Underwater target tracking is a fundamental technology for marine development, providing real-time position estimates of the interested targets. However, due to the harsh underwater environment and the noncooperativity of targets, improving tracking accuracy remains a challenge, especially for highly-maneuvering targets. To address this problem, based on multi-autonomous underwater vehicle (multi-AUV) networks, this paper extends the idea of interacting multiple models (IMM) and designs a bearing-only cooperative tracking algorithm in the consideration of the harsh underwater acoustic channels. Specifically, in position prediction, the combination of historical information and the concept of IMM reduces the severe time-lagged effect in traditional prediction methods and the model reliance in standard IMM filters. Then, during position update, a rigidity-assisted relative position representation is designed based solely on bearing measurements, which alleviates the impact of information loss due to communication interruptions, significantly enhancing the continuity of target tracking. Moreover, the algorithm design also considers various uncertainties that may concurrently occur underwater (e.g., error accumulation and model mismatches), and robust optimization strategies with the principle of maximum entropy are designed to enhance the environmental adaptability. Through various simulations and field experiments, the advantages of the proposed method have been validated.

RGB-Thermal (RGB-T) multimodal tracking has gained widespread attention due to its robustness in handling complex scenarios. Some existing methods focused on fully fine-tuning RGB-based trackers, which was parameter-inefficient and prone to overfitting due to the scarcity of multimodal data. Therefore, recent studies have explored multimodal prompting strategies, which mainly treat RGB as the dominant modality and TIR as an auxiliary prompting modality. This asymmetric framework lacks adaptability in changing dominant modalities, resulting in reduced robustness. To address these limitations, we propose LRPD: a novel multimodal Low-Rank (LoRA) Prompting and Distillation tracking framework. Specifically, the framework consists of two distinct stages. In the first stage, we pre-train a teacher (ViT-L encoder) and a student model (ViT-B encoder) using our LoRA-Prompting (LoRA-P) module. LoRA-P adopts a symmetric architecture that enables bidirectional cross-modal interaction in a parameter-efficient manner. In the subsequent stage, we design a prompt-driven knowledge distillation framework to transfer knowledge from the large teacher model to the lightweight student model. Task-specific designs, including an enhanced patch masking strategy, a feature alignment projector, deep visual prompts, and specialized distillation losses, are tailored to optimize student's tracking performance. Extensive experiments on three popular RGB-T tracking benchmarks demonstrate our method achieves new state-of-the-art performances.

Accurate positioning is essential for meaningful data collection in underwater acoustic sensor networks (UASNs), and localization has become a fundamental technology that provides real-time position estimates for sensor nodes. However, due to the harsh underwater environment as well as the difficulties and high expenses in network maintenance, localization in UASNs has always been a challenging problem. Different from previous works that rely on fixed anchors (e.g., buoys), this article uses an autonomous underwater vehicle (AUV) as a mobile anchor and proposes a reinforcement-learning-based trajectory planning method that allows the AUV navigation to meet the localization requirements of all sensor nodes. Specifically, based on gridded scenarios, this work models the node position uncertainties with information entropy and formulates AUV trajectory planning as a process of reducing the entropy of the whole network. Moreover, a modified actor-critic-based deep deterministic policy gradient (DDPG) reinforcement learning algorithm is designed to shorten AUV trajectory on the premise of ensuring a certain localization accuracy for UASNs. Through various numerical comparisons, the advantages of the proposed method have been validated in terms of efficiency and localization accuracy.

Cooperative localization aims to continuously provide position estimates for multiple-autonomous underwater vehicle (multi-AUV) systems during task execution such as marine monitoring; it is preferable over noncooperative schemes due to its high accuracy and strong robustness. However, various uncertain factors underwater, including model mismatches, accumulated errors, measurement noises and biases, time-varying communication channels, etc., still challenge the accuracy and robustness of cooperative localization. When such uncertainties arise, the performances of traditional methods degrade significantly. Therefore, this paper proposes a robust multi-AUV cooperative localization method that is able to combat these uncertainties by leveraging the principle of maximum entropy. To be explicit, a message-passing scheme is established using factor graphs, over which a distributed position estimation strategy for AUVs is designed based on belief propagation. To reduce the damages of uncertainties, maximum-entropy distributions are designed respectively for the prediction and correction processes of localization and are realized by particles. Specifically, through enlarging the particle coverage, uncertainty-induced misleading in position estimation is alleviated, and hence higher robustness is achieved. Simulations and field experiments show the advantages of the proposed algorithm over the state-of-the-art methods in terms of localization accuracy, robustness, and scalability.Note to Practitioners—In practical applications, due to the unavailability of global positioning systems underwater, AUV localization still lacks mature and stable solutions. In harsh underwater environments, most theoretical models often fail to accurately describe the practical conditions, leading to widespread mismatches, which severely degrade localization performance. Moreover, uncertainties such as accumulated errors, measurement noises, and position deviations would further reduce the position estimation accuracy. Existing methods usually consider these uncertainties independently, while, in practical uses, uncertainties often emerge in combination, making it challenging to maintain localization accuracy. What is worse is that most sophisticatedly designed algorithms pursue high accuracy, and their adaptability in engineering applications is difficult to guarantee. Hence, this work provides a robust solution for multi-AUV cooperative localization that can handle uncertainties simultaneously. By reducing the sensitivity to different uncertainties, the proposed algorithm can provide sustained high-accuracy position estimates for AUVs in complex underwater environments. The proposed method is experimentally verified and suitable for multi-AUV applications such as oceanic rescue, resource development, and marine monitoring.

Simultaneous target approaching is a fundamental ability for multiple autonomous underwater vehicle (multi-AUV) systems to continuously move toward the target for surrounding, capturing, information gathering, etc., and the efficiency and accuracy of the approaching heavily rely on the qualities of planned trajectories. This work presents a cooperative trajectory planning method for the simultaneous approaching of multi-AUV systems, and further provides strategies for connectivity maintenance and position errors to enhance the adaptability in practical uses. To be specific, a cooperative planning framework is designed with the inspiration of sensor selection, where trajectory planning is converted to a successive candidate selection problem and modeled as convex optimization. Under the framework, the considered simultaneous approaching, connectivity maintenance, and position errors are formulated as constraints and integrated into one convex optimization problem, which can be easily solved via well-established solvers. Various simulations and field experiments have validated the advantages of the proposed method by comparisons with alternative methods.

In the operation of multiple autonomous underwater vehicle (AUV) networks, AUV self-localization, inter-AUV communication, and AUV trajectory planning (collectively referred to as LCP) are fundamental technologies. Among them, accurate position estimation is often a prerequisite for effective planning, while communication is essential in both parts. Previous studies have typically addressed these three areas independently, yet their conflicting demands on resources, objectives, etc., frequently pose significant challenges to performance improvement. This paper investigates the LCP from an integrated perspective and a co-designed approach jointly addresses all of them is proposed based on graphical models and belief propagation. Specifically, guided by the various underwater communication conditions, a cooperative localization algorithm for AUVs is designed with the assistance from trajectory planning. Different from the previous, planning results are directly used for localization to resist harsh communications, rather than indirectly affecting localization accuracy by, for example, optimizing geometric relationships. As a result, the feedback enhancement from trajectory planning to AUV localization is realized and the mechanism of their mutual promotion for different communication conditions is achieved. Moreover, considering the changing trajectories, the proposed cooperative localization algorithm is further refined for underwater obstacle-avoidance scenarios. Various simulations and field experiments validate that the proposed algorithm can improve the overall performance of LCP by comparisons with state-of-the-art alternative methods.

This letter investigates the cooperative localization of autonomous underwater vehicle (AUV) systems in underwater anchor-free environments. Different from previous works, the impacts of mutual asynchronization among AUVs and acoustic wave bends are collectively considered to enhance localization accuracy. Specifically, a novel relative measurement representation called modified range difference (MRD) is firstly designed to mitigate the measurement errors jointly caused by the above two factors, which can be obtained with every two different underwater acoustic communications and further corrected with sound speed profiles. Corresponding to MRD, an extended state is also designed to explore the use of historical position information. As a result, an improved cubature Kalman filter is proposed, which can tackle the nonlinearity of MRD and the commonly unknown measurement variance. Consequently, the proposed method can provide higher localization accuracy by mitigating the combined effects of asynchronization and acoustic wave bends without synchronization operations and anchor assistances, which delivers superior adaptability for anchor-free scenarios. Simulations verify that the proposed method outperforms state-of-the-art ones in terms of localization accuracy.

The asynchronization among autonomous underwater vehicles (AUVs) is inevitable due to the inherent offsets and drifts in clocks, which critically degrades the accuracy of AUV cooperative localization (CL), especially in underwater anchor-free environments. This article designs a novel relative measurement representation called time difference of communication (TDOC) to eliminate the impact of asynchronization, which only requires one-way inter-AUV communications at different time steps even in the presence of large asynchronization. By exploiting TDOC, a joint spatial-temporal CL model is established with specially constructed measurement vectors, and a CL method, termed TDOC-CL, is proposed. It transforms the nonlinear observations into a weighted least-square (WLS) formation and is solved via iterative optimization, where various coupled uncertainties (i.e., measurement biases, measurement noises, and position errors) are involved and analyzed. Moreover, to compensate the accuracy degradation caused by the uncertainties, the TDOC-CL with bias reduction and error compensation, named iTDOC-CL-BREC, is further designed. In addition, the hybrid Cramér-Rao lower bound (HCRLB) is derived as the performance benchmark to evaluate the localization behaviors. Simulations and field experiments show that iTDOC-CL-BREC can provide higher localization accuracy compared with state-of-the-art methods, especially for long-duration tasks.

This article presents a hybrid time-of-arrival (TOA) and angle-of-arrival (AOA) cooperative localization algorithm, termed HTACL, for the multiple-autonomous-underwater-vehicle (multi-AUV) navigation in the absence of anchors (devices with known positions). Compared with traditional TOA-based methods, the joint use of TOA and AOA measurements improves the quality of position prediction and inter-AUV relative position information in localization. Specifically, by exploring the relationship between AOA measurements and AUV attitudes, the impact of the accumulated errors in inertial measurements on the position prediction is alleviated. Moreover, the directivity and resolution of relative position representation is improved through involving AOA information. As a result, the localization accuracy can be improved by HTACL. The design of HTACL also considers the influence of the long propagation delay during AUV cooperation to enhance its adaptability to harsh underwater environments. Moreover, HTACL is naturally distributed and with good scalability. The abovementioned features make it applicable and suitable for multi-AUV localization issues. Through various simulations, the advantages of HTACL are verified and it could stably provide better localization accuracy compared with the state-of-the-art alternative methods under different conditions.

Underwater communications are fundamental techniques for the Internet of Underwater Things (IoUT) to establish information links among underwater devices. Recently, orthogonal chirp division multiplexing (OCDM) has drawn great attention in underwater communications due to its advantages in dealing with burst interference in both time and frequency domains. However, the harsh underwater channel conditions (e.g., multipath propagation, temporal variations, and significant Doppler effects) have not been systematically considered in current underwater OCDM systems. In this article, two transmission block structures (Structure A and Structure B) based on OCDM are designed, and a unified OCDM receiver framework is proposed under a single scale multipath lag (SSML) channel model. In the receiver framework, the Doppler scaling factor and carrier frequency offset (CFO) are sequentially compensated, and receiver algorithms that apply the separate or superimposed transmission features (for the pilot and data) of Structure A and Structure B are proposed, respectively. To be specific, a multipeak Doppler scaling factor estimation algorithm and a closed-form CFO estimator are designed for the received signal of Structure A. Moreover, a null symbol-based CFO estimation algorithm is presented for the received signals of Structure B. The effectiveness and advantages of the proposed methods are analyzed and validated through simulations and channel data by comparisons with existing methods under different conditions.

The noncooperative target tracking is an important issue for the Internet of Underwater Things (IoUT). Autonomous underwater vehicles (AUVs) are preferred options to achieve the target tracking especially in anchor-free environments, where no equipments with known positions, named anchors, are deployed. The self-organized mobile network of multiple AUVs can localize and continuously monitor the target. Thus, in this article, we investigate the problem of the noncooperative target tracking using multiple AUVs in anchor-free environments. In the target tracking, AUVs play as references and their positions need to be estimated first. We propose a multi-AUV cooperative localization and target tracking (MCLTT) framework based on belief propagation (BP). Under MCLTT, BP-based underwater cooperative localization (BPUCL) and noncooperative mobile target tracking (NcMTT) algorithms are designed. Gaussian approximations are used to reduce communication costs among AUVs. The designed BPUCL alleviates the impact of the accumulated errors in the inertial measurements of AUVs and slows down the growth of the localization error. In NcMTT, model-free position prediction processes are proposed and a novel form of the particle-based BP message is designed using time-difference-of-arrival (TDOA) measurements. The simulation results validate the proposed algorithms by comparing with state-of-the-art methods.