# Payload Description

MBARI operates two working-class ROVs \cite{ROVDRspecs} \cite{ROVVNspecs} that can accommodate what are referred to as ``tool-sleds’’. These are structural frames that mount underneath the ROV and extend the existing frame of the vehicle, with space available for science tools specific to each ROV mission. The initial development of the survey system required that all components be mounted in such a way that could later transition into deployment onto an AUV. Thus all items were mounted in-line, fore to aft, and with their principal observation direction oriented vertically down. Swaths and widest aspects are arranged port to starboard. Figure \ref{fig:sledimage} shows two configurations of the sled mounted on MBARI ROVs.

![The imaging arrays mounted on two ROVs: Doc Ricketts (left), and Ventana (Right) \label{fig:sledimage} ](./img/RovSleds.jpg){width=45%}

## Multibeam Sonar

Mounted centrally in the sled, and collecting cross-track scans of the seafloor, lies the Teledyne-Reson Seabat 7125 multibeam. This 400-kHz sonar has a swath angle of 140$\degree$, extending along track less than 1$\degree$. At a typical 3-m altitude, individual beams can resolve detail at 5-cm horizontally, and 6-mm vertically. This is the same system that is often used to collect data from our 0.55-m (21-in) diameter Dorado class mapping AUVs \cite{Martin2013UUST}. This model provides a very accurate and highly configurable hydrographic mapping tool. In application, the data collected using this sensor can help in co-locating previous maps collected by the Dorado AUVs.

## Stereo Camera System

The photographic imaging data are sourced by a pair of stereo cameras. The digital industrial cameras are both Allied Vision model GX1920C. These color cameras have 5mm lenses and corrective optics attached. Imaging through a domed glass port, the diagonal field of view of the camera is 90$\degree$. Since typical surveys are conducted near 3-m altitude, these cameras have been placed cross-track, spaced 0.2-m apart. Software that operates these cameras is executed remotely aboard the ship, and also collects images synchronously from both imagers. Leveraging the vendor programming interface, this software was developed initially at the University of Michigan’s Perceptual Robotic Laboratory \cite{UMichiganPeRL} and has been modified to fit our needs. 


Illumination of the photography is provided by a pair of Ocean Imaging Systems Model 3831 Strobes. Available off-the-shelf, these 200 watt-second xenon lamps provide adequate illumination of seafloor and are hardware synchronized to image acquisition through isolated trigger signals generated in the camera electronics. Their 2 second recharge cycle is the limiting factor in defining the maximum frame rate of the camera system. In typical seawater conditions, this system can image at altitudes from 2-m to 5-m. At 3-m altitudes, individual pixels represent a 2-mm square.  

## Lidar

The lidar sensor used by this payload is the SL-1, offered by 3DatDepth Inc. This laser ranging device is an extremely accurate time-of-flight system that employs 2-axis galvanometer mirror control of a pulsed laser. Since this payload is a moving platform, the lidar is operated in a back and forth scanning mode to collect data along a fixed swath. However, the option to tilt the swath has been employed to collect better data in certain terrain. This range-gate device has been mounted on moving platforms in the past \cite{6741175}, used for equipment inspection purposes. 

The swath width of the SL-1 is 30$\degree$, clearly the narrowest observation of the three systems. This is currently the limiting case for our primary line spacing when designing surveys. Surveys conducted at 3-m altitude produce 1-cm lateral resolution topographic data. 

 
## Navigation

Position and attitude data are provided by a SeaDeViL, manufactured by the Kearfott Corporation. This unit is an inertial navigation system (INS) with an internal laser ring gyro coupled directly to a Doppler velocity logger (DVL) which produces velocities relative to seafloor. It employs a Kalman filter to estimate attitude and position, which is aided by external data sources. Those external sources include GPS position while the ROV is on deck, ultra-short baseline (USBL) acoustic ship-mounted tracking, and a Paroscientific DigiQuartz pressure sensor (vehicle depth). This position is critical to the software used to station keep the ROV, discussed later in Section \ref{fine-scale-surveys}. 
 
When aided properly, the navigation solution will experience drift less than 0.05% of distance traveled. This is a critical data source for co-locating the individually collected survey data sets. All acquisition platforms have their main processor clocks disciplined by a single source clock via network time protocols (NTP). Some items have also been upgraded to also discipline their clocks via GPS time strings coupled to 1PPS (Pulse Per Second) clock signals, passed down from the surface vessel.