorithms. Two algorithms, Wilson’s and a least-squares optimization routine, produced promising results with both techniques locating the cart reliably.
The accuracy of laser and GPS based positioning technologies required for anomaly detection and reacquisition degrades rapidly in non-ideal conditions such as in wooded terrain or underwater. A previous ESTCP project, MR-200129, studied the accuracy of various positional systems in a wooded environment, including GPS, Robotic Total Station (RTS), the ENSCO Ranger System (based on time-modulated ultra-wideband communications) and a GPS system with inertial aiding. These systems were only able to be positioned with greater than 10 cm accuracy on average, with positional errors above 30 cm experienced by all tested systems. A system that can more accurately position active EM sensors in obstructed environments would improve efficiency of munitions response projects in non-ideal conditions.