One of the most exciting areas of development for underwater remote-operated vehicles (ROVs) is their use in marine research. For decades, purpose-built (and generally large) ROVs like POODLE (the first real ROV, deployed in 1953) were on the forefront of oceanographic research work. Marine scientists did not generally use commercially-developed mass-produced ROVs until relatively recently – mainly because the mass-produced ROVs didn’t exist yet! They exist now, and the unique combination of affordability and standardization ROVs offer is now helping researchers push back the frontiers of knowledge in a number of exciting research fields.
One key area of research is the field of robotics itself – coordinating fleets of tiny ROVs is a new and unprecedented challenge for scientists at sea. Even more challenging is processing the terabytes of data that even a relatively modest quantity of ROVs can collect in a short period of time. A 2015 expedition by the Schmidt Ocean Institute’s Falkor to the remote Scott Reef in the Timor Sea, led by a team of University of Sydney scientists, deployed an eclectic flotilla of robotic vehicles, including gliders, autonomous underwater vehicles (AUVs), autonomous surface vessels (ASVs) and Lagrangian floats, and autonomous surface vessels (ASVs). The AUVs were used to take water measurements at different depths, the ASVs and gliders collected data on surface conditions, and the floats measured currents, water salinity and temperature, and other data.
Just in visual imagery alone, the Falkor collected some 400,000 images over a two-week period, about a terabyte every day. To gain insight into the meaning of the data they had collected, the team developed a web-based tool named Squidle, which crowdsources data analysis and lets the general public help teach computers how to interpret visual imagery. (You can learn more about Squidle at https://squidle.acfr.usyd.edu.au/) Another development of note from the Falkor expedition was the creation of a web-based tool to allow researchers to visualize the known positions of an entire fleet of ROVs in real-time using any Internet-connected device, such as a PC or smartphone.
Oceanic research on climate is one of the most critically important areas of science operating today. The challenge of global warming is tightly bound to our understanding of the Earth’s global ocean, and more visually spectacular catastrophic events like tsunamis and hurricanes emphasize the pressing need for deeper understanding of the oceanic climate. As one example, a Wave Glider ROV was collecting routine environmental data in the South China Sea in July of 2014 when it encountered Typhoon Rammasun, a lethal ocean storm with 10-meter high waves and winds approaching 200 miles per hour. The Wave Glider, tiny yet extremely rugged, rode out the storm without trouble and collected amazing data about the behavior of the ocean in response to the event. Fleets of ROVs could collect many times that amount of information, allowing unprecedented progress in areas like storm prediction and tracking. It’s not just data that can be collected – ROVs can retrieve water samples from anywhere in the ocean. (Aquabotix offers a water sample collector on the Endura line.)
The use of ROVs to monitor the health and size of fish populations has been under development for quite some time in the aquaculture field. Now researchers are drawing lessons from that work and applying it to help save the Great Barrier Reef. The GBR, a collection of almost 2000 reefs scattered across more than 130,000 square miles of ocean, faces a number of threats, including climate change and water quality problems, but scientists agree that the incursion of the Crown of Thorns starfish (COTS), a nightmare coral predator, is the Reef’s greatest immediate challenge. Scientists estimate that in the last twenty years, COTS population explosions have led to the destruction of about 40 percent of the Reef’s coral.
Monitoring the COTS populations was an important step, but monitoring alone can’t get anything done. Australians have operated hunting vessels to try to slow the advance of COTS populations, but even killing 400,000 COTS per year, as one anti-COTS diving team has done, is barely holding the COTS threat at a status quo level. Robotics researchers at the Queensland University of Technology, supported by a $750,000 AUS grant from the Google Impact Challenge Australia, have developed an ROV capable of hunting down and annihilating COTS. The prototype, dubbed “COTSBot,” is capable of operating autonomously, cruising a few feet above the coral reef using five integrated thrusters, scanning the surface of the coral and looking for COTS going about their nefarious business. When the robot spots a COTS – with a 99% accuracy rate – it swoops down and uses a robotic arm to inject the creature with a 20 ml vinegar solution, which kills it instantly. By mass deploying COTSBots in threatened areas of the reef, scientists hope not to just stem the incursion of the COTS, but to keep their population down to a manageable level.
The scientific work, both theoretical and applied, that oceangoing ROVs can support is critically important to both the health of the global environment, and to our growing ability to explore and understand the 95+ percent of the oceans that have not yet been genuinely explored. As always, human researchers and explorers will be irreplaceable in those efforts, but the addition of robotic assistance and tools in the form of ROVs will make their work vastly more effective, affordable, and effective.
One of the most common uses for commercial underwater remote-operated vehicles (ROVs) is in conducting inspections of water tanks, ship hulls, and submerged infrastructure such as bridge components or dams. A critical element of these inspections is measuring the thickness of metal components like hulls or girders. How do ROVs conduct this type of measurement?
On the Aquabotix Endura line of high-performance commercial ROVs, we offer the Cygnus NDT Metal Thickness gauge as an optional accessory. These gauges use ultrasound technology to measure the thickness of metal objects underwater. By emitting an ultrasonic beam into the surface of the metal and analyzing the return sound, the Cygnus NDT can measure metal of thicknesses up to 10”, even through coatings such as paint up to 0.787” thick.
The Cygnus NDT is extremely easy to use. The included CygLink software allows the ROV operator to visualize the tool’s measurements remotely on the video feed. To make things even easier, the optional Cygnus Probe Handler automatically aligns the probe to the wall or item being measured, with 15 degrees of movement, even if the operator has not perfectly approached the measurement subject.
Tools like the Cygnus greatly enhance the utility of our Endura ROVs. As ROVs take on more responsible roles in things like underwater inspections, the need for tools such as the Cygnus will continue to grow.
Many ROV applications require that the operator know the position of the ROV with varying degrees of precision. However, precise navigation for small ROVs is actually a very difficult task to achieve.
There are three basic positioning techniques that are practical for use on ROVs: dead reckoning, Global Positioning System (GPS), and Ultra-Short Baseline (USBL). We will look at each of them in turn.
Dead reckoning is the navigational method used by mariners all over the globe before the invention of compasses, chronometers and sextants. In dead reckoning, the navigator starts with a known (or estimated) position, and from there attempts to gauge the vehicle’s position by applying the vehicle’s heading and speed. Depending on the tools available, dead reckoning can be a surprisingly effective form of navigation. In the modern era, where the starting position of an ROV may be known with great precision and the heading and speed of the vehicle are also known quantities, dead reckoning can be sufficient for many applications. However, dead reckoning has one fatal defect: it cannot account for currents in the water. If the current is significant, the position of the ROV at the end of a run can be vastly different than the reckoned position.
The widespread availability and accuracy of global positioning systems (GPS) has made terrestrial navigation a breeze for motorists and surface vessels. Anyone with a line of sight to four or more GPS satellites – which effectively means anyone on the surface of the Earth – can use an inexpensive computerized navigation device to locate themselves, with an accuracy of as close as ten feet. GPS is cheap, effective, and nearly foolproof – so it’s the perfect solution for ROV navigation, right? Unfortunately, no – GPS is almost useless for most ROV applications. The problem is that the satellite communication signal is severely blocked by fresh water, and totally blocked by sea water. Underwater vehicles cannot use GPS without coming to the surface. There are some specialized use cases for GPS on an ROV, but they are few.
Ultra-Short Baseline (USBL) is a technique for determining a vehicle’s position underwater. It is a fairly complex system which involves a base station/transceiver, usually mounted on a ship, a transponder/transceiver mounted on the ROV, and a sophisticated computer system (usually located with the base station). The base station sends out a powerful acoustic pulse, similar to a sonar “ping”. This pulse is picked up by the transponder on the ROV, and a response pulse is sent from the ROV back to the base station. When the base station receives the response pulse, it calculates the duration between the initial signal and the response, and determines the range between the two transceivers based on that time. A set of transducers in the transceivers permits the calculation of a bearing between the two transceivers; the combination of the range and the bearing gives a precise position for the ROV relative to the base station.
USBL is extremely effective and accurate, but it is also expensive - a USBL installation for an ROV can cost upwards of $20,000. It also has a technical weakness in that under “busy” sonar conditions – a crowded harbor, for example – can cause problems for the system. However, for now, USBL represents the state of the art in ROV navigation systems.
Remote Operated Vehicles (ROVs) are of great utility in conducting underwater surveillance, data collection, sonar mapping, and searches for underwater objects. However, information gathering is only one of the things these versatile underwater craft can accomplish. With the addition of manipulator and grabber arms, ROVs can perform a wide array of useful tasks. ROVs have used manipulator arms since the first ROVs were deployed by the US Navy in the 1960s to recover objects on the sea bottom (including, famously, a hydrogen bomb that sank to the bottom of the western Mediterranean following a B-52 crash in 1966). Today’s manipulator arms are considerably more advanced than those early grippers.
Generally, modern ROVs use a master-slave manipulation model, where a human operator based on a surface vessel controls the ROV’s arm(s) either with a control panel or with an actual miniature manipulator matching the manipulator on the ROV. Remote control techniques originally developed in the space program allow any operator to work the arm easily, without the need for specialized ‘crane operator’-type skills. This allows the operator to be a mission specialist or a scientist skilled in what the manipulator needs to be doing, rather than a specialist who can control the arm well but needs to be told what to do with it.
Manipulator arms can be deployed to use a wide variety of tools. These tools are often simple or complex grippers, but can also be power tools like drills, saws, cutters, wrenches, or others. The earliest ROVs were limited to very simple grippers, and it was a major step forward when ROVs became capable of doing simple jobs like opening or closing a valve handle on an underwater platform.
There are three broad general categories of manipulator use for ROVs: industrial, military/law enforcement, and scientific. Industrial ROVs can use manipulators to cut and lay pipe, make and break hydraulic connections, lay and retrieving cable, lubricate underwater machinery, carry heavy tools for human divers, conduct trenching operations, and operate machinery like valves and levers. Military/law enforcement ROVs can use manipulator arms to retrieve or tag items, to place charges (for example, to neutralize underwater mines), and to cut cables. Scientific ROVs can use manipulator arms to select samples, to gather materials for analysis, to drill into the sea bottom, to sieve lighter materials, to take core samples, and to deploy and retrieve sensors.
ROVs equipped with manipulating arms have successfully assisted in exploring sunken vessels, raising ships and objects from the ocean floor, finding victims of crimes and criminal evidence, operating oil and gas platforms, and conducting pure scientific research. The use of manipulators makes these useful tools even more effective.
In a recent entry we talked about the importance of monitoring aquatic environments. Today we are going to talk about the various ways in which that monitoring can be performed. Each way has advantages and disadvantages.
Perhaps the simplest method of monitoring is manual observation, either with measuring instruments or with the naked eye. An example of this type of monitoring is a marine biologist who manually counts the number of whales surfacing off the coast in a particular spot in a particular time. The advantages of this type of monitoring are that it requires little investment in equipment, and has the highest degree of flexibility – if the biologist sees something else of interest at the survey site, he or she can make direct, first-hand observations without the need to redesign or redeploy instrumentation. Human observers may also be able to fix errors in the monitoring protocol “on the fly” in a way that a more instrument-focused approach cannot. For example, if a site survey has a particular set of GPS coordinates and the human observer arrives at those coordinates and realizes that the site location is off, the observer can fix the mistake and carry out the survey. The primary disadvantages of manual observation are that recurring observations quickly become very expensive, and there may also be reliability issues. For example, a graduate student can forget to come in to conduct a set of observations. Manual observations may also be subject to greater biases and subjectivity effects.
Another fairly simple, though powerful, technique for monitoring is the deployment of fixed sensors, whether singly or as part of a sensor network. An example of this type of monitoring is the global network of fixed data collection buoys and shore stations maintained by the National Oceanic and Atmospheric Administration’s National Data Buoy Center. This network of more than 1300 stations collects ocean temperature, current, conductivity, wave height, and wave frequency data, along with meteorological data, for use in weather forecasting, tsunami detection and alerts, and scientific and commercial research. The advantages of a fixed sensor network are in the acquisition of historical data sets which can be continuously compared to new data, the relative ease of maintenance of a network of fixed sites, and the ability to develop a comprehensive network. Indeed, the NDBC’s network has achieved global reach, making it one of the most valuable and useful fixed sensor networks in existence. The disadvantage of fixed sensors is that, being fixed, they cannot easily be redeployed to cover a new area of the aquatic region being monitored.
A more flexible approach to sensor deployment is the use of drop sensors, or temporary sensors for a particular location, which are then collected for re-use or allowed to degrade in place. An example of this type of sensor deployment would be the temporary placement of water-quality sondes along a flooding river system; the data collected from the sondes would be used for a limited time and then the sondes would be picked up and put back into storage or used on a different project. The advantage of this type of deployment is that it is extremely flexible; the sensors can be placed at the specific points where data collection is most important, and easily relocated as needs shift or if the initial deployment was less than optimal. The disadvantages of a drop sensor approach are that the sensors are generally not in a hardened or protected structure and are subject to theft, vandalism, or damage or destruction from environmental causes, and there is also potential for data loss owing to the ad hoc nature of the network infrastructure. For example, a set of inexpensive temporary buoys might be deployed to record ocean currents over a period of time, with the data to be collected at the end of the study period, only to have a major storm hit the network on the last day before collection, sinking half of the buoys and compromising the entire project.
The most sophisticated approach to monitoring the aquatic environment is to place the survey instruments on a ship, boat, or other craft and take the desired measurements from that vessel. For example, an aquaculture facility might have small craft with water quality sensors mounted in the hull to patrol the edge of the fish pens, testing the level of contaminants in the water. This approach has the advantages of most of the other approaches described here; it is flexible, it is protected from the environment, it has human observers on-site to address problems or concerns, etc. The main disadvantage of this approach is that ships and boats tend to be very expensive. The NBDC would need a much larger budget if it wanted to keep 1300 crewed ships on station to watch for tsunamis!
One way of greatly reducing the costs of this approach is to utilize ROVs as the data collection platforms. As ROVs are small, often portable by just one person, they can be deployed very flexibly in any aquatic environment, from water tanks to the deep ocean. Since they are controlled by a human operator, they maintain great flexibility and ability to adapt on the spot to problems that may arise, and are also protected against theft or vandalism by the proximity of the operator. When adverse weather or environmental conditions arise, the ROVs can be pulled out of the water, reducing the risk of data loss or loss of the vehicle itself owing to those kinds of conditions. Most importantly, ROVs can be equipped with a very wide array of top-quality scientific sensors, enabling them to monitor aquatic conditions just as well as sensor buoys or shore stations.
Every data collection mission has its own objectives and special needs, and finding the right type of monitoring solution for the mission requires judgment. ROV-based data collection is one option among many, but it is an option that is very flexible and has the potential to keep costs low without compromising the data collection process.
A vast array of organizations, governments, and individuals conduct monitoring in aquatic environments for an equally vast array of purposes. Monitoring is done for ecological reasons, for example, to record changes in the volume of a lake or stream so that managers will know whether to fill or drain a particular reservoir. Monitoring is done for legal reasons, for example, a factory that discharges waste into a river will be required to measure the levels of pollutants in order to comply with water quality laws. Monitoring is done for scientific and research reasons, for example, to assess changes in the temperature or salinity of the water in an area. Monitoring is also done for commercial reasons, for example, to check the bacterial counts in an aquaculture tank to ensure that the fish population is staying healthy and free of disease.
Some types of monitoring focus on the water itself, and its physical, chemical, and biological characteristics. This type of monitoring may look at the temperature of water over time, changes in its rate of flow or total volume in an area such as a lake or pond, or the pH (acidity/alkalinity) of the water. The monitoring may be done in order to assess the safety of the water for use by humans for drinking; is water from this stream low enough in harmful microorganisms to be directly usable without treatment, or does it require biological filtration? Monitoring may be done to ensure that water used in irrigation stays below a certain level of salinity and does not contain any of a number of pollutants. Monitoring could also be done in order to make sure that a pollution-control system upstream is working as intended.
Other types of monitoring are focused more on the species living in the aquatic environment. Marine biologists often engage in population surveys in a particular region, measuring the prevalence and population of a specific species or group of species. One famous example of this type of research is the tracking of salmonid numbers in the 20th century, which led to the discovery that increasing acidification of streams and oceans was impacting salmon and trout populations. This type of monitoring is done in several different ways; for large species such as oceangoing mammals, scientists may actually attach tracking devices to specific members of the species and gather data from the individuals’ movements, while for smaller creatures like fish, techniques such as gill-netting an area and then counting the fish in the net may be employed. Less invasively, water sampling might be done to measure bacterial levels when that is the focus of the research goal.
The specific environment being studied is of course highly relevant to how that environment can be monitored. Environmental monitoring of an aquatic environment can be done at a very “micro” level – a single sensor taking readings from a small aquatic habitat like a fish tank is technically monitoring that aquatic environment. At the opposite extreme, satellites in space can monitor vast expanses of ocean, at least coarsely. Most commonly, the scope of the aquatic environment to be monitored will dictate the range of appropriate monitoring solutions, whether that be simple cameras or sophisticated orbital satellites.
In next week’s blog we will take a look at the various ways in which monitoring of the aquatic environment is actually carried out.