**1. Introduction**

Natural or artificial shallow water environments, such as harbors, coastal areas, waterways, lakes and lagoons, are in general affected by anthropogenic pressures. For this reason, they would require periodic monitoring, to mitigate the effects of environmental crises caused by human activity or natural processes. However, to date, geophysical studies in shallow water areas (shallower than a few meters) are not a consolidated practice for various reasons, including the following: they present difficult access, even using small boats, in absence of accurate bathymetric maps, the shallow water represents an efficient waveguide for acoustic and ultrasonic noises that limits penetration of the signals into the substrate and the quality of echographic and seismic data, the effect of noise due to propellers, or other natural and artificial causes, is amplified, and the rapidity of environmental changes would require repeated investigations (4D), which is not economically viable with conventional methods.

The economic and social importance of shallow water environments, therefore, calls for the development of new technologies and methods, that could open their study to a wider range of researchers and environmental protection agencies: progresses and developments in the field of marine robotics could be an interesting opportunity to achieve this goal. In fact, the relatively recent availability of miniaturized although accurate sensors, as well as the development of innovative hardware architectures (Arduino ®, Raspberry ™, etc.) simplify design and implementation of low cost but highly performing Autonomous Surface Vehicles (ASV), which can operate in a variety of aquatic environments. This is the case of SWAP (Shallow Water Autonomous Prospector) a class of vehicles developed by ISMAR-CNR and Proambiente Scrl, characterized by limited size, high versatility, and low cost, and operating with a variety of di fferent payloads. OpenSWAP is the follow-up of that original project, developed under NAIADI (New Autonomous/automatIc systems for the study AnD monitoring of aquatic envIronments, https://www.consorzioproambiente.it/en/projects/terminati/51- naiadi-por-fesr-2014-2020), in the frame of a POR-FESR Emilia Romagna initiative [1]. The intensive use of "open" technologies and software packages for data acquisition and processing [2–4], as well as the low cost of production, have the potential to extend the use of these techniques and methods to a growing public of scientists studying geological processes in these rapidly changing environments. Although to date several ASV are available on the market, we believe that OpenSwap is innovative in many respects, as summarized in Table 1.



Together with a technical description of the vehicles, we present here some examples of data acquisition, which include single-beam echograms, side-scan sonar images, seismic reflection profiles, as well as multibeam data from di fferent shallow water environments.

#### **2. OpenSWAP, Philosophy and Motivations**

OpenSWAP is a class of ASVs (autonomous surface vehicles) developed with the aim of providing flexible and easily operating autonomous aquatic vehicles (Figure 1), from both hardware and software point of view, allowing to perform data acquisition in the shallow water environment.

**Figure 1.** An OpenSWAP vehicle during a test phase.

Although these ASVs are suitable for a number of di fferent payloads, including video cameras, current-meters, chemical and physical water sensors, water samplers, etc., we focused their design on acquisition of marine bathymetric and seismo-stratigraphic data. Our first target was implementing the ability of performing repeated surveys, i.e., following navigation paths with centimeters accuracy during subsequent runs. In fact, such performances are mandatory to analyze and monitor time-variant environmental processes and variables. Another functionality considered important was the possibility of planning in advance every practical aspect of a survey, such as: type of sensor employed and coverage to be obtained in a study area; routes to be followed by the vehicle during the run, avoiding obstacles; acquisition parameters (data sampling rate, maximum depths, etc.); time needed to complete the survey, in relation to batteries duration. This would allow for optimizing workflows in logistically di fficult environments, reducing risks and deployment time. The development of the OpenSWAP vehicles was focused from the very beginning considering implementation of two embedded geophysical sensors: (1) a single-beam echosounder (SBE), to perform bathymetric (repeated) surveys, and (2) a chirped sub-bottom profiler (SBP), allowing for the acquisition of high-resolution stratigraphic data.

The vehicle was designed small enough to be easily transported and deployed, also in di fficult conditions (steep shores; absence of docks; waters too shallow; etc.), but large enough to host suitable payloads and batteries. In a short summary, the design of the ASV was performed considering the following characteristics: housing as many batteries as possible, in relation to floating and navigation performances, to ensure long, self-running surveys with minimum returns to the base point; hosting in a safe waterproof container the on-board electronics; gathering implementation of di fferent propulsion systems (i.e., waterborne vs. aerial propellers); providing additional space for hosting most common third-party sensors. Other features considered important were: low acoustic noise and low water turbulence induced by propellers and hulls close to the acoustic sensors; compatibility with AI (artificial intelligence) stay-on-route algorithms, to gather repeated surveys with centimeters errors, also in presence of water and air turbulence caused by waves, currents, and/or strong wind under acceptable weather and sea conditions. During our tests, we have found that navigation paths are followed up to Beaufort 4, but data quality under such conditions is very poor. We conclude that Beaufort 3 would be the limit for single-beam acoustic surveys.

The final version of the vehicles should have been easily customizable, allowing for installation of additional proprietary sensors, such as commercially available multibeam echosounders (MBES), SBP, water samplers etc.

To obtain such performances, the main electronic board has been developed to include several low-level serial RS232 I/O ports, as well as higher level network supports (i.e., Ethernet, Wi-Fi etc.), allowing for interfacing any proprietary instruments with the in-house positioning systems in the frame of a local-area network architecture.

The software and firmware routines were developed in form of separate modules, to allow the end-users implementing new functionalities through the use of internal scripting. This feature was improved to enable low-level software customizations, including new navigation algorithms, to enhance, for example, navigation accuracy under special condition/environment, or to adapt the current course to incoming parameters (wind speed and direction, wave height, etc.). Finally, the choice of an open hardware architecture for the electronics allowed to reduce costs, favoring distribution of those parts not available on the consumer market, also in the form of self-assembled kits (see next section).

#### **3. Design of the Vehicles**

#### *3.1. Nautical Aspects*

For several reasons, we decided to base the design of the vehicles on a multi hull scheme, and in particular a catamaran. In fact, compared to mono hull boats, catamarans are characterized by: wider beams, which increase stability; shallower draught, a major plus in shallow waters; smaller hydrodynamic resistance, implying less power delivered to the propellers. The design was carried

out also considering some specific requirements of geophysical surveys, which employ acoustic and ultrasonic signals generated and received by transducers that should be perfectly coupled with the water. For these reasons, the vehicles would have been able to minimize the tradeo ff between "speed of the vehicle", which influences the time required to complete a survey, and "low acoustic noise in the water", caused by turbulence generated by propellers and viscous drags along the hull. The first technical solution to limit this negative e ffect, particularly evident in shallow waters, was installing the acoustic sensors between the two hulls of the catamaran, in an area characterized by relatively low water turbulence and shielded by the bubble carpet produced by propellers.

The main task of the OpenSWAP vehicles is performing geophysical surveys for study morphology and stratigraphy of the seafloor and penetrate the first tens of meters in unconsolidated or poorly consolidated sediments. For such objectives, acoustic and ultrasonic transducers are generally used, with alternating emission and detection of acoustic signal towards the bottom and below. Two instruments constitute very basic sensors for such surveys: the SBE (101–102 kHz), for accurate determination of depth and bottom reflectivity; the SBP, generating lower frequency (100–101 kHz) impulsive or frequency-modulated (chirped) signals, penetrating the subbottom and being reflected by acoustic impedance contrasts.

These instruments, particularly the SBP, are very sensitive to acoustic and electric noise, as well as to the presence of air bubbles in the water originated by turbulence close to the hulls or by propellers. To minimize such e ffects, we developed a catamaran with asymmetric hulls. Once designed, the fluid dynamic behavior was CFD (computational fluid dynamics) simulated, to iteratively optimize their shapes and to determine positions a ffected by minimum noise. As shown in Figure 2, a CDF plot of turbulence simulated at 7 km/h (3.7 knots) speed, the region between the hulls is very silent, and hence the best place for deploying the acoustic transducers. At higher speeds, noise slightly increases, but the region between the hulls remains the most favorable.

**Figure 2.** CFD simulation of the hulls behavior during the design phase of the vehicle at 7 km/h of speed. Note minimum turbulence between the two hulls.

The drawback of this design is that a catamaran with asymmetric hulls shows a slight stronger friction relative to the symmetric ones, but a very low noise between the hulls (Figure 2).

The design of hulls and vessel containing the electronics was carried out through the implementation of di fferent standalone vehicles, whose tests carried out during di fferent acquisition trials led to a final design implementing an electric powered plastic catamaran, made by linear low-density polyethylene (LLDPE) with two asymmetrical lateral hulls and a central case housing the electronics (Figure 3). The frame is made of aluminum profiles, which provide connection between different parts of the vehicle and could be used as supports to deploy other instruments and sensors within the low noise area between the hulls.
