AUVs explore the last great wilderness
E&T takes a look at the latest technologies that can be used to explore the world’s oceans, which remain the last great unexplored wilderness.
Autonomous underwater vehicles (AUVs) have come a long way in the last ten years or so with new battery technologies extending their endurance and programming development improving their reliability.
However, as Justin Manley, director of Scientific and Commercial Business at Liquid Robotics, points out: "Power for AUVs is still an issue and they have advanced but not in any breakthrough way.
"I was just talking to a colleague and they thought that the field had reached a plateau and I think in terms of technical capabilities they are right, we are at something of a plateau and I think two things might change the pace of development," says Manley, who is also chair of the Marine Technology Society UMV Committee.
Manley adds: "One is a radical advance in power systems or a radical new sensor, some new kind of multi-spectral camera for example - a new payload or new power source that would accelerate the field.
"The other thing that I think will happen is the funding picture will change in a couple of ways. In the US, we are applying new resources to climate change and environmental problems. That might drive new developments and applications."
Manley highlights two new exciting developments: "The first is that there are groups looking to combine glider AUVs with propeller AUVs. MBARI (see p48) is one of those groups. The idea here is that AUVs will have a range of thousands of kilometres as opposed to hundreds, but they will be much more precisely controlled than the gliders. It is a new hybrid concept that is coming together - a concept that we are talking about within the community.
"The other new thing that is out there which has had a lot of coverage recently is the Nereus hybrid remotely-operated vehicle (ROV - see p20)," Manley continues. "However, the key thing here is all that work that they did in the trench was actually done in ROV mode, so they did it connected by a very thin fibre-optic tether. They are also able to remove the tether to let it operate autonomously.
"The fact that you can get that deep is really exciting but you don't have to go deep for it to be exciting. The idea that you can use it to map out an area and then plug in a module that makes it an ROV with which you can take pictures and pick up rocks - that is a whole new innovative concept."
However, the weight of the umbilical cord on the ROV becomes a problem in extreme deep water.
"That's where the hybrid ROV comes in," explains Manley.
He explains that with a thin tether, you don't need a vessel with dynamic positioning. "You only have to look at the day rates of a ship with dynamic positioning compared with a ship without to realise there are huge cost savings.
"The tether does not supply power, it is a fully self-powered vehicle that carries its own batteries. It is essentially an AUV with a manipulator, so it can do a little light work. It can definitely inspect and it has about a two-day endurance."
To save battery power and increase endurance they use LEDs as a lighting source. "As you turn the camera it turns off the lights shining where the camera isn't looking. They have done all kinds of things to optimise the energy usage."
There are concepts where an AUV is equipped with a docking station that is connected by umbilical to the surface. Once the AUV starts running low on power, it can find its garage and recharge without having to go all the way back up to the surface. "That's a technical concept that has been demonstrated several times but nobody has yet found a compelling reason to do it all the time," says Manley. "It is cost which is the problem - we know we can do it, but it takes a lot of money to do it right."
However, it does offer an advantage if you are carrying out a survey with multiple vehicles because you don't have to keep coming back to lift them out of the water to recharge them. "The community that is going to make the best use of that early on is offshore oil because they already have infrastructure out there. I have heard that there are several companies looking at that kind of application for fully autonomous vehicles," Manley confides.
Instead of surfacing when a malfunction occurs, it has been proposed that the vehicle sits on the bottom to await rescue. "That is an idea that has been talked about - the way that issue has evolved is that software has got better and better and AUVs are more reliable," says Manley. However, I am aware that in the past year or so two fairly expensive AUVs have been lost. The analysis of the loss seems to indicate human error, which is to say that operators may have gotten complacent.
Manley tells E&T: "The story is that the software has got much better and is now so good that human error is a concern. AUV operators are now taking a more robust approach to operations manuals and team training."
Reliability was extremely important on the Autosub6000 when, in 2008, it carried out five missions to investigate potential threats to coastal communities along the Western European margin from giant landslides, earthquakes and tsunamis.
The Autosub6000 was developed by the Underwater Systems Laboratory at the National Oceanography Centre in Southampton, UK.
"When an AUV is sent 30km beneath an ice-shelf, there is no hope of recovery if anything goes wrong," explains Stephen McPhail, head of Platforms Development at the Underwater Systems Laboratory. "We tend to operate our AUVs in a totally autonomous mode. Once launched from the ship and dived, the AUV is very much expected to carry out the survey without any further intervention from the operators."
Communication is also a problem. The only way of communicating with the AUV when underwater is by using sound waves. So-called acoustic modems have a range of up to 7km at best, so the AUV is out of range for most of its mission.
GPS does not work underwater, so navigation is another big problem. When dived, the AUV must 'dead reckon' by measuring its velocity relative to the seafloor, using Doppler velocity measurements and an Inertial Navigation system.
Naturally, the greatest engineering challenges are due to pressure, which is half a tonne per cm2 at 5,000m deep. Electronic systems must be housed in titanium pressure-resistant enclosures. A novel feature of the Autosub6000 is that these enclosures are not needed for the rechargeable lithium polymer batteries. The centre has developed pressure-tolerant batteries, able to withstand the pressure. This enables more batteries to be carried, allowing a greater operating range and, with its current sensors, it has an endurance of 36 hours, or 180km at a speed of three knots.
LonWorks, with its networked and completely distributed architecture, was chosen for system design due to its modular approach. This meant that each engineer was able to independently develop individual devices and be assured that each device would be able to communicate with the others and function within the system when they were all finally installed in the vehicle.
Another major advantage is that the devices are 'intelligent' and communicate with each other. This ensures that when the AUV dived and became out of contact with the topside (research ship), the research mission can be carried out autonomously. The AUV's onboard mission control can, for example, receive information from the position control, velocity sensor and the collision avoidance sonar and pass on instructions to the motor control and depth control.
Simplicity is ensured because each node is not particularly complicated. The testability of the system is paramount and, as each node has a built-in test routine, individual nodes can be easily located and tested. Extensibility is straightforward and new nodes, such as new sensors, can be added with ease.
Fourteen CNS DV03TPT modules were used for the AUV's sensing and control applications. These interface with the various sensors and actuators, which make up the robot submarine. The system is very distributed with all nodes doing an equal amount of work.
The functions of some nodes in the system are:
- Depth sensor;
- Altimeter (SONAR) - measure height above the sea bed;
- Depth control - controls the depth and height off the seabed of the vehicle by moving the sternplane;
- Doppler Velocity Sensor (SONAR) - measures speed of AUV relative to the seabed;
- INS - inertial navigation system;
- GPS - gets global position fixes when the AUV is floating on the surface;
- Position control - controls the horizontal position and heading of the vehicle by moving the rudder;
- Emergency weight dropper - drops emergency ballast weight when fault detected;
- Collision avoidance sonar;
- Seawater conductivity and temperature sensor;
- Camera controller;
- Power monitor - checks on power used and condition of the batteries;
- Motor control - controls the propulsion motor;
- Mission control - the boss: coordinates all the control and stores the mission plan.
A PC 104-based data logger uses a Gesytec Easylon PC/104 Interface for accessing the LonWorks network. There is a Wi-Fi link from the ship to the AUV, with a CNS eNode Bridge linking Ethernet and the LonWorks twisted pair network. LonMaker and Nodebuilder software were used for developing and configuring the system.
At the end of each of the five 24-hour missions, the processes were analysed by the engineers. Scientific data obtained included water temperature, density, and chemical content. The seabed features of interest were formed when giant submarine flows ripped out huge volumes of seafloor creating seabed scours hundreds of metres across and up to 100m deep.
Autonomous market gap
Researchers at Monterey Bay Aquarium Research Institute (MBARI) in the US realised that there was a gap in the market when it came to AUV for scientific research. "There was a giant gap in capabilities between the existing propeller driven vehicles - which last about a day and run reasonably fast, at maybe three or four knots - and the gliders which last many months but run very slow and carry very few instruments," explains James Bellingham, chief technologist at MBARI.
They have developed an AUV that they believe fills this void and is a vehicle "that runs at about two knots which is still pretty fast and carries more payload than a glider", says Bellingham. It also has eight times the power allocated for payload that a glider would use and can operate for weeks at a time.
"We have learned a lot in the process of building previous vehicles and we are now able to make them a lot more efficient, so this is a lower drag vehicle and it has a more efficient propulsion system," says Bellingham. "Perhaps most importantly, the power consumed by the onboard electronics is a small fraction - about one-fortieth of what it is in a Durado class AUV - so we greatly reduce the power required for the vehicle and management systems. There has also been a lot of effort made to find optimum methods to control it.
"Gliders are really slick systems but they always fly at a certain angle of attack," continues Bellingham. "They need to generate lift off their wings for forward motion, which means their body is not flying at minimum drag."
While the MBARI AUV may be propeller driven, it still has a number of the features of a glider in it to allow it to change its buoyancy and shift its battery weight forwards and backwards. "That lets you fly the vehicle more efficiently than if you were just steering it with fins. It also lets us fly it slower, so we have a slow-speed mode which should let it fly many thousands of kilometres."
It has initially been targeted at two different activities. The first is climate science, making chemical and biological measurements in the boundary currents, which can be fairly strong. You need to have a fast vehicle to punch through currents like the Gulf Stream.
"We are also increasingly interested in the chemical and biological ocean, which requires larger and more power hungry sensors. The big buzzword here is ocean acidification but also monitoring the chemical and biological response of the ocean through changes," Bellingham explains.
The other activity revolves around process experiments, which is focused on enabling predictions. Bellingham asks: "How do you know what is going on now, and how can I make a model that models the chemical and biological ocean more accurately?
"It turns out that most of the productivity happens in up-welling regions and we happen to have one right off [the coast] here. It's all about understanding the dynamics of why particular organisms suddenly bloom; why does their population suddenly run away at a particular time; then why and how do they die off? What are the mechanisms for them dying off?
Bellingham continues: "These are all unanswered questions and until you have answers to them you cannot expect your models to be accurate for extrapolating to a different ocean condition.
"What would be ideal is to be at the right place and time so when the bloom occurred everything is in place and you measure that bloom as it is increasing. You could take water samples capturing the organisms and carry out analysis in the lab later.
"As the bloom begins to die you would continue to follow it as it dies off and then be able to relate the changes in population to changes in environmental conditions, predation and mortality."
However, to do that requires a much more intensive measurement capability than has been available in the field and typically they have to be done from a ship.
"A big part of our plan is operating these vehicles in large numbers and have them coordinate with each other to follow the bloom," adds Bellingham. "Biological populations tend to be patchy and right now our thinking is to follow the patch and monitor that patch through its entire lifecycle.
"I am focused on the vehicle part of it but lots of other people are contributing to this enterprise developing biological and chemical sensors."