Environmental remediation in nearshore environments is significantly complicated by the dynamics of the environment. Applications related directly to detection and characterization of unexploded ordnance (UXO) in nearshore environments as well as those required to support UXO surveys are challenging due to the nature of the environment. There are a wide variety of problems presented: 

• Limited mobility, traction, and control; 
• Poor visibility and overall situational awareness; 
• Various physics-based challenges for sensing, such as acoustic backscatter and multi-path propagation, lack of radio frequency transmission, and moving free surface; and 
• Accurate positioning and navigation. 

Despite these challenges, there is a need for geophysical surveying in very nearshore areas. Underwater munitions are becoming increasingly problematic as ports and harbors, seashores, and other underwater environments are commercially developed or utilized for work or recreational activities. These environments vary significantly with respect to depth, morphology, geology, munitions density, and human exposure scenarios.

WRT's main goal for this demonstration project was to assess a combination of platform and sensing systems to deliver a UXO mapping survey and characterization technology in very challenging nearshore environments. Nearshore environments such as swash zones, surf zones with breaking waves, and shallow tidal areas provide hydrodynamic and bottom conditions that are particularly challenging from the standpoint of mobility and stability of sensing platforms to acquire high quality geophysical data. To overcome the limitations of current diver/manportable or ship-towed configurations in these nearshore regimes, WRT evaluated both platform and sensor performance to demonstrate and characterize a tailored and integrated robotic bottom crawler-towed sensor solution in representative shoreline UXO sites. The tests and demonstrations reported are among the first of their kind in terms of quantification of UXO detection survey performance metrics for a system that can traverse back and forth between fully submerged and dry land environments.

The final objective of this demonstration was to identify shortcomings and areas of improvement in the hardware, software, and operation of the integrated system.

For this demonstration, WRT integrated and implemented a robotic amphibious crawler platform with a towed EMI sensor array designed to provide a stable and mobile geophysical sensing platform for seafloor investigations. The crawler system is the SeaView SurfROVer, an amphibious robot with radio telemetry or integrated fiber optic tether system, track controller, lights and cameras, and Global Positioning System (GPS)/Inertial Navigation System (INS). The crawler is transported in the self-powered/self-contained operator control station (OCS) trailer system. The OCS contains the subunits for radio communication and data interface with the robotic crawler, GPS base station, helmsman navigation and control units and displays and sensor analyst computing facilities. The towed EMI sensor payload technology is the WRT FlexEM 3D system suitable for underwater UXO mapping, detection, and dynamic target classification. It comprises two transmitters and six triaxial “cube” receivers across its two-meter swath width. The integrated system contains subsea power supplies in the form of two 7.5 kWh smart battery pods in addition to an isolated and independent battery power supply for the EMI array, which enables six to eight hours of continuous operations. Command and control of the crawler-electromagnetic (EM) unit is provided through a wireless radio link from the crawler antenna mast to the OCS (located either on shore or on a vessel). Positioning of crawler and tow sled is provided through a mast-mounted realtime kinematic (RTK) GPS rover antenna on the crawler and a set of inertial measurement units (IMU) combined with a tow-point optical encoder for translating yaw as well as roll and pitch motions from the GPS locations.

The crawler platform and integrated tow sled system were successfully deployed multiple times and proved to be a stable operating platform with adequate tractive control on all substrates on which it was tested (dry grass and gravel, soft sand, mud and silt, shelly sands, and dry and saturated fine to coarse sand). Initial deployment testing exposed the need for improvements to the fiber optic tether system, the tow-point encoder and positioning system, and to the EM array sled. Analyses of EM data acquired during these preliminary tests were used to optimize system configuration and data acquisition parameters. Specifically, WRT made iterative modifications to the mechanical tow sled assemblies and EM system electronics to reduce the overall noise floor of the system by a factor of six. Additionally, these tests yielded significant improvements for the mission operations and methods used to survey with the system. This included waypoint following and user interface navigation guidance software as well as assessments and planning tools for turn radius and traction/trafficability potential of the system.

To assess system stability, WRT analyzed and compared inertial data from the crawler and tow sled IMUs. Typical target area transects began on the beach and progressed directly into the surf perpendicular to the shoreline. Once in the surf, the crawler-EM system rate of forward advance would slow due to hydrodynamic resistance and possibly also due to softer sediments on the seafloor. Among the most dynamic and challenging regimes was deep swash and surf inside of the wave break where strong wave induced currents’ ebb and flow.

Data from WRT’s full coverage surf zone surveys were analyzed using post-processing software to measure area coverage, target detection and localization, and classification performance. All targets were detected at signal to noise ratio (SNR) > 20 decibels (dB) for a 100% probability of detection. This far exceeded its detection metric of 100% detection using a nine dB threshold. The target localization accuracy was better than 30 centimeter (cm) circular error probability (CEP) for all target encounters exceeding its performance metric of 100cm CEP. However, the variance in target positions were approximately 50cm, which was higher than its target metric of 35cm. This was found to be mainly due to wobble and sway of the GPS mast and motion of the GPS rover antenna atop of the mast. WRT also observed and analyzed data to assess the overall stability of the system in the relatively dynamic surf zone conditions during its operations. WRT found the crawler-towed system to be very stable in the face of crashing waves up to 1.8 meters tall and forceful currents in the swash and surf zones during ebb and flood cycles between wave breaking events. In some cases, the EM sled appeared to lift from the bottom (perhaps related to heavy cavitation) and sway before the motion of the crawler was enough to pull it back in track. This added to the uncertainty of the array position, although WRT was able to mitigate this through analysis of tow point encoder and relative differences between crawler- and sled-mounted IMU data.

WRT also showed that the demonstrated configuration of the EM array was capable of target classification when multiple transects were aggregated together and inverted for magnetic polarizability curves. Inverted polarizability curves were matched to library curves for a subset of targets and analyzed. Although WRT was only able to compile a limited number of target encounters and even fewer clutter encounters, the consistency of the inverted polarizabilities yield promise for reducing the number of clutter by as much as 50-75%. Additional experiments with methods to switch the two side-by-side transmitter coils on the array indicate potential paths to improved classification that are less dependent on (or completely independent of) relative positioning between adjacent transect overpasses. This holds great promise for dynamic classification implementations in the marine environment.

In addition, WRT conducted assessments of the cost and logistical complexities for potential deployment and operation of the technology. Projected daily rates of approximately $8,500 for the integrated demonstration system (including a crawler helmsman operator) lead to considerable savings relative to deployment of an explosive ordnance disposal (EOD)-trained dive team searching the seafloor for UXO. Estimation of incurred labor and equipment costs estimated during survey mode operations yields 100% real coverage costs of approximately $2,500/acre or $1,250/hour. WRT assessed potential cost savings using this technology for a particular UXO site study where 500 survey contacts required reacquisition and further investigation. In this case, the crawler-EM system reveals as much as a 50% cost savings relative to conventional diver-based methods. Previous assessments have identified as many as 420 underwater ranges at over 120 different military sites comprising approximately 10 million acres of marine or lacustrine environment potentially contaminated with UXO. Of these 420 sites, it projected that 100 or more contain water depths that prohibit the use of towed geophysical survey systems or EOD divers. In many highly dynamic or otherwise challenging nearshore sites, it is likely not cost effective or just plain impossible to use dive teams for geophysical surveying.

In general, WRT experienced relatively smooth operations from the set-up stage through to completing surveys to demobilization at the end of the day. Very little calibration of the system was required. A functional test of power and communications of all systems was conducted without any issues. The crawler platform was mobilized and positioned initially through a small handheld controller that attaches directly to the crawler platform. The demonstration surveys were conducted without any particular issues; however, WRT noted some observations toward future implementation and optimization for potential cost-effective production operations. These include: (i) added perception and control for obstacle avoidance and navigational awareness; (ii) automation of the grid surveys to reduce operator demands and potential fatigue; and (iii) stability of the antenna mast and sensor sled to maximize positioning accuracy and tracking.

While the demonstrations of this crawler-based EM technology showed effective surveying over moderately sized (one to two acres) areas, larger areas may prove more challenging. Implementation effectiveness will depend highly on the hydrodynamic and bottom conditions at a site as well as the overall shape and size of the survey area. In general, WRT might anticipate a need to clear areas from a beach or shoreline to a prescribed closure depth. This is generally the depth of water along a shoreline profile at which sediment transport is very small or essentially non-existent and coincides with the foreshore region including swash, surf, and longshore current dominated nearshore. In this case, it would likely be more effective to plan surveys along shore-parallel transects and, thus, stability to shore-normal hydrodynamics such as currents from impinging waves and ebb/flow in swash areas is critical to maintaining high quality towed EM data.

Overall, the demonstrations reported illustrated the degree to which the crawler-towed EM system can fill gaps in current nearshore geophysical mapping, detection, and UXO classification. Specifically, WRT proved that the system could: (i) be mobilized and deployed in a cost-effective manner for shore-based operations; (ii) effectively cover surf zone target areas with 100% coverage transects; (iii) detect and localize UXO targets from 60 millimeters (mm) to 105mm in size buried beneath the seafloor to within a 50cm radius; and (iv) reduce clutter through analyses of inverted polarizabilities with moderate levels of confidence and effectiveness.