AOVs currently perform several offshore operations in maritime environments.
These operations need accurate navigation and localization to ensure the accuracy of the data acquisition and processing. The navigational accuracy is the precision to reach a predetermined waypoint, while the localization accuracy relates to the error when localizing the AOV within a map.
The most common practices in autonomous systems above the water surface are the global positioning system (GPS) and spread spectrum or radio communications.
However, these signals can only propagate in short distances while performing in an underwater scenario and are not suitable for autonomous underwater systems.
Thus, acoustic-based sensors and communications are selected for underwater applica-tions because they have better performance. Nevertheless, acoustic communicaapplica-tions still suffer from many shortcomings, such as the low data rate, small bandwidth, high latency, or unreliability. Hence, the communication system must manage its transmissions without losing data.
2.1.1 Unmanned Surface Vehicles
Situational awareness is crucial in the design of high levels of autonomy in USVs.
Wolfet al. [86]developed situational awareness during USV patrol missions based on change detection and object-level tracking method for detecting targets, establishing
their position, and identifying fluctuations in the nearby environment. For the USV state estimations, a vector GPS compass is commonly used as the primary sensor to obtain accurate heading and position for USV navigation. In this case, the GPS compass uses a satellite-based augmentation system for differential GPS position, providing a low-cost and highly accurate vehicle pose. Apart from the GPS compass, active ranging sensor methods, such as LiDAR and radar, can be utilized for state estimation. These methods are notably effective when there is a loss or jamming of GPS signals. These GPS signals may become weak and unreliable when the USV navigates near bridges or other covered environments. Additionally, a suitable choice for these scenarios is SLAM, which is becoming increasingly important in research applications due to the possibility to detected contours and employ them as landmark features[31].
USVs require the capabilities of obstacle detection and recognition, tracking targets, and mapping environments to accomplish real-world applications. There are two categories when grouping the environmental perception approaches for USVs based on the characteristics of the intended applications: passive perception methods, which adopt the infrared or visual sensors employed in numerous environment perception applications, and active ranging sensor methods, with LiDAR, radar, and sonar as the main sensors. LiDARs are the sturdiest sensors for acquiring depth data in obstacle detection techniques. Halterman and Bruch[29]studied the performance of three-dimensional (3D) scanning LiDAR installed in a USV. Another active perception sensor in USV applications is marine radar, which is the most used obstacle detection method for far-field applications[40]. The primary use of pulse radar sensors is still in the military area, but it is becoming more important in research applications.
Zhuanget al.[89]developed an embedded collision avoidance system in a USV based on a marine radar sensor. Additionally, Hanet al. [32] addressed the algorithm development for multiple target detection and tracking for a USV in the sensor fusion framework by integrating LiDAR and marine radar.
2.1.2 Autonomous Underwater Vehicles
Most underwater applications still use old technologies, such as long baseline and ultra-short baseline (USBL), requiring support infrastructure. However, dynamic multi-agent system approaches are more often being used in these applications because they
allow for flexibility and rapid deployment using minimal infrastructure. Regarding these dynamic approaches, underwater systems increasingly include the use of SLAM techniques based on above-ground robotics applications[63]. Thus, more accurate AUV navigation is becoming possible in a more cost-efficient way.
Accurate localization and navigation are essential in data acquisition and process-ing for autonomous applications. As mentioned above, most autonomous systems count on GPS systems and spread-spectrum communications above the water’s sur-face. However, those signals only propagate over short distances because of the rapid attenuation of higher-frequency signals in the underwater environment. Thus, acoustic-based systems are used in AUV applications because their performance is better in the underwater scenario[58]. The underwater navigation and localization techniques are categorized as the following main categories[41]:
• Inertial/dead reckoning: Inertial navigation uses gyroscopes and accelerometers to disseminate the current AUV state. Nonetheless, each of these methods has unbounded position error growth.
• Acoustic transponders and modems: These navigation techniques measure the time of flight between signals from acoustic beacons or modems to the other platform.
• Geophysical: These techniques utilize external environmental information as references for the AUV’s navigation. The underwater sensors need to detect, identify, and classify some surrounding environment features.
The underwater localization and navigation methods need specific navigation and survey sensors placed in the AUV platform. The most basic sensors for AUV navigation are the compass, which provides a globally bounded heading reference, and the barometer or pressure sensor, which measures the underwater depth of the AUV.
Regarding acoustic navigation techniques, USBL navigation enables the underwater localization of the AUV relative to a support platform, offering an efficient and stable acoustic communication network[53]. The phase differencing across transceivers determines a relative bearing, while the time of flight determines the range. These transceivers, also known as model and transponder units, form the USBL navigation system, with its range being a major limitation. The modem is usually installed on the AUV’s nose, while there is an acoustic transponder placed on a support platform that acts as the target because its position known and fixed. Batistaet al.[3]proposed the
use of a USBL positioning system for sensor-based integrated guidance and control.
This approach is used in the aColor AUV in the current thesis, employing the USBL navigation system for localization and possible communication with the support platform. Additionally, a mechanical imaging sonar allows underwater situational awareness capabilities.
Regarding AUV navigation algorithms, Milleret al. [48]considered the navigation problem for an AUV using an error state estimation based on a Kalman filter. The sensors used for this state estimation were a doppler velocity log (DVL), a pressure sensor, an long baseline system, and an attitude sensor. Ribaset al. [62]addressed the development of the Girona500 AUV that implements dead reckoning navigation based on a solid-state attitude and heading reference system (AHRS) and a DVL.
Additionally, their study included the absolute position through a USBL system while the vehicle is underwater and using a GPS signal while it is on the water surface.
The high-accuracy USBL system enables underwater localization and communication between the support and AUV platforms. This thesis employs the Girona500 AUV as the advanced underwater platform, which involves the sensor integration for localization and situational awareness capabilities of the AUV. Additionally, Font et al. [20]addressed a USBL-aided navigation method in an AUV. Their method included the state estimation based on a two-parallel extended Kalman filter with the data gathered from a pressure sensor, a GPS, a DVL, an inertial measurement unit, and a visual odometer.