Modular robotics for ocean assessment - A research programme at INESC TEC

Anibal Matos

  (1)INESC TEC & Faculty of Engineering of the University of Porto

Nuno Cruz

  (2)INESC TEC & Faculty of Engineering of the University of Porto

The adoption of autonomous robotic platforms for the exploration and exploitation of the ocean has been steadily increasing throughout the last decades. The benefits of these platforms in terms of costs reduction, gains in amount and quality of gathered data, time and space extension of operations, or the increase of operations safety are already field validated, highly contributing to the maturity of this technology.

There are already multiple commercial solutions, addressing different applications and several operational scenarios. Hundreds of end-users worldwide are currently employing and benefiting from platforms. At the same time, they bring novel requirements and opportunities pulling new developments and novel research questions.

The continuous progress of autonomous underwater robots mainly lies on the innovation results on the underlying technologies. Major advances in electronics (miniaturisation, power efficiency, etc.), batteries (increase of energy to volume or energy to mass ratios), perception sensors (increased resolution and sensitivity in digital cameras, higher resolution and frequency bands in hyperspectral cameras, and better resolution in sonar systems.), or navigation sensors (improved accuracy of inertial systems and acoustic based speed sensors), among others, are at the basis of the ever increasing capabilities of these platforms.

But taking fully advantage of such underlying technologies is not a trivial task. It requires extensive and in-depth research leading to the proposal of novel solutions to fundamental robotic operation questions: navigation – determination of position, attitude, and velocity of the robot; guidance – definition of a plan towards the attainment of a mission goal; and control – implementation of actions defined in the plan. Additionally, as these platforms have specific mission goals, research efforts also address issues related to the acquisition of payload data or with end-effector operations.

Furthermore, as oceanic environments are harsh and not yet fully characterised, all these research efforts require experiments, ultimately in real operational scenarios.

Considering these aspects, a sustainable research programme in underwater robotics must not only address the more fundamental research questions, that naturally evolve with time and are partially driven by the available underlying technologies, but also to be supported in an experimental setup that enables a systematic validation of proposed solutions.

AUV Modules

The roots of these developments emerged during the 1990s, when a research group explored the development of control algorithms and acoustic tracking systems for partner institutions' autonomous underwater vehicles (AUVs). The experience gained from utilising these vehicles played a crucial role in identifying key characteristics for a new generation of AUVs. Firstly, it was essential for the vehicle to possess a highly modular design, enabling easy reconfiguration (such as sensor swapping or addition) and independent subsystem development. Additionally, a streamlined construction was necessary to minimise drag in the desired direction of movement. Size and weight reduction were also prioritised to facilitate deployment and recovery from smaller support boats, eliminating the need for complex and costly logistics.

Regarding manoeuvrability, conventional portable AUVs featured a torpedo-shaped body with rear propellers and control surfaces (fins) for diving and steering. Although efficient during motion, these systems required a minimum forward velocity for the control surfaces to be effective. Consequently, people realised that this new generation of vehicles should possess full attitude control, even at zero velocity, allowing for unrestricted approaches to underwater targets.

The development of these small-sized AUVs relied on the use of modular building blocks. This modularity encompassed various aspects, including hardware construction, electronic subsystems, software, and control. The most clear sign of modularity in these building blocks stemmed from the design of the hull sections. To assemble AUVs using modular components and achieve a seamless overall profile, the sections were designed with matching edges and consistent cross sections, featuring an outer diameter of 200mm. This dimension struck a balance between a compact size for manageability and enough space to accommodate dry compartments for electronics and a diverse range of wet sensors. Additionally, a variety of materials, such as plastics and common metals, available off the shelf as rods or tubes, could be machined to meet this dimension.

The modules were engineered with an identical male/female coupling system, secured in place by evenly distributed radial set screws along the perimeter to allow rotation. They were divided into three classes: 1) Dry compartments, which were cylindrical enclosures designed to withstand hydrostatic pressure and featured connectors in the end caps; 2) Wet extensions, consisting of water-filled cylindrical sections with a male termination at one end and a female termination at the other, facilitating stacking, interchangeability, or direct attachment to the dry compartments; 3) End sections, encompassing water-filled nose cones and tails that formed the bow and stern of the AUV, accommodating various configurations of propulsion thrusters.

Modular AUVs

Using these building blocks, the initial vehicle was the MARES AUV (Modular Autonomous Robot for Environment Sampling), a portable AUV capable of hovering. Since its inception, in 2007, the MARES AUV has undergone continuous updates and has been deployed in various configurations in the field. The primary objective behind its design was to create an open architecture system for conducting underwater robotics research, specifically tailored for operations in coastal waters, with the ability to hover within the water column. The MARES AUV consisted of a single dry compartment, measuring 60cm in length, housing all electronic boards, while the batteries were positioned at the bottom to lower the centre of gravity. Attached to each side of the dry compartment were the wet extensions, which incorporated payload sensors, communication devices, and through-hull thrusters, enabling independent control of vertical velocity and pitch angle. This setup allowed for operations in highly confined areas, with virtually separate horizontal and vertical motion even at 0 m/s velocities. This unique feature distinguished the MARES AUV from other AUVs of similar size and weight: depending on configuration, MARES versions ranged from 1.5 to 2 meters in length, with 35 to 45kg of dry mass. Furthermore, the system’s modularity facilitated the integration of additional sensors or thrusters, such as those used for sway control. Additionally, a module with vertical and horizontal fins could be added to enable more traditional control modes.

Over the course of 15 years, the adaptable MARES architecture has been instrumental in collecting oceanic data across diverse application scenarios, spanning from pollution monitoring to traditional and cutting-edge mapping techniques. These endeavours have taken place not only in Portugal, but also internationally. Simultaneously, this programme has provided support for the experimental validation of outcomes derived from PhD theses and MSc dissertations over the past decade.

In 2011, a significant milestone was achieved in testing the maturity and versatility of the system components with the development of TriMARES. TriMARES, a 75kg, 3-body system, was designed to meet the specific requirements of a Brazilian consortium seeking a hybrid ROV/AUV solution. This innovative system incorporated high-definition cameras and sonars, enabling efficient inspections of hydroelectric dams. Leveraging the modularity of the MARES architecture and the simplicity of the building blocks, the development of TriMARES was successfully completed and transported to Brazil within an impressively short timeframe of just over six months.

Subsequently, in 2017, another significant milestone was accomplished through the development of DART (Deep Autonomous Robotic Traveler), a portable AUV designed for operations up to 4000 meters deep. A notable breakthrough was replacing the original dry compartment, constructed from polyacetal copolymer (POM), with a pressure housing composed of borosilicate glass, carefully adapted to align with the coupling of the wet extensions. Additionally, all sensors and actuators that came into direct contact with water were substituted with versions capable of withstanding pressures of at least 400 bar. This vehicle played a pivotal role within the system and notably contributed to the successful participation of the Shell Ocean Discovery XPrize, resulting in the Portuguese team receiving a share of the $1 million prize awarded to the finalists of the competition.

Expanding the Uses of AUVs

One prevalent constraint of AUVs is the limited onboard energy, which imposes a maximum endurance for the vehicle. Another limitation pertains to communication, as electromagnetic signals experience significant attenuation in saltwater, effectively blocking RF-based technologies. To overcome these limitations, the concept of underwater docking stations has gained popularity. These stations function as underwater garages, providing mechanical protection against the elements while also serving as wireless charging stations and facilitating short-range, high-data-rate communications for AUVs. INESC TEC has conducted extensive testing of diverse docking station versions, including both the conventional "funnel-shaped" design and the innovative "cradle" configuration, capitalising on the hovering capability. These enable secure docking and recharging for AUVs assembled using the building blocks. In fact, the inherent modularity of the building blocks provides significant advantages in accommodating independent developments that evolve at different paces. In the context of these docking stations, bespoke modules were specifically designed and validated to seamlessly integrate close-range communication devices, underwater structure perception sensors, and wireless power transfer functionalities.

Within a broader scope, INESC TEC is leveraging the modularity of the vehicles and docking stations to validate an innovative ocean observation paradigm. This approach involves combining the mobility of the vehicles with the advanced capabilities of a new generation of "smart cables." These enhanced fibre optic communication cables span across the oceans, equipped with distributed sensors that enable ocean observations on a basin scale. Under this new paradigm, the repeaters positioned along the communication cables serve as anchors for local observatories and function as "service stations" for AUVs. These repeaters not only provide energy to recharge the AUVs' batteries but also serve as information gateways, facilitating internet access for data exchange. This integration of modular vehicles, docking stations, and smart cables establishes a symbiotic relationship that promises to revolutionise ocean observation capabilities on a global scale.