One of the most significant challenges in the underwater domain is to retrieve the autonomous underwater vehicle (AUV) position within the surrounding environment. Indeed, reliable navigation systems are fundamental to perform complex tasks and missions. Most of the navigation filters for AUVs are based on Bayesian estimators such as the linear Kalman Filter (KF), the extended KF, the unscented KF, or the particle filter where, usually, different instruments including a Doppler velocity log (DVL) contribute to the localization task. The usage of forward‐looking SONARs (FLS) in navigation‐aiding is, most of the time, devoted to limiting the navigation drift of the AUV by using simultaneous localization and mapping methods. Therefore, these devices are commonly employed with a standard navigation sensors set comprising an attitude heading reference system and a DVL. In this contribution, the authors propose a novel navigation strategy specifically tailored to AUVs based on an adaptive unscented KF, where linear speed estimations are obtained with a 2D FLS instead of with a DVL and therefore promoting the employment of FLSs as an aid for underwater navigation. The marine robotics community could gain significant benefits from reliable navigation achieved with an FLS‐based navigation architecture. Most importantly, a single FLS can be used for imaging‐related applications (i.e., sonograms acquisition) and navigation, where, instead, different dedicated devices are currently employed for the two tasks. Smaller AUVs usually possess reduced payload carrying capabilities; thus, multitasking use of onboard sensors, which leads to compactness, is a desirable feature. Navigation data obtained during sea trials performed in La Spezia (Italy) at the NATO STO Centre for Maritime Research and Experimentation has been used for offline validation. Afterward, the online results of real autonomous underwater missions undertaken in La Spezia (Italy) and at Vulcano Island, Messina (Italy), are reported.

中文翻译：

---

带有2D前瞻性SONAR的水下导航：用于AUV的基于自适应无味卡尔曼滤波器的策略

水下领域中最重大的挑战之一是在周围环境中找回自主水下航行器（AUV）的位置。确实，可靠的导航系统对于执行复杂的任务和任务至关重要。用于AUV的大多数导航滤波器都基于贝叶斯估计器，例如线性卡尔曼滤波器（KF），扩展KF，无味KF或粒子滤波器，通常，包括多普勒速度测井（DVL）在内的不同仪器会本地化任务。在大多数情况下，前瞻性SONAR（FLS）在导航辅助中的使用致力于通过使用同时定位和制图方法来限制AUV的导航漂移。所以，这些设备通常与包含姿态航向参考系统和DVL的标准导航传感器一起使用。在这项贡献中，作者提出了一种基于自适应无味KF专门针对AUV量身定制的新颖导航策略，其中使用2D FLS而不是DVL获得线速度估计，从而促进FLS的使用，以帮助进行水下导航。借助基于FLS的导航体系结构实现可靠的导航，海洋机器人技术社区可以从中受益匪浅。最重要的是，单个FLS可以用于成像相关的应用程序（即，超声图采集）和导航，而目前，这两个任务采用了不同的专用设备。较小的AUV通常具有减少的有效载荷运载能力；因此，车载传感器的多任务处理使用会导致结构紧凑，这是一个理想的功能。北约STO海事研究与实验中心在拉斯佩齐亚（意大利）进行的海上试验中获得的导航数据已用于离线验证。之后，报告了在拉斯佩齐亚（意大利）和墨西拿的武尔卡诺岛（意大利）进行的实际自主水下任务的在线结果。