Prototype Design and Experimental Evaluation of Autonomous Collaborative Communication System for Emerging Maritime Use Cases

sensors Article Prototype Design and Experimental Evaluation of Autonomous Collaborative Communication System for Emerging Maritime Use Cases Jiri Pokorny 1,2,* , Khanh Ma 2, Salwa Saafi 1,2 , Jakub Frolka 1, Jose Villa 3, Mikhail Gerasimenko 2, Yevgeni Koucheryavy 2and Jiri Hosek 1   Citation: Pokorny, J.; Ma, K.; Saafi, S.; Frolka, J.; Villa, J.; Gerasimenko, M.; Koucherayvy, Y.; Hosek, J. Prototype Design and Experimental Evaluation of Autonomous Collaborative Communication System for Emerging Maritime Use Cases. Sensors 2021,21, 3871. https://doi.org/10.3390/ s21113871 Academic Editor: Andrey V. Savkin Received: 30 April 2021 Accepted: 29 May 2021 Published: 3 June 2021 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). 1 Department of Telecommunications, Faculty of Electrical Engineering and Communication, Brno University of Technology, Technicka 12, 616 00 Brno, Czech Republic; [email protected].cz (S.S.); [email protected].cz (J.F.); [email protected].cz (J.H.) 2Unit of Electrical Engineering, Tampere University, Korkeakoulunkatu 7, 337 20 Tampere, Finland; [email protected] (K.M.); [email protected] (M.G.); evgeny[email protected] (Y.K.) 3Mechatronics Research Group, Tampere University, Korkeakoulunkatu 6, 337 20 Tampere, Finland; [email protected] *Correspondence: [email protected] Abstract: Automated systems have been seamlessly integrated into several industries as part of their industrial automation processes. Employing automated systems, such as autonomous vehicles, allows industries to increase productivity, benefit from a wide range of technologies and capabilities, and improve workplace safety. So far, most of the existing systems consider utilizing one type of autonomous vehicle. In this work, we propose a collaboration of different types of unmanned vehicles in maritime offshore scenarios. Providing high capacity, extended coverage, and better quality of services, autonomous collaborative systems can enable emerging maritime use cases, such as remote monitoring and navigation assistance. Motivated by these potential benefits, we propose the deployment of an Unmanned Surface Vehicle (USV) and an Unmanned Aerial Vehicle (UAV) in an autonomous collaborative communication system. Specifically, we design high-speed, directional communication links between a terrestrial control station and the two unmanned vehicles. Using measurement and simulation results, we evaluate the performance of the designed links in different communication scenarios and we show the benefits of employing multiple autonomous vehicles in the proposed communication system. Keywords: autonomous vehicles; maritime use cases; USV; UAV; collaborative communication system; directional wireless links; prototype design 1. Introduction Digitalization and digital transformation are shaping industries across the world. Among these, the maritime industry can benefit from the process of modernizing its existing practices to improve operational efficiency [ 1 ]. As part of the systems and solutions in which vessel operators are investing, autonomous vehicles and robotics are being utilized not only in ports and harbors, but also to aid the sea transport services in offshore areas [ 2 ]. The deployment of surface, aerial, and underwater autonomous systems open possibilities for improvements in current maritime operations and assist in several services, such as maritime search and rescue, onboard applications, navigation, and fleet management [ 3 ]. However, these benefits are conditional upon the utilization of a wireless communication system that involves these entities and supports the communications between them. This trend in unmanned-vehicle-aided offshore systems is the main motivation behind launching research in this area. Specifically, the goal is to construct a prototype of a collaborative communication system composed of different autonomous vehicles operating Sensors 2021,21, 3871. https://doi.org/10.3390/s21113871 https://www.mdpi.com/journal/sensors Sensors 2021,21, 3871 2 of 20 in an offshore environment. This system is built upon surface, aerial, and underwater components represented by a Unmanned Surface Vehicle (USV), Unmanned Aerial Vehicle (UAV), and Autonomous Underwater Vehicle (AUV), respectively. As part of the final prototype, a collaborative communication system should be implemented to connect these three subsystems using different radio access technologies. The overall communication layout of the system is depicted in Figure 1. Prospective applications of the AUV include underwater mapping, environmental surveying, and identification of underwater hazards to navigation. However, mainly the USV and UAV are considered in our previous and current works on the autonomous collaborative communication system. In [ 4 ], the design of a high-speed directional communication link between the Ground Control station (GC) and USV was outlined and its performance was evaluated. To further enhance the proposed automated offshore system, alternative solutions should be taken into account to maintain the GC–USV connection in cases where the Line-Of-Sight (LOS) link cannot be established. Hence, we build upon the work in [ 4 ] by involving the aerial component represented by the UAV to complement the previously proposed GC–USV setup and, thus, to enable a fully fledged automated offshore system [5]. GC USV AUV UAV Directional Wi-Fi (Links 1, 2, 3) Omni-directional Wi-Fi (Link 4) GPS (Links 5, 6) Landing beacon (Link 7) 1 1 2 2 3 3 5 5 6 6 4 4 7 7 Figure 1. Overall layout of the proposed wireless communication system. 1. GC–USV directional Wi-Fi; 2. USV–UAV directional Wi-Fi; 3. GC–UAV directional Wi-Fi; 4. UAV–AUV nondirectional Wi-Fi; 5. USV–satellite link; 6. UAV–satellite link; 7. Beacon for UAV landing. The rest of this article is organized as follows: Section 2provides a description of the state-of-the-art of the industrial applications of UAVs, specifically in the maritime industry, their different communication scenarios, and the challenges of their deployment with the review of relevant related research works. In Section 3, we describe the considered wireless communication modes between the GC, USV, and UAV and discuss the communication enablers of these modes, specifically in terms of location-based beam-steering capabilities and radio propagation models. We then detail, in Section 4, the prototype architecture design of the GC, USV, and UAV with their different communication modules. Further, Section 5describes the mechanical, power, and networking components and methods utilized in the prototype implementation of the automated offshore system. The performance of the proposed system is evaluated in Section 6, where we compare the evaluation results obtained from measurement campaigns to those obtained from simulations of selected analytical models. This comparison will allow us to identify the models that can better characterize the radio propagation in the studied scenarios and that can, thus, be used for simulation-based evaluation of future enhancements of the communication system. Additionally, the benefits of employing a UAV in the proposed communication system are evaluated on a real scenario and the technique for determining the optimal UAV location is described. Sensors 2021,21, 3871 3 of 20 2. State-of-the-Art and Related Work UAVs, commonly known as drones, are characterized by their mobility, adaptive altitude, advanced wireless communication, and sensing capabilities [ 6 ]. Such characteristics motivate several industries to employ UAVs in different operating scenarios, including but not limited to the manufacturing industry [ 7 ], agriculture [ 8 ], and public safety [ 9 ]. In fact, the global market of drones for industrial applications is projected to grow to around USD 43 billion by 2024, reaching the compound annual growth rate of 20% between 2018 and 2024 [ 10 ]. Similarly to other industries, the applications of UAVs in maritime use cases can vary from search and rescue procedures, data gathering for navigation and fleet management, and remote monitoring via video surveillance [ 11 ] to charging shipborne sensors using wireless power transfer [12]. As part of the deployment options, UAVs can be used as aerial base stations to enhance the coverage [ 13 ] and capacity of wireless networks [ 14 ]. This option can be valuable in offshore areas where LOS links with the terrestrial access network cannot be established. Additionally, UAVs can operate as flying mobile terminals and be used as relay nodes in out-of-coverage and Non-Line-Of-Sight (NLOS) scenarios [15]. In such situations, vessels can make use of the existence of proximate UAVs to access the network via relaying. In brief, UAV-assisted communications can be utilized to empower the performance of wireless networks toward reliable, massive data transmissions, and additional coverage ranges for maritime applications [16]. However, the abovementioned benefits come with various challenges in the deployment of UAVs. These challenges are in the scope of several research works, such as air-to-ground channel modeling [ 17 ], trajectory planning for UAVs performing exploration tasks in disaster scenarios [ 18 ], dynamic placement of relay UAVs [ 19 ], and the challenge of energy-efficient operation in UAV networks due to the limited onboard energy [ 20 , 21 ]. To tackle some of these challenges, certain research works have proposed the use of Software-Defined Networking (SDN) for several control functions in Flying Ad Hoc Networks (FANETs), including but not limited to topology management and collision avoidance [22,23]. Although involved in various applications, UAVs are utilized as the only type of unmanned vehicle in most of the systems studied in the reviewed research works. In this article, we propose the use of two types of unmanned vehicles, namely, USV and UAV, in a collaborative communication system for maritime use cases. We detail the UAV deployment scenarios in the proposed autonomous communication system in the following sections. 3. Proposed Communication Modes and Solution Components Based on the LOS/NLOS scenarios, two communication modes between the GC and the USV in the presented automated offshore system are proposed, namely, a LOS communication mode and a relay communication mode. The first mode is based on LOS communications between the GC and the USV and is depicted in Figure 2a. The second communication mode utilizes relaying via UAV in an NLOS scenario, as demonstrated in Figure 2b. GC USV Directional antenna GC-USV LOS link (a) GC–USV LOS communications. GC USV UAV Directional antenna Omni-directional antenna GC-UAV LOS link GC-USV NLOS link UAV-USV LOS link (b) GC–USV communications via UAV relaying. Figure 2. Illustration of the proposed communication modes in the autonomous collaborative system. Sensors 2021,21, 3871 4 of 20 To establish high-speed links, the GC, USV, and UAV are equipped with directional antennas and dynamic beam-steering capabilities. In Section 3.1, the location-based beamsteering algorithm utilized in our communication system is introduced. In addition, propagation models are utilized to predict the channel path losses and, thus, to characterize the behavior of radio propagation in the communication system with the consideration of certain environment effects. Further details on the selected propagation models are provided in Section 3.2. 3.1. Location-Based Beam-Steering Algorithm As depicted in Figure 2, the utilization of directional antennas depends on the communication mode. In the LOS communication mode, both the USV and GC are equipped with directional antennas. In order to align the beams of these antennas in the system, an algorithm for beam-steering based on devices’ locations was developed in [ 4 ]. This algorithm included two main steps: (1) calculation of the actual beam-steering angle; (2) transmission of the control commands with the new angle value. In the relay communication mode, both the USV and the UAV are equipped with one directional antenna and one omnidirectional antenna while the GC carries one directional antenna. Using the same algorithm proposed in [ 4 ], the drone’s yaw can be set in a way that the installed antenna is horizontally aligned with the GC. For tilting of the antennas on the UAV and the GC, Equation (1) , obtained from [ 24 ], can be used to calculate the elevation angle between two points with elevations elev1 and elev2 above the ground: Elevation angle =180 πelev2−elev1 d−d 2R, (1) where R is the Earth’s radius and d is the distance between two points ( lat 1, lon 1) and ( lat 2, lon2), which can be calculated using the following Haversine formula [24]: d=2rsin−1 ssin2lat2−lat1 2+cos(lat1)cos(lat2)sin2lon2−lon1 2!. (2) Finally, to find the tilting for the directional antennas on the UAV and the GC, the horizontal angle calculation part in the algorithm proposed in [ 4 ] should be replaced by the elevation angle using Equation (1). 3.2. Radio Propagation Models The choice of radio propagation models in the presented communication system depends on the radio links in both LOS and relay communication modes, namely, GC–USV and GC–UAV–USV links. Specifically, the Free Space Path Loss (FSPL) model is used for the simulation of the communications between the UAV and the other devices. The GC–USV link can be characterized with LOS communications in the near sea-surface environment. The selected propagation models are adjusted for the 5 GHz frequency bands since the IEEE 802.11ac standard is utilized as the connectivity solution in this communication system. On top of the offered capacity and coverage, the choice of Wi-Fi technology can be justified by its low cost and high availability that make it suitable for our experimental prototyping. 3.2.1. FSPL Model The FSPL model describes an ideal radio condition between a transmitting and a receiving antenna, where only LOS link exists without other sources of scattering, diffraction, or reflection [ 25 ]. It is one of the commonly used models for characterization of the ground-to-air radio propagation [ 26 ]. Therefore, this model is used to determine path loss in the communication from surface devices to the UAV based on Equation (3) [25]. LFSPL =32.44 +20log10(fc) + 20log10(d), (3) Sensors 2021,21, 3871 5 of 20 where LFSPL—free space path loss (dB); fc—carrier frequency (MHz); d—distance between transmitter and receiver (km). 3.2.2. Near Sea-Surface Propagation Models To characterize the radio propagation over the GC–USV link in the near sea-surface environment, the channel model from the study of Lee et al. [ 27 ] is utilized. Authors in [ 27 ] proposed a combination of 2-ray and 3-ray propagation models and showed that the utilization of each model depends on a break point dbreak . This metric indicates the transition point between the two models. In detail, the path loss prediction ability of the 2-ray model in a near sea-surface LOS environment deteriorates as the propagation distance increases beyond dbreak . In this latter case, the 3-ray model becomes a better option for the path loss prediction. The break point dbreak can be estimated using Equation (4) , where λ is the wavelength in meters, and ht and hr are the transmitting and receiving antenna heights in meters, respectively [27]. dbreak =4hthr λ. (4) The representation of the utilized 2-ray path loss model is depicted in Figure 3and the predicted path loss L2-ray (dB) can be calculated using Equation (5). L2-ray =−10log10(λ 4πd22 sin2πhthr λd2). (5) Figure 3. Representation of the 2-ray radio propagation. On top of the direct and reflected rays, the 3-ray model takes into consideration the refracted ray caused by the ducting effects, as shown in Figure 4. This model assumes that the evaporation duct layer is horizontally homogeneous and he is its effective height. Based on Equations (6) and (7), the 3-ray path loss L3-ray (dB) can be predicted. L3-ray =−10log10(λ 4πd2 [2(1+∆)]2), (6) with ∆=sin2πhthr λdsin2π(he−ht)(he−hr) λd. (7) Sensors 2021,21, 3871 6 of 20 Figure 4. Representation of the 3-ray radio propagation. 4. System Prototype Architecture To enable LOS and relay communication modes, the communication modules of all three platforms (USV, GC, and UAV) must be equipped with long-range high-throughput transceivers. The long-range aspect can be enabled with dynamic steering of directional antennas. Each module combines various mechanical, networking, and electrical components to enable the antenna steering capability. The USV, GC, and UAV have communication modules with different architecture designs that are better depicted in Figures 5–7, respectively. The USV consists of a single-board computer, a motion sensor, two servo motors, and a servo controller installed on the compound plate together with the directional antenna. These two servo motors are responsible for the vertical and horizontal steering of the antenna. For horizontal beam-steering of the USV antenna, the positioning data (i.e., position and orientation) is collected from a Global Positioning System (GPS) compass module by the USV main computer using Robot Operating System (ROS). Subsequently, the data from the GPS compass is sent to the single-board computer via Ethernet. By assuming that the location of the GC is fixed, this computer calculates the steering angle by using the positions of the vessel and the GC, then sends a control signal in the ROS format to the servo controller. The controller then activates a servo motor to rotate the USV antenna towards the GC. When the transceivers on both sides are connected through a Wi-Fi link, the USV’s GPS coordinates are sent to the ground system and collected by the controller board in GC. The GC steering system also includes a DC motor, a motor driver, and a servo motor. The DC motor steers the antenna horizontally to the direction of the autonomous vessel. Single-board computer Servo Motor (Azimuth) Servo Motor (Elevation) Servo controller Motion Sensor GPS module USV Central Computer USV Directional Antenna Power Regulator External power source Ethernet Serial I2C Ethernet Power line Data line Control line USV Omni-directional Antenna Ethernet Figure 5. Architecture design for the USV. The antenna systems on both the USV and GC are capable of vertical rotation. Since the vessel is floating on water, the tilting angle of the antenna may vary. To solve this, a 6-axis sensor is utilized to estimate the compensated vertical angle. The embedded computer controls the other servo to rotate the antenna in order to keep it stable against Sensors 2021,21, 3871 7 of 20 the waves. The servo motor in the GC system steers the antenna toward the flying UAV in the case when the system works in relay communication mode. Single-board computer Servo Motor (Elevation) DC Motor (Azimuth) GPS module USV Central Computer GC Directional Antenna Power regulator External power source Ethernet DC Motor Driver Rotary Encoder Power line Data line Control line Figure 6. Architecture design for the GC. In NLOS scenarios, a UAV can be deployed to establish a relay link. The UAV steering system includes an embedded computer and a servo motor. The computer receives GPS coordinates of the UAV, including latitude, longitude, and altitude, to calculate the tilting angle. Then, it controls the servo to steer the antenna vertically. For horizontal rotation, the yaw-control capabilities of the drone are used to keep the antenna constantly directed toward the GC. All the subsystems have a separate power regulator to provide a suitable power supply to each component. Single-board computer Servo Motor (Elevation) UAV Directional Antenna Power regulator UAV’s Batteries Ethernet Power line Data line Control line UAV Omni-Directional Antenna GPS module Ethernet USB Figure 7. Architecture design for the UAV. 5. System Prototype Implementation The mechanical structure for the communication part of the USV and GC was designed in our previous work [ 4 ]. The design of the antenna rotation mechanism for the USV is described in detail in [ 28 ]. In this work, the USV communication system was modified by using more rigid materials and more precise steering parts to improve the resistance against the wind and to increase the precision of rotation, as shown in Figure 8a. Two aluminum brackets were used to hold the antenna. This prevents the antenna from swinging when the wind is strong or when the autonomous vessel is moving at high speed. Two gears were placed on the bracket for the vertical rotation. One was connected to the servo motor through a shaft and the other was installed on the shaft of the antenna. A belt linked those two gears to transmit the rotary motion from the servo motor to the antenna. In this Sensors 2021,21, 3871 8 of 20 frame version, two ROBOTIS Dynamixel MX-28 servo motors were used for horizontal and vertical rotation of the vessel antenna. The MX-28 servo features the tracking capabilities of its speed, temperature, shaft position, voltage, and load. The shaft position can be maintained and modified accordingly for each individual servo with the control algorithm on the AX-12 actuator. It allows controlling the motor’s response in terms of speed and strength. The servo’s built-in microcontroller manages the control of sensors and antennas. The servo produces a high stall torque of 1.5 Nm and high no-load speed of 60 RPM, which are suitable for beam-steering rotation. (a) (b) Figure 8. Final mechanical design of both antennas on the USV and UAV. ( a ) Mechanical platform for the directional antenna on the USV; (b) Mounting platform for the directional antenna of the UAV. The directional antenna on the UAV was attached to a carbon-fiber platform. The design of this frame followed the structure of the UAV so that it could be mounted onto it, as demonstrated in Figure 8b. This mounting platform consists of several plates and has enough space to place other electrical devices on it. The brackets holding the antenna and the gears were 3D printed. The servo motor shaft was connected to one gear and the antenna was installed to the adjacent gear, which enabled the vertical rotation. 5.1. Single-Board Computers and Microcontrollers The USV and GC are both controlled by the Beaglebone Green single-board computers. They are responsible for exchanging GPS messages, running beam-steering algorithms and executing measurement scripts. Table 1shows the hardware specification of the Beaglebone Green board. Table 1. Hardware specifications of the Beaglebone Green board. Feature Value Processor AM335x 1 GHz ARMR Cortex-A8 RAM 512 MB DDR3 On-board storage 4 GB eMMC Accelerator support NEON floating-point and 3D graphics accelerator Micro USB 1, for Powering and data communication USB 1, for Hosting GPIO 2 ∗46 pin headers Networking 1 Ethernet port Operating temperature 0 to 75 °C Sensors 2021,21, 3871 9 of 20 The UAV system is controlled by the Udoo X86 ULTRA version single-board computer. One task of the controller on the UAV is forwarding the network packets in the relay communication mode. The Mikrotik SXT AC antenna and the Wi-Fi USB dongle were, respectively, connected to the Ethernet and USB ports of this board. Udoo X86 was selected due to its Gigabit Ethernet network interface and USB 3.0 to provide a high data-transfer rate. In addition, the chosen embedded board is compatible with Arduino 101 platform so that it can control the servo motor using Arduino software. Besides, this board has powerful hardware that can be utilized in other autonomous tasks such as video processing. The technical specifications of the Udoo X86 board are presented in Table 2. Another controller used in the system was the Arbotix-M Robocontroller. This robot controller is an advanced solution for Dynamixel servos and other high-accuracy robotic actuators. It incorporates a robust AVR microcontroller, a wireless interface, dual motor drivers, and 3-pin headers for hobby servos with digital and analog I/O. The Arbotix-M Robocontroller is for controlling the two, newly installed Dynamixel MX-28 servo motors. Table 2. Hardware specifications of the Udoo X86 board (Arezzo, Italy). Feature Value CPU Intel®Pentium N3710 up to 2.56 GHz GPU Intel®HD Graphics RAM 8 GB DDR3L Dual Channe Video interfaces 1∗HDMI 1.4 (CEC), 2∗Mini DisplayPort ++ On-board storage 32 GB eMMC soldered on-board Networking 1∗Gigabit Ethernet LAN interface 1∗M.2 Key E slot for optional Wireless Module Audio interfaces HD Audio Codec ALC283CG Microphone + Headphone Combo Connector (TRRS) Preamplified stereo speaker output, S/PDIF output USB 3∗USB 3.0 type-A sockets Other interfaces 2∗HSUART ports, 2∗I2C interface, 1∗SDIO interface 1∗LPC interface Platform compability Arduino™ 101-Compatible through standard Arduino™ Pins layout, compatible with Arduino™ shields 5.2. Power Management The power supply input varies for all the devices of the system. Thus, it is necessary to install different power regulators for each specific device. For the USV system, power was supplied by the ship accumulator and connected to the voltage regulators in the communication module of the moving vessel. The components on the UAV were provided with the power from the UAV’s battery. In order to show the power demands of utilized equipment, an overview on the input power requirements of each component is presented in Table 3. The GC and USV power supply systems have not changed from the previous design except for the installation of AX-12 servos and the Arbotix-M Robocontroller, as shown in Figure 8a. The final electrical design of the UAV is presented in Figure 9. Sensors 2021,21, 3871 16 of 20 GC UAV USV Figure 16. UAV relay scenario in Tampere University, Hervanta campus. While the UAV was airborne, the two antennas on the ground (USV and GC) were successfully connected and exchanged packets to each other using the relay link. The Udoo board installed on the UAV played the important role of a routing device and forwarded packets between the USB Wi-Fi dongle and Gigabit Ethernet interface connected to the MikroTik antenna. 6.2.2. Measurement Results The RSS levels at the UAV directional antenna and omnidirectional antenna on the ground are shown in Figure 17. The height of the UAV was constant at 22.5 m throughout the whole flight. There are four positions shown in Figure 17. At the beginning of the flight (position 0), the UAV was already at the height of 22.5 m above the ground. However, the directional antenna on the UAV was not aligned with the directional antenna on the GC until position 2. Before that, the links between all three parts (GC, UAV, and USV) had already been established, even though the RSS of the directional antenna on the UAV was relatively low. This was due to the fact that the distance between GC and UAV is short compared to the scenario at a lake. The RSS of the omnidirectional antenna is higher than that of the directional antenna on the UAV. As the beam started to be aligned at position 2, the signal level of the directional antenna continued increasing until position 3. Figure 17. RSS level in the relay communication mode measurement. The connection between the GC and USV depends on two ground-to-air radio links— between GC and UAV, and between USV and UAV. The measurement results show that one of them (with lower RSS) will always be a bottleneck link, which limits the throughput of the overall relay system. Therefore, the position of the UAV should be well selected to optimize the bottleneck radio link. 6.2.3. UAV Positioning Analysis In commercial solutions, the UAV moves autonomously; however, the optimal relaying position still needs to be calculated. The optimal location can potentially increase the RSS levels. In this section, the simulation reflecting the RSS with varying UAV positions is discussed and evaluated. Sensors 2021,21, 3871 17 of 20 The relay scenario with the position of the UAV are illustrated in Figure 18. As discussed in Section 3.2, the channel model used to simulate the GC–UAV and UAV–USV connections is the FSPL model, since the ground-to-air radio links in this situation are in an LOS and open-space environment. In the simulations, the scenario from Section 6.1 is used as well as the parameters. The l1 and l2 are set to 3 km and 1 km, respectively. The UAV is assumed to fly at an altitude h = 22.5 m above the ground, which is higher than the average level of trees and other infrastructures on the island. The values d1 and d2 are calculated based on the parameters above. In the simulation, the two values l1 and l2 vary from 0 to 4 km, which indicates the installed location of the GC and movement of the USV. The RSS levels of the GC–UAV and UAV–USV links are calculated based on the FSPL model and then compared to each other. The one with a lower signal level is the bottleneck, which will be plotted against d1and d2. Figure 18. Simulation scenario for obtaining the optimized UAV position between GC and USV. The bottleneck RSS plot of the communication system is shown in Figure 19. The latter shows that the RSS increases as the d1 and d2 distance decreases. Area A in the figure implies the cases when the bottleneck RSS only depends on d1 , regardless of the d2 value. In area B, the RSS values are subject to the UAV–USV distance. For each constant d1 , as the distance d2 between the UAV and the USV increases (crosses the red line), the UAV–USV link becomes the bottleneck for the overall system and the RSS values become lower than those in area A. Hence, it is necessary to keep the d2 distance to be on the left of the red line. The red line represents the optimal RSS while the GC–UAV–USV distances are growing. To find the relation between d1 and d2 on that line, each (d1 , d2) pair on the red line was collected and parsed to a curve fitting function in Matlab. The function poly f it() in Matlab is used for this purpose to calculate the coefficients of a polynomial that best fits the input data. Then, the line can be estimated as d2= 0.6309 d1− 0.1954, with d1 and d2 measured in km. Thus, with the current antenna setup and transmitter power settings, the UAV should keep its distance to the USV to be equal to or shorter than 0.6309 d1− 0.1954 to optimize the overall RSS of the system. The relation for optimal d1 and d2 can be expressed as d2⩽0.6309d1−0.1954. Figure 19. Bottleneck RSS of the overall radio link. Sensors 2021,21, 3871 18 of 20 Using the estimated optimal position of the UAV against the USV, the simulated bottleneck RSS of the relay link was compared to the measured RSS. The x-axis features the horizontal distance l between the GC and the USV, which means l=l1+l2 . Using that, along with the height and the relation d2= 0.6309 d1− 0.1954, a pair of (l1 , l2) and (d1 , d2) can be found for each distance l . From those values, the path loss values for GC–UAV and UAV–USV were calculated and the lower ones were collected as a basis for RSS to be plotted. Figure 20 presents the mentioned comparison. The USV loses the LOS at a distance beyond 3.5 km; the relay link (red curve) works better than the direct link (blue curve) and becomes a solution to maintain the connection between the GC and the USV. However, the direct link performs better than the relay link at distances below 3.5 km. The reason for this is the lower reception gain of the USV and UAV antennas used in the relay communication mode. In the LOS communication mode with direct GC–USV connection, the antenna on USV is directional and the gain of the main lobe is 23 dBi while in relay mode, the reception gain is 16 dBi for the directional antenna on the UAV and 12 dBi for the omnidirectional antenna on USV. The use of 12 dBi omnidirectional antenna on USV also limits the possible UAV–USV distance, which creates imbalances between optimal distance values for the relay mode components. While the UAV has to keep its proximity to the autonomous vessel, the range between the GC and the UAV becomes significantly higher than the UAV–USV distance. The 16 dBi directional antenna on the UAV cannot compensate that distance to provide an adequate signal strength compared to the 23 dBi one in the LOS mode, which leads to an unfair comparison of the LOS mode and relay mode in terms of equipment settings. Figure 20. Comparison of measured RSS and simulated relay RSS. Therefore, another comparison between LOS mode and relay mode was made with equal conditions of GC and USV. Assuming that the effective radiated power from the transmitter (GC) are the same for both cases, the USV will now be assumed to use the 23-dBi directional antenna on a UAV–USV link. Using the same approach mentioned previously, the simulated relay RSS is illustrated with a yellow curve. It is clear that the relay mode system with the 23-dBi directional antenna on USV performs better than the LOS mode system when the vessel is 2-km farther than the GC. In this case, with the higher received gain in USV, the range of UAV against the USV is extended and the GC–UAV and UAV–USV links in relay mode are now more balanced in terms of optimal range. Knowing the optimal position of the UAV in reference to the vessel provides ideas on how the UAV should be used in an autonomous collaborative offshore network. To prevent the signal loss, or session interruption in the NLOS case, the vessel should predict that there will be obstacles blocking the channel based on the map calculations and then deploy the UAV in advance to maintain the connection between the GC and the USV. Sensors 2021,21, 3871 19 of 20 After being launched, the UAV follows the vessel within a range that is calculated based on the distance between the USV and the GC using the optimal position investigated in this section. Another usage of the UAV is to extend the coverage of the wireless system. With higher USV antenna gain, the relay mode system can perform better than the LOS mode link in farther distances. Therefore, the vessel can deploy the UAV to improve the connection quality when the RSS drops below a defined value. 7. Conclusions In this article, we presented the results of our research work, which aimed to design, implement, and test a wireless autonomous collaborative communication system for emerging maritime use cases. The core components of the proposed system are GC, USV, and UAV. To support offshore monitoring and unmanned operations, the designed system is based on high-speed, long-range communications between these components. Taking into account the LOS/NLOS scenarios between the GC and USV, two communication modes were defined: an LOS communication mode, which implies a direct LOS link between the GC and USV; and a relay communication mode, where a UAV is deployed to act as a relay in NLOS scenarios. Each unmanned vehicle in the proposed autonomous collaborative system is equipped with communication modules that incorporate wireless interfaces, directional antennas, rotation mechanisms for mechanical beam-steering, and controllers for data collection and processing. Further details on the design and implementation of our communication system were provided in this article. We also assessed the performance of the designed and implemented links using measurement campaigns. Additionally, simulations were utilized—first, to compare the measurement data to analytical results based on the reviewed propagation models; second, to determine the optimal UAV position in the relay communication mode. Potential enhancements of our communication system include conducting the measurements on the relay communication mode in a maritime environment, and developing a handover mechanism between the two defined communication modes to maintain continuous connections between the unmanned vehicles and the control station. A next step in the implementation of the overall automated offshore system depicted in Figure 1can be the design and implementation of communication links with the AUV. Including the underwater component will enable a fully integrated collaborative system with more autonomous vehicles and further support of emerging maritime use cases. Author Contributions: Conceptualization, J.P., K.M., M.G.; methodology design, S.S., J.F.; prototype implementation, J.P., K.M., J.F., J.V., M.G.; experimental evaluations, K.M., J.V., M.G.; results validation, J.P., S.S., J.H.; writing—original draft preparation, J.P., K.M., S.S., J.F.; writing—review and editing, J.V., M.G., J.H.; supervision, M.G., Y.K., J.H.; project administration, Y.K.; funding acquisition, M.G., Y.K. All authors have read and agreed to the published version of the manuscript. Funding: This paper is based on the Autonomous and Collaborative Offshore Robotics (aCOLOR) project funded by the Technology Industries of Finland Centennial and Jane & Aatos Erkko Foundations under Future Makers Funding Program 2017. The authors gratefully acknowledge the contributions of the company Alamarin-Jet Oy. The described research was financed by the Ministry of Industry and Trade of Czech Republic project No. FV40309. The authors also acknowledge funding from the European Union’s Horizon 2020 Research under the Marie Sklodowska Curie grant agreement No. 813278 (A-WEAR project). Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Not applicable. Conflicts of Interest: The authors declare no conflict of interest. Sensors 2021,21, 3871 20 of 20 References 1. Sullivan, B.P.; Arias Nava, E.; Desai, S.; Sole, J.; Rossi, M.; Ramundo, L.; Terzi, S. Defining Maritime 4.0: Reconciling principles, elements and characteristics to support maritime vessel digitalisation. IET Collab. Intell. Manuf. 2021. [CrossRef] 2. Sanchez-Gonzalez, P.L.; Díaz-Gutiérrez, D.; Leo, T.J.; Núñez-Rivas, L.R. Toward digitalization of maritime transport? Sensors 2019,19, 926. [CrossRef] 3. Zolich, A.; Palma, D.; Kansanen, K.; Fjørtoft, K.; Sousa, J.; Johansson, K.H.; Jiang, Y.; Dong, H.; Johansen, T.A. Survey on communication and networks for autonomous marine systems. J. Intell. Robot. Syst. 2019,95, 789–813. [CrossRef] 4. Khosravi, Z.; Gerasimenko, M.; Urama, J.; Pyattaev, A.; Escusol, J.V.; Hosek, J.; Andreev, S.; Koucheryavy, Y. Designing high-speed directional communication capabilities for unmanned surface vehicles. In Proceedings of the 2019 16th International Symposium on Wireless Communication Systems (ISWCS), Oulu, Finland, 27–30 August 2019; pp. 651–655. 5. Ma, K. Implementation and Evaluation of Communication System for Autonomous Offshore Vehicles. Master’s Thesis, Tampere University, Tampere, Finland, 2020. 6. Valavanis, K.P.; Vachtsevanos, G.J. Handbook of Unmanned Aerial Vehicles; Springer: Berlin, Germany, 2015; Volume 1. 7. Salhaoui, M.; Guerrero-González, A.; Arioua, M.; Ortiz, F.J.; El Oualkadi, A.; Torregrosa, C.L. Smart industrial IoT monitoring and control system based on UAV and cloud computing applied to a concrete plant. Sensors 2019 ,19, 3316. [CrossRef] [PubMed] 8. Delavarpour, N.; Koparan, C.; Nowatzki, J.; Bajwa, S.; Sun, X. A Technical Study on UAV Characteristics for Precision Agriculture Applications and Associated Practical Challenges. Remote Sens. 2021,13, 1204. [CrossRef] 9. Mehallegue, N.; Djellab, M.; Loukhaoukha, K. Efficient Use of UAVs for Public Safety in Disaster and Crisis Management. Wirel. Pers. Commun. 2021,116, 369–380. [CrossRef] 10. Global Commercial Drone Market Size in 2018 and 2024. Available online: https://www.statista.com/statistics/878018/globalcommercial-drone-market-size/ (accessed on 28 May 2021). 11. Mozaffari, M.; Saad, W.; Bennis, M.; Nam, Y.H.; Debbah, M. A tutorial on UAVs for wireless networks: Applications, challenges, and open problems. IEEE Commun. Surv. Tutor. 2019,21, 2334–2360. [CrossRef] 12. Carrillo, D.; Mikhaylov, K.; Nardelli, P.J.; Andreev, S.; da Costa, D.B. Understanding UAV-Based WPCN-Aided Capabilities for Offshore Monitoring Applications. IEEE Wirel. Commun. 2021. [CrossRef] 13. Alzenad, M.; El-Keyi, A.; Lagum, F.; Yanikomeroglu, H. 3-D placement of an unmanned aerial vehicle base station (UAV-BS) for energy-efficient maximal coverage. IEEE Wirel. Commun. Lett. 2017,6, 434–437. [CrossRef] 14. Lai, C.C.; Chen, C.T.; Wang, L.C. On-demand density-aware uav base station 3d placement for arbitrarily distributed users with guaranteed data rates. IEEE Wirel. Commun. Lett. 2019,8, 913–916. [CrossRef] 15. Li, M.; Yu, F.R.; Si, P.; Yang, R.; Wang, Z.; Zhang, Y. UAV-Assisted Data Transmission in Blockchain-Enabled M2M Communications with Mobile Edge Computing. IEEE Netw. 2020,34, 242–249. [CrossRef] 16. Zeng, Y.; Zhang, R.; Lim, T.J. Wireless communications with unmanned aerial vehicles: Opportunities and challenges. IEEE Commun. Mag. 2016,54, 36–42. [CrossRef] 17. Khawaja, W.; Guvenc, I.; Matolak, D.W.; Fiebig, U.C.; Schneckenburger, N. A survey of air-to-ground propagation channel modeling for unmanned aerial vehicles. IEEE Commun. Surv. Tutor. 2019,21, 2361–2391. [CrossRef] 18. Sánchez-García, J.; Reina, D.; Toral, S. A distributed PSO-based exploration algorithm for a UAV network assisting a disaster scenario. Future Gener. Comput. Syst. 2019,90, 129–148. [CrossRef] 19. Magán-Carrión, R.; Camacho, J.; García-Teodoro, P.; Flushing, E.F.; Di Caro, G.A. A Dynamical Relay node placement solution for MANETs. Comput. Commun. 2017,114, 36–50. [CrossRef] 20. Zhan, C.; Zeng, Y.; Zhang, R. Energy-efficient data collection in UAV enabled wireless sensor network. IEEE Wirel. Commun. Lett. 2017,7, 328–331. [CrossRef] 21. Xu, J.; Zeng, Y.; Zhang, R. UAV-enabled wireless power transfer: Trajectory design and energy optimization. IEEE Trans. Wirel. Commun. 2018,17, 5092–5106. [CrossRef] 22. E Silva, T.D.; de Melo, C.F.E.; Cumino, P.; Rosário, D.; Cerqueira, E.; De Freitas, E.P. STFANET: SDN-based topology management for flying ad hoc network. IEEE Access 2019,7, 173499–173514. [CrossRef] 23. Zhao, Z.; Cumino, P.; Souza, A.; Rosario, D.; Braun, T.; Cerqueira, E.; Gerla, M. Software-defined unmanned aerial vehicles networking for video dissemination services. Ad Hoc Netw. 2019,83, 68–77. [CrossRef] 24. How To Calculate Distances, Azimuths and Elevation Angles Of Peaks. Available online: http://tchester.org/sgm/analysis/ peaks/how_to_get_view_params.html (accessed on 28 May 2021). 25. Parsons, J.D. The Mobile Radio Propagation Channel; Wiley: Hoboken, NJ, USA, 2000. 26. Cui, Z.; Briso, C.; Guan, K.; Matolak, D.W.; Calvo-Ramírez, C.; Ai, B.; Zhong, Z. Low-altitude UAV air-ground propagation channel measurement and analysis in a suburban environment at 3.9 GHz. IET Microw. Antennas Propag. 2019 ,13, 1503–1508. [CrossRef] 27. Yee Hui, L.; Dong, F.; Meng, Y.S. Near sea-surface mobile radiowave propagation at 5 GHz: Measurements and modeling. Radioengineering 2014,23, 824–830. 28. Morón Alguacil, C. Design and Analysis of Directional Antenna Structure for Unmanned Surface Vessel. Master’s Thesis, Tampere University, Tampere, Finland, 2019.