From Heights to the Deep Sea,a Review of Robots Interacting with Offshore Structures

Abstract

An increasing energy demand forces companies to use the space available at sea, causing the number of offshore structures to increase. These offshore structures require maintenance. However, the harsh offshore environment makes this dangerous and costly for human workers. A solution would be the use of robots with manipulation capabilities. This literature review identifies robots that have manipulation capabilities and can be used to perform inspection, maintenance, and repair tasks on offshore structures. The environment of an offshore structure changes along its height. Therefore, the robots in this review are categorized based on the height they operate on. Starting at the top of wind turbines all the way down to the ocean floor. At great heights, unmanned aerial manipulators have some manipulation capabilities but are limited in their flight time and are vulnerable to disturbances. Climbing robots avoid the problem of strong winds at heights, but their manipulation capabilities are still limited. Legged robots are already used on the superstructure of offshore oil and gas platforms but would benefit from more autonomous operations. There is a limited number of robots operating in the splash zone. However, this number can be increased by making some adjustments to existing robots. Finally, various underwater robots with manipulation capabilities exist and are commercially available. However, they lack the ability to perform manipulation tasks autonomously. Overall, there are several robots that can or have the potential to perform manipulation tasks on offshore structures, but challenges need to be overcome before robots can be used on offshore structures on a large scale. Future research should focus on flexibility, durability, and autonomy. Overcoming these limitations will improve the safety, efficiency, and cost-effectiveness of offshore operations.

Introduction

The world’s reliance on the ocean for its energy needs is increasing. The worldwide offshore wind capacity has been growing by an average of 28% for the last 10 years.¹ In 2021, 21.1 GW of new offshore wind capacity was connected to the grid, a record for offshore wind.² Furthermore, offshore oil and gas are also expected to keep growing in the coming years.³ The increase in offshore activities inevitably leads to a rise in the number of offshore structures. Consequently, the demand for inspection, maintenance, and repair (IMR) tasks will increase.

The offshore environment is very harsh. Temperatures can vary between −40°C and +40°C,⁴ wind and waves constantly blast across the lower levels of offshore structures, and there is extreme pressure at the foundation of the structures. Owing to these inherent complexities of offshore environments, IMR tasks in offshore environments are complex, costly, and pose significant safety risks to human workers. Fortunately, using robots could provide a solution to the problem. These robots should be able to withstand a variety of disturbances. An offshore structure spans multiple environments with widely different conditions that robots performing IMR tasks need to face. Under the surface of the ocean, robots experience forces from ocean currents and high pressures at great depths. Although high above the surface, robots must resist wind gusts and rain. Different robots are needed for each environment.

Robots have been used for IMR tasks on offshore structures. However, their use is mostly limited to inspection, lacking a physical interaction between the robot and the offshore structure. Although inspection robots improve the speed and efficiency of the maintenance process, technicians are still required to perform more complex IMR tasks. Robots that are capable of physical manipulation could further reduce the need for human workers, potentially improving the safety, efficiency, and cost-effectiveness of offshore operations. Therefore, the goal of this work is to show the advancements and challenges for robots that need to perform physical manipulation on offshore structures, such as repairs, in different offshore environments such as the deep sea and at heights.

This literature review aims to find robots that interact with offshore structures starting with the robots that work high above the water level and working down to the seabed. This starts by defining the different environments and the search approach in Methods section. The current technologies are listed in Current Technologies section. Afterward, the challenges and opportunities are discussed in Discussion section, followed by the Conclusion.

Methods

This literature review aims to find the gaps in robot technology that interact with offshore structures.

Operating domains

Offshore environments are very harsh conditions to operate in. The conditions are not the same for every part of an offshore structure. Therefore, the operating area has been divided into three operating domains (see Fig. 1):

Atmospheric environment

Splash zone

Subsea environment

In each of these domains, different challenges arise for the robots operating in them. This chapter gives an overview of these operating domains and what challenges they pose to robots.

FIG. 1.

Overview of the different operating domains in the offshore environment.

Atmospheric environment

The atmospheric environment encompasses everything between the highest tip of a wind turbine and the area not affected by waves. Currently, the highest offshore structure is the H260-18MW wind turbine developed by China State Shipbuilding Corporation.⁵ The highest point on this turbine is 280 m above the sea level.

In this domain the most important challenges are wind and temperature. Above 100 m wind speeds of over 10 m/s are easily encountered.⁶ High wind speeds complicate the control of unmanned aerial vehicles (UAVs).⁷ Furthermore, extreme wind conditions such as hurricanes can cause oil platform to be shut down and the crew will be evacuated.⁸ Finally, extreme temperatures further complicate the operational conditions. The most extreme temperature can be found in the Kashagan Field in the northern Caspian Sea. The temperature ranges from −40°C in winter to 40°C in summer.⁴

Splash zone

The splash zone is situated between the atmospheric environment and the subsea environment—basically, all regions that are affected by waves.⁹ It poses challenges to robots from both the atmospheric and the subsea domains. The combination of wind and water poses harsh conditions for robots.

Waves create a significantly unstable and unpredictable environment. Operating near an offshore structure creates a large risk of collisions. Furthermore, the constant change in water level, as well as water spray, creates a mixture of salty water and oxygen that heavily promotes corrosion and the growth of marine biofouling. The wave force on the structure is a function of the structure’s diameters.¹⁰ Failing to remove marine biofouling from offshore structures increases the risk of structural failure. Specialized materials and sufficient propulsion methods, as well as proper control strategies, are necessary to deploy robots in this domain.

Subsea environment

The last operating domain is the subsea environment. There is a wide variety of submerged offshore structures—primarily, the foundation of wind turbines and oil and gas platforms. The foundation of fixed oil platforms can be as deep as 535 m below the sea level.¹¹ However, the deepest floating oil platform is tethered to the seabed at 2450 m below the sea level.¹² Wind turbines are usually located in shallower waters. The tallest fixed foundation stands at 58 m.¹³ There are wind farms under construction with wind turbines on floating foundations; however, the foundations of the deepest farms so far are located 120 m below the surface.¹⁴

This variety in depth leads to a multitude of challenges. As the robot moves down into the ocean, the pressure increases. From 200 meters below the sea level, the deep sea starts.¹⁵ In the deep sea, there is no light, very high pressure, and the temperature drops drastically. Besides the challenges related to depth, there is always the threat of corrosion and strong underwater currents.

Search approach

The search in this review was performed using Scopus and Google Scholar. Scopus was used as an initial search engine, whereas Google Scholar provided additional information on the results from Scopus. Scopus was used because of the quality of its database and the functionalities of its search engine.¹⁶

The search started with a search of reviews, to create a broader understanding of the robots used for offshore IMR tasks. As the scope of this article is relatively broad, other reviews have been written that highlight different parts of this review. The reviews found are listed in Table 1. All operating domains and a variety of robots are mentioned; however, the number of robots with manipulation capabilities is still very limited. Only in the subsea environment, a significant number of robots with manipulation capabilities were mentioned.

Table 1.

Summary of Existing Reviews on Offshore Robotics and the Number of Robots with Manipulation Capabilities

Source	Year	Operating domain	Robot types	UAM	Climb	Leg	USV	UUV
¹⁷	2022	All	UAV, climb, leg, UUV	—	—	—	—	1
¹⁸	2019	All	UAM, leg, UUV	1	—	—	—	1
¹⁹	2018	Subsea	UAV	—	—	—	—	—
²⁰	2022	Atmospheric, subsea	UAV, climb, UUV	—	—	—	—	5+
²¹	2018	Atmospheric	UAM	3	—	—	—	—
²²	2022	Subsea	UUV	—	—	—	—	3
²³	2019	Atmospheric	Climbing	—	2	—	—	—
²⁴	2023	Atmospheric	Climbing	—	3	—	—	—
²⁵	2021	Splash zone	Climbing	—	—	—	—	—
²⁶	2023	Atmospheric	Legged	—	—	4	—	—

UAM, unmanned Aerial Manipulator; USV, unmanned surface vessel; UUV, unmanned underwater vehicle.

Table 2.

Summary of UAMs

Source	Year	Name	Type	#^a	Nav.	Man.	TRL
³¹	2014	AMUSE	Octoquad	1	A^b	A	6
³²	2017	n.a.	Hexacopter	2	R^c	R	6
³⁴	2019	SAM	Octocopter	1	n.a.	n.a.	4
³⁰	2014	MM-UAV	Quadcopter	2	R	R	4
³⁵	2016	n.a.	Hexarotor	1	R	R	6
^36,37	2018	DRAGON	Snake	2	A	A	4
³⁸	2019	n.a.	Quadcopter	1	A	A	6
²⁸	2020	n.a.	Hexarotor	1	A	A	7^d
²⁹	2019	AeroX	Octocopter	1	R	A	7^d

Number of manipulators.

Autonomous.

Remote operated.

Not tested in offshore environment.

SAM, suspended aerial manipulator; TRL, technology readiness level.

Table 3.

Summary of Climbing Robots

Source	Year	Name	Locomotion	Adhesion	#^a	Surface	Nav.	Man.	TRL
⁴⁰	2014	BladeBUG	Legged	Suction	n.a.	Turbine blade	n.a.^b	n.a.	7
⁴¹	2016	n.a.	Wire driven	Wires	2	n.a.	A^c	A	6
⁴²	2020	n.a.	Wheeled	Magnetic	1	Metal	R^d	n.a.^e	4
⁴³	2016	Climbot	Legged	Clamping	1	Poles	R	R	6^f
⁴⁴	2019	CMBOT	Legged	Clamping	1	Handrails	A	R	5

Number of manipulators.

Limited information available.

Autonomous.

Remote operated.

Welding has not been tested.

^fNot tested in offshore environment.

The search was split into operating domains, and by performing multiple searches for each domain, a systematic search was conducted. Each different search has a different set of search terms, consisting of general search terms applicable to all domains and domain-specific search terms (see Fig. 2). The general search terms are applicable to all domains and are used to steer the search engine into the right subject area, that is, robots that interact with offshore structures. The domain-specific search terms are used to further specify the search. Finally, all articles before 2010 were removed from consideration, as almost 75% of the articles found on Scopus using the keywords “offshore” and “robot” were published in 2010 or later. Combining the general and domain-specific search terms creates a comprehensive but targeted set of information.

FIG. 2.

Search terms used in the research for this literature review.

The search terms used are shown in Figure 2. Note that not all general search terms were used in each search. This is due to a lack of articles on a certain subject. For example, articles on unmanned aerial manipulators (UAMs) do exist; however, the number designed for offshore environments is limited. To discover potential future developments, the search was then broadened, by omitting “offshore” to allow for a more wide variety of robots.

Current Technologies

The results for each operating domain have been summarized in a table. The table includes common information across all categories, such as the year of publication, name, technology readiness levels (TRLs), and the ability to navigate and manipulate autonomously. TRLs are a measurement indicator system that allows for the comparison of maturity between different technologies. Each robot is assigned a number from one to nine that indicates the maturity of the technology. TRL 1 means the technology is only a fundamental concept, whereas TRL 9 is a commercially available fully developed robot. The TRLs were determined by comparing the information provided in the articles with the definition of TRLs, first developed by National Aeronautics and Space Administration.²⁷ Some tables contain additional information, relevant to the robots summarized in the table, such as the maximum diving depth for underwater robots. This structured presentation allows for a clear overview and comparison of the different robots operating within each domain.

Atmospheric environment

UAMs

Inspection of offshore structures has been extensively performed by UAVs. Their maneuverability and speed make them an excellent option to visually inspect offshore structures. They can carry various sensors such as cameras, thermal sensors, and LiDAR. However, new technologies allow UAVs to manipulate and touch offshore structures. These robots are called UAMs and are equipped with some form of manipulator. UAMs can be used to place sensors,^28,29 turn valves,³⁰ and other manipulation tasks. These UAMs are summarized in Table 2.

The AMUSE is an octoquad aerial manipulator with a large 7-DOF manipulator arm.³¹ The arm is capable of lifting 1.5 kg. Considering the weight limitations imposed on UAMs, this is a relatively good performance. Alternatively, an AUM can be equipped with two manipulators.³² Two arms allow an AUM to perform more complex tasks. Unfortunately, the UAM developed by Suarez et al. can only lift 0.3 kg per arm, considerably limiting the robot’s capabilities. Both of these UAMs are designed to counter external disturbances, that is, allow outdoor operation. Nevertheless, the outdoor conditions simulated during testing do not approach the conditions present in offshore environments.

UAMs have to operate in close proximity to offshore structures. When operating in complex environments, where space is limited, UAMs are extra vulnerable to crashes.³³ A potential solution is the cable-suspended aerial manipulator or SAM.³⁴ A main aerial carrier equipped with winches and cables can lower and raise the SAM. The SAM is equipped with propulsion units, but these are only used to change the orientation and compensate for disturbances. Therefore, propulsion units do not need to be as powerful as they do not have to lift the weight of the robot, thus making the robot smaller. A robotic arm is attached underneath the SAM. Because the main aerial carrier can be quite large, weight is less of a problem compared to other UAMs. This allows the SAM to use a 15 kg manipulator, considerably heavier than other UAMs. The lifting capacity of this robotic arm is not mentioned.

The UAMs mentioned before are capable of picking up and moving objects. However, more complex tasks need to be performed in offshore IMR. A common task in the offshore industry is the turning of valves and bolts. The MM-UAV was designed to turn valves.³⁰ Using two grippers, the UAM attaches itself to the valve. The UAM rotates, thus turning the valve. The torque needed to turn the valve comes directly from the rotors. Generating the amount of torque needed is difficult as there is no stable base to exert the force on. Therefore, the valves used in the experiment with MM-UAV had little friction. In practical situations, corrosion increases the difficulty of operating the valves, making it more challenging to turn them. To generate the additional torque required to open those valves, an impact device can be used. This impact device is implemented on a UAM.³⁵ This UAM is designed to perform torsional work at high altitudes and has an impact device to unscrew light bulbs. The impact device allows the robot to accelerate independent of the gripper. After a full turn, the UAM reconnects with the static gripper, transferring all the energy to the gripper and unscrewing the light bulb. Although this technology is only used for a light task, it could also be used on valves and bolts.

The two robots mentioned earlier are specifically designed for (valve) turning. However, in an offshore environment, flexibility is crucial. A robot that offers this flexibility is the DRAGON. The DRAGON is a flexible UAM capable of changing its shape.³⁶ The DRAGON consists of four links connected by joints, allowing it to take a stable shape when manipulating and a narrow shape when navigating complex environments. This makes the DRAGON very well suited to operate in complex environments such as offshore oil and gas platforms.

Because of the shape of the robot, a different thruster configuration was needed compared to the more common quad and hexacopters. Each link has two propulsion units that are capable of turning along the longitudinal axis of the link, allowing the robot to hover in any configuration.

The DRAGON has performed manipulation tasks such as valve turning. A fork-like attachment was placed on one end of the robot. This fork is placed in between the spokes of the valve wheel. The robot then turns itself to generate the torque to open the valve. Owing to its flexible shape, the DRAGON is capable of turning valves at any angle.³⁷

Another maintenance task would be spot repairs to prevent leaks. In the work done by Chermprayong et al.,³⁸ a UAV is equipped with a delta manipulator. The end-effector of the manipulator contains a foam extruding mechanism. The foam has a large expansion ratio, making it an excellent material to fill gaps in pipelines or other structures. A delta manipulator was used because this yields a higher accuracy compared with serial manipulators.³⁹ The delta manipulator can compensate for both translation and rotational offsets caused by the UAV. In experiments, the robot was capable of autonomously filling two holes and a crack using the foam, proofing the feasibility of this concept.

The Universidad de Sevilla developed a UAM designed for physical inspection and sensor placement on bridges.²⁸ Bridges are located in environments that share a lot of similarities with the offshore environment, that is, wind and water. Therefore, this robot has the potential to be used in the offshore sector.²⁸ The platform was customized to allow more space for avionics and realize the tilting of the rotors. By tilting the rotors, the drone becomes fully actuated.²⁸ The UAM docking gear consists of four supports that are pressed against the surface of the offshore structure. The ends of the supports are lined with rubber. The friction between the rubber and the surface of the offshore structure in combination with the force exerted by the rotors secures the UAM to the surface. Once docked, the manipulator arm can carry out physical inspections or place sensors. The system is currently designed for the inspection of bridges. Therefore, the docking gear is located on the top of the UAM, allowing only vertical docking. However, there are plans to allow horizontal docking for the inspection of pillars and walls.

A UAM that is already capable of physically inspecting and repairing vertical surfaces is the AeroX.²⁹ This octocopter, designed for the oil and gas industry, is equipped with a robotic arm with rubber wheels at its end-effector (see Fig. 3). The wheels are placed directly on the to-be-inspected surface. The arm can be tilted to reach both horizontal and vertical surfaces. Multiple devices can be placed at the end-effector, including sensors, gel injection devices, and a system for placing and retrieving sensors. This allows the AeroX to perform not only inspection but also small maintenance and repair tasks.

FIG. 3.

AeroX.²⁹ Published under the Creative Commons Attribution (CC BY 4.0) license.

During the flight, the AeroX is controlled remotely by a human pilot. Once the robotic arm’s wheels touch the inspection site’s surface, the AeroX switches to contact-flight mode. In contact-flight mode, the UAM fully autonomously keeps the robot steady with respect to the contact point on the surface. It exerts a force on the surface to ensure the wheels stay in contact. In contact-flight mode, a human inspector can remotely change the location of the end-effector. The UAM will automatically follow the motion of the end-effector; the human pilot does not need to intervene.

Contrary to Sanchez-Cuevas et al.,²⁸ this robot cannot stabilize itself using a docking gear. Although the rubber wheels provide some form of antislip, it relies fully on its internal control system for stabilization. Tests were performed on the stability of the AeroX. Using wind and an external force (pulling a rope), the stability of the UAM was validated. However, the AeroX is designed for the onshore oil and gas industry. The conditions offshore are far harsher; thus, it is not clear if this UAM could also operate in this environment.

Climbing robots

Not only UAMs are capable of IMR tasks at height, but climbing robots have also shown promise. A major advantage of climbing robots is the limited effect of wind on the robot, especially compared with UAMs. Various systems have been designed to traverse offshore structures. Climbing robots always have physical interaction with offshore structures, but only robots that have the ability to do more than visual inspection are selected, see Table 3.

A robot designed for inspection and small repair tasks on offshore wind turbines is the BladeBUG. This six-legged climbing robot uses suction cups to stick to turbine blades⁴⁰ (see Fig. 4). It was designed by a former turbine engineer who was convinced that advanced robotics could increase the efficiency and cost of wind turbine IMR. The BladeBUG has been tested on a wind turbine in Scotland, where it successfully walked along the turbine blade.⁴⁰ The website also mentions the test of repair capabilities; however, limited information is available.

FIG. 4.

BladeBUG. Reprinted with permission from BladeBUG.⁴⁰

Big advantage of the BladeBUG approach is the elimination of wires. Wired robots are attached to a wind turbine’s nacelle and lower themselves along the wind turbine blade.⁴¹ The advantage of this system is increased stability and a larger payload. This robot was designed for the cleaning and inspection of wind turbines. The robot is capable of switching rotor blades by lowering itself below the tip of the turbine blade. The turbine turns until the next blade is lined up vertically; the robot is then lifted along the length of the blade. The robot is capable of cleaning all blades; however, problems arise when deploying this robot. The robot needs to be attached to the nacelle, and this is a problem not solved by Lee et al.⁴¹

A robot can also use magnets to secure itself on the surface of an offshore structure. A robot with magnetic spherical wheels was designed for the navigation of nonuniform curved surfaces, vertical walls, and traversing of 90° corners.⁴² This robot aims to weld steel structures, such as gas tanks. The robot has two sets of two magnetic spherical wheels. Each set of wheels is attached to the robot’s main body using a rocker arm suspension. This suspension in combination with the spherical shape of the wheels allows the robot to traverse nonuniform curved surfaces with all wheels in contact with the structure. The contact between the wheels and the surface is necessary for the magnets in the wheel to secure the robot to the structure.

A drawback of magnets and suction cups is that the surface needs to be relatively smooth and consist of the right material. The surfaces of oil and gas rigs and the lower part of wind turbines are complex. It is scattered with pipelines, machinery, and a lot of handrails. These are obstacles for rolling robots but offer an advantageous handhold for robots that use a clamping mechanism. Two of these robots are in development: the Climbot and CMBOT. The Climbot consists of four links connected by joints and, at each end, a gripper module.⁴³ These grippers are used both as clamping devices and manipulators. The grippers are sufficiently strong to allow one gripper to secure the robot to a post, whereas the other gripper takes any position. This allows for multiple different gaits, further enhancing the flexibility of the Climbot. Once a gripper grabs a pole, it locks and can maintain this position without a power supply, allowing the robot to cut power to one gripper during manipulation and save energy.

Although the Climbot uses its grippers both as a climbing mechanism and as manipulators, the CMBOT separates these functions.⁴⁴ The CMBOT has two legs equipped with clamping devices that can grab handrails. The two legs are connected through a central axle. This axle is also attached to the robotic arm used for manipulation tasks. The arm consists of four links allowing for a long reach and great flexibility. This robot was designed for IMR tasks on railway bridges. Unfortunately, the clamping device can only be attached to flat surfaces as that was the type of handrail used on the railway bridge. However, oil and gas platforms are filled with handrails; thus, the CMBOT might find an application on offshore structures.

Legged robots

Legged robots have found their way into the offshore industry, mostly in an inspection capacity.^45,46 New technologies open the door to legged robots with manipulation tasks. Legged robots are capable of traversing complex terrain such as stairs and doors. Although there are legged robots with climbing capabilities,⁴⁷ their movement is mostly limited to the floor. Most legged robots are quadrupedal or four legged.²⁶ This setup provides a stable base for a manipulator arm. However, it is not always necessary to use a nonlocomotion arm as a manipulator. Legged robots that can manipulate objects with their legs exist, but the complexity of their manipulation is limited. To perform more complex tasks, a separate manipulator arm is needed.²⁶ These robots are the focus of this review and are listed in Table 4.

Boston Dynamics are at the cutting edge of robotics, and it is no surprise that they developed the most advanced robot in this section. The Spot arm is a modified version of the Spot robot.⁴⁸ The Spot is a quadrupedal robot capable of full autonomous inspection of offshore structures.⁴⁹ It has the ability to climb stairs and visually inspect areas that are too dangerous for humans. The Spot arm has a manipulator arm attached to the top of the robot. This manipulator can lift 5 kg off the floor and gives the robot the capability to open doors, turn valves, and move objects. The manipulator is not fully autonomous, but can fully autonomously open doors. The Spot is already deployed on offshore oil and gas platforms, opening the door for the use of the Spot arm.

A different legged robot that has been used on offshore structures is the ANYmal.⁴⁶ This robot is similar to the Spot, but the implementation of a manipulation arm is lacking. A project combines the ANYmal with a robotic arm designed by Kinova.⁵⁰ This combination has been tested by picking up a bottle from a table while standing on different terrains. This included uneven surfaces but also the placement of the back legs on a skateboard.⁵⁰

The advantage of the robots listed earlier is that the platform on which the manipulator is placed is a proven concept, allowing for faster implementation on offshore structures.

The DogGet was designed for the autonomous identification and transport of an object.⁵¹ Finding and moving an object is easy for a human, but robots struggle with this task. Most robots in this review are not (fully) autonomous in their manipulation tasks. The DogGet aims to perform these tasks in a human environment without the intervention of humans. It independently maps its environment and can avoid collisions with moving objects, at the same time trying to identify the object it needs to find. This is highly complex and not entirely feasible. To decrease the level of complexity, the general location of the object was known to the robot. However, it was capable of identifying the object and avoiding obstacles on its way. The manipulation part of this task does not make this robot a great asset for IMR tasks at offshore structures. However, the ability to identify an object autonomously is one of the biggest challenges for all robots in this review, making the DogGet an asset nonetheless.

Splash zone

Robots similar to the wheeled climbing robot in Eto and Asada⁴² can be used to clean the jackets in the splash zone and below; such a robot is presented by Fan et al.⁵² This robot is equipped with actuated magnetic wheels that secure the robot to the jackets. The robot can operate on different diameter pipes, both above and below the water level. This makes the robot well suited for cleaning in the splash zone. The cleaning equipment is a cavitation jet that can clean a surface 70 mm wide.⁵² This allows the robot to clean up to 21.11 m²∕h, which is comparable with the speed of manual cleaning.

Alternatively, the ROVMS can be used. This ROV also attaches itself to a jacket using magnets. However, its locomotion is proved by thrusters.⁵³ This allows the ROVMS to independently move to and between jackets. Unfortunately, the robot can only move in water, thus not allowing it to clean above the water line. This ROV is equipped with a cavitation jet pistol to clean the surface of the jacket.⁵³ Once secured to the jacket, a manipulator arm moves the cavitation pistol along the surface, removing any biofouling. The robot remains stationary until the surface in reach of the robotic arm is cleaned. The properties of these robots are summarized in Table 5.

Unmanned surface vessels

The use of unmanned surface vessels (USVs) in the splash zone seems an obvious next step in offshore robotics. Several USVs are being developed for surveillance and transport. However, after a comprehensive search for USVs that interact with offshore surfaces, no significant results were found. Several USVs are being developed but all for inspection and transportation tasks.

USVs could be used as support vessels for other robots that have the ability to physically interact with offshore structures. Several examples of concepts that combine USVs, ROVs, and even UAVs and climbing robots are being developed.^20,54,55 For example, Peng et al.⁵⁶ provide a concept to transport ROVs to a work site using an USV. The USV is equipped with a manipulator arm that can lift the ROV out of the water.⁵⁶ Moreover, a system to deliver a BladeBUG to a wind turbine using a UAV and USV is discussed by Jiang et al.⁵⁷ Using USVs not only increases the operational range of other robots but also increases communication with onshore control centers.

Subsea environment

The robots operating in the underwater domain have existed for quite some time. However, most robots are designed for observation tasks. Intervention tasks still require divers or ROVs operated by a human. This means people and materials need to be moved to the work site. Several AUVs are developed that have manipulation, intervention, and maintenance capabilities. These robots can be divided into several categories.

Shape-shifting

A potential solution is shape-shifting robots.^58–60 These robots change their shape or orientation depending on the task at hand. To reach the offshore structure they will take a more hydrodynamic shape. This saves energy and time, allowing for more extensive intervention tasks to be executed at a longer range. Once arrived at the location, the robot will change its shape into a more stable form. When performing manipulation, it is crucial that the robot maintains the correct position. Underwater currents and the force applied by the manipulation arm on the structure could cause the robot to move. A more stable form helps maintain the correct position.

The Aquanaut is an excellent example of a shape-shifting AUV developed by Nauticus Robotics.⁵⁸ This, transformer-like, AUV travels in a torpedo shape until it reaches the maintenance site. On site, it unfolds its two manipulator arms. Because of its change in shape, the Aquanaut is capable of travelling for 250 km at 3 knots. At the same time, its arm can lift over 60 kg while being fully extended.⁶¹

It can fully autonomously travel to the working location; however, the arms still need to be operated remotely. The first two versions of the aquanaut are being commissioned as of 2023.⁶² A third version is planned; however, whether that version allows for autonomous intervention tasks is unclear.

A different shape-shifting AUV is being developed at the University of Florence.⁵⁹ This AUV has a diamond shape in intervention mode and folds into a linear shape when travelling.⁵⁹ This AUV is still a concept, and although the article mentions the use of manipulators, there is no explicit example of it. It does however create an alternative concept of the shape-shifting technology.

Finally, the Cuttlefish AUV does not change its shape but rather changes its orientation to allow for a more stable intervention mode. The Cuttlefish has a slender cuboid shape. It travels in a horizontal orientation, in the direction of its smallest cross-section to reduce hydrodynamic drag.⁶⁰ Once arrived at the intervention site, it tilts 90° into a vertical orientation. This new orientation is a more stable position to hover and exposes the two manipulator arms that are located underneath the AUV. One of the manipulator arms is used to dock the AUV, and the other arm is used for manipulation tasks. The two arms have a combined lifting capacity of 17 kg, which is considerably lower compared with Aquanaut. However, at 4 knots, it is slightly faster. Once the Cuttlefish has taken its vertical position, a human operator takes over control to perform the manipulation task.

Subsea residency

A different proposal is subsea residency. The goal of subsea residency is to extend the duration of the robot’s mission and allow additional activities such as recharging, mission planning, and data exchange.²² Subsea residency requires additional equipment at the seabed, such as docking stations and network cables. However, there is no longer a need for a ship to be in proximity to the robot.

Saab developed an AUV called Sabertooth for inspection and intervention. This AUV can operate valves using a torque tool.⁶³ It was designed to remain at the bottom of the sea for over six months without needing service. After the tasks are completed, the Sabertooth returns to its docking station. The docking station can remain at the bottom for over 5 years.⁶³ The AUV can travel at 5 knots and has a limited range of 50 km. This is considerably shorter than the Aquanauts range; however, owing to its ability to remain at the work site for long periods of time, this is less of a problem. The rated depth of the Sabertooth is 3000 m. This makes it one of the deepest diving AUVs in this review.

Another AUV designed with subsea residency in mind is the Eely500.⁶⁴ This robot is shaped like a snake (or eel as the name suggests) (see Fig. 5). The robot consists of multiple modular units connected by flexible joints. The modules can be battery packs and thrusters and, at each end, a payload module. The composition and length of the AUV can be adjusted according to the customer’s needs. Owing to its narrow shape (20 cm) and flexibility, it is especially helpful in accessing hard-to-reach and confined spaces.⁶⁵ To increase the functionality of the AUV, both ends of the Eely500 can be exchanged at a special tool station. The AUV places each end into a tool spot and grabs the attachment needed for its task. It can switch between inspection, brush, grabber, torque, and cutter tools depending on the mission requirements.⁶⁴ This makes the Eely500 a highly versatile AUV. The navigation from the docking station to the intervention site can be done autonomously; however, manipulation still needs to be performed remotely. The AUV can remain at the bottom for over 6 months and communicates wireless to the docking station. The docking station can be connected to the operator via network cables present at the sea bed or via a communications buoy.

FIG. 5.

Eely500. Reprinted with permission from Eelume AS.⁶⁷

Multi-robot

Operating large and heavy AUVs around offshore structures creates the risk of damaging offshore structures and robots. Furthermore, heavy AUVs require larger ships to be deployed. This reduces the applicability of the AUV. However, precise long-distance travel and heavy lifting require big robots to accommodate navigation equipment, batteries, and sensors. A potential solution is the use of multiple robots. These robots can have different tasks or simply split the load between them.

A concept design to work close to an offshore structure with less risk of damage is the parent–child underwater manipulation system. This system uses two different robots and takes advantage of the best of both worlds. The larger parent robot has precise navigation capabilities and advanced sonar equipment, whereas the child robot is used for light intervention tasks.⁶⁶ The parent robot called Cyclops is a hovering type AUV that carries the child to the maintenance location.⁶⁷ Because the child does not need advanced navigation and communication system, it can be smaller, allowing it to reach confined spaces. Furthermore, the mass of the child robot (8.6 kg) is far smaller compared with the parent robot (210 kg). This lowers the risk of damaging the structure. The parent and child robot are connected through a tether that can be extended up to 20 m.⁶⁶ The tether is naturally buoyant, limiting the drag on the smaller child robot.

Alternatively, a multitude of the same AUV could be used. The TWINBOT concept was designed to divide the load between multiple AUVs.⁶⁸ This not only allows for more complex manipulation tasks but also simplifies the deployment of the AUV. Large robots require specialized ships with strong enough cranes to lift the robots in and out of the water. By splitting the load, smaller, easier deployable AUVs can be used. During the mission, one of the AUVs is assigned the master, and the other robot is the slave. The master robots send commands to the slave through underwater acoustic communication channels.⁶⁸ However, the control is decentralized, meaning each robot can operate independently. This allows continuation to the next checkpoint even if the connection between slave and master is lost. Furthermore, the decentralized control allows the use of a nonhomogeneous fleet of AUVs.⁶⁸ The robots used to test the TWINBOT concept are GIRONA500 AUVs.⁶⁹ The GIRONA500 AUV is one of the earlier developed AUVs with intervention capabilities. Its design allows the AUV to adapt its configuration depending on the mission. The propulsion system, buoyancy, and mission payload can all be adjusted.⁷⁰ In the test scenario, two GIRONA500 AUVs lift a pipe, move it along a path, and return it to its original position. This task was completed successfully, validating the concept.⁶⁸

Humanoid

The Ocean One is a humanoid underwater robot designed to perform discovery and exploration tasks in deep-sea environments that would otherwise be performed by human divers.⁷¹ The Ocean One is roughly the size of a full-grown adult, equipped with two manipulator arms and a head resembling a human shape. This head houses two cameras. The two cameras allow the operator to experience the surroundings of the robot in 3D. Its grippers are soft enough to handle gentle marine life but strong enough to grip handholds. They were developed to operate with the same dexterity as a human hand. The robot receives control inputs from the pilot through a bimanual haptic device. The robot mimics the movement of the teleoperating pilots’ hands, and the pilot receives force feedback. The only input is the movement of the hands; the body of the robot autonomously follows the movement of the hands while avoiding obstacles. The movement of the body is controlled using eight thrusters.

The functionality of the Ocean One was validated with a mission to a sunken 17th-century ship. At a depth of 91 m, the Ocean One retrieved a vase from the ship and brought it to the surface.

All underwater robots mentioned in this chapter are summarized in Table 6.

Discussion

This literature review has identified numerous robots that are capable or have the potential to interact with offshore structures. Although the potential is prominent, there are still challenges to overcome for effective deployment in offshore environments.

Atmospheric environment

Reaching the highest areas of offshore structures is easiest using UAMs. Several UAMs with manipulation capabilities are under development and show promise. However, most of these robots were not designed for the offshore environment. Several of the UAMs mentioned have been tested for disturbances^31,34,38 or designed for environments that have similarities with the offshore environment.^28,29 Nevertheless, a significant number of challenges need to be overcome before these robots can be deployed in an offshore environment.

Disturbances: The biggest challenge for UAMs in the offshore environment is the wind. External disturbance makes it difficult for UAMs to accurately position themselves.³³ Unfortunately, UAMs were only tested in wind speeds up to 1 m/s, which is well below the average wind speed at wind farms in the North Sea (8 m/s).⁶

Flight time: The flight time of the AUMs mentioned in this review is very limited. The manipulators add extra weight which reduces the flight time. None of the UAMs are able to fly for more than 16 minutes.⁷⁴ This would not be sufficient to perform manipulation missions around large offshore structures.

Manipulation: The manipulation capabilities of UAMs are promising. However, the amount of payload that can be carried and the torque generation are still very limited.

Climbing robots are able to climb both wind turbine blades,^40,41 oil and gas rigs,^43,44 and large metal structures in general.⁴² Because of their physical connection to the offshore structure, they are less vulnerable to wind. However, climbing robots are still in the development stage. To advance into a phase that allows more practical use, the following problems need to be solved.

Transportation: The biggest challenge is getting the climbing robots on the offshore structure. The BladeBUG needs to be placed on a wind turbine blade, whereas the wire-driven robot in Lee et al.⁴¹ has to be attached to the nacelle of the turbine. How these robots are transported to that location still needs investigation.

Manipulation: All robots in this article mention the ability to perform physical repair tasks. However, specifics are often limited. The wire-driven robot has good cleaning capabilities, but for most other climbing robots, the manipulation capabilities are still in the early stages.

Multienvironment: The climbing robots can be divided into two categories: smooth surfaces and complex structures. The robot operating in one type of environment is not capable of operating in the other. If the capabilities of both robots could be combined, this would drastically improve the flexibility of the robot.

Energy supply: Weight is an issue for most climbing robots. Carrying a heavy battery requires a stronger adhesion method. There are robots that use a wired connection to an energy supply; however, this limits their freedom of movement.²³

Autonomy: The climbing robots are still dependent on human operators. Improving the autonomous capabilities of the robots would increase their flexibility.

Legged robots are one of the most advanced types of robots in this report. They are capable of working in environments with humans, are being deployed on offshore structures, and can execute certain manipulation tasks autonomously. In order to expand their capabilities, even more some challenges need to be solved.

Autonomy: Legged robots have quite advanced autonomous capabilities, such as autonomously opening doors and picking up an object. Improving these capabilities too, for example, turning valves, would drastically improve the usability of the robots.

Flexibility: Legged robots are excellent at navigating areas that are designed for humans; they can use stairs and open doors. However, legged robots would be even more versatile if they could also climb walls. Hong et al. showed that this is possible, but manipulation capabilities should be added.⁴⁷

An alternative to these legged robots is the Taurob Operator. This tracked robot has an arm capable of lifting up to 75 kg.⁷⁵ It has been tested on an offshore structure, and an inspection version is already commercially available.⁷⁶ As no scientific articles were found on this robot, it was left out of the overview of scientific literature in Table 4.

Table 4.

Summary of Legged Robots

Source	Year	Name	Arm load [kg]	#^a	Nav.	Man.	TRL
^48,49	2021	Spot arm	5	1	A^b	A/R^c	9
^50,77	2021	ANYmal + Kinova Jaco arm	4.4	1	A	R	5
⁵¹	2022	DogGet	n.a.	1	A	A	5

Number of manipulators.

Autonomous.

Remote operated.

Splash zone

A limited number of robots were found that operate in this environment. Crawlers could potentially be used to clean the jackets of offshore structures.^52,53 No USVs that could interact with offshore structures were found in this research.

Maneuverability: Both robots have limited maneuverability but in different domains. Combining the locomotion of both robots allows a single robot to clean multiple jackets, above and below the waterline. This type of robot could service multiple domains of the offshore environment—potentially, from far below the waterline to significant parts above.

Autonomy: Like most robots, the autonomous capabilities are still limited, and both robots mentioned in Table 5 are remotely operated.

Cleaning equipment: The flexibility of the cleaning equipment is still fairly limited. The robotic arm on the ROVMS does provide some flexibility; however, the robots still struggle with complex and unique shapes.

Table 5.

Summary of Robots That Can Operate in the Splash Zone

Source	Year	Name	Adhesion	Locomotion	#^a	Nav.	Man.	TRL
⁵²	2018	n.a.	Magnetic	Wheeled	1	R^b	n.a.^c	7
⁵³	2021	ROVMS	Magnetic	Thrusters	1	R	R	7

Number of manipulators.

Remote operated.

The cavity jet pistol has a fixed position on the side of the robot.

Table 6.

Summary of Underwater Robots

Source	Year	Name	Depth (m)	Shape	#^a	Nav.	Man.	TRL
^58,61	2018	Aquanaut	3000	Torpedo/Humanoid	2	A^b	R^c	9
⁵⁹	2018	—	1500	Torpedo/Isotropic	—	A	—	3
⁶⁰	2022	Cuttlefish	1000	Box	2	A	A	4
⁶³	2011	Sabertooth	3000	Box	1	A	R	9
^64,65	2017	Eely500	500	Snake	1	A	R	8
⁷²	2018	USM	—	Snake	1	A	A	2
^66,67	2023	Cyclops × child	100	Box	1	A	A	7
⁶⁸	2021	TWINBOT	500	Multi-Torpedo	1	A	A	5
^69,70	2011	GIRONA 500	500	Multi-Torpedo	1	A	A	9
^71,73	2017	Ocean One	91^d	Humanoid	2	A	R	8

Number of manipulators.

Autonomous.

Remote operated.

Claims have been made that the robot can dive to 1000 m.

Opportunities lie in the use of robots that are capable of independently moving to and between offshore structures while also being capable of cleaning above the waterline. Potentially, USV could be used to assist when traveling large distances between structures.

Subsea environment

The development of underwater robots with manipulation capabilities is further advanced than most of the other operational domains. Multiple robots are in advanced testing phases or commercially available.^58,63,64 The variety of robots is great; thus, a large number of tasks and scenarios can be performed by these robots. However, to reach their full potential several issues still need to be addressed.

Autonomous manipulation: Most unmanned underwater vehicles (UUVs) in this review have the capability to autonomously navigate. However, for manipulation tasks, they are mostly still dependent on a human operator. Part of this problem is the identification of objects.⁶⁴

Collaboration: The use of UUV could be extended even further if they were able to collaborate with other robots—for example, extending the range using a USV as a long-range carrier.⁵⁴

Components

Every robot in this review needs to move and interact with offshore structures. Therefore, the locomotion mechanism and equipment of the robot are vital. Figure 6 shows that multiple forms of locomotion are being used by robots. This allows them to operate in multiple environments. Further improvement may be found in the use of nature-inspired locomotion system. This technology allows for unprecedented maneuverability and durability.⁷⁸

FIG. 6.

Overview of the locomotion mechanisms and equipment used by the robots in this review.

The types of equipment are less diverse. Most robots are equipped with some sort of gripper. Grippers can perform a wide variety of tasks making them popular. However, the capabilities of grippers can vary greatly; from simple single hinge grippers to grippers with high dexterity and haptic feedback.⁷¹ New possibility arises when using soft robotic grippers. They allow for improved dexterity and are less affected by the pressure at great depths.⁷⁹

Conclusion

Overall, there is significant progress in the development of robots that are capable of interacting with offshore structures. Several robots are already commercially available or in advanced test phases. The main opportunities lie in enhancing the autonomy and flexibility of robots to be able to interact with the offshore structures.

The flexibility could be improved using a multi-environment robot. For example, a robot with thrusters and magnetic wheels could operate not only underwater but also in the splash zone and above. In addition, this would allow the robot to navigate between multiple structures. Alternatively, a multi-robot could be used to operate in multiple environments. The heavier parent could supply long-range travel and underwater operations, whereas the lighter child robot operates above the surface. Combining the strengths of multiple robots creates a system that pushes the innovation of robots that can interact with offshore structures. Furthermore, the flexibility of the grippers has a large opportunity for improvement. Most robots are equipped with grippers. These grippers are often basic and rigid in the context of offshore applications. Research in the use of soft robotics and grippers with enhanced dexterity would boost innovation. Autonomous operation is restricted by the lack of autonomous manipulation. Research should address the lack of object identification to reduce the need for human operators.

Focusing future research on these issues allows robots to enhance the safety, efficiency, and cost-effectiveness of IMR tasks in the offshore environment.

Footnotes

Abbreviations Used

References

IRENA. Renewable energy statistics 2023. International Renewable Energy Agency, 2023.

Williams

, Zhao

, Lee

. Global offshore wind report 2022. GWEC; 2022.

Florence

. Offshore technology. Report: Offshore oil and gas set for highest growth in a decade, March 2023. Available from: https://www.offshore-technology.com/news/offshore-oil-gas-highest-growth-decade/

NASA earth observatory. Kashagan oil and gas field, Kazakhstan. NASA Earth Observatory, Oct 2015 Available from: https://earthobservatory.nasa.gov/images/86890/kashagan-oil-and-gas-field-kazakhstan.

CSSC. CSSC Haizhuang h260-18mw offshore wind turbine giant emerges, 2023. Available from: http://cssc-hz.com/?en/enNews/NewsReleases/148.html

Ronda

, Wijnant

, Stepek

. Inter-annual wind speed variability on the North Sea. KNMI, 2017.

Da Costa Siqueira

. Modeling of wind phenomena and analysis of their effects on UAV trajectory tracking performance. West Virginia University, 2017; doi: 10.33915/etd.7347

Kaiser

. The impact of extreme weather on offshore production in the Gulf of Mexico. Applied Mathematical Modelling, 2008; 32(10):1996–2018; doi: 10.1016/j.apm.2007.06.031

Kaplan

, Silbert

. Impact forces on platform horizontal members in the splash zone. OTC Offshore Technology Conference, All Days: OTC–2498–MS, May 1976; doi: 10.4043/2498-ms

10.

El-Reedy

. Chapter 8—Risk-based inspection technique. In Offshore Structures. Gulf Professional Publishing, 2012. pp. 563–634; doi: 10.1016/B978-0-12-385475-9.00008-0

11.

Offshore technology. Petronius Field Project, Gulf of Mexico, Jun 2000. Available from: https://www.offshore-technology.com/projects/petronius/

12.

Shell global. Perdido, 2023 Available from: https://www.shell.com/about-us/major-projects/perdido.html.

13.

SSE. World’s deepest offshore wind turbine foundation installed in Scottish waters, 2023 Available from: https://www.sse.com/news-and-views/2023/04/world-s-deepest-offshore-wind-turbine-foundation-is-installed-in-scottish-waters/#:: text=The%20world’s%20deepest%20wind%20 turbine,off%20the%20coast%20of%20Angus.

14.

Equinor. Hywind Scotland, 2017 Available from: https://www.equinor.com/energy/hywind-scotland.

15.

Snelgrove

. Ecosystems of the deep oceans (ecosystems of the world 28). EoS Transactions, 2003; 84(39):398–398; doi: 10.1029/2003EO390014

16.

Pranckutė

. Web of science (WOS) and Scopus: The titans of bibliographic information in today’s academic world. Publications, 2021; 9(1):12–12; doi: 10.3390/publications9010012

17.

Mitchell

, Blanche

, Harper

, et al. A review: Challenges and opportunities for artificial intelligence and robotics in the offshore wind sector. Energy and AI, 2022; 8:100146–100146; doi: 10.1016/j.egyai.2022.100146

18.

Bogue

. Robots in the offshore oil and gas industries: A review of recent developments. IR, 2019; 47(1):1–6; doi: 10.1108/IR-10-2019-0207

19.

Liu

, Hajj

, Bao

. Review of robot-based damage assessment for offshore wind turbines. Renewable and Sustainable Energy Reviews, 2022; 158:112187; doi: 10.1016/j.rser.2022.112187

20.

Enrica

, Bibuli

, Nikola

, et al. Challenges and future trends in marine robotics. Annual Reviews in Control, 2018; 46:350–368; do: 10.1016/j.arcontrol.2018.10.002

21.

Ruggiero

, Lippiello

, Ollero

. Aerial manipulation: A literature review. IEEE Robot Autom Lett, 2018; 3(3):1957–1964; Doi: 10.1109/lra.2018.2808541

22.

Song

, Marburg

, Manalang

. Resident subsea robotic systems: A review. Mar Technol Soc j, 2020; 54(5):21–31; do: 10.4031/MTSJ.54.5.4

23.

Bogue

. Climbing robots: Recent research and emerging applications. IR, 2019; 46(6):721–727; doi: 10.1108/IR-08-2019-0154

24.

Fang

, Cheng

. Advances in climbing robots for vertical structures in the past decade: A review. Biomimetics, 2023; 8(1):47–47; doi: 10.3390/biomimetics8010047

25.

Tecchio

, Vieira

, Batista

, et al. Overview of robotic applications on offshore OGI IMR: Splash zone. OCEANS 2021: San Diego—Porto, Sep 2021; doi: 10.23919/oceans44145.2021.9706133

26.

Gong

, Sun

, Nair

, et al. Legged robots for object manipulation: A review. Front Mech Eng, 2023; 9; doi: 10.3389/fmech.2023.1142421

27.

Mankins

John C

. Technology Readiness Levels—A White Paper. Office of Space Access and Technology, NASA, April 1995.

28.

Sanchez-Cuevas

, Gonzalez-Morgado

, Cortes

, et al. Fully-actuated aerial manipulator for infrastructure contact inspection: Design, modeling, localization, and control. Sensors (Basel), 2020; 20(17):4708–4708; doi: 10.3390/s20174708

29.

Trujillo

, Martínez de

, Martin

, et al. Novel aerial manipulator for accurate and robust industrial NDT contact inspection: A new tool for the oil and gas inspection industry. Sensors (Basel), 2019; 19(6):1305–1305; doi: 10.3390/s19061305

30.

Orsag

, Korpela

, Bogdan

, et al. Valve turning using a dual-arm aerial manipulator. 2014 International Conference on Unmanned Aircraft Systems (ICUAS), May 2014; doi: 10.1109/icuas.2014.6842330

31.

Heredia

, Jimenez-Cano

, Sánchez

, et al. Control of a multirotor outdoor aerial manipulator. 2014 IEEE/RSJ International Conference on Intelligent Robots and Systems, Sep 2014; doi: 10.1109/iros.2014.6943038

32.

Suarez

, Ramon Soria

, Heredia

, et al. Anthropomorphic, compliant and lightweight dual arm system for aerial manipulation. 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), September 2017; doi: 10.1109/iros.2017.8202266

33.

Chen

, Zhan

, He

, et al. Robust control for unmanned aerial manipulator under disturbances. IEEE Access, 2020; 8:129869–129877; doi: 10.1109/access.2020.3008971

34.

Sarkisov

, Kim

, Bicego

, et al. Development of SAM: Cable-suspended aerial manipulator. 2019 International Conference on Robotics and Automation (ICRA), May 2019; doi: 10.1109/ICRA.2019.8793592

35.

Shimahara

, Leewiwatwong

, Ladig

, et al. Aerial torsional manipulation employing multi-rotor flying robot. 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), October 2016; doi: 10.1109/IROS.2016.7759258

36.

Zhao

, Anzai

, Shi

, et al. Design, modeling, and control of an aerial robot dragon: A dual-rotor-embedded multilink robot with the ability of multi-degree-of-freedom aerial transformation. IEEE Robot Autom Lett, 2018; 3(2):1176–1183; doi: 10.1109/lra.2018.2793344

37.

Zhao

, Nagato

, Okada

, et al. Forceful valve manipulation with arbitrary direction by articulated aerial robot equipped with thrust vectoring apparatus. IEEE Robot Autom Lett, 2022; 7(2):4893–4900; doi: 10.1109/lra.2022.3154018

38.

Chermprayong

, Zhang

, Xiao

, et al. An integrated delta manipulator for aerial repair: A new aerial robotic system. IEEE Robot Automat Mag, 2019; 26(1):54–66.

39.

Hernández

, Altuzarra

, Salgado

, et al. Designing parallel manipulators: From specifications to a real prototype. Industrial Robot, 2012; 39(5):500–512; doi: 10.1108/01439911211249797

40.

Bladebug: Advanced robotics for turbine maintenance. 2014 Available from: Bladebug.co.uk.

41.

Lee

, Oh

, Son

. Wire-driven parallel robotic system and its control for maintenance of offshore wind turbines. 2016 IEEE International Conference on Robotics and Automation (ICRA), May 2016; doi: 10.1109/icra.2016.7487221

42.

Eto

, Asada

. Development of a wheeled wall-climbing robot with a shape-adaptive magnetic adhesion mechanism. 2020 IEEE International Conference on Robotics and Automation (ICRA), May 2020; doi: 10.1109/icra40945.2020.9196919

43.

Guan

, Jiang

, Zhu

, et al. Climbot: A bio-inspired modular biped climbing robot—system development, climbing gaits, and experiments. ASME. J. Mechanisms and Robotics, 2016; 8(2); doi: 10.1115/1.4028683

44.

Chang

, Luo

, Qiao

, et al. Design and motion planning of a biped climbing robot with redundant manipulator. Applied Sciences, 2019; 9(15):3009–3009; doi: 10.3390/app9153009

45.

Maslin

. Offshore Engineer. ROBOTICS: Meet Your New Offshore Robotic Co-workers; Charles, Eddie, ANYmal & Spot, August 2021 Available from: https://www.oedigital.com/news/489613-robotics-meet-your-new-offshore-robotic-co-workers-charles-eddie-anymal-spot.

46.

Gehring

, Fankhauser

, Isler

, et al. Anymal in the field: Solving industrial inspection of an offshore HVDC platform with a quadrupedal robot. In Field and Service Robotics. ( Ishigami

, Yoshida

, eds.) Springer Singapore: Singapore; 2021. pp. 247–260.

47.

Hong

, Um

, Park

, et al. Agile and versatile climbing on ferromagnetic surfaces with a quadrupedal robot. Sci Robot, 2022; 7(73):eadd1017; doi: 10.1126/scirobotics.add1017

48.

Boston Dynamics. Spot arm Available from: https://bostondynamics.com/wp-content/uploads/2020/10/spot-arm.pdf.

49.

Bouman

, Ginting

, Alatur

, et al. Autonomous spot: Long-range autonomous exploration of extreme environments with legged locomotion. 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), October 2020; doi: 10.1109/iros45743.2020.9341361

50.

Ferrolho

, Merkt

, Ivan

, et al. Optimizing dynamic trajectories for robustness to disturbances using polytopic projections. 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), October 2020; doi: 10.1109/iros45743.2020.9341788

51.

Linghan

, Cong

, Shengguo

, et al. DogGet: A legged manipulation system in human environments. 2022 12th International Conference on CYBER Technology in Automation, Control, and Intelligent Systems (CYBER), July 2022; doi: 10.1109/cyber55403.2022.9907688

52.

Fan

, Yang

, Chen

, et al. An underwater robot with self-adaption mechanism for cleaning steel pipes with variable diameters. IR, 2018; 45(2):193–205; doi: 10.1108/IR-09-2017-0168

53.

, Li

, Wang

, et al. Optimization of tool orientation for improving the cleaning efficiency of offshore jacket-cleaning systems. Applied Ocean Research, 2021; 112:102687–102687; doi: 10.1016/j.apor.2021.102687

54.

Lachaud

, Monbeig

, Nolleau

, et al. Opportunities and challenges of remote operating a ROV embarked on a USV. 2018 OTC Offshore Technology Conference, April 2018; doi: 10.4043/29000-ms

55.

Moles

, Ladreux

, Karl

. Operational adoption challenges to USV/ROV services. 2019 OTC Offshore Technology Conference, April 2019; doi: 10.4043/29569-MS

56.

Peng

, Chen

, Lu

, et al. Design and application of a 3-DoF manipulator for launch and recovery system. 2018 International Conference on Reconfigurable Mechanisms and Robots (ReMAR), June 2018; doi: 10.1109/remar.2018.8449869

57.

Jiang

, Jovan

, Moradi

, et al. A multirobot system for autonomous deployment and recovery of a blade crawler for operations and maintenance of offshore wind turbine blades. Journal of Field Robotics, 2022; 40(1):73–93; doi: /10.1002/rob.22117

58.

Manley

, Halpin

, Radford

, et al. Aquanaut: A new tool for subsea inspection and intervention. OCEANS 2018 MTS/IEEE Charleston, October 2018; doi: 10.1109/oceans.2018.8604508

59.

Pagliai

, Ridolfi

, Gelli

, et al. Design of a reconfigurable autonomous underwater vehicle for offshore platform monitoring and intervention. 2018 IEEE/OES Autonomous Underwater Vehicle Workshop (AUV), November 2018; doi: 10.1109/auv.2018.8729776

60.

Christensen

, Hilljegerdes

, Zipper

, et al. The Hydrobatic Dual-Arm Intervention AUV Cuttlefish. OCEANS 2022, Hampton Roads, October 2022; doi: 10.1109/oceans47191.2022.9977150

61.

Nauticus robotics. Aquanaut—the future of underwater robotics: Specification Available from: https://nauticusrobotics.com/wp-content/uploads/2023/05/Aquanaut-Product-Sheet-v7.pdf.

62.

Oitzman

. The Robot Report. Nauticus begins commissioning of Mark 2 Aquanaut, April 2023 Available from: https://www.therobotreport.com/nauticus-begins-commissioning-of-mark-2-aquanaut/.

63.

Johansson

, Siesjo

, Furuholmen

. Seaeye Sabertooth, A Hybrid AUV/ROV Offshore System. SPE Offshore Europe Oil and Gas Conference and Exhibition, September 2011; doi: 10.2118/146121-ms

64.

Liljeback

, Mills

. Eelume: A flexible and subsea resident IMR vehicle. OCEANS 2017—Aberdeen, June 2017; doi: 10.1109/oceanse.2017.8084826

65.

Eelume

. Eelume—Subsea Intervention Available from: https://eelume.com/.

66.

Kim

, Kim

, Song

, et al. Parent-child underwater robot-based manipulation system for underwater structure maintenance. Control Engineering Practice, 2023; 134:105459–105459; doi: 10.1016/j.conengprac.2023.105459

67.

Pyo

, Cho

, Joe

, et al. Development of hovering type AUV “Cyclops” and its performance evaluation using image mosaicing. Ocean Engineering, 2015; 109:517–530; doi: 10.1016/j.oceaneng.2015.09.023

68.

, Cieslak

, Ridao

, et al. TWINBOT: Autonomous Underwater Cooperative Transportation. IEEE Access, 2021; 9:37668–37684; doi: 10.1109/access.2021.3063669

69.

Ribas

, Palomeras

, Ridao

, et al. Girona 500 AUV: From Survey to Intervention. IEEE/ASME Trans Mechatron, 2012; 17(1):46–53; doi: 10.1109/TMECH.2011.2174065

70.

Ribas

, Ridao

, Magí

, et al. The Girona 500, a multipurpose autonomous underwater vehicle. OCEANS 2011 IEEE—Spain, June 2011; doi: 10.1109/oceans-spain.2011.6003395

71.

Khatib

, Yeh

, Brantner

, et al. Ocean one: A robotic avatar for oceanic discovery. IEEE Robot Automat Mag, 2016; 23(4):20–29; doi: 10.1109/mra.2016.2613281

72.

Sverdrup-Thygeson

, Kelasidi

, Pettersen

, et al. The underwater swimming manipulator—a bioinspired solution for subsea operations. IEEE J Oceanic Eng, 2018; 43(2):402–417; doi: 10.1109/joe.2017.2768108

73.

Brantner

, Khatib

. Controlling ocean one. Springer Proceedings in Advanced Robotics, November 2017. pp. 3–17; doi: 10.1007/978-3-319-67361-5_1

74.

DJI Official. Matrice 100 specs, 2016 Available from: https://www.dji.com/nl/matrice100.

75.

Taurob a Dietsmann Group Company. The Taurob Operator Available from: https://taurob.com/operator/#operator-specs.

76.

Taurob a Dietsmann Group Company. The Inspector robot Available from: https://taurob.com/inspector/.

77.

Campeau-Lecours

, Lamontagne

, Latour

, et al. Kinova modular robot arms for service robotics applications. International Journal of Robotics Applications and Technologies, 2017; 5(2):49–71; doi: 10.4018/IJRAT.2017070104

78.

Aracri

, Giorgio-Serchi

, Suaria

, et al. Soft robots for ocean exploration and offshore operations: A perspective. Soft Robot, 2021; 8(6):625–639; doi: 10.1089/soro.2020.0011and

79.

Jovanova

, Var

SCS

. Design of a soft underwater gripper with SMA actuation. In Proceedings of ASME 2023 Conference on Smart Materials, Adaptive Structures and Intelligent Systems, SMASIS 2023. American Society of Mechanical Engineers, 2023; doi: 10.1115/SMASIS2023-111702